The material is based upon work supported by the STC Program of the National Science Foundation No. DMR 0120967.

All rights reserved. No part of the Review may be reproduced in any form or by any means without written permission.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Printed in the United States at the University of Washington, Seattle, WA

Inquires should be addressed to:

Center on Materials and Devices for Information Technology Research Educational Partnership Programs University of Washington Department of Chemistry Box 351700 Seattle, WA 98195-1700 [email protected] http://stc-mditr.org

Volume 2, Number 1 Welcome to the Second Edition of CMDITR Review of Undergraduate Research

This volume of the Reviews features extended abstracts of students who participated in the National Science Foundation (NSF) Center on Materials and Devices in Information Technology Research (CMDITR) Summer 2005 Research Experiences for Undergradu- ates (REU) program. The REU experience often acts as a launching point or catalyst for entry by undergraduates into technical fields of study. The REU experience is symbiotic in nature as it supports not only the undergraduate participant who experiences research first-hand prior to com- mitting to graduate study, but also the students mentors, be they faculty members, research scientists, post-doctoral fellows, or graduate students. While working with REU students graduate students and post-doctoral fellows learn teaching and mentoring skills needed to manage labs they will be responsible for in the future. Faculty members and other researchers, who share their expertise with these enthusiastic learners, also benefit as they reflect on the excitement upon which their careers have been built. The CMDITR REU Summer Program placed undergraduate students from across the United States in CMDITR state-of-the-art research labs at the University of Washington, University of Arizona and Georgia Institute of Technology. The 2005 program expanded on the previous year’s program with a doubling of the number of undergraduates to 32 participants. Several of these undergraduates in the 2005 program were part of a new, collaborative exchange program between Norfolk State University and Georgia Tech. All of these undergraduates worked on authentic interdisciplinary research contributing to advancements in information technology with researchers in the fields of chemistry, physics, optics, materials science and engineering. The REU Program emphasized the teamwork nature of sci- entific research and was supplemented by a collection of activities including ethics training and workshops in scientific communication. The role of the Review is to offer a forum for participants involved in the CMDITR REU to share their research with their REU peers, future REU students, CMDITR graduates students and faculty members, and others interested in the work of CMDITR. The Review is also a forum that depicts the breadth and depth of CMDITR research. To learn more about the CMDITR REU program and opportunities please visit http://stc-mditr.org/REU. A special thanks to all the REU participants for their work, their mentors for their time and patience, and the REU program coordina- tors (Maggie Harden, Olanda Davidson Bryant, and Kristin Wustholz) for their efforts to make the program successful. The extended abstracts included in the Review are presented in alphabetical order by the participant’s last name.

Sara Selfe, Ph.D., Editor TABLE OF CONTENTS

Synthesis of Dendrimer Building Blocks with Crosslinkable Moieties 7 KATHY L. BECKNER, University of Kentucky

Contact Angle Measurements on Diamond Surfaces for a Study on Hydrophobic Forces 9 ARIEL BEDFORD, Florida A&M University

Barium Titanate Doped Sol-Gel for Electro-Optic Devices 13 DENIZ CIVAY, University of Washington

Synthesis of Norbornene Monomer of PPV-C60 dyads (TDC-I-056) 17 TAÍNA D. CLEVELAND, Georgia Institute of Technology

Development of Efficient Two-Photon Radical Initiators and Low Shrinkage Materials 21 AMANDA COOPER, Georgia Institute of Technology

Synthesis of Nonlinear Optical-Active Materials 25 DAN DARANCIANG, Washington University in St. Louis

Behavioral Properties of Colloidal Crystals at and Around Thermally Induced Defect Sites 29 MALLORY DAVIDSON, University of Washington

Optical Properties of Metal Nanoparticle Composits 31 KRYSTLE L. DZIENIS, Pennsylvania State University

Electro-Optic Properties of Hybrid Sol-Gel Materials in Fabry-Perot Modulators 35 BRENDA EBY, University of Idaho

Quantized Hamilton Dynamics Applied to Condensed Phase Spin-Relaxation 39 STEPHEN T. EDWARDS, Harvey Mudd College

Investigating New Cladding and Core Materials for Hybrid Electro-optic Modulators 45 PARISSA FATHALIPOUR, University of Arizona

Synthesis of TPD-Based Compounds for Use in Modification of ITO Surfaces and Metal Nanoparticles 49 AARON D. FINKE, University of Arizona

Optimizing Hybrid Waveguides 53 ANDREW GARDNER, Highline Community College

Synthesis and Analysis of Thiol-Stabilized Nanoparticles 57 EDDIE HOWELL, Norfolk State University

Quinoxaline-Containing Polyfluorenes: Enhanced Blue Electroluminescence by Addition of a Hole Blocking Layer 61 KELLI A. IRVIN, Montana State University

Synthesis of Dendron-Functionalized Chromophores: An Approach to Supramolecular Assembly for Electro-optic Applications 67 ZERUBBA LEVI, Gonzaga University

4 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Building an Optical Oximeter to Measure the Oxygen Content of Blood Non-invasively 71 JOANN LIN, University of Washington

Moisture Uptake of Thin Polymer Films 75 EPHRIAM LUCAS, Georgia Institute of Technology

Toward Molecular Resolution c-AFM with Carbon Nanotube Tips: Development of Carbon Nanotube Growth Techniques 77 AUSTIN MCLEOD, Northern Arizona University

Synthesis and Characterization of Extended Squaraine Compounds 81 TEHETENA MESGANAW, Georgia Institute of Technology

Enhanced Heat Dissipation Substrates for Organic Semiconductor Devices 83 AARON MONTGOMERY, University of Virginia

1,1-Diphenyl-2,3,4,5-tetrakis(9,9-dimethylfluoren-2-yl)silole Properties in Organic Light-Emitting Diodes and Organic-Field Effect Transistors 85 SARAH MONTGOMERY, Purdue University

Effects of Surface Chemistry on Cadmium Selenide Nanocrystal Fluorescence 89 MARSHA S. NG, University of Hawaii

Synthesis of a Polyene EO Chromophore Using a Diels-Alder Reaction to Form a Side-Chain Structure 95 DENIS NOTHERN, Cornell University

Spectroscopic Investigations of Chromophores in Dyed Salt Crystals 99 STACY A. OLIPHANT, Edmonds Community College

Characterization of the Molecular Parameters Determining Charge-Transport in a Series of Substituted Oligoacenes 103 ROBERT SNOEBERGER, University of Washington

Optimization of Semiconductor Nanoparticle Synthesis and Integration into Sol-Gel Monoliths 107 CINDY TAYLOR, University of Arizona

Characterization of the Photodecomposition of the CF3-FTC Chromophore 111 JILLIAN THAYER, Olympic College

Electroluminescent Properties of Organic Light-Emitting Diodes (OLEDs) with 2,5-Bis(9,9-dimethylfluoren-2-yl)-1,1,3,4-tetraphenylsilole 115 EVANS THOMPSON, Georgia Institute of Technology

Determination of Molecular Orientation of Self-Organized Aggregates of New Liquid-Crystal Perylene Dye 119 NATALIE THOMPSON, Georgia Institute of Technology

Hydrogel Materials for Two-photon Microfabrication 123 MAYEN UDOETUK, Norfolk State University

The Design of a Fluid Delivery System for Micro-Core Optical Fiber 127 GREG WINCHELL, Everett Community College

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 5 6 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis of Dendrimer Building Blocks with Crosslinkable Moieties

Kathy L. Beckner Dominic V. McGrath and David Sisk University of Kentucky McGrath Lab, Department of Chemistry The University of Arizona

INTRODUCTION RESULTS AND DISCUSSION Dendrimers have been employed in various applications pri- In this project, several synthetic routes to prepare insulat- marily because of their ability to insulate materials and display ing and cross-linking dendrimers were explored. The synthe- functional groups.1 Dendritic encapsulation of chromophores has sis of two types of dendrons were attempted: 3,5-di-tert-butyl been shown to enhance optical and electronic properties, while dendrons and ester cinnamate dendrons. Two different methods peripheral functional groups can be used for further reactivity.2 were explored in the synthesis of 3,5-tert-butyl dendrons, while Potential applications for dendrimers include organic light emit- an effective synthetic route was established for ester cinnamate ting devices, light harvesting and potentially drug delivery or dendrons by employing a pentyl group. gene therapy.3 Dendritic cross-linking of peripheral functional The synthesis of the 3,5-di-tert-butyl dendrons was per- groups can enhance dendrimer-based insulation,4 create stable formed using previously published procedures (Scheme 1).7 The dendrimer films,5 or generate dendritic nanocapsules.6 In organic original synthesis produced the 3,5-di-tert-butyl 4–methoxyben- electronic devices, films are prepared by spin-casting, a process zyloxy terminated dendrons using a four step series of reactions that requires the polymers to be soluble. However, in order to that was performed without purification until the final product cast another layer over the previous one, the materials need to be which was simply recrystallized. When the sequence was carried insoluble after being deposited. Crosslinking has been shown to out with an allyl group however, it was not possible to isolate the be a potential solution to this problem.7 product using similar methods. When the product could not be The McGrath group has demonstrated that dendrimers with isolated by standard methods of purification, stepwise purifica- cinnamate moieties on their periphery can be photo-crosslinked tion was attempted, however this proved unsuccessful. to provide stable dendrimer films. Originally ethyl cinnamate esters were used to fabricate the photocrosslinkable dendrons. However the synthesis of these dendrons has proven difficult be- cause of their insolubility at lower generations. The goal of this research is to replace the ethyl ester with a pentyl group in or- der to improve solubility. Increasing the solubility of these ester cinnamate dendrons in this way should allow the preparation of Scheme 1. Attempted synthetic route to 3,5-di-tert-butyl building blocks containing allyl ether end groups. higher generation dendrons in greater quantities. Increasing the dendron’s solubility by altering the ester in this fashion is also Subsequently a different approach was taken to try and pro- important for the spin casting process used in device fabrication. duce 2 (Scheme 2). Instead of protecting the aldehyde, it was first Previously, an improved synthesis of 3,5-di-tert-butyl den- reduced selectively to obtain benzyl alcohol 3. drons has been reported starting from 3,5-di-tert-butyl-4-hy- Attempts to alkylate the X residue proved fruitless. Howev- droxytoluene (BHT).8 In addition to the tert-butyl groups, these er, this methodology produced similar results in that the resulting dendrons have a peripheral methoxy group on each ring. To in- product could not be isolated by standard purification methods. crease the versatility of these dendrons, experiments were per- formed to vary the length and functionality of these peripheral alkoxy chains. For example, replacing the methoxy with an al- lyloxy. This particular end-group would allow for more versatil- ity in the molecule, by making cross-linking a possibility. Den- drimers possessing this particular end-group has been shown to successfully cross-link in a ring closing metathesis reaction by Scheme 2. Alternate route used in attempt to produce 9 the Zimmerman group. 3,5-di-tert-butyl terminated dendrons

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 7 SYNTHESIS OF DENDRIMER BUILDING BLOCKS WITH CROSSLINKABLE MOIETIES

Crosslinkable cinnamate dendrons were prepared with a pen- varying functionality. A synthetic route was developed to pro- tyl ester similar to the ethyl ester previously reported (Scheme duce up to [G-2]-cinnamate dendrons with a pentyl group, which 3).10 was found to be an improvement over the previous synthesis of ethyl ester cinnamate dendrons. In the future film studies will be performed on both polymers produced from these dendrons as well as the dendrons themselves to determine the effects of crosslinking.

REFERENCES 1 Fréchet, J.M.J. J. Polym. Sci. Part A: Polym Chem 2003, 41, 3713-3725. 2 Grayson, S.M.; Frechet, J.M.J.; Chem. Rev. 101, 3819-3868. 3 Grayson, S.M.; Fréchet, J.M.J. Chem. Rev. 101, 3819-3868. 4 Guo, W.J.; Li, J.J.; Peng, X.; Wang, A. J. Am. Chem. Soc. 2003, 12, 3901-3909. Scheme 3. Synthetic route used to produce [G-1]-cinnamate dendrons. 5 D’Ambruoso, G.D. Ph.D. Thesis, University of Arizona, 2004. 6 Lencoff, N.G.; Spurlin, T.A.; Gewirth, A.A.; Zimmerman, S.C.; In the first step of the synthesis, which was taken from modi- Beil, J.B.; Elmer, S.L.; Vandeveer, H.G. J. Am. Chem. Soc. 2004, 11 fied published procedures, removal of the remaining pentanol 126, 11420-11421. used to produce 4 proved difficult and therefore was not an im- 7 Zimmerman provement over previous methodology. However, the solubility 8 McGrath, D.V.; Shanahan, C.S. J. Org. Chem. 2005, 70, 1054. in the consecutive steps of the synthesis was greatly increased as 9 Zimmerman, S.C.; Elmer, S.L. J. Org. Chem. 2004, 69, 7363- a result of incorporation of the pentyl group. Compound 6 was 7366. found to be much more soluble than its ethyl ester counterpart. 10 D’Ambruoso, G.D. Ph.D. Thesis, University of Arizona, 2004. While the Mitsunobu reaction used to produce compound 5 still 11 Cernerd gave low yields, it was an improvement over the previous synthe- sis of ethyl ester cinnamate dendrons. ACKNOWLEDGEMENTS Compound 6 was then used to produce the [G2]-cinnamate -Dr. Dominic McGrath, Professor of Chemistry, University of dendron (7). Arizona -McGrath Research Group -Funding provided by the Center on Materials and Devices for Information Technology Research (CMDITR), an NSF Science and Technology Center No. DMR 0120967

Scheme 4. Synthesis of [G-2]-cinnamate dendron.

CONCLUSIONS Kathy Beckner will graduate from the University of Kentucky in Fall 2005 with a B.S. in chemistry. She would like to thank Dr. In this project various synthetic routes were explored for gen- Dominic McGrath and the McGrath group for their support. erating both 3,5-di-tert-butyl terminal dendrons as well as cross- linkable cinnamate dendrons. It is possible that this reaction will work under less harsh conditions by using a weaker base. After an effective synthetic route is established for these dendrons, an entire class of BHT derived compounds may be synthesized with

8 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Contact Angle Measurements on Diamond Surfaces for a Study on Hydrophobic Forces Ariel Bedford, Florida A&M University Dr. Elisa Riedo, Lina Merchan, Dr. Robert Szoszkiewicz Riedo Lab, School Of Physics, Georgia Institute of Technology

Introduction character of the surfaces. When measuring the The hydrophobic force is the attractive advancing contact angle of a surface, the contact force between hydrophobic surfaces in water angle steadily increases and peaks before having a solutions. It is usually measured at the steady decreasing period. This maximum contact nanoscale, and its origin and nature constitutes a angle is called advancing contact angle. The peak major area of study presently. The entropy driven has to be seen in unison with a slight increase of force that causes oil to separate from water is an the baseline length of the water droplet, and then example of the hydrophobic effect. Hydrophobic a return to plateau-like recordings. On the other forces are crucial in many physical and biological hand when measuring receding contact angles, the phenomena such as protein folding. Currently, measurements were sensed to have slight research dealing with these forces on surfaces that decreases, a low point peak, and then steady are hydrophobic and stable, such as diamond, is increases, while having simultaneous baseline limited. The goal of this work was to study the length measurements decreasing with the peaks, hydrophobic behavior of diamond surfaces with and having plateau-like recordings follow as well. different roughness and grain size and to relate The first duty in the research process consisted of Contact Angle Measurementsthis behavioron Diamond to the range and magnitude Surfaces of the taking static contact angle measurements of each hydrophobic forces generated by these surfaces. In surface. The second step was to then take for a Study on Hydrophobicparticular we studied Forces the relationship between advancing and receding contact angle hydrophobic force and surface characteristics such measurements that were isolated into small as static and dynamic contact angle, contact angle intervals of time (usually 500 millisecond hysteresis, grain density and roughness. intervals over a 3 second period). The third task Ariel Bedford Dr. Elisa Riedo, Lina Merchanwas to calculate any possible contact angle Florida A&M University Theory and Dr. Robert Szoszkiewiczhysteresis on the surfaces, and the fourth assignment was to compare these results with Static contact angle is the angle made Riedo Lab, School Of Physics previously recorded grain density and roughness by a tangent line between the gas and liquid Georgia Institute of Technology data. phases of a water droplet on a surface. Decisive patterns of contact angle advancing and receding were used in order to view the hydrophobic INTRODUCTION on the surfaces, and the fourth assignment was to compare these results with previously recorded grain density and roughness The hydrophobic force is the attractive force between hy- Advancing/Receding Contact Angles vs. data. drophobic surfaces in water solutions. It is usually measured at Baseline Length vs. time Graphs Definitions the nanoscale, and its origin and nature constitutes a major area Contact angle grows steadily, peaks, then steadily declines of study presently. The entropy driven force that causes oil to separate from water is an example of the hydrophobic effect. Hy- Baseline length constant, increases Advancing/receding slightly, reverts to constant drophobic forces are crucial in many physical and biological phe- theory definitions diagram nomena such as protein folding. Currently, research dealing with Contact angle declines steadily, peaks, these forces on surfaces that are hydrophobic and stable, such as then steadily increases diamond, is limited. The goal of this work was to study the hy- drophobic behavior of diamond surfaces with different roughness Baseline length constant, decreases slightly, reverts to constant and grain size and to relate this behavior to the range and mag- Advancing/receding contact angles vs. baseline lengths vs. time graph definitions nitude of the hydrophobic forces generated by these surfaces. In particular we studied the relationship between hydrophobic force RESEARCH METHOD and surface characteristics such as static and dynamic contact angle, contact angle hysteresis, grain density and roughness. Materials Nine diamond surfaces were used in total throughout the THEORY work period. Five surfaces were nanocrystalline, had a thick- ness range of about 2-3μm, and had diamond grain in the scale Static contact angle is the angle made by a tangent line be- of 100nm. The nanocrystalline surfaces were grown with vary- tween the gas and liquid phases of a water droplet on a surface. ing concentrations of methane, which affected the grain size. Decisive patterns of contact angle advancing and receding were The other four surfaces were two polycrystalline disks, having a used in order to view the hydrophobic character of the surfaces. thickness of around 180μm and totaling 11mm in diameter. Both When measuring the advancing contact angle of a surface, the sides of the polycrystalline surfaces were used for research; one contact angle steadily increases and peaks before having a steady side was deemed smoother due to being polished mechanically decreasing period. This maximum contact angle is called advanc- before being sent for this particular research use. The nanocystal- ing contact angle. The peak has to be seen in unison with a slight line surfaces were signified from al25-al29, and the polycrystal- increase of the baseline length of the water droplet, and then a line surfaces sd1u, sd1d, sd2u, and sd2d, respectively. return to plateau-like recordings. On the other hand when mea- suring receding contact angles, the measurements were sensed to Procedure have slight decreases, a low point peak, and then steady increas- Contact angle measurements were taken using a CAM 100 es, while having simultaneous baseline length measurements contact angle machine with a 50mm USB camera. For static decreasing with the peaks, and having plateau-like recordings contact angle measurements, the metal platform of the machine follow as well. The first duty in the research process consisted was adjusted in a manner so that a single water droplet could be of taking static contact angle measurements of each surface. The lowered onto a sample surface. Producing a single water droplet second step was to then take advancing and receding contact an- for static contact angles and a continuous water flux for dynamic gle measurements that were isolated into small intervals of time contact angles were manipulated using a water syringe suspended (usually 500 millisecond intervals over a 3 second period). The above the platform. The syringe was attached to an apparatus on third task was to calculate any possible contact angle hysteresis the CAM that would allow for adjustment in height of the syringe

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 9 Analysis contact angles was around ± .394, and the Generally, advancing and receding contact angle average standard error for advancing baseline graphs followed the definition described lengths was around ± .002. The average range previously. The peaks for contact angles and between receding contact angle measurements for plateaus for baseline length were observed all diamond surfaces was accordingly. The average range of advancing between 22°-66°; the range of receding baseline contact angle measurements for all diamond lengths averaged between 4.4-4.28mm. The surfaces was between 70°-89°; the range of average standard error for receding contact angle Research Method was adjusted in a manner so that a single wateradvancing baseline lengths averaged between 3.8- measurements was around ± 2.95, and the Materials droplet could be lowered onto a sample surface.5.2mm. The average standard error for advancing average standard Producing a single water droplet for static errorcontact for receding contact angle measurements Nine diamond surfaces were used in total was around ± .14. The range for static contact throughout the work period. Five surfaces were angles and a continuous water flux for dynamicangle measurements for all diamond surfaces was contact angles were manipulated using a waterbetween 71°-90°. The range of contact angle nanocrystalline, had a thickness range of about 2- hysteresis for all diamond surfaces was quite The grain density and roughness were 3 m, and had diamond grain in the scale of syringe suspended above the platform. Thelarge, between 8°-74°. The range of grain density inadvertently omitted for all the polycrystalline 100nm. The nanocrystalline surfaces were grown syringe was attached to an apparatus on thefor CAM all diamond surfaces was between 7-15 m-2; samples except one, surface sd1u, which had a that would allow for adjustment in height theof rangethe for roughness was between 5.2-41nm. roughness of sample recording of 5.2 nm sans a with varying concentrations of methane,CONTACT which ANGLE MEASUREMENTS ON DIAMOND SURFACES FOR A STUDY ON HYDROPHOBICgrain density FORCESrecording. affected the grain size. The other four surfaces syringe by a winding, spindle-like metal rod. were twoby polycrystalline a winding, disks, spindle-like having a metal rod.Dynamic Dynamic contact contactangle measurements angle were thickness of around 180 m and totaling 11mm recorded by adjusting the metal platform so that in diameter.measurements Both sides of the were polycrystalline recorded by adjustingthe water thesyringe metal was nearly platform touching a sample surface. Therefore, the water flux could be better surfacesso were that used the for research;water syringeone side was was nearly touching a sample surface. deemed smoother due to being polished adjusted so that a droplet could grow or shrink for mechanicallyTherefore, before being the sentwater for fluxthis particular could be betteradvancing adjusted contact so thatangle a and droplet receding contact research use. The nanocystalline surfaces were angle measurements, respectively. After taking signifiedcould from al25-al29, grow or and shrink the polycrystalline for advancing contactan image angleof a single and water receding droplet, camera frames surfacescontact sd1u, sd1d, angle sd2u, andmeasurements, sd2d, respectively. respectively. and corresponding After taking data werean image exported from the CAM software into Microsoft Excel software to Procedureof a single water droplet, camera framesplot and graphs corresponding and chart pertinent data data. The same were Contact exported angle measurements from the were CAM taken software format into Microsoftwas followed Excelfor multiple soft images- of a using a CAM 100 contact angle machine with a water droplet growing or shrinking on a sample 50mm USBware camera. to plot For static graphs contact and angle chart pertinentsurface. data. The same format measurements,was followed the metal platform for multiple of the machine images of a water droplet growing or shrinking on a sample surface.

RESULTS/CONCLUSION Surface sd1u, as shown above, had the lowest recorded mea- surement for roughness; it also had the lowest recorded measurement out of all the surfaces for contact angle hysteresis (8°). Generally, it

Generally, it was observed that the surfaces with wasResults/Conclusion observed that the surfaces with highGenerally, hysteresis it was observed seemed that the surfaces to withhave Surface sd1u, as shown above, had the highhigh hysteresis hysteresis seemed seemed to tohave have high high roughness roughness high roughness Surface sd1u, asrecordings. shown above, had Surfaces the withrecordings. high Surfaces roughness with high roughness recordings lowestlowest recorded measurementmeasurement for for roughness; roughness; it it recordings. Surfaces with high roughness recordings also seemed to have high static contact alsoalso had the lowest recordedrecorded measurementmeasurement out out of of recordings also seemed to have high static contact also seemed to have high static contactangle angle measurements. measurements. allall thethe surfaces for contactcontact angleangle hysteresis hysteresis (8°). (8°). angle measurements. Water syringe placing water droplet on sample surface

Water syringe placing water droplet on sample surface Rouugghhnneessss ooff SSuurrfafacceess vvs.s. SStattaitcic CoCnotnatcatctAnAgnlegle Series1Series1 f f o o 86 )

ANALYSIS e ) l e s l s g 84 e g 84 e n al28 n e al27 al28

e al27 A r A r

g 82

g 82 t

Generally, advancing and receding contact angle graphs fol- e t e c c d d a a

( 80 t

( 80 t s n s n e o

lowed the definition described previously. The peaks for contact e o 78 c 78 C c C a

a al29al29

f sd1u f sd1u c r c i r 76 i 76 al25al25 t

angles and plateaus for baseline length were observed accordingly. u t u al26al26 a a S S t t 74

S 74 The average range of advancing contact angle measurements for S 72 all diamond surfaces was between 70°-89°; the range of advanc- 0 5 10 10 15 15 20 20 25 25 30 30 35 35 40 40 45 45 ing baseline lengths averaged between 3.8-5.2mm. The average RRoouugghhnneses s ofof SuSrufrafcaecse(sn(n standard error for advancing contact angles was around ± .394, and the average standard error for advancing baseline lengths Roughnneessss ooff SSuurrffaacceess vvss. . SStattaitcic CoCnotnatcatctAnAgnlegleo o was around ± .002. The average range between receding contact Series1Series1 f f 86

o 86 angle measurements for all diamond surfaces was between 22°- o ) e ) e l l s

s 84 g g e al27 al28 e al28

n al27 n e e

66°; the range of receding baseline lengths averaged between 4.4- A r A r 82

g 82 g t e t e c c d a d a 4.28mm. The average standard error for receding contact angle ( 80 t

( 80 t s n s n e o e o

c 78 C c 78 C a al29 measurements was around ± 2.95, and the average standard error a f sd1u al29 f c

r sd1u i c r i

t 76 al25 u

t 76 al25 u al26 a

S al26 a t S

for receding contact angle measurements was around ± .14. The t S S 74 range for static contact angle measurements for all diamond sur- 72 faces was between 71°-90°. The range of contact angle hysteresis 0 5 10 10 15 15 20 20 25 25 30 30 35 35 40 40 45 45 for all diamond surfaces was quite large, between 8°-74°. The RRoouugghhnneess s ofof SuSrufrafcaecse(sn(n range of grain density for all diamond surfaces was between 7- References -2 nd 15μm ; the range for roughness was between 5.2-41nm. [1][1] J. Israelachvili, IntermolecularIntermolecular & & Surfaces Surfaces Forces, Forces, Academic Academic Press. Press. 2 2Ed.,nd Ed., 2003 2003

The grain density and roughness were inadvertently omit- REFERENCES ted for all the polycrystalline samples except one, surface sd1u, [1] J. Israelachvili, Intermolecular & Surfaces Forces, Academic which had a roughness of sample recording of 5.2 nm sans a grain Press. 2nd Ed., 2003 density recording.

10 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

Greetings, I am Ariel bedford, chemical engineering student from Atlanta, Georgia. I attend Florida State University and hope to partake in more research in the nanotechnology sector.

Greetings, I am Ariel Bedford, chemical engineering student from Atlanta, Georgia. I attend Florida State University and hope to partake in more research in the nanotechnology sector.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 11 12 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Barium Titanate Doped Sol-Gel for Electro-Optic Devices

Deniz Civay Nasser Peyghambarian, Robert Norwood, University of Washington and Chris DeRose Peyghambarian Lab, Optical Science Center The University of Arizona

The goal of my research was to successfully dope BaTiO3 nanocrystals into a sol-gel ma- trix, creating an electro-optic (EO) composite. After poling I determined the EO properties of the composite material. The resulting material exhibits an EO coefficient of 4pm/V however poling conditions are not yet optimized. The advantage of using sol-gel is that the composite material will be photopatternable.

RESEARCH METHODS Another important material property is the dielectric con- stant. The dielectric constant limits operational bandwidth. Ide- I began my research by reviewing relevant literature. The ally the dielectric constant for optical and electrical waves should nanocrystals that have had some success in this field are PbTiO 3 be matched. The third important material property to consider and BaTiO . However PbTiO and BaTiO nanocrystals have not 3 3 3 is the electro-optic coefficient. The electro-optic coefficient is yet been doped into sol-gel. BaTiO has a high EO coefficient but 3 a measure of how much the index of refraction changes with an its high dielectric constant limits high speed and wide bandwidth applied electric field. operation. The low dielectric constant of sol-gel served to lower Before starting the experimental procedure I used the Max- the overall dielectric constant of the composite improving the well-Garnet Theory to model the scatting loss and composite di- bandwidth of the material. I purchased some BaTiO nanocrys- 3 electric constant. The key variables in the Maxwell-Garnet The- tals from Sigma-Aldrich and TPL, inc. in order to perform the ory are the nanocrystal radius, the wavelength of the incoming experiments. The kind of testing that I performed on the BaTiO 3 light, the dielectric constant of the nanocrystal and sol-gel, the nanocrystal composite films was to determine the EO coefficient refractive index of the nanocrystal and sol-gel, and the volume and loss. fraction. INTRODUCTION TO THEORY CALCULATIONS I am investigating electro-optic nanocomposites because sol- The following equations are based on the Maxwell-Garnet gel is photopatternable. Doping sol-gel with an EO nanocrys- Theory. tal results in a photopatternable EO material. Barium titanate nanocrystals are being used because of the high EO coefficient of 150pm/V. Barium titanate nanocrystals are also easily accessible Equation 1: π λ 4 2 2 2 π 2 and cheap. Csca = (8/3)*[(2* *nc*r)/( )] *{[( nc / nm) -1]/[( nc / nm) +2]} * *r Some important material properties to consider are the scat- tering loss, dielectric constant and electro-optic coefficient. The Where: scattering loss is a measure of how much energy is lost from a nm = index of refraction of the sol-gel = 1.5 beam of light passing through a material due to inhomogeneities. np = index of refraction of the particle = 2.3 See Figure 1 below for an illustration of this phenomenon. r = average particle radius = 20nm λ = wavelength of light = 1550nm

Csca = .389657701

Equation 2: Figure 1. An illustration of scattering loss. 3 αsca = (3*η*Csca)/(4*π*r )

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 13 BARIUM TITANATE DOPED SOL-GEL FOR ELECTRO-OPTIC DEVICES

Where: Based on Figure 1 it can be determined that the loss is direct- η = fill fraction of the spheres in the polymer host = 0-1 ly dependent on the fill fraction; the equation for n = .2 is exactly twice the equation for n = .1. A low fill fraction is ideal according Equation 3: to these graphs because it will minimize losses, however a low fill fraction will also reduce the electro-optic effect. A particle radius L3db = 1/ αsca of 20nm with a fill fraction of .1 would have a loss of 50 dB/cm. Where: The next relationship analyzed the effect of a small change in refractive index between the matrix and the particle. This was ac- L3db = propagation length for 1/e scattering loss

complished by graphing the loss versus particle size for an nm of 1.6, Equation 4: 1.7, 1.8, 1.9, and 2.0. This graph can be seen in Figure 2 below.

αdb = [-10 log (1/e)]/L3db

Where:

αdb = loss in db/cm

See Table 1 below for the αdb and L3db values corresponding to a fill fraction from 0-1 in increments of .02.

αdb @ 1550 n L 1550 (cm) db (db/cm) 0 0 0 0.02 0.42978 10.1051 0.04 0.21489 20.2102 0.06 0.14326 30.3152 0.08 0.10744 40.4203 Figure 2. Loss versus particle size for various sol-gel indices of refraction. 0.1 0.08596 50.5254 0.12 0.07163 60.6305 By re-graphing one loss value from each line in Figure 2 ver- 0.14 0.06140 70.7355 0.16 0.05372 80.8406 sus the corresponding sol-gel index it can be determined that the 0.18 0.04775 90.9457 relationship between the difference in refractive index and the loss 2 0.2 0.04298 101.0508 is governed by the equation y=172.23x -736.12x+790.82, where x

is the nm. See Figure 3 for an illustration of this equation. Table 1. αdb and Ldb values corresponding to a fill fraction from 0-.2.

Because the loss values calculated were too high to be use- ful, a graph of loss versus particle radius was created where three different fill fractions were used. See Figure 1 below.

Figure 3. Sol-gel index versus loss.

Based on Figure 3 the index of the matrix needs to be greater than or equal to 1.9 in order to minimize loss. The index of the sol-gel currently being used is 1.5. Figure 1. Loss versus particle radius for fill fractions of .2, .15, and .1.

14 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 CIVAY

Equation 5: k’ = vm*k’m*[2/3+k’p/(3*k’m)]+vp*k’p

vm*[2/3+k’p/(3*k’m)]+vd

Where: k’ = composite dielectric constant k’m = dielectric constant of the matrix = 5 k’p = dielectric constant of the particle = 1525 vm = volume fraction of the matrix v p = volume fraction of the particle

The data for Figure 5 was obtained by varying the volume Figure 4. The dielectric constant versus the volume fraction of barium titanate. fraction of the particle from 0 to 1 and setting vm+ vp = 1. Accord- • A ratio of 1:20, solute to solvent, was used for 1%. 1% BaTiO ing to Figure 4 the relative dielectric constant when the vp = .15 is 3 equal to 7.6. was dissolved overnight in .8mL of water. A ratio of 1:5.5 was

used for 5%: 5% BaTiO3 was dissolved overnight in 1.15mL SUMMARY of isopropyl-alcohol. A ratio of 3.5:1 was used for 10%: 10% BaTiO3 was stirred overnight in 1.5mL of water. • The α for a fill fraction of .15 with the current assumptions db Note: All solutions appeared equally cloudy and white after is 75.8 db/cm (useful for Fabry-Perot devices but not wave- stirring for one day. guides) • The 1% and 5% solutions then had a pre-mixed sol-gel added • If the sol-gel index was increased to 1.9 the loss would be re- to it. The sol-gel added to the 10% solution had not yet been duced to 14.2 db/cm formed (the ingredients to form the sol-gel were added directly • At a fill fraction of .15 the composite dielectric constant is 7.6 to the dissolved 10% barium titante). All samples were left to • 5 nm particle radius would result in 1.2 db/cm loss for a .15 fill stir for 2 days. fraction (useful for waveguide devices) • Each sol-gel solution was spun onto a half-etched ITO slide. • The final result is photopatternable • The samples were then baked at 350 for 1hr. Note: The 10% slide looked the worst, then the 1%. The 5% EXPERIMENTAL PROCEDURE made a decent quality slide. This difference is most likely due • The corresponding amount of grams for 1 wt%, 5 wt% and 10 to the alcohol solvent and the fact that the sol-gel was added wt% barium titanate were measured using the table below. after it had been mixed. • After making films containing 1%, 5% and 10% barium tita- Wt% BaTiO (g) 3 nate an easy way to determine if the calculated loss agrees with 1 0.0397 the actual loss is to test the samples with a spectrophotometer. 2 0.0803 By using a spectrophotometer the actual loss will be recorded 3 0.1217 4 0.1639 for various wavelengths. The data can then be compared to 5 0.2070 Figure 5. 6 0.2511 7 0.2961 8 0.3421 9 0.3890 10 0.4371 11 0.4862 12 0.5364 13 0.5878 14 0.6404 15 0.6942 16 0.7493 17 0.8057 18 0.8635 19 0.9227 20 0.9834 Table 2. The amount of grams of barium titanate needed Figure 5. Loss versus wavelength for three different fill fractions. for each corresponding weight percent.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 15 BARIUM TITANATE DOPED SOL-GEL FOR ELECTRO-OPTIC DEVICES

RESULTS • We now have two new solutions that have been spun and baked and are almost ready to be tested. One solution contains .21g The absorptivity versus the wavelength for 1 wt%, 5 wt% and of barium titanate and .4mL of HCl. The BaTiO was added to 10 wt% of barium titanate was determined using a spectropho- 3 the amount of HCl required for making the sol-gel. The second tometer. The theoretical and the experimental data do not appear solution contained .43g of barium titanate mixed into 1.5mL to have any correlation other than a general shape; the trendlines of HCl, and then added to the sol-gel. The first solution is the from Figure 5 do not fit Figure 6. only one that looks testable. • In order to prevent the gold from peeling I am soft baking the film and then corona poling it at 150 degrees Celsius for 2 hours with 4kV applied. The gold will be sputtered after pol- ing is complete.

ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

Figure 6. Absorptivity versus wavelength for three different fill fractions.

A possible explanation for the discrepancy between Figure 5 and Figure 6 is film quality. Because the spectrophotometer only analyzes a small portion of the sample and the 10 wt% was of poor quality, it is possible that the section analyzed did not con- Deniz Civay will be graduating from the University of Washington tain any of the BaTiO3/sol-gel. It is also possible that the barium in December with a BS in materials science engineering. titanate/water solution was not homogeneous. She then plans to pursue her PhD. Because the 5 wt% sample turned out the best I decided to do the simple reflection test on it to determine the electro-optic coefficient. There wais some electro-optic activity; however the numbers were very low. A common problem was the gold elec- trode peeling off of the sample. In the case where an R33 of 38.2 was achieved the gold was damaged. See Table 3 below.

BaTiO3 Solvent Voltage Held for Temperature R (g) (mL) (V) (min) (oC) 33 0.22 1.15 200.00 3.00 150.00 1.40 0.22 1.15 200.00 3.00 160.00 OL 0.22 1.15 200.00 4.00 150.00 4.00 0.22 1.15 150.00 4.00 160.00 2.00 0.22 1.15 100.00 5.00 150.00 38.2* 0.22 1.15 100.00 3.00 150.00 0.20

Table 3. 5 wt% BaTiO3 in IPA. *Irregular modulation curve.

WORK IN PROGRESS

• I have recalculated the amount of BaTiO3 according to volume percent rather than weight percent to see if that is a more ac- curate description of the fill fraction. • Rather than using IPA or water as the solvent I am now using a mild acid, .1N HCl, which the barium titanate is somewhat miscible in: However after one day in the refrigerator the solution separated.

16 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis of Norbornene Monomer of PPV-C60 Dyads (TDC-I-056)

Taína D. Cleveland Dr. Jian-Yang Cho (JD) and Dr. Seth Marder Georgia Institute of Technology School of Chemistry & Biochemistry Georgia Institute of Technology

INTRODUCTION Most PPVs have wide molecular weight distribution, mak- ing them less applicable to control morphology in the solid. A As the need for a renewable, clean energy source is expand- synthetic procedure for the preparation of PPV reported by ing, a promising technology based on organic or polymeric ma- Kretzschmann and Meier2 is adopted in this research plan because terials may offer a lightweight, flexible, cost-effective renewable the synthesized PPV shows low polydispersity index (PDI). solar energy solution.1 Conducting polymers like PPV (poly(1,4- phenylenevinylene)) functioning as donors and C60 derivatives as acceptors have been widely used for photovoltaic cell applica- RESEARCH METHODS tions. However, composite films made by mixing PPV and C60 The research plan will be approached through organic syn- usually leads to phase separation because C60 tends to form crys theses followed by characterization of organic products in each tals which makes the donor and acceptor molecules incompatible step. The first step of the synthesis follows a Williamson ether in composite films. It results in poor homogeneity and low opti- synthesis. The second step involves formylation of an aromatic cal quality of the films leading to inefficient intramolecular and compound. The third step is a condensation reaction to form an intermolecular energy or electron transfer in solar cells. imine. The fourth step is a poly-condensation reaction where po- tassium tert-butoxide is used. The fifth step is to regenerate the OBJECTIVE OF THE RESEARCH PLAN aldehyde through the hydrolysis of imine. The sixth step is the The objective of this investigation is to synthesize polymeric reduction of the aldehyde to a primary alcohol. The seventh step materials where donors and acceptors are covalently linked to combines PPV, C60 and a polymerizable norbornene moiety to polymer backbone, which presumably can avoid phase separation yield a norbornene-containing PPV-C60 dyads. The eighth step problems. In addition, these polymeric materials can be dissolved follows Williamson ether synthesis. A proposed synthetic scheme in common organic solvents, “wet” methods such as spin coating is shown in Figure 1. can be used for device fabrication.

Figure 1. Proposed synthetic scheme for preparation of ROMP monomers.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 17 SYNTHESIS OF NORBORNENE MONOMER OF PPV-C60 DYADS (TDC-I-056). SYNTHESIS (21.17 g, 54.4%). 2,5-bis-dodecyloxy-4-methyl-benzaldehyde.

1H NMR (300 MHz, CDCl3 δ): 10.39 (s, 1H), 7.20 (s, 1H), 6.77 Synthesis of 1,4-bis(dodecyloxy)-2-methyl-benzene (s, 1H), 4.01 (t, J = 6.3 Hz, 2H), 3.94 (t, J = 6.3 Hz, 2H), 2.25 (s, (TDC-I-001). 3H), 1.71-1.83 (m, 4H), 1.24-1.57 (m, 36H), 0.86 (t, J = 6.6 Hz, + 6H). HRMS-EI (m/z): M calcd for C32H56O3, 488.42295; found, 488.42373.

Synthesis of (E)-N-(2,5-bis(dodecyloxy)-4-methyl benzylidene)aniline (TDC-I-026). A mixture of 1-iodoundecane (101.0g, 0.34 mol), 2-meth- ylbenzene-1,4-diol (10.47 g, 0.08 mol), and potassium hydroxide grinded pellets (37.88 g, 0.67 mol) in 200 mL of DMSO was stirred at room temperature for 1 day. The reaction mixture was poured into H2O and extracted with methylene chloride. The or- ganic layer was washed with water ( 3 X 300 mL) and dried over A mixture of 2,5-bis-dodecyloxy-4-methyl-benzaldehyde magnesium sulfate. Excess of solvent was remove under reduce (TDC-I-020) (15.73 g, 0.03 mol), freshly distilled aniline ( 2.92 pressure. The crude product was recrystillized from hot ethanol mL, 0.032mol), 250 mL of ethanol, and 40 mL of benzene. A to give white solid. A second crop was further purified using Dean-Stark and condenser were used to remove water generated charcoal to eliminate undesired brown color (38.64 g, 99.48%). in the reaction and shift the reaction to completion. The reaction 1H NMR (300MHz, CDCl3 δ): 6.60-6.72 (m, 3H), 3.87 (t, J = mixture was refluxed at 95°C for 2 days. Excess solvent were 6.3 Hz, 2H), 3.86 (t, J = 6.6 Hz, 2H), 2.18 (s, 3H), 1.67-1.80 (m, remove under reduce pressure. The crude product was recrys- 4H), 1.20-1.46 (m, 39H), 0.86 (t, J = 6.8 Hz, 6H). tallized from hot hexanes and charcoal was added to discolor the solution. The crude product was off-white solid (16.08 g, Synthesis of 2,5-bis dodecyloxy-4-methyl-benzal- 88.6%). 1H NMR (300 MHz, C D , δ): 9.31 (s, 1H), 8.15 (s, 1H), dehyde (TDC-I-020). 6 6 7.34-7.35 (m, 2H), 7.15-7.21 (m, 2H), 6.97-7.01 (m, 1H), 6.68 (s, 1H), 3.78 (t, J = 6.6 Hz, 2H), 3.70 (t, J = 6.3 Hz, 2H), 2.35 (s, 3H), 1.57-1.66 (m, 4H), 1.20-1.40 (m, 36H), 0.92 (t, J = 6.6 Hz, 13 1 6H). C{ H} NMR (75 MHz, C6D6, δ): 156.24, 153.93, 152.22, 132.41, 129.41, 128.32, 127.99, 127.67, 125.63, 123.93, 121.42, 116.08, 109.27, 69.43, 68.38, 32.31, 30.10, 30.03, 29.99, 29.80, A Schlenk tube was charged with 1,4-bis-dodecyloxy-2- 29.77, 29.74, 29.66, 26.45, 23.09, 17.19, 14.35. methyl-benzene (TDC-I-001, JYC-VI-044, JYC-VI-0046) (36.66 g, 79.5 mmol). The Schlenk tube was evacuated and refilled with Synthesis of PPV (TDC-I-042). N2. Under a nitrogen counterflow, add 50 mL of dry CH2Cl2. The reaction mixture was allowed to cool down to 0°C in an ice bath. Titanium tetrachloride (17.4 mL, 0.15 mol) was added dropwise using a syringe. α,α- dichloromethyl methyl ether (98%) (7.1 mL, 0.027 mol) was added dropwise via a syringe. Hydrogen chloride gas was generated and neutralized with 10% solution of NaOH. The reaction was stirred at 0°C for an hour. The reaction A Schlenk tube was charged with (E)-N-(2,5-bis(dodecyloxy)- mixture was poured into ice and H2O. The aqueous layer was adjusted to pH~7. The organic layer was extracted with dichlo- 4-methylbenzylidene)aniline (TDC-I-026) (4.5 g, 7.9 mmol). The romethane and dried over magnesium sulfate. Excess of solvent Schlenk tube was evacuated and refilled with nitrogen. Under was removed under reduce pressure. The crude product was re- counterfolw of nitrogen 50 mL of dry DMF were added, and po- crystallized from hot hexanes to give light brown solid (33.83 g, tassium tert-butoxide (1.79 g, 16 mmol). The reaction mixture 87.0%). Proton NMR indicates the presence of isomer mixture of was refluxed at 90°C for 3 days. The reaction mixture was moni- 3,6-bis-dodecyloxy-2-methylbenzaldehyde and 2,5-bis-dodecy- tored through UV-VIS. The reaction mixture was treated with loxy-4-methyl-benzaldehyde in 9:1 ratio. The crude product was HCl to neutralize the potassium tert-butoxide. An orange solid purified by repeated column chromatography (three times) (silica was recovered and it was washed with water, ethanol and acetone. gel, hexanes:dichloromethane = 7:3) to give pale yellow solid The crude product was fluorescence. Proton NMR indicates that

18 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 CLEVELAND the reaction had not reached completion, because of the presence of imine in the region of 9.3 ppm. Excess of HCl was added to mixture of crude product and methanol, to complete the hydro- lysis of imine to an aldehyde. The polymeric product was again collected by vacuum filtration, washed with water, ethanol, and acetone, and dried under vacuum to give an orange solid (3.42 g). Based on proton NMR integration indicates n equals to 13 which is consistent to the reported value in literature (n = 13).

Synthesis of Norbornene monomer of PPV-C60 dyads (TDC-I-056).

A Schlenk round bottom flask (500 mL) was charged with PPV (0.71 g, 0.1 mmol), (6-bicyclo[2.2.1]hept-5-en-2-yl-hexylamino)- acetic acid (51 mg, 0.2 mmol), C60 (72.4 mg, 0.1 mmol), and dry toluene (100 mL). The reaction mixture was refluxed at 120 °C overnight. Excess of solvent was remove under reduced pressure. Methanol was added reaction mixture and a dark brown solid was collected by vacuum filtration. 1H NMR spectrum of the dark brown solid shows the disappearance of aldehyde resonance. The dark brown solid did not show fluorescence properties.

ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 19 20 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Development of Efficient Two-Photon Radical Initiators and Low Shrinkage Materials

Amanda Cooper Takashi Okada and Seth Marder Georgia Institute of Technology Marder Lab, School of Chemistry and Biochemistry Georgia Institute of Technology

INTRODUCTION EXPERIMENTAL FOR TWO The development of highly efficient two-photon induced PHOTON RADICAL INITIATOR radical initiators and the development of low shrinkage polymers 1) 2,5-bis(chloromethyl)-1,4-dimethoxybenzene for use in 3-D microfabrication are two areas of current interest 1, 4-dimethoxybenzene (27.6g, 0.20 mol), paroformalde- in the field of photonics. An efficient form of two-photon radical hyde (18.1g, 0.60 mol), acetic acid (330 mL) and HCl (600 mL) initiators that has been developed involves a dye unit that is con- were combined and stirred for five minutes. The reaction was nected by a bridge to an initiator (Structure A). We propose that then sonicated for one hour. The resulting solid was filtrated and this molecule can be made more efficient by introducing electron washed several times with water. The crude product was recrys- donating groups to the dye unit thereby making the unit more tallized with acetone and methylene chloride. The pure product electron rich (Structure B). By creating a more electron rich dye was dried to yield a white solid (13.5g, 28.7%). unit, the overall difference between the LUMO of the excited dye and the initiator is increased. An increase in the difference leads 2) tetraethyl-1,5-bismethoxyl-p-xylene phosphonate to a greater rate of electron transfer. 2,5-bis(chloromethyl)-1,4-dimethoxybenzene (13.5g, 0.057 Efficient low shrinkage materials for 3-D microfabrication mol) and triethylphosphite (115 mL, ) were combined and re- need to exhibit two characteristics: a minimal loss in volume dur- fluxed overnight. The reaction was cooled to room temperature. ing cross-linked polymerization and good solubility in organic An equal amount of hexane was added and the mixture was stirred solvents of the non-cross linked material. Currently in the area of for 25 minutes. The product was filtrated and washed several low shrinkage materials, units that polymerize by chain polymer- times with hexane. The product was dried and a white solid was ization have been found to lose around 1-5% volume mass in the collected (22.9g, 90.8%). polymerization reaction. This loss of volume mass, we believe, can be reduced to nearly 0% by using a ring opening reaction as 3) diethyl 4-{(E)-2-[4-(dibutylamino)phenyl]ethenyl} opposed to a chain linking reaction. This can be achieved by at- -2,5-dimethoxybenzylphosphonate taching an epoxide or oxetane group, which can then be initiated Tetraethyl-1,5-bismethoxyl-p-xylene phosphonate (15.6g, to cross-link (Structure C). To maximize the solubility in organic 0.035mol) and 4-N,N-dibutylaminobenzaldehyde (6.00g, solvents, it has been proposed to use single molecules that are 0.039mol, 91%pure) were added together in anhydrous THF capable of cross linkage as opposed to polymers. This minimizes (300mL). The reaction mixture was cooled to 0°C. After one the additional interactions at sites on the molecule that are not hour, tBuOK(1M, 23.4mL) was added. The reaction ran over- related to the cross linkage. night. The reaction was quenched with water (121mL) and the

Structure A Structure B Structure C

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 21 DEVELOPMENT OF EFFICIENT TWO-PHOTON RADICAL INITIATORS AND LOW SHRINKAGE MATERIALS product was extracted three times with EtOAc and then washed EXPERIMENTAL FOR LOW three times with water. The organic layer was dried. The solid SHRINKAGE OLIGOMER product was purified by column chromatography using EtOAc as the eluent. The purified product was dried and a bright yellow 1) p-t-butyl-calix[8]arene Paraformaldehyde (17.5g, 0.55mol), p-t-butylphenol (50.0g, solid was collected (7.15g, 50.0%). 0.33mol), and NaOH (10M, 1mL) were added together in xylene (300mL). The reaction was heated to reflux. The reaction was 4) 4-piperazinobenzaldehyde left overnight and refrigerated for several hours, but no product Piperazine (20.8g, 0.242mol) was added to a solution of formed. K2CO3 (11,3g, 0.082mol) in 30mL DMSO. The reaction set to reflux under nitrogen. Once the reflux temperature was reached, 4-fluorobenzaldehyde (8.5mL) in DMSO (10mL) was added. 2) (1-bromo)butyl-(2-[3-oxetane])propyl ether 1,4 dibromobutane (78.7mL, 0.58mol), 3-methyl-3-oxetane The reaction was refluxed overnight. Water was added (1150mL) methanol (20.0g, 0.20mol), NaOH (23.5g, 0.60mol), and TBAB and the solid product was filtered (9.02g, 58.9%). (1.26g, 0.020mol) were added together in hexane (200mL) and water (100mL). The reaction was heated to reflux and left over- 5) 4-N-t-butoxycarbonyl-N’-(p-formylphenyl)piperazine night. The product was extracted with hexane and washed with 4-piperazinobenzaldehyde (9.02g, 0.047mol) was dissolved water. The organic layer was dried, leaving a pale brown liquid. in anhydrous DCM (112mL) under nitrogen. The solution was The liquid was distilled using a full vacuum twice, but the prod- cooled to 0°C. After 20 minutes, triethylamine (6.6mL, 0.047mol) uct could not be purified. and di-t-butyldicarbonate (10.8mL, 0.047mol) was added. The reaction was allowed to warm to room temperature and left over- night. The reaction mixture was washed three times with water 3) octakis(t-butyl)-octakis[2-(3-oxetane)propyloxy] and dried. No product was obtained. butoxycalix[8]arene P-t-butyl-calix[8]arene (0.500g, 0.00061mol), K2CO3 (1.34g, .0098mol), and KI (0.563g, .0033mol) were added togeth- 6) N-(4-{(E)-2-{4-[4-(1-t-butoxycarbonyl)piperazin- er in anhydrous DMF (15mL). The reaction was heated to 100°C. 1-yl]phenyl}ethenyl}-({[2,5-dimethoxyphenyl]etheny (1-bromo)butyl-(2-[3-oxetane])propyl ether (2.60g, 0.0098mol) l}phenyl)-N,N-dibutylamine in anhydrous DMF (10mL) was added dropwise. The reaction Diethyl 4-{(E)-2-[4-(dibutylamino)phenyl]ethenyl}-2,5-di was left overnight. No product formed. methoxybenzylphosphonate (7.15g, 0.014mol) and 4-N-t-bu- toxycarbonyl-N’-(p-formylphenyl)piperazine (4.01g, 0.014mol) CONCLUSIONS AND OUTLOOK were added together in anhydrous THF (160mL). The mixture was cooled to 0°C. After 20 minutes, tBuoK (1M, 14.5mL) The synthesis of the donor portion of the two photon radical was added. The reaction was run overnight. The reaction was initiator is nearly complete. The final crude product will be puri- quenched with water (130mL). The product was then extracted fied through recrystallization and the yield will be determined. with EtOAc and washed several times with water. The organic The pure product will then attached to an acceptor molecule con- layer was dried. A column was run to purify the product using sisting of a fluoro-substituted triphenyl sulfur compound. The hexane:EtOAc (5:1) as an eluent. The purified product was dried resulting molecule will then be tested for its ability as a two pho- and a bright yellow solid was collected (6.07g, 66.4%) ton radical initiator using UV-vis spectrometry. We believe that the molecule will absorb at a high peak wavelength, which will 7) 4-N,N-(2,5-dimethoxy-4-{(E)-2-[4-piperazin-1- indicate a large amount of energy dissipated upon irradiation. ylphenyl]ethenyl}ethenyl)aniline The synthesis of the oligomer did not present any yield in N-(4-{(E)-2-{4-[4-(1-t-butoxycarbonyl)piperazin-1- any of the steps. For the first synthesis, a possible explanation yl]phenyl}ethenyl}-({[2,5-dimethoxyphenyl]ethenyl}phenyl)- is that the reaction is an equilibrium reaction with water as a by- N,N-dibutylamine (6.07g, 0.0093mol) was dissolved in THF product. In order to drive the reaction to completion, water must (200mL). HCl (2M, 60mL) was then added and the reaction was be removed. This was not done the first time, so the synthesis heated to reflux. After four hours, the reaction was allowed to cool will be repeated using a Dean-Stark trap to remove water from to room temperature and left overnight. NaOH (2M, 65.6mL) was the reaction mixture. The purification of the second synthesis then added. The product was extracted with ether:DCM (8:2) and will be reattempted, and the temperature during the vacuum dis- then dried. Product was found to contain very few impurities, but tillation will be monitored more closely. After the first two prod- will be further purified before yield is determined. ucts have been obtained, the third experiment will be reattempted

22 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 COOPER using different reaction conditions, specifically a different base. Once the final oligomer has been synthesized, it will be tested for volume shrinkage during crosslinking. The irradiated film will then be treated with organic solvent and the solubility of the oligomer will be tested. Future research will focus on the synthesis of other potential efficient two photon radical initiators and low shrinkage materi- als. The results of the efficiencies of the products will be tested and compared to one another. Once we obtain an efficient radical initiator and low volume shrinkage material, the products will be synthesized on a large scale and distributed to various research groups to be used for experiments in other fields of photonics.

ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 23 24 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis of Nonlinear Optical-Active Materials Dan Daranciang, Washington University in St. Louis Donald Responte and Werner Kaminsky SynthesisDepartment of of Nonlinear Chemistry, University Optical-Active of Washington Materials Introduction The key characteristic of all NLO active materials Since the first observationDan Daranciang of second harmonic is seen Donald in their Responte packed and lattice Werner geometry. Kaminsky In order to Washington University in St. Louis Department nof Chemistry generation (SHG) in 1961 by Franken, multiple exhibit second-, Universitythird-, or of Washingtonth-order optical effects, the nonlinear optical materials unit cell of the material must be non- have been characterized and centrosymmetric. Otherwise, upon excitation of the studied. Franken irradiatedINTRODUCTION material, dipole-dipoleMETHODS vibrations occur in equal quartz with 694 Since nm the light, first observation and of second harmonic generation magnitude The key but characteristic opposite of direction, all NLO active and materialstherefore is cancelseen to his surprise,(SHG) in 1961 he by detected Franken, multiple nonlinear optical materials eachin their other packed out. lattice Another geometry. consideration In order to exhibit in our second-, design emitted photonshave been at characterized 347 nm, and studied. Franken irradiated quartz wasthird-, how or nth-order to ensure optical that effects, our the products, unit cell inof the addition material to half the originalwith 694 nm wavelength light, and to his surprise, he detected emitted pho- beingmust be NLO non-centrosymmetric active, could. Otherwise, potentially upon haveexcitation photonics of the tons at 347 nm, half the original wavelength and therefore twice material, dipole-dipole vibrations occur in equal magnitude but and therefore twice the Figure 1. Schematic of applications. We decided to use aromatic functional the energy of the incoming photons. This singular observation, the opposite direction, and therefore cancel each other out. Another energy ofresult theof a second incoming order phenomenon SHG. Twoinvolving photons the absorption of of groupsconsideration in in our our design products. was how Aromatic to ensure that rings our products, contain photons. two This photons, spawned singular a new wavelength classification of materials:(red) are nonlin- regionsin addition of tohigh being NLO-electron active, density,could potentially which have would photon give- observation,ear optical the result(NLO) materials. of a NLOabsorbed materials to promise reach ato have great ourics applications.compounds We the decided ability to use to aromatic readily functional transfer groupscharge. in Synthesis of Nonlinear Optical-Active Materials utility in developing tomorrow’s"virtual" technology. excited Taken state. together, A the our products. Aromatic rings contain regions of high π-electron second order phenomenonDan Daranciang, Washington University in St. Louis With this in mind, we chose to study the reaction involvingsheer the quantity absorption and variety of of DonaldNLOsingle effectsResponte photon that and have Werner at been Kaminsky/2 observed ofdensity, aryl which isothiocyanates would give our compounds with (S)-2-butanol. the ability to readily (S)-2- allow us to harness the propertiesDepartment of Chemistry,light in exciting University and of unexpect Washington- transfer charge. two photons, spawned a new (blue) is emitted butanol is a relatively low molecular weight chiral Introductioned ways. Photonics-based technology sees use todayThe in everything key characteristic of all With NLO activethis in materials mind, we chose to study the reaction of aryl iso- classification Sincefrom the pulse of first materials:measurement observation of nonlineardevices second to harmonic fiber optical optics. is (NLO) seen in their packed ligand latticethiocyanates geometry. with with In only order (S )-2-butanol. toone possible (S)-2-butanol nucleophilic is a relatively reaction, low generation (SHG) in 1961 by Franken, multiple exhibit second-, third-, or nmolecularth-order optical weight effects, chiral the ligand with only one possible nucleophilic materials.nonlinear NLO optical materials materials promise to haveunit great cell of the and material therefore must be itsnon- product with a substituted aryl utility have in developing been characterized tomorrow’s and technology. centrosymmetric Taken . Otherwise,isothiocyanatereaction, upon excitationand therefore is of theeasily its product predictable. with a substituted Since arylit is isothio chiral,- studied. Franken irradiated material, dipole-dipole vibrationscyanate is occur easily in predictable. equal Since it is chiral, the product would together,quartz the with sheer 694 nm quantity light, and and variety of NLO magnitudeeffects but opposite thedirection, product and therefore would cancel have a chiral carbon center, and we to his surprise, he detected each other out. Another considerationhave a chiral in carbon our design center, and we were therefore assured that the that haveemitted been photons observed at 347 nm, allow us to harnesswas how the to ensure thatwere ourunit products, celltherefore of the in additionproduct assured to could that not havethe aunit center cell of symmetry.of the product propertieshalf the of original light wavelength in exciting and unexpectedbeing ways. NLO active, couldcould potentially not havehave photonics a center of symmetry. and therefore twice the Figure 1. Schematic of applications. We decided to use aromatic functional Photonics-basedenergyFigure of 1 the. Schematic technology incoming of SHG. SHG. Two Two seesphotons photons of use wavelength of today λgroups (red) are in in absorbed our products. Aromatic rings containN S photons. to reach This a "virtual" singular excited wavelength state. A single (red) photon are at λ/2 (blue)regions is emitted.of high -electron density, which would giveC OH � everythingobservation, from the pulse result of measurement a absorbed to reach devices a toour fibercompounds the ability to readily transfer charge. + second order phenomenon "virtual" excited state. A With this in mind, we chose to study the reaction optics. S involving While the absorption many such of materialssingle photon have at /2been discoveredof aryl (two isothiocyanates ex- withR (S)-2-butanol. (S)-2- two photons, spawned a new (blue) is emitted butanol is a relatively low molecular weight chiral While amples many are such KTiO materials4 and KDP, havepotassium been dihydrogen discovered phosphate), classification of materials: nonlinear optical (NLO) ligand with only one possiblesubstituted nucleophilic phenyl reaction, isothiocyanate (S)-2-butanol (two materials. examplesa great NLO deal are materialsof work KTiO promiseremains.4 and to For have KDP, one, great though potassium theand “classic” therefore itsex - product with a substituted aryl dihydrogenutilityamples in phosphate), developing of NLO tomorrow’s materials a great technology.have deal certainly of Taken work been provenremains.isothiocyanate SHG active, is easily predictable. Since it is chiral, together,experimentation the sheer quantity we and have variety performed of NLO effects shows that otherthe product compounds would have a chiral carbon center, and we For one,that have though been observed the allow“classic” us to harness examples the ofwere NLO therefore assured that the unit cell of the product H propertiescan have of light even in larger exciting SHG and efficiencies. unexpected ways. In addition,could the not current have a ex center- of symmetry. NS materialsPhotonics-based have certainly technology been sees use proven today in SHG active, N S ample of chromophores has recently attracted much experimentalC OH everything from pulse measurement devices to fiber � experimentationinterest, but we these have compounds performed are difficult shows to that synthesize, other involve + optics. S R O compoundsWhilelengthy can many reaction have such materialseven sequences, larger have beenand SHG discoveredafford efficiencies. increasingly smaller In per- R (two examples are KTiO substituted phenyl isothiocyanate (S)-2-butanol S cent yields at each step.4 and KDP, potassium addition,dihydrogen the phosphate), current a example great deal of work of chromremains. ophores has For one, We though proposed the “classic” this summer examples to of synthesize NLO and characterize H recently attracted much experimental interest, but NS materialsby X-ray have certainly diffraction been and proven SHG SHG emission active, experiments a certain phenyl-substituted i-butyl-N-phenyl thiocarbamate these compoundsexperimentation we are have difficult performed shows to synthesize, that other involve class of materials, i-butyl-N-phenyl thiocarbamates. Through this O compounds can have even larger SHG efficiencies. In R S lengthyaddition, reactionwork, the we current hypothesized sequences, example of that chrom and weophores could afford lay has increasinglyout a basis for the sim- FigureFigure 2. 2. The The reaction reaction of aryl of isothiocyanates aryl isothiocyanates with (S)-2-butanol. with When R(S )-2- recently attracted much experimental interest, but = EWG, the isothiocyanate undergoes electrophilic attack at the cumulene ple, rational synthesis and design of NLO active materials,phenyl-substituted which i-butyl-N-phenyl thiocarbamate smallerthese percent compounds yields are difficult at each to synthesize, step. involve butanol.carbon, When affording Rthe =phenyl-substituted EWG, the i-butyl-isothiocyanateN-phenyl thiocarbamate. undergoes demonstrated significant activity compared to the compounds We lengthy proposed reaction sequences, this summer and afford increasingly to synthesize Figure and 2. The reaction of electrophilicaryl isothiocyanatesThe geometry attack with at ( Sthe)-2- atindicated the cumulenecarbon on the butanolcarbon, is preserved. affording the smallercurrently percent yields in wide at each use. step. butanol. When R = EWG, the isothiocyanate undergoes characterizeWe proposed by X-ray this summer diffraction to synthesize and SHG and emissionelectrophilic attack at thephenyl-substituted cumulene carbon, affording thei-butyl-N-phenyl thiocarbamate. The characterize by X-ray diffraction and SHG emission phenyl-substituted i-butyl-N-phenyl thiocarbamate. The geometryCMDITR Review at the of Undergraduate indicated Research carbon Vol. 2 onNo. 1 theSummer butanol 2005 25 is experimentsexperiments a a certain certain class class of materials, of materials, i-butyl-N- i-butyl-geometryN- at the indicated carbon on the butanol is phenylphenyl thiocarbamates. thiocarbamates. Through Through this work, this we work,preserved. we preserved. hypothesized that we could lay out a basis for the Aryl isothiocyanates were selected because of their hypothesizedsimple, rational that synthesis we could and design lay of NLOout active a basis formolecular the features. The aromaticAryl ring isothiocyanates provides a high were selected because of their materials, which demonstrated significant activity number of mobile electrons, which was key for simple,compared rational to the compounds synthesis currently and indesign wide use. of NLO potential active nonlinear opticalmolecular applications. features. Unlike The aromatic ring provides a high materials, which demonstrated significant activitythiocyanates (the –S=C=Nnumber linkage), ofiso thiocyanates mobile electrons, which was key for Methods feature a –R=S linkage, which builds out selected compared to the compounds currently in wide use. potential nonlinear optical applications. Unlike thiocyanates (the –S=C=N linkage), isothiocyanates Methods feature a –R=S linkage, which builds out selected hydrogen bonds, structures molecular packing, and also easier to obtain a powder than to grow crystals of enhances crystal growth. In addition, the sulfur atom acceptable size and purity for other studies. The scan scatters well, allowing absolute X-ray structure area was (20 µm)2 with a step size of 1 µm, and we determinations (XRD). The crystalline nature of these measured the average SHG counts (emitted photons of products made for easy characterization by XRD. half the wavelength of the incoming photons) over To study the feasibility of usingSYNTHESIS this OF NONLINEAR reaction OPTICAL-ACTIVE to each MATERIALS area, omitting areas that did not achieve obtain Aryl crystals, isothiocyanates racemic were 2-butanol selected because was of used their molecuto test- the In ordersignificant to “get an idea” of SHG our compounds’ activity. NLO Theseactivity, we averages were reactionlar features. with The aaromatic variety ring of provides aryl isothiocyanates,a high number of mobile many decided to subjectcompared powder forms to of giveour compounds rough to an relativeSHG activity ofelectrons, which had which an was electron-withdrawing key for potential nonlinear group optical (EWG) appli- at activity test determinations.using a SHG microscope (Figure 3), with two stan- thehydrogencations. para Unlike bonds, position:thiocyanates structures (the phenyl –S=C=N molecular isothiocyanate,linkage), packing,isothiocya - and 4- dards: KAP alsoand KDP. easier to obtain a powder than to grow crystals of fluorophenylenhancesnates feature crystal a –R=S growth.linkage, isothiocyanate, which In addition, builds out selected the 4-chlorophenyl sulfur hydro atom- Using aacceptableResults powder form was size a stepand we purity took to forminimize other the studies. ori- The scan 2 isothiocyanate,scattersgen bonds, well, structures allowing 4-bromophenyl molecular packing, absolute and isothiocyanate, enhances X-ray crystal structure 4- entation dependenceareaWe was of succeeded SHG (20 in µm) crystals. in with It synthesizingwas a also step easier size to obtain and of 1 characterizing µm, and we determinationsgrowth. In addition, (XRD). the sulfur The atom crystallinescatters well, allowingnature absoof these- a powder thanmeasured to grow crystals the average of acceptable SHG size counts and purity (emitted for photons of nitrophenyllute X-ray structure isothiocyanate, determinations (XRD). and The crystalline 3-chlorophenyl nature other studies.two The scan target area was compounds:(20 µm)2 with a stepi-butyl- size of 1 Nµm,-(4-nitrophenyl) products made for easy characterization by XRD. half the wavelength of the incoming photons) over isothiocyanate.of these products made We for then easy characterization performed a by second XRD. reaction and we measuredthiocarbamate the average SHG (Fig. counts 4) (emitted and i-butyl- photons ofN -(4-chlorophenyl)half with To To the studystudy costlier the the feasibility feasibility (S)-2-butanol of using of this using reaction to obtain this to obtain reaction the crys - non- to the wavelengtheachthiocarbamate of the area, incoming omitting photons) (Fig. over 5). areas each Crystalline area, that omitting did products not achieve were centrosymmetricobtaintals, racemic crystals, 2-butanol racemic structure. was used 2-butanol to test the reaction was used with ato variety test the areas that didsignificantobtained not achieve forsignificant SHG the SHG other activity. activity. reactions These These averages we averages attempted were (in reactionofThe aryl isothiocyanates, reaction with a variety involved many of of which mixingaryl had isothiocyanates, an orelectron-withdraw dissolving many - the were comparedcomparedparticular, to give rough torelative the give activity determinations. 4- rough relative activity isothiocyanateofing which group (EWG)had an in atelectron-withdrawing 2-butanol the para position: with phenylthe aid group isothiocyanate, of vortexing. (EWG) at determinations.bromophenyl Thethe4-fluorophenyl reactionpara position:isothiocyanate, mixture was phenyl 4-chlorophenyl heated isothiocyanate, in isothiocyanate,a sand bath 4-at isothiocyanate,RESULTS 4- 4-bromophenyl isothiocyanate, 4-nitrophenyl isothiocyanate, aboutfluorophenyl 110° C for 3 isothiocyanate, h, then allowed to 4-chlorophenyl cool to room We succeededResultsfluorophenyl in synthesizing and characterizing two target isothiocyanate,and 3-chlorophenyl isothiocyanate. 4-bromophenyl We then isothiocyanate,performed a second 4-compounds: i-butyl-We N succeeded-(4-nitrophenyl) in synthesizing thiocarbamate (Fig. and 4) characterizing temperature. CrystalsS were grown by slow reaction with the costlier ( )-2-butanol to obtain the non-centro- and i-butyl-N-(4-chlorophenyl) thiocarbamate (Fig. 5). Crystal- evaporationnitrophenylsymmetric structure. of the isothiocyanate, solvent. and 3-chlorophenyl two target compounds: i-butyl-N-(4-nitrophenyl) hydrogen bonds, structuresisothiocyanate. molecular packing, We then and performed also a second easier reactionto obtain line a productspowderthiocarbamate were than obtained to grow for (Fig. the crystals other 4) reactionsand of i-butyl- we attemptedN-(4-chlorophenyl) In The order reaction to “get involved an mixing idea” or of dissolving our com thepounds’ isothiocya NLO- (in particular, the 4-bromophenyl isothiocyanate, 4-fluorophenyl enhances crystal growth. Inactivity,with nate addition, in the2-butanol we costlierthe decided with sulfur the ( S toaid)-2-butanolatom of subject vortexing. powder toTheacceptable obtainreaction forms mixture the size of non- ourand purity forthiocarbamate other studies. (Fig. The 5). scan Crystalline products were 2isothiocyanate, and 3-chlorophenyl isothiocyanate with (S)-2-bu- scatters well, allowing compounds absolutecentrosymmetricwas heated in X-ray a to sand an bathstructure. structure SHG at about activity 110° C for test3area h, then using was allowed (20 a to SHG µm) tanol), with but a unfortunately, stepobtained size offor due 1 to the µm, time other andconstraints, we reactions these were we not attempted (in determinations (XRD). Themicroscope crystallinecoolThe to room reaction temperature.nature(Figure involved of 3), Crystalsthese with mixingwere two grown standards:measured or by dissolvingslow evapora KAPthe average- and the characterized. SHG countsparticular, The quality (emitted of the photons the XRD structures of 4- was excellent tion of the solvent. isothiocyanate, and products made for easy characterizationKDP.isothiocyanate by XRD.in 2-butanol with thehalf aid theof vortexing. wavelength in most of cases. thebromophenyl incoming As expected, photons)the products overwere chiral, which the To study the feasibility The of using reaction this mixture reaction was to heated each in a sand area, bath omitting at space groups areasisothiocyanate, confirmed. that did not achieve 4- about 110°Scan C for 3 h, then allowed to cool to room fluorophenyl Figure 5. i-butyl-N-(4- obtain crystals, racemic 2-butanol wasstage used to test the significant SHG activity. These averages were chlorophenyl) temperature. Crystals were grown by slow reaction with a variety of aryl isothiocyanates, many compared to give roughFigure relative 4. i-butyl- N activity-(4- thiocarbamate XRD of which had an electron-withdrawingevaporation100x 1.3 group ofN.A. the (EWG) solvent. at determinations. nitrophenyl) thiocarbamate structure. Space group: P the para position: phenylIn orderObjective isothiocyanate, to “get an idea” 4- of our0º comor 90ºpounds’ NLO XRD structure. Space group: 21. P 21. fluorophenyl isothiocyanate,activity, we 4-chlorophenyl decided to subject powderResults forms of our compounds to an SHG activity test using a SHG 3-chlorophenyl isothiocyanate, 4-bromophenyl isothiocyanate, 4- We succeeded in synthesizing and characterizing microscope (Figure 3), with two standards: KAP and isothiocyanate with (S)-2-butanol), but unfortunately, nitrophenyl isothiocyanate, Dichroic and 3-chlorophenyl two800 target nm compounds: i-butyl-N-(4-nitrophenyl) isothiocyanate, and KDP. due to time constraints, these were not isothiocyanate. We then performedmirror a second reaction thiocarbamateexcitation (Fig. 4) and i-butyl-N-(4-chlorophenyl) with the costlier (S)-2-butanol to Scan obtain the non- thiocarbamate (Fig. 5). Crystalline products were Figure 5. i-butyl-N-(4- Figure 4. i-butyl-N-(4-nitrophenyl) thiocarbamate chlorophenyl) Emissionstage XRD structure. Space group: P 21. centrosymmetric structure. obtained for the other reactionscharacterized.Figure we4. i attempted The Nquality (in of the XRD structures was filters -butyl- -(4- thiocarbamate XRD The reaction involved mixing100x or 1.3 dissolving N.A. the particular, the 4- excellentnitrophenyl) in most thiocarbamate cases. As structure. expected, Space the productsgroup: P isothiocyanate in 2-butanol with theObjective aid of vortexing. 0ºbromophenyl or 90º wereXRD chiral, structure. which Space the group: space groups21. confirmed. The reaction mixture was heated in a sand bath at isothiocyanate, 4- PWhen 21. we attempted to form i-butyl-N-phenyl 3-chlorophenyl about 110° C for 3 h, then allowed to cool to room fluorophenyl thiocarbamate, we were surprised to obtain a isothiocyanate with (S)-2-butanol), but unfortunately, temperature. Crystals were grown by slow previously unseen product Dichroic Singl 800 nm due to time constraints, these were not excitation (according to searches of evaporation of the solvent. mirror e- the Cambridge Structural In order to “get an idea” of our compounds’ NLOPhoto Database and SciFinder) activity, we decided to subject Emission powder forms of our Us characterized. The quality of the XRD structures was containing seven phenyl compounds to an SHG activity filters test using a SHG excellent in most cases. As expected, the products FigureFigure 3. 3.SHG SHG microscope microscope diagram. diagram.Light of 800 nm Light wavelength of is 800 nm rings (Fig. 7). microscope (Figure 3), with two standards: KAP and were chiral, which the space groups confirmed. wavelengthpulsed onto is the pulsed sample (P onto= 1 or 2 the mW) sampleand the number (P =of 1emitted or 2 400 mW) and FigureIt 5 is. iisothiocyanate,-butyl- heavilyN-(4-chlorophenyl) conjugated thiocarbamate and KDP. nm photons is recorded by the single-photon photomultiplier tube. WhenXRD structure. we Space attempted group: P 21. to form i-butyl-N-phenyl the number of emitted 400 nm photons is recorded by the and also chiral, and therefore predicted to have Scan single-photon photomultiplier tube. thiocarbamate,Figure 5. i-butyl- we N-(4- were surprised to obtain a 26 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 nonlinear optical activity. However, its crystals were stage previouslychlorophenyl) unseen product ing a powder form was aSingl step we took to minimize Figure 4. i-butyl-N-(4- (accordingnot thiocarbamate large enough to searches XRD to perform of SHG studies on. In the orientation dependence ofe- SHG in crystals. It was 100x 1.3 N.A. nitrophenyl) thiocarbamatetheaddition, structure. Cambridge attemptsSpace Structural group: toP Photo Objective 0º or 90º XRD structure. Space group:Database 21. and SciFinder) Figure 7. P 21. Us Ph N C H •PhNCS•C H containing 3-chlorophenyl seven phenyl 7 7 3 2 6 6 Figure 3. SHG microscope diagram. Light of 800 nm XRD structure. Space isothiocyanate with (S)-2-butanol),rings (Fig. but 7). unfortunately, group: P 21/c. wavelength is pulsed onto the sample (P = 1 or 2 mW) and It is heavily conjugated Dichroic 800 nm due to time constraints, these were not mirror the number excitationof emitted 400 nm photons is recorded by the and also chiral, and therefore predicted to have single-photon photomultiplier tube. nonlinear optical activity. However, its crystals were ing a powder form was a step we took to minimize not large enough to perform SHG studies on. In Emission characterized. The quality of the XRD structures was the orientation dependence of SHG in crystals. It was addition, attempts to filters excellent in most cases. As expected, the products Figure 7. were chiral, which the space groups confirmed. Ph7N7C3H2•PhNCS•C6H6 When we attempted to form i-butyl-N-phenyl XRD structure. Space thiocarbamate, we were surprised to obtain a group: P 21/c. previously unseen product Singl (according to searches of e- the Cambridge Structural Photo Us Database and SciFinder) containing seven phenyl Figure 3. SHG microscope diagram. Light of 800 nm rings (Fig. 7). wavelength is pulsed onto the sample (P = 1 or 2 mW) and It is heavily conjugated the number of emitted 400 nm photons is recorded by the and also chiral, and therefore predicted to have single-photon photomultiplier tube. nonlinear optical activity. However, its crystals were ing a powder form was a step we took to minimize not large enough to perform SHG studies on. In the orientation dependence of SHG in crystals. It was addition, attempts to Figure 7.

Ph7N7C3H2•PhNCS•C6H6 XRD structure. Space group: P 21/c. hydrogen bonds, structures molecular packing, and also easier to obtain a powder than to grow crystals of enhances crystal growth. In addition, the sulfur atom acceptable size and purity for other studies. The scan scatters well, allowing absolute X-ray structure area was (20 µm)2 with a step size of 1 µm, and we determinations (XRD). The crystalline nature of these measured the average SHG counts (emitted photons of products made for easy characterization by XRD. half the wavelength of the incoming photons) over To study the feasibility of using this reaction to each area, omitting areas that did not achieve obtain crystals, racemic 2-butanol was used to test the significant SHG activity. These averages were reaction with a variety of aryl isothiocyanates, many compared to give rough relative activity of which had an electron-withdrawing group (EWG) at determinations. the para position: phenyl isothiocyanate, 4- fluorophenyl isothiocyanate, 4-chlorophenyl Results isothiocyanate, 4-bromophenyl isothiocyanate, 4- We succeeded in synthesizing and characterizing nitrophenyl isothiocyanate, and 3-chlorophenyl two target compounds: i-butyl-N-(4-nitrophenyl) isothiocyanate. We then performed a second reaction thiocarbamate (Fig. 4) and i-butyl-N-(4-chlorophenyl) with the costlier (S)-2-butanol to obtain the non- thiocarbamate (Fig. 5). Crystalline products were centrosymmetric structure. obtained for the other reactions we attempted (in The reaction involved mixing or dissolving the particular, the 4- isothiocyanate in 2-butanol with the aid of vortexing. bromophenyl The reaction mixture was heated in a sand bath at isothiocyanate, 4- about 110° C for 3 h, then allowed to cool to room fluorophenyl temperature. Crystals were grown by slow evaporation of the solvent. In order to “get an idea” of our compounds’ NLO activity, we decided to subject powder forms of our compounds to an SHG activity test using a SHG microscope (Figure 3), with two standards: KAP and isothiocyanate, and KDP. Scan Figure 5. i-butyl-N-(4- stage chlorophenyl) Figure 4. i-butyl-N-(4- thiocarbamate XRD 100x 1.3 N.A. nitrophenyl) thiocarbamate structure. Space group: P Objective 0º or 90º XRD structure. Space group: 21. P 21. 3-chlorophenyl isothiocyanate with (S)-2-butanol), but unfortunately, Dichroic 800 nm due to time constraints, these were not mirror excitation

Emission characterized. The quality of the XRD structures was DARANCIANG filters excellent in most cases. As expected, the products When we attempted to form i-butyl-N-phenyl thiocarbamate, We were not discouraged by the failure to obtain the predict- were chiral, which the spacewe were groups surprised to confirmed.obtain a previously unseen product (accord- ed product. On the contrary, due to this unexpected discovery, we When we attempteding to to searches form of the iCambridge-butyl- StructuralN-phenyl Database and Sci- were able to characterize triphenylguanidine by SHG, and it gave Finder) containing seven phenyl rings (Fig. 7). counts four times as large as KDP and six times as large as KAP, thiocarbamate, we were surprised to obtain a two far better-known NLO materials. We believe that prolonged previously unseen product heating caused an overreaction—the thiocarbamate was probably Singl (according to searches of formed, but went on to react with other phenyl isothiocyanate e- units. the Cambridge Structural Raster images of the powder forms of the crystals revealed Photo Database and SciFinder) unmistakable inhomogeneities in the crystals, which limited our Us ability to find a meaningful average. Therefore, our reported av- containing seven phenyl erages only consider those areas where high SHG intensity was Figure 3. SHG microscope diagram. Light of 800 nm rings (Fig. 7). observed. Despite this limitation, we did obtain average SHG wavelength is pulsed onto the sample (P = 1 or 2 mW) and counts of KAP, KDP, triphenylguanidine, and 4-chlorophenyl It is heavily conjugated isothiocyanate powder samples in the method described previ- Figure 7. Ph N C H •PhNCS•C H XRD structure. Space group: P 21/c. the number of emitted 400 nm photons is recorded by the and also chiral, and therefore7 7 3 predicted2 6 6 to have ously. KAP and KDP were our two standards, and 4-chlorophenyl isothiocyanate was a negative control. As previously stated, tri- single-photon photomultiplier tube. It is heavily conjugated and also chiral, and therefore pre- nonlinear optical activity. However, its crystals were phenylguanidine was highly active relative to our standards. dicted to have nonlinear optical activity. However, its crystals ing a powder form was a step we took to minimize not large enough to performwere not large SHGenough to studiesperform SHG on. studies In on. In addition, the orientation dependence of SHG in crystals. It was addition, attemptsreplicate thisattempts to synthesis to replicate have this met synthesis with havefailure, met giving with failure, giv- triphenylguanidine byCONCLUSIONS SHG, and it gave counts four ing thereplicate less complex this synthesisthree-ringed have compound: met with triphenylguanidine failure, giving triphenylguanidine by SHG, and it gave counts four the less the complex less complexthree-ringed three-ringed compound: compound: times times as Triphenylguanidine large as large as KDPas KDP and shows and six six high times times SHG as activity largelarge as as relativeKAP, KAP, to (Fig. 8). Figure 7. triphenylguanidinetriphenylguanidine (Fig. 8). (Fig. 8). twoKDP fartwo andbetter-known far KAP, better-known but further NLO NLOcharacterization materials. materials. ofWeWe the believe believethiocarbamates that that Ph7N7C3H2•PhNCS•C6H6 prolongedis needed.prolonged There heating heating is still causedmuch caused work an anto do overreaction—the toward developing a XRD structure. Space thiocarbamatestrongthiocarbamate theoretical was understanding was probably probably formed, of formed, what gives but but rise went went to on the on toSHG to group: P 21/c. reactactivity withreact of withother these other phenylcompounds. phenyl isothiocyanate isothiocyanate We have already units.units. attempted to react Figure 8. Figure 8. other Raster chiral nucleophilic images of the ligands powder with forms the aryl of theisothiocyanates, crystals Triphenylguan Raster images of the powder forms of the crystals Triphenylguan suchrevealed as methylbenzylamine unmistakable and inhomogeneities l-menthol, and some in of the these idine XRD revealedcrystals, unmistakable which limited inhomogeneities our ability to find in a the idine XRD structure. crystal products currently await characterization. Many optical crystals,meaningful which average. limited Therefore, our our ability reported to averages find a structure. Space group: P properties of this family of compounds remain unexplored. meaningfulonly consider average. those Therefore, areas where our high reported SHG intensity averages Space group: P na 21. As a whole, the research lays out a basis for the rational design was observed. Despite this limitation, we did obtain na 21. onlyof NLO consider active thosematerials. areas The whereunderstanding high to SHG be gleaned intensity from average SHG counts of KAP, KDP, wasthis observed. series of experiments Despite should this limitation, shed additional we light did on obtainthis fas- averagetriphenylguanidine, SHG counts and 4-chlorophenyl of KAP, isothiocyanate KDP, cinatingpowder problem. samples in the method described previously. triphenylguanidine, and 4-chlorophenyl isothiocyanate We were not discouraged by the failure to obtain KAP and KDP were our two standards, and 4- powder samples in the method described previously. the predicted product. On the contrary, due to this chlorophenyl isothiocyanate was a negative control. Figure 8. Triphenylguanidine XRD structure. Space group: P na 21. We were notunexpected discouraged discovery, by the we failure were able to obtain to characterize KAP As and previously KDP werestated, triphenylguanidineour two standards, was highly and 4- the predicted product. On the contrary, due to this chlorophenyl isothiocyanate was a negative control. unexpected discovery, we were able to characterize As previously stated, triphenylguanidine was highly

Figure 9. From left to right, (20 µm)2 raster images of KAP, KDP, and triphenylguanidine powder with a step size of 1 µm (P = 2 mW). Even at this small Figurescan 9. From size, observed left to SHGright, intensity (20 µm) is not2 raster constant. images The boxed of KAP, yellow KDP, area in and each triphenylguanidine image is the area where powderthe averages with in aFig. step 10 aresize taken of 1from. µm (P = 2 mW). Even at this small scan size, observed SHG intensity is not constant. The boxed yellow area in each image is the area where the averages in Fig. 10 are taken from. CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 27

active relative to our standards. Figure 9. From left to right, (20 µm)2 raster images of KAP, KDP, and triphenylguanidine powder with a step size of 1 µm (P = 2 mW). Even at this small scan size, observed SHG intensity is not constant. The boxed yellow area in each image is the area where the averages in Fig. 10 are taken from.

active relative to our standards.

Figure 10. . Average SHG counts/100 ms for powder samples over the boxed areas in Fig. 9.

Figure 10. . Average SHG counts/100 ms for powder samples over the boxed areas in Fig. 9. replicate this synthesis have met with failure, giving triphenylguanidine by SHG, and it gave counts four the less complex three-ringed compound: times as large as KDP and six times as large as KAP, triphenylguanidine (Fig. 8). two far better-known NLO materials. We believe that prolonged heating caused an overreaction—the thiocarbamate was probably formed, but went on to react with other phenyl isothiocyanate units. Figure 8. Raster images of the powder forms of the crystals Triphenylguan revealed unmistakable inhomogeneities in the idine XRD crystals, which limited our ability to find a structure. meaningful average. Therefore, our reported averages Space group: P na 21. only consider those areas where high SHG intensity was observed. Despite this limitation, we did obtain average SHG counts of KAP, KDP, triphenylguanidine, and 4-chlorophenyl isothiocyanate powder samples in the method described previously. We were not discouraged by the failure to obtain KAP and KDP were our two standards, and 4- the predicted product. On the contrary, due to this chlorophenyl isothiocyanate was a negative control. unexpected discovery, we were able to characterize As previously stated, triphenylguanidine was highly

Figure 9. From left to right, (20 µm)2 raster images of KAP, KDP, and triphenylguanidine powder with a step size of 1 µm (P = 2 mW). Even at this small scan size, observed SHG intensity is not constant. The boxed yellow area in each image is the area where the averages in Fig. 10 are taken from.

active relative to our standards. SYNTHESIS OF NONLINEAR OPTICAL-ACTIVE MATERIALS

Figure 10. . Average SHG counts/100 ms for powder samples over the boxed areas in Fig. 9. Figure 10. Average SHG counts/100 ms for powder samples over the boxed areas in Fig. 9.

REFERENCES 1. Bauman, R. J. Eng. Chem. Data 1966, 11, 274-276. 2. Kaminsky, W.; Fitzmaurice, A.; Glazer, A. J. Phys. D.: Appl. Phys. 1998, 31, 767-775. 3. Smith, P. The Organic Chemistry of Open-Chain Nitrogen Compounds; W. A.Benjamin: New York, 1965; pp 234-247.

ACKNOWLEDGEMENTS The authors wish to thank all the members of the 2005 Hooked on Photonics REU program at the University of Wash- ington, and are especially indebted to Bart Kahr for lab space and guidance, the NSF for funding, the MDITR-STC, Kristin Wustholz, and Sara Selfe.

Dan Daranciang, a senior at Washington University in St. Louis, is completing a double major in chemistry and Germanic languages and literatures. He hopes to be accepted to graduate school in the fall.

28 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Behavioral Properties of Colloidal Crystals at and Around Thermally Induced Defect Sites

Mallory Davidson Ashlee St. John and Dr. L. Andrew Lyon University of Washington Lyon Lab, School of Chemistry and Biochemistry Georgia Institute of Technology

Soft sphere interactions are described for PNIPAm (Poly-N-isopropylAcrylamide) micro gel particle assemblies doped with gold nanoparticles as a laser hits a small region of the crystal. The gold nanoparticles will act as a local heater excited by the laser and will in- troduce a defect into the assembly. The thermoresponsivity of the microgel particles will be utilized to study the behavioral properties the system. Observation of the assembly as it melts to form a defect at a particular spot and how the microgel particles will reorder them- system. The particlesselves after the deswellheat is removed as will be analyzed. SpectrophotometryThis will be done for a more fundamental was run to the temperatureknowledge increases of colloidal crystalsto the with soft sphere determineinteractions. In addition, how it many will provide gold a lower critical solutionbetter understanding temperature of the kinetics and thermodynamicsnanoparticles of the crystallization were process. needed in the of the polymer; allowing the system assemblies. Enough gold to act more as a fluid.2 nanoparticles were needed to INTRODUCTION RESULTS AND DISCUSSIONS absorb photons from the laser, but Experiment There is a lot of known information on hard sphere interac- Goldnot Incorporationtoo many nanoparticles so that Assembliestions with crystal assemblies,of PNIPAm but not a(Poly-N- lot about soft sphere in- there In order were to induce no a defect perturbed the crystal assemblies phase needed to teraction systems.1 Colloidal crystals with soft interactions will be isopropylAcrylamide) microgels bebehaviors. doped with gold At nanoparticles. the first UV VISnoticeable Spectrophotometry researched for fundamental understanding of the kinetics and ther- doped with gold nanoparticles will waspeak run to of determine 520nm how many (this gold isnanoparticles where were gold needed modynamics of the crystallization process. With the information in the assemblies. Enough gold nanoparticles were needed to ab- have a defect induced by particles of this size absorb) from provided through this research the knowledge gained can be ap- sorb photons from the laser, but not too many nanoparticles so positioning a frequency doubled Nd: plied for later scientific applications of optic and sensing devices. thatthe there UV were VIS no perturbed read phaseout behaviors. determined At the first notice- YAGNamely (Yttrium thermoresponsivity Aluminum of the particles Garnet, will be utilized to study ablehow peak much of 520nm gold (this is was where putgold particles into ofthe this size ab- the =532nm) behavioral properties laser of atthe system. a small The particles section deswell as the sorb)sample. from the UVIt wasVIS read found out determined to be how 500 much� goldL. was oftemperature the crystal. increases to theAs lower the critical laser solution hits temperature a of putSee into thefigure sample. 1.It was found to be 500μL. See figure 1. sectionthe polymer; of allowing the the crystalline system to act more the as a fluid. gold2 nanoparticles will absorb the EXPERIMENT 1.6 photonsUV Vis grafromphs of 6.8 w eitheght perce ntlaserof particles because of their Assemblies high extinction of PNIPAm (Poly-N-isopropylAcrylamide) coefficient. mi- 1.4 crogels doped with gold nanoparticles will have a defect induced Particles with no The1.6 energy gained willParticles bewithout released 1.2 gold nanoparticles by positioning1.4 a frequency doubledgold Nd: YAG (Yttrium Alumi- in thee form of heat. This released 1 numc 1.2 Garnet, λ=532nm) laser at a small section of the crystal. Particles with n Particles with 200 300MicroL of gold heata will warm up the crystal, and As the1 laser hits a section of the crystalline the gold nanopar- 0.8 b MicroL of Gold nanoparticles oner 0.8 can observe what happens to ticleso will absorb the photons fromParticles the with laser 300 because of their high Particles with s 0.6 0.6 Absorbance thatextinctionb part coefficient. of the The energycrystalMicroL gained of Gold as will thebe released in the 500MicroL of gold

A 0.4 nanoparticles 0.4 defectform0.2 of heat. is This induced. released heat Also willParticles warm with one 400 up thecan crystal, and one MicroL of Gold observecan observe0 whatthe happens crystal to that aspart ofit the cools crystal as the defect 0.2 is induced.300 400 Also 500 one 600 700can 800observe 900 theParticles crystal with 500 as it cools down after down after the laserMicroL is of Gold removed to 0 seethe laser how is removed Wtheavelength (n defectmto) see how thereorders defect reorders into into a acrystal. 350 450 550 650 750 850 crystal. Differential interference contrast microscopy was used to Wavelength (nm look at the microgel particles. Afterwards the images attained were analyzed using IDL software. IDL was used to obtain par- Figure 1. Determination of the Differential interference contrast Figure 1. Determination of the amount of gold nanoparticles ticle trajectories of the images. amount needed to dope the crystalline assemblies. microscopy was used to look at the of gold nanoparticles needed to microgel particles. Afterwards the CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 29 dope the crystalline assemblies. images attained were analyzed using IDL software. IDL was used to In addition, comparisons of the obtain particle trajectories of the assemblies with gold nanoparticles images. were observed under heat cycling Results and Discussions between 35 C and room temperature to ensure no Gold Incorporation In order to aggregation occurred. See Figure 2. induce a defect the crystal This was done in a series of ten assemblies needed to be doped with temperature cycling. If they were gold nanoparticles. UV VIS BEHAVIORAL PROPERTIES OF COLLOIDAL CRYSTALS AT ANDmost AROUND crystalline THERMALLY INDUCED sample DEFECT so SITES that References clean trajectory images can be 1. Torquato, S,; Truskett, T.M., In addition, comparisons of the assemblies with gold theobtained. particles of the In microgel addition, that wereone out could of phase. take This out of Debenedetti, ° nanoparticles were observed under heat cycling between 35 C phaselonger transition movies occurred so in thatthe viewing the ofan the microscope im- P.G. Is random close packing of and room temperature to ensure no aggregation occurred. See age that the camera took. So clear crystalline trajectories of the observation of both the melting and spheres Figure 2. This was done in a series of ten temperature cycling. defect were hard to obtain. In the future it is suggested to use the recrystalliation process could be well defined? Phys. Rev. Lett. 2000, If they were to aggregate then the spectra would look different most crystalline sample so that clean trajectory images can be 84, shifting closer to a wavelength around 560nm. It was found that obtained.made. In Inaddition, this oneexperiment could take longer one movies could so that the the slopes all had similar trends having a distinct point at 520nm an onlyobservation look of at both fluid the meltingto crystalline and recrystalliation process 2064. where the gold is visible. Thus no aggregation was observed. couldtransition be made. In thiswith experiment the camera one could that only look at fluid to crystallinewas used. transition with the camera that was used. 2. WU, J.; Zhou, B.; Hu, Z. Phase behavior Phase Behavior Another test for these particles was done to ensure that there of thermally responsive microgel was no perturbing of the phase behavior trend between the two colloids. crystalline assemblies. It was observed that the assemblies doped Phys. Rev. Lett. 2003, 90, 048304 with gold nanoparticles tended to start off in a glassy phase until they were heated. Yet the two assemblies followed the same trend of melting at the same temperature.

Inducing the Defect After ensuring the assemblies were relatively the same the Figure 3. Left is trajectory of a microgel assembly obtained from IDL. assemblies with the gold nanoparticles were tested with the mi- FigureRight is image 3. obtained Left from is the trajectory camera on the microscope. of a Circled croscope and laser. The assembly was maintained at a set temper- microgelareas are the locations assembly of the laser obtainedand the induce defect from in the sample. ature usingsystem. an objective The heater particles and temperature deswell stage, asand then Spectrophotometry was run to the temperature increases to the determineIDL. Right ishow image many obtained gold from the laser waslower introduced critical to heat solution up a particular temperature area of the sample. nanoparticlesACKNOWLEDGEMENTS were needed in the of the polymer; allowing the system assemblies.the camera onEnough the microscope. gold In addition, laser intensity was used to vary the2 amount of energy to act more as a fluid. nanoparticlesCircled areas arewere the needed locations to of added to the system. The energy ranged from barely exciting a absorb photons from the laser, but Experiment notthe lasertoo many and nanoparticlesthe induce defect so that in particular Assembliesspot to melting a ofcertain PNIPAm area of the(Poly-N- assembly. It was there were no perturbed phase isopropylAcrylamide) microgels behaviors.the sample. At the first noticeable Mallory Davidson is currently seen that asdoped the melting with temperature gold nanoparticles was approached it will took less peak of 520nm (this is where gold intensity fromhave the alaser defect to melt orinduced excite the sample.by most The crystalline opposite sample so particlesthat References of this size absorb) from studying chemical engineering at positioning a frequency doubledclean trajectory Nd: images cantheAcknowledgements be UV VIS1. Torquato, read S,;out Truskett, determined T.M., was also observedYAG (Yttrium that when theAluminum sample was Garnet,farobtained. away from In addition, the one couldhow take much Debenedetti, gold was put into the the University of Washington. Upon =532nm) laser at a smalllonger section movies so that the an P.G. Is random close packing of melting temperature it took more energy from the laser to excite sample. It was found to be 500�L. graduation in June 2006 she plans of the crystal. As the laserobservation hits a of both the meltingSee and figure spheres 1. REFERENCES the sample.section of the crystalline recrystalliationthe gold process could be well defined? Phys. Rev. Lett. 2000, to get a full time job either doing nanoparticles will absorb themade. In this experiment one could 84, 1. Torquato,1.6 S,; Truskett, T.M., Debenedetti, P.G. Is random close photons from the laser becauseonly look at offluid to crystalline 2064. marketing or consulting for their high extinction coefficient.transition with the camerapacking that of1.4 spheres well defined? Phys. Rev. Lett. 2000, 84, 2064. UV Vis graphs of 6.8 weight percent of particles Particles with no engineering companies. However, The energy gained will be releasedwas used. 1.2 2. WU, J.; Zhou, B.; Hu, Z. Phasegold nanoparticles in the form of heat. This released 2. WU, J.; 1Zhou, B.; Hu, Z. Phase behavior of thermally respon- 1.6 Particles without behavior Particles with graduate school is a considered an heat1.4 will warm up the goldcrystal, and 300MicroL of gold e 0.8 of thermally responsive microgelnanoparticles c 1.2 sive microgel colloids. Phys. Rev. Lett. 2003, 90, 048304 onen can observe whatParticles happens with 200 to

a Particles with future option 1 0.6 colloids. Absorbance

b MicroL of Gold 500MicroL of gold thatr 0.8 part of the crystal as the o Particles with 300 Phys. Rev. Lett. 2003, 90, 048304nanoparticles s 0.6 0.4 defectb is induced. AlsoMicroL one of Gold can

A 0.4 observe0.2 the crystal asParticles it with 400cools 0.2 0 MicroL of Gold down after the laser is removed to 0 300 400 500 600 700 800 900 Particles with 500 350 450 550 650 750 850 see how the defect reordersMicroL of Gold into a crystal. Wavelength (nm) Wavelength (nm Figure 1. Determination of the FigureDifferential 2. Annealing process interference of a microgel showing nocontrast aggregation amount microscopy was used to lookFigure at 3. theLeft is trajectory of a after 10 heat cycles between 35°C and room temperature. of gold nanoparticles needed to microgel particles. Afterwardsmicrogel assemblythe obtained from dope the crystalline assemblies. images attained were analyzedIDL. Right is image obtained from using IDL software. IDL wasthe usedcamera onto the microscope. CONCLUSIONS AND RECOMMENDATIONS In addition, comparisons of the obtain particle trajectoriesCircled of areasthe are the locations of assemblies with gold nanoparticles images. the laser and the induce defectMallory in Davidson is currently studying chemical engineering at the It was shown that a defect could be introduced from the were observed under heat cycling the sample. University of Washington.Mallory Davidson Upon isgraduation currently in June 2006 she plans to Results and Discussions between studying35 C chemicaland room engineering at YAG laser. For defect and trajectory images see Figure 3. How- temperatureget a full time job either to doing ensure marketing orno consulting for engineering Gold Incorporation In orderAcknowledgements to the University of Washington. Upon aggregationcompanies. However, occurred. graduate school is Seea considered Figure a future 2. option ever analysisinduce of the crystala defect assemblies the proved crystal to be inconclusive graduation in June 2006 she plans This was done in a series of ten assemblies needed to be doped with to get a full time job either doing temperature cycling. If they were because asgold the sample nanoparticles. would be heated UV up theVIS crystal would be marketing or consulting for excited out of phase in the microscope image and the software engineering companies. However, graduate school is a considered an that traces the particles could not produce accurate images of future option

30 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Optical Properties of Metal Nanoparticle Composites

Krystle L. Dzienis Wojtek Haske and Dr. Joe Perry Pennsylvania State University Perry Labs, Department of Chemistry and Biochemistry Georgia Tech

Metal nanoparticle films and composites were investigated to study their electronic and optical properties in hopes to develop procedures for their use in electronic and photonic devices. Much emphasis was placed on studying the digestive ripening process under ar- gon flow and studying the behavior of nanoparticles in a polymer matrix forming a nano- composite thin film.

INTRODUCTION gands) or gold (DT) (gold nanoparticles with dodecane thiol li- gands) nanoparticles were dissolved in various solvents to study There is a strong interest in metal nanoparticles and their the effect of solvent on the nanoparticle’s UV-vis absorption composite films because of their potential use in advanced mate- peak. Solvents used to prepare these solutions included dichloro- rials with interesting electronic and optical properties1. In order to benzene, dichloromethane, toluene and cyclohexane. utilize these composites in future device applications, the ability to manipulate the size, arrangement, and morphology of metal nanoparticles has to be achieved2. In order to better understand Film Preparation and use them in these devices, nanoparticles should be mono- All of the films investigated were prepared on glass substrates. dispersed and highly soluble. The nanocomposite films consisting of Au or Ag nanoparticles Metal nanoparticles absorb electromagnetic waves by inducing mixed with various polymers (PS = polystyrene, PMMA = poly oscillation of electrons, so called plasmons, which result in the en- (methyl methacrilate), PVK = poly (vinyl carbozole)), were spin hancement of the electric field in the vicinity of the particles. This cast at 1000rpm for 2 minutes using a toluene based solution. phenomenon can contribute to enhancements of nonlinear optical properties of molecules close to the nanoparticle surface. Metal RESULTS nanoparticles will be modified with two-photon chromophores to Digestive Ripening investigate changes in optical properties of the chromophores.

EXPERIMENTAL Synthesis Gold (Au) and silver (Ag) nanoparticles have been synthe- sized using a one phase method. Either hydrogen tetrachloroaurate or silver nitrate is reduced by sodium borohydride in the presence of thiol ligands in an ethanol solution. Thiol ligands used in this study include octane thiol (OT) and dodecane thiol (DT). Figure (a) – Au before Figure (b) – Au after

Digestive Ripening Metal nanoparticles were dissolved in a toluene solution with excess thiol ligands2. The solution was then refluxed for ap- proximately three hours in an argon atmosphere.

Solution Preparation Small amounts of silver 3:1 (OT:DT) (silver nanoparticles Figure (c ) – Ag before Figure (d) – Ag after coated with a ratio of 3 to 1, octane thiol to dodecane thiol, li-

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 31 absorption spectra of Ag nanoparticles were The maximum of the absorption spectrum was similar in all solvents investigated. Au shifted from 500nm in solution to 545nm in nanoparticles had two distinct sets of absorption polystyrene films. Films with higher weight spectra, with particles in toluene and ratios of nanoparticles exhibited larger red shifts cyclohexane showing slightly narrower plasmon of the plasmon band (570nm). bands than particles in dichlorobenzene and An investigation of ligand effect on OPTICAL PROPERTIES OF METAL NANOPARTICLE COMPOSITSdichloromethane. nanocomposite films was also conducted. Thin films of Ag 3:1 (OT:DT) nanoparticles (filtered Transmission electron microscopy has been used to inves- two distinct sets ofThin absorption Film spectra, Composites: with particles Ag in 3:1 toluene (OT:DT) with 0.45µm filters) and Ag 1:1 (OT:DT) tigate the effects of digestive ripening. Figure (a) and (b) shows and cyclohexane showingnanoparticles slightly narrower were combined plasmon bands with than various nanoparticles (filtered with 0.2µm filters) in polymers to cast nanocomposite films. Before polystyrene were prepared with various weight Au nanoparticles before and after digestive ripening, respectively. particles in dichlorobenzene and dichloromethane. casting the films the solutions were filtered using ratios, (Graph f and g nanoparticle ratios Before digestive ripening, the average size of the particles was a 0.2µm filter. decreased by � for each solution). 7.1±5nm, which is a much larger size range compared to 4.9±0.8nm Thin Film Composites Graph (f) – Absorption spectra of Ag 3:1 (OT:DT) for particles after digestive ripening. Figure (c) and (d) shows Ag Ag 3:1 (OT:DT) nanoparticles were combined with various nanoparticle films in polystyrene. The weight ratio of 3:1 (OT:DT) nanoparticles before and after digestive ripening, re- nanoparticles was decreased by � for every sample from Film Preparation: All of the films investigated polymers to cast nanocomposite films. Before casting the films sample 1 to sample 4. were prepared spectively. on glass Before substrates. digestive ripening The the particles were 4.6±1.5nm the solutions were filtered using a 0.2µm filter. nanocomposite filmsin size. consisting After digestive of Au ripening or Ag the particles were slightly smaller A red shift of surface plasmon resonance is seen showing a nanoparticles mixed with various polymers (PS = and more uniform with an average size of 4.0±1.2nm. polystyrene, PMMA = poly (methyl maximum absorption peak around 500nm (Graph d) compared to methacrilate), PVK Absorption = poly (vinyl spectra carbozole)), in the UV-visible range were measured on silver nanoparticles in solutions, which show a maximum absorp- were spin cast at 1000rpmAu (DT) for nanoparticles 2 minutes before using anda after digestive ripening, in a tion peak around 420nm. Since polystyrene showed the largest toluene based solution. toluene solution. Before digestive ripening the spectrum shows a red shift and was the most readily available, it was used for all of broader absorption peak with a higher baseline (Graph a) than the the subsequent nanocompositeGraph (d) film – Absorption preparations. spectra of Ag 3:1 (OT:DT) absorption spectra of Ag nanoparticlesRESULTS were The maximum of the absorption spectrum was nanoparticle films similar in all solvents investigated. solution Au after digestiveshifted ripening. from 500nm in solution to 545nm in A similar red shift of surface plasmon resonance with respect nanoparticles had two distinctDigestive sets of absorptionRipening: Ag nanoparticlespolystyrene were not films.soluble after Films digestive with ripening higher and weightto solution measurementsA red shift was ofalso surface seen forplasmon Au (DT) resonance nanopar is- seen spectra, with particles in toluene and ratios of nanoparticles exhibited larger red shifts showing a maximum absorption peak around an absorption spectrum in solution could not be measured. ticles in polystyrene (Graph e) (filtered with .1µm filter). Graph (g) – Absorption spectra of Ag 1:1 (OT:DT) cyclohexane showing slightly narrower plasmon of the plasmon band (570nm). Graph (a) – UV-vis of Au nanoparticle 500nm (Graph d) compared to silver The maximum of the absorption spectrum was shifted from nanoparticle films in polystyrene. The weight ratio of bands than particles in dichlorobenzene and An investigation of ligand effect on nanoparticles in solutions, which show a nanoparticles was decreased by � for every sample from Ag nanoparticles were not soluble after digestive dichloromethane. Nanoparticlenanocomposite Solutions films was also conducted. Thin500nm in solution maximum to 545nm absorption in polystyrene peak films.around Films 420nm. with Since sample 1 to sample 4. Film Preparation: All of the films investigated ripening and an absorption spectrum in solution An investigationfilms of of Ag solvent 3:1 effects(OT:DT) was nanoparticles conducted on Au (filtered higher weight ratiospolystyrene of nanoparticles showed exhibited the largest larger red red shiftshifts and of was were prepared on glass substrates. The could not be measured. In both cases a blue shift in the absorption peak Thin Film Composites: Ag 3:1 (OT:DT) with 0.45µm filters) and Ag 1:1 (OT:DT) the most readily available, it was used for all of nanocomposite films consisting of Au or Ag (DT) and Ag 3:1 (OT:DT) nanoparticles. the plasmon band (570nm).the subsequent nanocomposite film preparations. was observed when the nanoparticle nanoparticles were combined with various nanoparticles (filtered with 0.2µm filters) in nanoparticles mixed with various polymers (PS = UV-vis spectroscopy was performedNanoparticle on the various Solutions: solu- An An investigationinvestigation of of Aligand similar effect redon nanocomposite shift of surface films was plasmon concentration was lowered. The Ag 1:1 spectrum polystyrene,polymers PMMA to cast= polynanocomposite (methyl films. Before polystyrene were prepared with various weight tions (Graph b and c). The absorption spectrasolvent of Ag effects nanoparticles was conducted also conducted.on Au (DT) Thin resonanceand films of Agwith 3:1 respect (OT:DT) to solutionnanoparticles measurements (fil- also shows a shoulder around 400nm. methacrilate),casting PVK the = polyfilms (vinyl the solutions carbozole)), were filtered using ratios, (Graph f andAg g3:1 nanoparticle(OT:DT) nanoparticles. ratios were similar in all solvents investigated. Au nanoparticles had tered with 0.45µm filters)was also and Ag seen 1:1 for (OT:DT) Au (DT) nanoparticles nanoparticles (fil- in Film werePreparation: spin casta 0.2µm at All 1000rpm of filter. the forfilms 2 minutesinvestigated usingFigure a (a) – Au before Figure (b)decreased – Au after by � for each solution). polystyrene (Graph e) (filtered with .1µm filter). CONCLUSION were toluene prepared based solution. on glass substrates. The Completing the digestive ripening process under nanocomposite films consisting of Au or Ag Graph (f) – Absorption spectra of Ag 3:1 (OT:DT) absorption spectra of Ag nanoparticles were an Theargon maximum flow shows of many the absorptionimprovements spectrum to the was RESULTS nanoparticle films in polystyrene. The weight ratio of nanoparticles mixed with various polymers (PS = similar in all solvents investigated. Au Aushifted nanoparticles. from 500nmThese improvements in solution include to 545nm a in nanoparticles was decreased by � for every sample from more uniform size and better solubility of the Au polystyrene,Digestive Ripening: PMMA = poly (methyl sample 1 to sample 4. nanoparticles had two distinct sets of absorption polystyrene films. Films with higher weight methacrilate), PVK = poly (vinyl carbozole)), nanoparticles. The effect of digestive ripening on spectra, with particles in toluene and Agratios 3:1 of (OT:DT) nanoparticles nanoparticles exhibited was larger not red as shifts were spin cast at 1000rpm for 2 minutes using a Graph (a) – UV-vis of Au nanoparticle cyclohexane showing slightly narrower plasmon of the plasmon band (570nm). toluene based solution. beneficial. After digestive ripening, the Ag bands than particles in dichlorobenzene and nanoparticles An investigation were no longer of soluble ligand and only effect a on Ag nanoparticles were not soluble after digestive dichloromethane. slightnanocomposite change in their films size range was was also seen. conducted. Thin RESULTS ripening and an absorption spectrum in solution filmsFor Au of (DT) Ag 3:1and (OT:DT)Ag 3:1 (OT:DT) nanoparticles in various (filtered could not be measured. Figure (c ) – Ag before Figure (d) – Ag after Thin Film Composites: solvents, no major shifts in the surface plasmon Digestive Ripening: Ag 3:1 (OT:DT) with 0.45µm filters) and Ag 1:1 (OT:DT) band was seen. Nanoparticle Solutions: An investigation of nanoparticles were combined with various nanoparticles (filtered with 0.2µm filters) in Transmission electron microscopy has been used When particles are placed in a polymer matrix Graph (d) – Absorption spectra of Ag 3:1 (OT:DT)solvent Grapheffects (a) was – UV-vis conducted of Au nanoparticle on Au (DT) andGraph (b) - Absorption spectra of Ag 3:1 (OT:DT)polymers Graph (e) –to Absorption cast nanocomposite spectra of Au (DT) nanoparticles films. Before in polystyrene were prepared with various weight to investigate the effectsGraph of (a) digestive - UV-vis of Au ripening. nanoparticle Graph (c) - Absorption spectra of Au a red shift of surface plasmon resonance occurs. nanoparticle films Ag 3:1 (OT:DT) nanoparticles. Graphnanoparticles (b) - Absorption in solution spectra of Ag castingpolystyrene the films films. the The solutions Au nanoparticle were amount filtered was using ratios, (Graph f and g nanoparticle ratios Figure (a) and (b) shows Au nanoparticles 3:1 (OT:DT) nanoparticles in solution (DT) nanoparticles in solution When Ag 3:1 (OT:DT) nanoparticles were Figure (a) – Au before Figure (b) – Au after a 0.2µmdecreased filter. by � in every film (top to bottom). decreased by � for each solution). before and Agafter nanoparticles digestive ripening, were notrespectively. soluble after digestive placed in various polymers, there were A red shift of surface plasmonBefore resonance digestiveripening is seenripening, and an the absorption average size spectrum of the in solution showing a maximum absorptionparticles peak wascould around7.1±5nm, not be measured. which is a much larger Graph (f) – Absorption spectra of Ag 3:1 (OT:DT) Graph (g) – Absorption spectra of Ag 1:1 (OT:DT) nanoparticle films in polystyrene. The weight ratio of 500nm (Graph d) comparedsize range to compared silver to 4.9±0.8nmnanoparticle for particles films in polystyrene. The weight ratio of Nanoparticle Solutions: nanoparticles was decreased by � for every sample from nanoparticles in solutions, after which digestive show ripening. a Figure (c)nanoparticles and (d) An shows investigation was decreased by of � for every sample from sample 1 to sample 4. maximum absorption peak aroundAg 3:1 420nm. (OT:DT)solvent Since nanoparticles effects was before conducted and afteron Au sample (DT) 1 to and sample 4. polystyrene showed the largestdigestive red shift ripening,Ag and 3:1 was (OT:DT)respectively. nanoparticles. Before digestive Figure (a) – Au beforethe most Figure readily (b) –available, Au after it ripeningwas used the for particles all of were 4.6±1.5nmIn both cases in size. a blue shift in the absorption peak the subsequent nanocompositeAfter film preparations. digestive ripening the was particles observed were when the nanoparticle A similar red shift of slightly surface smaller plasmon and more uniformconcentration with an was lowered. The Ag 1:1 spectrum Figure (c ) – Agresonance before with Figure respect (d) – Ag toafter solution average measurements size of 4.0±1.2nm. also shows a shoulder around 400nm. was also seen for Au (DT) nanoparticlesAbsorption spectra in in the UV-visible range were measured on Au (DT) nanoparticles before Transmissionpolystyrene electron microscopy (Graph e)has (filtered been used with .1µm filter). CONCLUSION and after digestiveGraph (b) ripening, - Absorption in spectra a toluene of Ag 3:1 (OT:DT) Graph (c)- Absorption spectra of Au (DT) nanoparticles in to investigate the effects of digestive ripening. Completing the digestive ripening process under Graph (d) – Absorption spectra of Ag 3:1 (OT:DT) solution. Before digestiveGraph ripeningnanoparticles (d) – Absorption the inspectrum solution spectra of Ag Graph (e) – Absorptionsolution spectra of Au (DT) Graph (f) - Absorption spectra of Ag 3:1 (OT: Figure (a) and (b) shows Au nanoparticles 3:1 (OT:DT)an argonnanoparticle flow films shows many improvementsnanoparticles in polystyreneto the films. The DT) nanoparticlenanoparticle films in polystyrene. films The shows a broader absorption peak with a higher before and after digestive ripening, respectively. Au nanoparticle amount was decreased weight ratio of nanoparticles was decreased by baseline (Graph a) than theAu solution nanoparticles. after TheseUV-vis improvements spectroscopy include was a performed on the Before digestive ripening, the average size of the by ½ in every film (top to bottom). A red shift½ for everyof surface sample from plasmon sample 1 resonanceto sample 4. is seen digestive ripening. more uniform size and bettervarious solubility solutions of (theGraph Au b and c). The particles was 7.1±5nm, which is a much larger showing a maximum absorption peak around size range compared to 4.9±0.8nm for particles nanoparticles. The effect of digestive ripening on Graph (g) – Absorption spectra of Ag 1:1 (OT:DT) Figure (c ) – Ag before Figure (d) – Ag after 32 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 500nm (Graph d) compared to silver after digestive ripening. Figure (c) and (d) shows Ag 3:1 (OT:DT) nanoparticles was not as nanoparticle films in polystyrene. The weight ratio of nanoparticles in solutions, which show a TransmissionAg 3:1 (OT:DT) electron nanoparticlesmicroscopy has before been and used after beneficial. After digestive ripening, the Ag nanoparticles was decreased by � for every sample from maximum absorption peak around 420nm. Since sample 1 to sample 4. to investigatedigestive ripening, the effects respectively. of digestive Before ripening. digestive Graph (b) - Absorptionnanoparticles spectra of Ag were 3:1 (OT:DT) no longer soluble and only a nanoparticles in solution polystyrene showed the largest red shift and was Figureripening (a) and the particles(b) shows were Au 4.6±1.5nm nanoparticles in size. slight change in their size range was seen. In both cases a blue shift in the absorption peak beforeAfter and after digestive digestive ripening ripening, the respectively. particles were For Au (DT) and Ag 3:1 (OT:DT) in various the most readily available, it was used for all of Beforeslightly digestive smaller ripening, and the more average uniform size withof the an solvents, no major shifts in the surface plasmon the subsequent nanocomposite film preparations. was observed when the nanoparticle average size of 4.0±1.2nm. particles was 7.1±5nm, which is a much larger band was seen. A similar red shift of surface plasmon concentration was lowered. The Ag 1:1 spectrum Absorption spectra in the UV-visible range size range compared to 4.9±0.8nm for particles When particles are placed in a polymer matrix resonance with respect to solution measurements also shows a shoulder around 400nm. were measured on Au (DT) nanoparticles before after digestive ripening.Graph (e) Figure – Absorption (c) and spectra (d) shows of Au (DT) nanoparticles in a red shift of surface plasmon resonance occurs. was also seen for Au (DT) nanoparticles in and after digestivepolystyrene ripening, films. The in Au a nanoparticle toluene amount wasGraph (c)- Absorption spectra of Au (DT) nanoparticles in polystyrene (Graph e) (filtered with .1µm filter). CONCLUSION Ag 3:1solution. (OT:DT) Before nanoparticles digestivedecreased ripening beforeby � in theevery and spectrum film after (top to bottom). Whensolution Ag 3:1 (OT:DT) nanoparticles were Completing the digestive ripening process under digestiveshows ripening, a broader respectively. absorption Beforepeak with digestive a higher placed in various polymers, there were ripeningbaseline the particles (Graph a)were than 4.6±1.5nm the solution in size. after UV-vis spectroscopy was performed on the an argon flow shows many improvements to the Afterdigestive digestive ripening. ripening the particles were various solutions (Graph b and c). The Au nanoparticles. These improvements include a slightly smaller and more uniform with an more uniform size and better solubility of the Au average size of 4.0±1.2nm. nanoparticles. The effect of digestive ripening on Absorption spectra in the UV-visible range Ag 3:1 (OT:DT) nanoparticles was not as were measured on Au (DT) nanoparticles before beneficial. After digestive ripening, the Ag and after digestive ripening, in a toluene Graph (c)- Absorption spectra of Au (DT) nanoparticles in nanoparticles were no longer soluble and only a solution solution. Before digestive ripening the spectrum slight change in their size range was seen. shows a broader absorption peak with a higher For Au (DT) and Ag 3:1 (OT:DT) in various Graph a) UV-vis spectroscopy was performed on the baseline ( than the solution after solvents, no major shifts in the surface plasmon digestive ripening. various solutions (Graph b and c). The band was seen. When particles are placed in a polymer matrix Graph (e) – Absorption spectra of Au (DT) nanoparticles in a red shift of surface plasmon resonance occurs. polystyrene films. The Au nanoparticle amount was decreased by � in every film (top to bottom). When Ag 3:1 (OT:DT) nanoparticles were placed in various polymers, there were DZIENIS

ACKNOWLEDGEMENTS STC-MDITR, Summer 2005 REU – “Hooked on Photon- ics,” NSF, COPE, Georgia Institute of Technology, Dr. Joe Perry, Wojtek Haske, Perry Research Group

REFERENCES [1] T. Shimizu, T. Teranishi, S. Hasegawa, M. Miyake, J. Phys. Chem. B 2003, 107, 2719-2724. [2] X.M. Lin, C.M. Sorensen, K.J. Klabunde, Journal of Nanopar- Graph (g) - Absorption spectra of Ag 1:1 (OT:DT) nanoparticle ticle Research 2: 157-164, 2000. films in polystyrene. The weight ratio of nanoparticles was decreased by ½ for every sample from sample 1 to sample 4.

tered with 0.2µm filters) in polystyrene were prepared with vari- ous weight ratios, (Graph f and g nanoparticle ratios decreased by ½ for each solution). In both cases a blue shift in the absorption peak was observed when the nanoparticle concentration was lowered. The Ag 1:1 spectrum also shows a shoulder around 400nm. Krystle Dzienis is currently studying Materials Science and Engineering, specializing in polymers at the Pennsylvania State University. She is planning on continuing her education in the field of Materials Science by going on to CONCLUSION graduate school upon completing her undergraduate degree in May 2007. Completing the digestive ripening process under an argon flow shows many improvements to the Au nanoparticles. These improvements include a more uniform size and better solubility of the Au nanoparticles. The effect of digestive ripening on Ag 3:1 (OT:DT) nanoparticles was not as beneficial. After digestive ripening, the Ag nanoparticles were no longer soluble and only a slight change in their size range was seen. For Au (DT) and Ag 3:1 (OT:DT) in various solvents, no major shifts in the surface plasmon band was seen. When particles are placed in a polymer matrix a red shift of sur- face plasmon resonance occurs. When Ag 3:1 (OT:DT) nanopar- ticles were placed in various polymers, there were differences in the size and shape of the absorption spectra. These differences might be due to the different arrangements the nanoparticles have in the different polymers, and the possibility of aggregates forming could have varied depending on the polymer structure. Additionally, the local environment around the nanoparticles in different instances could affect the plasmon resonance. A blue shift occurs when the concentration of nanoparticles to polymer is lowered, which might be due to interactions between nanopar- ticles. Additional investigations are needed to provide further in- sight into these findings. Investigation of the digestive ripening of Ag nanoparticles in various solvents will be conducted to see if control of the size distribution of the nanoparticles can be ob- tained. Also, X-ray diffraction studies of the nanocomposite films will be performed to investigate the arrangement of nanoparticles in the polystyrene matrix.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 33 34 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 media with noElectro-Optic physical linkage. Properties of Hybrid Sol-Gel Introduction This method has Materials lead to such in Fabry-Perot Modulators The electro-optic (EO) effect problems as phase separation, occurs when an electric field is high loss, and poor thermal applied across a material that has stability. Brenda The second Eby method, Haiyong Gan, Hongxi Zhang and Mahmoud Fallahi an optical electromagnetic wave side-chain,University of links Idaho the Fallahi Lab, Optical Science Center Where _ is the wavelength of the incident traveling through it. The applied chromophores covalently to the ������University of Arizona wave, � is the thickness of the film, � is the silica backbone. This reduces the electric field causes changes in refractive index of the EO polymer, and _ the indices of refraction in the problems demonstrated in the Fabry-Perot Etalons, ������ �, are material that result in first method and has shown symmetric structures with the EO polymer is the internal incident angle of the light going through the device. The waves in modulations in the light wave promise in thermal and temporal (3-[5-(2-{4-[bis-(2-hydroxy-ethyl)-amino]-phenyl}-vinyl)-sandwiched between two layers each of thio- INTRODUCTION phase are reinforced through constructive passing through the material. stability. The chromophore, TCBD phen -2 -yl] -2 which,5 –dicyanocontain -4 -[3 -(3-hydroxya glass -propoxy)substrate, -phenyl]a This modulation phenomenon is The electro-optic(3-[5-(2-{4-[bis-(2-hydroxy- (EO) effect occurs when an electric field is interference, resulting in strong narrow –hexa -2 ,4-diene-dinitrile),transparent isconductive shown in Figureelectrode 1. Chromophoreand a useful as it can be used in high-applied acrossethyl)-amino]-phenyl}-vinyl)- a material that has an optical electromagnetic wave highly reflective (>99%) Distributed transmission bands that vary depending thiophen -2 -yl] -2 ,5 –dicyano -4 - material was provided by the University of Washington and pro- speed data transfer and opticaltraveling through it. The applied electric field causes changes in Bragg Reflective (DBR) mirror.are in phase.upon Thethe equationmultiple thatfactors shown in 1 [3 -(3-hydroxy -propoxy) -phenyl] cessing of hybrid sol-gel material was done using methods previ- communication applications. the As indices of refraction in the material that result in modulations Device describes ��������the phase shift��). becauseThey include film a result, materials exhibiting –hexa -2 ,4-diene-dinitrile), is ously established.4 in the light wave passing through the material. This modulation Fabry-Perot����������� Etalons, Figure of the DBRthickness, mirrorchange differencein refractive index due desirable electro-optic properties shown in Figure 1. Chromophore 2, are symmetric structures with between to successiveEO effect, and reflectionssurface roughness. In are currently in high demand.phenomenon material is useful as wasit can be provided used in high-speed by the data transfer 1 the EO polymerDEVICE sandwiched is: general transmission will occur whenever This research examined and the optical communicationUniversity of applications. Washington As a result, and materials between two layers each of which the phase shift satisfies the following electro-optic properties of hybridexhibiting desirableprocessing electro-optic of properties hybrid are sol-gel currently in high Fabry-Perot Etalons, Figure 2, are symmetric structures with 2p material was done using methods contain a glass substrate,���������� a Ê equation:ˆ sol-gel materials in Fabry-Perotdemand. This research examined the electro-optic properties of the EO polymer sandwiched between two layers each of whichd = Á ˜2nl cosq (1) 4 transparent conductive electrode Modulator devices. previously established. Ë l ¯ hybrid sol-gel materials in Fabry-Perot Modulator devices. contain a andglass asubstrate, highly a transparent reflective conductive (>99%) electrode and Where � is the wavelength of� the a highly reflectiveDistributed (>99%) BraggDistributed Reflective Bragg Reflective (DBR) � � � � ; Material incident wave, l is the thickness mirror. (DBR) ��� mirror. where m = any integer ��� EO polymers exhibit of the film, n is the refractive index of the EO polymer, and � is anisotropic properties because of DBR Mirrors����� A typical wavelength versus their uniaxial character. Hybrid the internal incident angle of the �������������������������������������� transmission spectra illustrating this is sol-gels are a unique type of EO light going through the device. shown in �������� below. polymer, which combine the rigid The waves in phase are reinforced matrix stability of Si/SiO2 with EO Polymer through constructive the polarizability of organic Fabry-Perot Etalons are commonly usedinterference,in resulting in strong chromophore molecules.2 Hybrid data transmission and have applicationsnarrowas transmission bands that sol-gels are attractive as EO modulators,ITO tunable filters, and opticalvary depending upon the multiple polymers because they are low switches. As light passes through factorsthe shown in Equation (1). cost, easy to process materials Glass They include film thickness, Figure 1: TCBD Chromophore Structure4 Figuredevice 1: Fabry-Perot a resonance Etalon Schematic 3 4 with low loss. There are two ways Figure 1. TCBD Chromophore Structure dueFigureto internal2. Fabry-Perotreflection Etalon Schematicoccurs inside changethe in refractive index due to EO effect, and surface to achieve this meshing of EO polymer cavity between the highly inorganic and organic compounds. roughness. In general MATERIAL Fabry-PerotFabry-Perot Etalons are commonly Etalons used in are data transmission The first is through doping (guest reflective DBR mirrors. Interferencetransmission will occur whenever commonly used in data host) which introduces the EO polymers exhibit anisotropic properties because of their and have applicationsbetween asthese modulators,two surface tunableresults filters, inandonly theoptical phase shift satisfies the transmission and have chromophores to the sol-geluniaxial character. Hybrid sol-gels are a unique type of EO poly- switches. As lightspecific passes wavelengthsthrough the deviceof a resonancelight being following equation: applicationstransmitted. asSuccessive modulators,reflections will mer, which combine the rigid matrix stability of Si/SiO2 with the due to internaltunable reflection filters, occurs inside and the optical EO polymer cavity 2 cancel each other out because of d=2 mp ; polarizability CMDITR ofReview organic of Undergraduate chromophore Research Vol. molecules. 1 No. 1 Hybrid sol- between theswitches. highly reflective As DBR light mirrors. passes Interference between destructive interference except at the where m = any integer (2) Summer 2004 2 gels are attractive as EO polymers because they are low cost, easy these twothrough surface resultsthe device in only a resonance specific wavelengths of light wavelengths were the light waves are in to process materials with low loss.3 There are two ways to achieve being transmitted.due to internalSuccessive reflection reflections occurswill cancel each other �������������������������������������������� insidephase. the The EO equation polymer that cavitydescribes the A typical wavelength������������������ versus this meshing of inorganic and organic compounds. The first is out because of destructive interference except at the wavelengths betweenphase theshift highlybecause reflectiveof the DBRDBR mirrortransmission spectraAdapted from illustrating Lambda Research Optics Inc. were the light waves are in phase. The equation that describes through doping (guest host) which introduces the chromophores mirrors.difference Interferencebetween successive betweenreflections this is shown in Figure 3 below. the phase shift because of the DBR mirror difference between to the sol-gel media with no physical linkage. This method has theseis: two surface results in only When voltage is applied to the material a lead to such problems as phase separation, high loss, and poor successivespecific reflections wavelengths is: of light being shift in the resonance wavelength occurs and correlates to the change in refractive thermal stability. The second method, side-chain, links the chro- transmitted. � �� � Successive mophores covalently to the silica backbone. This reduces the reflections� � will� cancel���� each��� � other index (��) due to the electro-optic effect out because of destructive������������������ problems demonstrated in the first method and has shown prom- � � � and the change in thickness (��) due to a interference except at the ise in thermal and temporal stability. The chromophore, TCBD wavelengths wereEquation the 1light waves

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 35 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 2 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 3 change in the piezo (mechanical) forces ������� acting upon the material. The resonance wavelength and the shift in that wavelength are measured by ������������������������ using a broadband source which outputs Indium tin oxide, a transparent through a fiber collimator into the Fabry- conductive material acts as the electrode in Perot Etalon. Light that is transmitted the device and is the first layer deposited from the cavity is focused onto a single onto the glass substrate. Then the DBR mode fiber detector which leads to a layer is coated on top followed by the optical spectrum analyzer (OSA). The TCBD sol-gel. The TCBD sol-gel is OSA measures the wavelength while the prepared and spin coated at varying speeds applied voltage is varied letting the to set the thickness of the EO polymer wavelength shift be directly observed. film. After baking, this layered material is By switching the broadband source to a corona poled above the glass transition narrowband, tunable laser the extinction temperature at a high voltage. Poling is ratio and dynamic modulation of the done to orient the chromophore molecules sample can be measured. The laser and force non-centro-symmetry in the propagates through a collimator, continues film. The combination of temperatures through the device cavity and finally the above the glass transition temperature, transmitted light is focused onto the fiber which causes the Si/SiO2 sol gel matrix to detector. The dynamic modulation can be Where _ is the wavelength of the incident relax, and the application of a high electric observed by switching the DC voltage field causes the highly polarizable ������ wave,Where� is_theis thethicknesswavelengthof theoffilm,the�incidentis the source to AC by inserting a function chromophore molecules to rotate and generator and changing to a photo detector ������Fabry-Perot Etalons, ������ �, are refractivewave, � isindexthe thicknessof the EOofpolymer,the film, and� is the_ refractive index of the EO polymer, and _ orient themselves with the electric field. which inputs into a digital oscilloscope. symmetricFabry-PerotstructuresEtalons,with the������EO polymer�, are is the internal incident angle of the light After poling another glass slide with ITO is the internal incident angle of the light See ������� �� �� ��� � below for testing sandwichedsymmetric structuresbetween twowithlayersthe EOeachpolymerof going through the device. The waves in and DBR is placed on top of the EO going through the device. The waves in schematics. whichsandwichedcontainbetweena glasstwo layerssubstrate,each aof phase are reinforcedELECTRO-OPTICthrough PROPERTIESconstructive OF HYBRID SOL-GEL MATERIALSpolymer IN FABRY-PEROTand electrical MODULATORSwires are connected which contain a glass substrate, a interference,phase are reinforcedresultingthroughin strongconstructivenarrow transparent conductive electrode and a Where λ is the wavelength of the incident wave, l is the with the electricto each field.ITO Afterlayer. poling anotherThe layers glass slideare withthen ITO ��������� transparent conductive electrode and a transmissioninterference,bandsresultingthat invarystrongdependingnarrow �������� highly reflective (>99%) Distributed thickness of the film, n is the refractive index of the EO polymer, and DBR is clampedplaced on topand ofepoxy the EO ispolymerapplied andto electricalseal the wires ������ highly reflective (>99%) Distributed upontransmissionthe multiplebands factorsthat varyshowndependingin Bragg Reflective (DBR) mirror. and θ is the internal incident angle of the light going through the are connecteddevice. to each ITO layer. The layers are then clamped and Bragg Reflective (DBR) mirror. ��������upon the ��multiple). Theyfactorsincludeshownfilmin device. The waves in phase are reinforced through constructive epoxy is applied to seal the device. ����������� thickness,��������change��).in refractiveThey includeindex duefilm ������� interference, resulting in strong narrow transmission bands that ����������� �������� ����������� tothickness,EO effect,changeand surfacein refractiveroughness.index Indue ������������� ������ vary depending upon the multiple factors shown in Equation (1). �������� ��������� generalto EO transmissioneffect, and surfacewill occurroughness.wheneverIn �������� They include film thickness, change in refractive index due to ������ ������������ thegeneralphasetransmissionshift satisfieswill theoccurfollowingwhenever ������� ������������������������������������������������ ���������� EO effect,equation:the andphase surfaceshift roughness.satisfies In generalthe followingtransmission will Sol-gel film ���������� occur wheneverequation: the phase shift satisfies the following equation: ������������� ������� �������� ���� � � �� � ; ����� ������ � � ��� ; ����������� ��� where m = any integer ��� ����������� ��� where m = any integer ��� ������� Equation 2. where m = any integer ��������� ����� A typical wavelength versus ����� ����� �������� ������ �������� ������������������������������������������� A typical wavelength versus ��������� A typicaltransmission wavelengthspectra versus transmissionillustrating spectrathis illustratingis ��������������������� �������������������������������������� transmission spectra illustrating this is �������������� this isshown shown inin Figure �������� 3. below. shown in �������� changebelow.change in inthe the piezo piezo (mechanical) (mechanical) forces forces �������������� ������������������������������������������� acting upon the material. Figure 4. Corona Poling Apparatus acting upon the material. TheThe resonanceresonance wavelength andand thethe Fabry-Perot Etalons are commonly used in shift in that���������������������������������wavelength are measured by Fabry-Perot Etalons are commonly used in ������������������������ shift in that wavelength are measured by data transmission and have applications as ������������������������ usingusingaa broadbandbroadbandMETHODSsourcesource which outputsoutputs modulators,data transmissiontunableandfilters,have applicationsand opticalas IndiumIndiumtintinoxide,oxide, a a transparenttransparent throughthroughaa fiberfiber collimatorcollimator into thethe Fabry-Fabry- conductive material acts as the electrode in conductive material acts as the electrode in ThePerotPerot resonanceEtalon.Etalon. wavelengthLightLight that andis transmittedthe shift in that wavelength switches.modulators,As tunablelight passesfilters,throughand opticalthe the device and is the first layer deposited � the device and is the first layer deposited fromfrom the the cavity cavity is is focused focused onto aa singlesingle CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 switches. As light passes through the onto the glass substrate. Then the DBRare measured by using a broadband source which outputs device a resonance onto the glass substrate. Then the DBR modemode fiberfiber detectordetector which leads toto aa layer is coated on top followed by thethrough a fiber collimator into the Fabry-Perot Etalon. Light that duedeviceto internal a resonancereflection occurs inside the layer is coated on top followed by the opticaloptical spectrumspectrum analyzeranalyzer (OSA). TheThe TCBDTCBDsol-gel.sol-gel.TheTheTCBDTCBDsol-gelsol-gel isisis transmittedOSA measures from the thecavitywavelength is focusedwhile ontothe a single mode fiber EOduepolymerto internalcavityreflectionbetweenoccurstheinsidehighlythe prepared and spin coated at varying speeds OSA measures the wavelength while the prepared and spin coated at varying speeds applied voltage is varied letting the EO polymer cavity between the highly to set the thickness of the EO polymerdetectorapplied which leadsvoltage to a isopticalvaried spectrumletting analyzerthe (OSA). The reflective DBR mirrors. Interference to set the thickness of the EO polymer wavelength shift be directly observed. film. After baking, this layered material isOSA measureswavelength the shiftwavelength be directly while observed. the applied voltage is varied betweenreflectivetheseDBRtwo surfacemirrors.resultsInterferencein only film. After baking, this layered material is By switching the broadband source to a corona poled above the glass transition By switching the broadband source to a between these two surface results in only corona poled above the glass transitionletting thenarrowband, wavelengthtunable shift belaser directlythe extinction observed. specific wavelengths of light being temperature at a high voltage. Poling is narrowband,ratio and dynamictunable modulationlaser the extinctionof the temperature at a high voltage. Poling is transmitted.specific wavelengthsSuccessive reflectionsof light beingwill done to orient the chromophore molecules Byratio sampleswitchingandcan dynamicthebe broadbandmeasured.modulation sourceThe toof lasera narrowband,the tunable done to orient the chromophore molecules transmitted. Successive reflections will Figure 3. Typical Transmittanceand Spectraforce for Fabrynon-centro-symmetry Perot Etalon in thelaser thesamplepropagates extinctioncanthrough ratiobe measured. anda collimator, dynamicThe continuesmodulationlaser of the sample cancel each other out because of and force non-centro-symmetry in the propagates through a collimator, continues cancel each other out because of Adapted from Lambdafilm. ResearchThe Opticscombination Inc. of temperaturescan be measured.through the Thedevice lasercavity propagatesand finally throughthe a collimator, con- destructive interference except at the film.aboveThethecombinationglass transitionof temperaturestemperature, throughtransmittedthe lightdeviceis focusedcavity andontofinallythe fiberthe above the glass transition temperature,tinues through the device cavity and finally the transmitted light wavelengthsdestructive wereinterferencethe light exceptwaves areat inthe ��������������������������������������������which causes the Si/SiO2 sol gel matrix to transmitteddetector. Thelightdynamicis focusedmodulationonto thecanfiberbe When voltage is applied whichto the causesmaterialthe aSi/SiO2 shift insol thegel resomatrix- to phase.wavelengthsThe equationwere thethatlightdescribeswaves arethein ��������������������������������������������������������������relax, and the application of a high electricis focuseddetector.observed onto theTheby fiberdynamicswitching detector.modulationthe TheDC dynamicvoltagecan be modulation can relax, and the application of a high electric observed by switching the DC voltage phase. The equation that describes the nance wavelengthAdapted occurs from������������������ Lambdaand correlates fieldResearchcauses Opticsto the Inc. thechangehighly in refracpolarizable- be observedsource by toswitchingAC by theinserting DC voltagea function source to AC by insert- phase shift because of the DBR mirror fieldchromophorecauses themoleculeshighly topolarizablerotate and source to AC by inserting a function tive index (∆n) dueAdapted to thefrom electro-optic Lambda Research effect Optics and Inc. the change in generator and changing to a photo detector differencephase shiftbetweenbecausesuccessiveof the DBRreflectionsmirror chromophoreorient themselvesmoleculeswith theto electricrotate field.anding a functiongeneratorwhich inputsgeneratorand changinginto anda digital changingto a photooscilloscope. todetector a photo detector which is:difference between successive reflections thicknessWhen (∆l )voltage due to a is change appliedorient inAfter the tothemselves polingpiezothe material another(mechanical)withglass thea electricslide forceswith field.ITOinputs intowhichSee a������� digitalinputs ��oscilloscope.into�� ���a digital� below Seeoscilloscope.for Figurestesting 5, 6, and 7 below After poling another glass slide with ITO is: actingshift uponWhen inthe the voltagematerial. resonance is applied wavelengthand DBRto theis materialplaced occurson atop of the EOfor testingSeeschematics. schematics.������� �� �� ��� � below for testing andshift correlates in the resonance to the andchangepolymer wavelengthDBR inandis refractiveplacedelectrical occursonwirestop areof connectedthe EO schematics. �� polymerto eachandITOelectricallayer. wiresThe layersare connectedare then ��������� � ���� ��� indexand DEVICE correlates(��) due ASSEMBLYto to the the electro-optic change / POLINGin refractive effect �������� � � � � ��� � � ������������������ to clampedeach ITOandlayer.epoxyTheis appliedlayerstoaresealthenthe ��������������� ��� ��� index (��) due to theclampeddevice. electro-opticand epoxy effectis applied to seal the ������ �������� � �� �� � � � ������������������ and the change in thickness (��) due to a � Indiumand tinthe oxide, change a transparent in thicknessdevice. conductive (��) due material to a acts as � � the electrode in the device and is the first layer deposited onto the ������� ����������� ��������������� ������������� ������ �������� glass substrate. Then the DBR layer is coated on top followed by ����������� ����������������� ������������� �������� ������ ������ ������������ �������� ��������� the TCBD sol-gel. The TCBD sol-gel is prepared and spin coated������� ������������������������������������������������ �������� Sol-gel film ������ ������������ 2 CMDITR Review of Undergraduate Research Vol. 1 No. 1 at Summer varying 2004 speeds to set the thickness of the EO polymer film.������� Af- ������������������������������������������������Figure 5. Shift Measurement Schematic Sol-gel film 2 CMDITR Review of Undergraduate Research Vol. 1 No. 1 ter Sbaking,ummer 2004 this layered material������������� is corona poled above the glass ������� �������� ���� ������ transition temperature at a high������������� voltage.����� Poling is done to orient ������� ����������� �������� ���� ����������� ������ the chromophore molecules and force����� non-centro-symmetry in ����������� ������� the film. The combination of temperatures above��������� the glass transi�����������- ����� ����� �������� ������ ������� ��������� �������� ��������� tion temperature, which causes the Si/SiO sol gel matrix��������������������� to relax, ����� ����� �������� 2 �������������� ������ �������� ��������� and the application of a high electric field causes ���������������������the highly po- �������������� ������������������������������������������� larizable chromophore molecules to rotate and orient themselves ��������������������������������� ��������� Figure���������������������������������� 6. Extinction Measurement Schematic ��������������������������������� 36 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 �

CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 � ������� �������� ���� Extinction measurements also show ������ promising results with an extinction ratio measured at 5 Volts. The extinction ratio ������������ is the log ratio of the intensity at 0 volts ������������ ������ compared to the intensity at an applied ��������� voltage.

���������� �������� ������ ���������

������������������������ ���������������������������������������� ���������

Additional measurements with a surface ����������������������������������� profiler were done to analysis the roughness of the individual layers of the Fabry-Perot Etalon device. Each surface was scanned in sequence with preceding material layers in place. For example, the third layer, DBR, was scanned with ITO

�������and glass�������� layers���� below it. Extinction measurements also show ������ promising results with an extinction ratio EBY measured at 5 Volts. The extinction ratio �������������������� is the log ratio of the intensity at 0 volts ������������ Surface analysis of the sample was shown to be a major fac- ������������ compared to the intensity at an applied �������������By varying��������the����voltage over a range of 60 Extinction measurements also show ��������� voltage.tor in the performance of the device. Scans of the different device ������ V, the shifts in ������ � were observed. promising�����������������������������������������results with an extinction ratio levels are shown along with the average standard deviation val- ���������� After normalizing��������the shift wavelengths, measured at 5 Volts. The extinction ratio ������ ��������� ues. See Figures 11, 12. the average shift was calculated.������������ is the logSurfaceratioanalysisof the intensityof the sampleat 0 voltswas shown ������������ to be a major factor in the performance of ������ compared to the intensity at an applied ������������������������������������������������� voltage. ��������� ��������������������������� ���������� �������� TCBD Sol-Gel Additional������measurements���������with a surface profiler were done to analysis the DBR roughnessFigure 7. Dynamicof the Modulationindividual Measurementlayers Schematicof the ����������������������������������������� Fabry-Perot����������������������������������������Etalon device. Each surface was scanned in ���������sequence with preceding Additional measurements with a surface profiler were done Surface analysis of the sampleITOwas shown material layers in place. For example, the to be a major factor in the performance of to analysisthird thelayer, roughnessDBR, of wasthe individualscanned layerswith ofITO the Fabry- the device. Scans of the differentGlass device Perot EtalonAdditionaland glass device. layers measurements Each below surface it. was scannedwith a insurface sequence with levels are shown along with the average preceding material layers in place. For example, the third layer, standard deviation� ���values.��� See ������������� ��� profiler��������������������were done to analysis the ������� DBR, roughnesswas scanned withof the ITOindividual and glass layerslayers belowof it.the ����������������������������� By varying the voltage over a range of 60 ����������������������������������������� Fabry-PerotV, the shiftsEtalonin ������device.� wereEachobserved.surface After normalizingRESULTS /the DISCUSSIONshift wavelengths, ��������������������������� was scanned in sequence with preceding Surfacetheanalysisdevice.of ScanstheFiguresample 8of: Surfacethewas Scansdifferentshown device the average shift was calculated. Bymaterial varying thelayers voltagein place.over a rangeFor example,of 60 V, thethe shifts in to be alevelsmajor arefactorshownin thealongperformancewith theofaverage Figurethird 8 werelayer, observed.DBR, Afterwas normalizingscanned the withshift wavelengths,ITO the device.standardScansdeviationof the differentvalues. deviceSee ������� ����������� the averageand glass shift was layers calculated. below it. levels are�������shown along����������������������������������with the average TCBD Sol-Gel Research in the future will examine ���� standard deviation values. See ������� avenues in which to increase the �������������������� ������� DBR ��� performance of the EO polymer Fabry- By varying the voltage over a range of 60 � ��� ��� ��� ��� ��� 4 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 ITO Perot Etalon devices. Plans to replace the V, the shifts in ������ � were observed. ��������������������������������� After normalizing the shift wavelengths, ���������������������������Glass ITO with ZnO are being examined. the average shift was calculated. ��� Improving surface consistency and material homogeneity also are avenues to ���� ������������������������ pursue in the near future to increase device �������������������������������� � performance. TCBD Sol-Gel �� �� �� �� �� �� �� � � � � � � � �������������� ���������� ����������������������������������� DBR ������� ExtinctionFiguremeasurements 9. Shift Measurement Resultsalso show ����������������������������������������������� �������� ���� Figure 12. Probability Density Distribution for TCBD Sol-Gel 1. A. Yariv, ������� ����������� �� ������ ������ promising results with an extinction ratio ������������ � ��� ��� ��� ��� ��� ��������������, 5th ed. Oxford Extinctionmeasured measurementsat 5 Volts. alsoThe showextinction promising ratioresults with ITO 4 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 TheThe average�����������������������������average standardstandard deviation showndeviation below is shownthe value ob- ������������ is the log ratio of the intensity at 0 volts University, New York. (1997) an extinction ratio measured at 5 Volts. The extinction ratio is the tained from averagingGlass the distribution of each surface scan. Six ������������ compared to the intensity at an applied below is the value obtained from 2. H.Goudket, M. Canva, and Y. Levy, ������ ��������� log ratio of the intensity at 0 volts compared to the intensity at an or moreaveraging scans for theeach distributionsurface were recorded.of each Thesurface DBR mirror appliedvoltage. voltage. Journal of Applied Phsics. 90, 6044 (2001) is thescan. roughestSix surfaceor morein the device.scans Thisfor haseach implicationssurface in the 3. S. Najafi, T. Touam, R. Sara, M. ���������� device performance as the roughness can be directly tied to the �������� were������������������������recorded. The DBR mirror is the ������ ��������� performance of the device. Andrews and M. Fardad, Journal of roughest surface in the device. This has Lightwave Technology. 16, 1640 (2000) implicationsSurfacein the deviceAverageperformance Standard Deviation as 4. H. Zhang, D. Lu, Nu.Peyghambarian, ���������������������������������������� the roughnessTCBD Sol-Gelcan be directly3.02 nmtied to the M. Fallahi, J. Lou, B. Chen and A.K.Y. ��������� DBR Mirror 5.36 nm performanceITO of the device. 1.36 nm Jen, Optics Letters 30 (2): 117-119 JAN ����������������������������������� Glass 2.31 nm 15 (2005) Additional measurements with a surface Table 1. Indication������� of Roughness from���������������� Surface Scanning of Materials ��������� profiler were done to analysis the ������������ ������� ���������������� roughness of the individual layers of the 4 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 �����������������������������������������Figure10: Extinction Results TCBD Sol-gel ���������� ������� National Science Foundation Fabry-Perot Etalon device. Each surface ��� ������� University of Arizona was scanned in sequence with preceding Surface analysis of the sample was shown CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 37 ����� ������� College of Optical Sciences material layers in place. For example, the to be a major factor in the performance of ��������������������������������������������� third layer, DBR, was scanned with ITO the device. Scans of the different device ��������������������� and glass layers below it. levels are shown along with the average ����������������������������������������� ����������������������������������� standard deviation values. See ������� ����������� �������������������� ����������������������������������������� ������� Fabry-Perot modulators with electro- ���������������������������������������� By varying the voltage over a range of 60 optic hybrid sol-gel were successfully ����������� V, the shifts in ������ � were observed. fabricated and tested. The samples After normalizing the shift wavelengths, ��������������������������� the average shift was calculated. examined demonstrate promising results with measured wavelength shifts and extinction ratios. This indicates low drive voltage requirement. Additionally TCBD Sol-Gel dynamic modulation showed a rapid response with low loss. Results indicate DBR that the hybrid sol-gel Fabry-Perot � ��� ��� ��� ��� ��� Modulators are promising devices for ITO ����������������������������� many optical applications. Glass

������������������������ CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 �

�����������������������������������

4 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 �����������ELECTRO-OPTIC PROPERTIES OF HYBRID SOL-GEL MATERIALS IN FABRY-PEROT MODULATORS ���������������������������������� Research in the future will examine ���� CONCLUSIONS avenues in which to increase the Fabry-Perot modulators with electro-optic hybrid sol-gel ��� were successfully fabricatedperformance and tested. Theof samplesthe EO examinedpolymer Fabry- demonstrate promisingPerot resultsEtalon with measureddevices. wavelengthPlans shiftsto replace the ���� and extinction ratios. ITO This indicateswith lowZnO drive voltageare beingrequire- examined. ��� ment. Additionally dynamicImproving modulation showedsurface a rapid responseconsistency and with low loss. Resultsmaterial indicate thathomogeneity the hybrid sol-gel Fabry-Pealso are- avenues to rot Modulators are promising devices for many optical applica- ���� pursue in the near future to increase device

�������������������������������� tions. � performance. �� �� �� �� �� �� �� � � � � � � � FUTURE WORK �������������� Research in the future���������� will examine avenues in which to in- �����������������������������������������������crease the performance1. ofA. theYariv, EO polymer������� Fabry-Perot����������� Etalon �� ������ ������������ devices. Plans to replace�������������� the ITO with ZnO are being, 5 examined.th ed. Oxford The average standard deviation Improvingshown surface consistency and material homogeneity also are avenues to pursue University,in the near future New to increase York. device (1997) perfor- below is the value obtainedmance.from 2. H.Goudket, M. Canva, and Y. Levy, averaging the distribution of each surface Journal of Applied Phsics. 90, 6044 (2001) scan. Six or more scans for each surface 3. S. Najafi, T. Touam, R. Sara, M. were recorded. The DBR mirror is the AndrewsREFERENCESand M. Fardad, Journal of roughest surface in the device. This1. A. hasYariv, Optical electronicsLightwave in Modern Technology. Communications 16,, 51640th (2000) implications in the device performanceed. Oxfordas University,4. NewH. York.Zhang, (1997) D. Lu, Nu.Peyghambarian, the roughness can be directly tied2.to H.Goudket,the M. Canva, and Y. Levy, Journal of Applied Phsics. 90, 6044 (2001) M. Fallahi, J. Lou, B. Chen and A.K.Y. performance of the device. 3. S. Najafi, T. Touam,Jen, R. Sara,Optics M. AndrewsLetters and M. Fardad,30 (2): Jour117-119- JAN nal of Lightwave Technology.15 (2005) 16, 1640 (2000) ������� ����������������4. H. Zhang, D. Lu, Nu.Peyghambarian, M. Fallahi, J. Lou, B. ���������Chen and A.K.Y. Jen, Optics Letters 30 (2): 117-119 JAN 15 ������������ ������� (2005) ���������������� ���������� ������� National Science Foundation ��� ������� ACKNOWLEDGEMENTSUniversity of Arizona ����� ������� National Science FoundationCollege of Optical Sciences ���������������������������������������������University of Arizona ��������������������� College of Optical Sciences����������������������������������������� ����������������������������������� ����������� Funding for this research����������������������������������������� provided by the Center on Materials Fabry-Perot modulators with electro-and Devices for Information���������������������������������������� Technology Research (CMDITR), optic hybrid sol-gel were successfullyan NSF Science and Technology����������� Center No. DMR 0120967 fabricated and tested. The samples examined demonstrate promising results with measured wavelength shifts and extinction ratios. This indicates low drive voltage requirement. Additionally dynamic modulation showed a rapid response with low loss. Results indicate that the hybrid sol-gel Fabry-Perot Modulators are promising devices for 38 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 many optical applications.

CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 � Quantized Hamilton Dynamics Applied to Condensed Phase Spin-Relaxation

August 19, 2005 Stephen T. Edwards Oleg Prezhdo Harvey Mudd College Department of Chemistry, University of Washington AugustAugust 19, 19, 2005 2005 August 19, 2005 QuantizedAugust 19, Hamilton 2005 Dynamics Applied to Condensed Phase Spin-Relaxation QuantizedQuantized Hamilton Hamilton Dynamics DynamicsQuantized Applied Applied Hamilton to to Condensed Condensed DynamicsINTRODUCTION Phase Phase Applied to Condensed Using Phase second order equations, QHD can reproduce important quantum effects such as zero-point energy and tunneling. These Quantized Hamilton Dynamics Applied to Condensed Phase Chemistry has a myriad of quantum effects. However, to treat Spin-RelaxationSpin-Relaxation StephenSpin-Relaxation T. Edwards effects are preserved while reducing the complexity of a quantum all of chemistry with quantum mechanics would be nearly impos- Spin-Relaxation problem to one that is nearly as simple than classical mechanics. sible. Many quantum systems have no closed form solution, and StephenStephen T. T. Edwards Edwards numerical techniquesStephen are often prohibitively T. Edwards computationally ex- Stephen T. Edwards I. Introduction THE SPIN-BOSON MODEL pensive. Classical mechanics offers numerically simple solutions, August 19, 2005 Chemistry has a myriadbut cannot of reproduce quantum quantum effects. results, However, such as to tunneling treat all or ofzero- chemistry with The Spin-Boson quantum model consists of a single two-state spin I.I. Introduction Introduction mechanicsI. Introduction would be nearlypoint energy. impossible. As an approximation Many quantum to the systemsHeisenberg haveIII. representa no The closed- Spin-Boson formsystem solution, coupled Model and bilinearly to a bath of harmonic oscillators I. Introduction numerical techniquestion are of often quantum prohibitively mechanics, computationally quantized Hamilton expensive. dynamics offers Classical through mechanics the vibrational offers coordinate. The model has applications The Spin-Boson model consists of a single two-state spin system coupled bilinearly to a bath of ChemistryChemistry has has a myriad a myriad of of quantum quantum effects.Chemistry effects. However,Quantized However, has a tomyriad to treat Hamilton treat of all quantum all of of chemistry Dynamics chemistry effects. with with AppliedHowever, quantum quantum to to treat Condensed all of chemistry Phaseto condensed with quantum phase nonadiabatic electron transfer, low temper- numerically simple solutions,a method but of treating cannot these reproduce systems quantumthat maintains results, bothharmonic suchquantum as oscillators tunneling through or zero-point the vibrational coordinate. The model has applications to condensed Chemistry has a myriad of quantummechanicsmechanics effects. would would However, be be nearly nearly to impossible. treat impossible. all ofmechanics Many chemistry Many quantum quantum would with besystems quantum systems nearly have impossible. have no no closed closedSpin-Relaxation Many form form quantum solution, solution, systems and and have1 no closed form solution, and energy. As an approximationeffectsIII. and The tothe the relative Spin-Boson Heisenberg simplicity representation Modelof classical mechanics. ofphase quantum nonadiabatic mechanics,ature electron proton quantized transfer, transfer, low and temperature macroscopic proton quantum transfer, coherence and macroscopic in quantum mechanics would be nearly impossible.numericalnumerical Many techniques techniques quantum are systems are often often prohibitively have prohibitively nonumerical closed computationally computationally form techniques solution, are expensive. and oftenexpensive. prohibitively Classical Classical mechanics computationally mechanics offers offers expensive. Classical mechanics3 3 offers Hamilton dynamics offers Quantum a method two-state of treating systems these are systems frequently that encounteredcoherence maintains in SQUIDS. bothSQUIDS. quantumThe systemThe effects system is described is described by the by Hamiltonian the Hamiltonian numerical techniques are often prohibitivelynumericallynumerically computationally simple simple solutions, solutions, expensive. but but cannot cannotnumerically Classical reproduce reproduce mechanics simple quantum quantum solutions, offersresults, results,The butSpin-Boson such such cannot as as tunnelingStephen tunneling reproduce model T. or consists1 Edwards or zero-point quantum zero-point of a results, single two-state such as tunneling spin system or zero-point coupled bilinearly to a bath of and the relative simplicitychemistry of classicaland physics, mechanics. for instance a spin-½ particle or the po- p2 1 g numerically simple solutions, butenergy. cannotenergy. reproduce As As an an approximation approximation quantum results, to to the such the Heisenbergenergy. Heisenberg as tunneling As representation an representation or approximation zero-pointharmonic of of quantum to oscillatorsquantum the Heisenberg mechanics, mechanics, through representation quantized the quantized vibrational of coordinate. quantum mechanics, TheH = modelh ¯Ωσ quantized+ has ¯hεσ applications+ j + toµ condensedω2q2 j (σ q + q σ ) (3) Quantum two-statelarization systems of arelight. frequently The behavior encountered of these systems in chemistry become very and physics, forx instancez 2µ 2 j j j − 2 z j j z (3) energy. As an approximation toHamilton theHamilton Heisenberg dynamics dynamics representation offers offers a method a method of quantum ofHamilton oftreating treating mechanics, these dynamics these systems systems quantized offersphase that that a nonadiabatic method maintains maintains of bothtreating bothelectron quantum quantum these transfer, systems effects effects low that temperature maintains proton both quantum transfer, effects and macroscopicj j quantum a spin- 1 particle or the polarization of light. The behavior of these systems become very complex  andand the the relative relative simplicity simplicity of of classical classical mechanics. mechanics.2 I.1 Introduction1 complex when interaction with3 their1 environment is considered.2 Hamilton dynamics offers a method of treating these systems that maintainsand the both relative quantum simplicity effectscoherence of classical in SQUIDS. mechanics.The system2 isIII. described Thewhere Spin-Boson byσ and theσ Hamiltonianare Model Pauli matrices. The term Ω is related to tunneling probability; 2¯hε is the 1 when interaction withThe their spin-boson environment system simulates is considered. the couplingThe of a two-state spin-boson xspin systemz where simulates σx and σz theare Pauli matrices. The term is related to tun- and the relative simplicity of classicalQuantum mechanics.Quantum two-state two-state systems systems are are frequently frequentlyQuantum encounteredChemistry encountered two-state has in a systems in chemistry myriad chemistry of are quantum and frequently and physics, effects. physics, encountered However, for for instance instance to treat in all chemistryenergy of chemistry gap and between with physics, quantum the two for spin instance states; and p , q , ω , µ , and g are respectively the momentum, 2 j j j j j 1 1 coupling1 of a two-statesystem spin to system an environment to an environment of harmonic oscillators. of harmonic This oscillators.system neling This probability; system has 2ħε is the energy gap between the two spin Quantum two-state systems area spin-a frequently spin-particleparticle encountered or or the the polarization polarization in chemistry of of light. and light. physics, Themechanics The behavior behavior for would instance of be of these nearly these systems impossible. systems become Many become quantum very very complex systems complex haveThe no Spin-Bosoncoordinate,p closedj form1 model frequency, solution,2 2 consists reducedg andj of a singlemass, two-stateand coupling spin strength system coupled of the jth bilinearly bath oscillator. to a bath of 2 2 a spin- 2 particle2,3 or the polarization of light.2,3 The behavior of these systems become very complex 1 been wellnumerical studied, techniques2 sohas2 it been is are a oftenwell good studied, prohibitively systemH so to= computationallyit compareh ¯ isΩ σa xgood+ ¯hεσ newsystemz expensive.harmonic+ solution to compare oscillators Classical methods+ newµ mechanics through jω to,j qstates;j such the offers asvibrationaland(σ quantizedz qpj, +qj, q coordinate.ωjσj, zμ)j, and g Thej are model respectively has(3) applications the momentum, to condensed a spin- particle or the polarizationwhenwhen of interaction light. interaction The behavior with with their their of environmentthese environment systemswhen is become interaction is considered. considered. very with complexThe theirThe spin-boson spin-boson environment system system is considered. simulates simulates the2 theThe spin-boson2µThisj system2 system has− been2 simulates analyzed and the solved in many ways, including use of the non-interacting 2 j 2 3 2 Hamiltonnumerically dynamics. simplesolution solutions, methods but cannot to, such reproduce as quantized quantum Hamilton results,phase dynamics. such nonadiabaticblip as approximationtunneling electron or zero-pointcoordinate, transfer,and with lowfrequency, Makarov temperature reduced and Makri’s proton mass, transfer, numerical and coupling and path macroscopic strength integrals. of quantumtheThe quantized when interaction with their environmentcouplingcoupling of is of a considered. two-state a two-state spin spinThe system system spin-boson tocoupling to an an environment system environmentenergy. of a simulates two-state As an of approximation of harmonic harmonic the spin system oscillators. to oscillators. the to Heisenberg an environmentThis This representation system system hasofcoherence has harmonic of quantum in SQUIDS.oscillators. mechanics,3 The quantized system This system is described has by the Hamiltonian 2,32,3 2,3 Hamilton dynamicsjth scheme bath oscillator. is applied to this system so that its results can be compared with exact coupling of a two-state spin systembeenbeen to well an well studied, environment studied, soso it of isit harmonic ais good a good system oscillators. systembeen to well to compareHamilton comparestudied, This dynamicssystem new newsowhere solution solutionit hasoffers isσ a ax methods good methodand methods systemσ ofz to,are treating to, such toPauli such compare these as matrices. as quantized systems quantized new that solution The maintainssolutions term methods both Ω as is quantum well related to, as such providing effects to as tunneling quantized an alternate probability; solution to 2¯ thishε problem.is the Additionally, this is the first 1 This system has2 been analyzed and solved in many ways, 2,3 II. Quantizedand the relative Hamilton simplicityQUANTIZED of Dynamics classical mechanics. HAMILTON DYNAMICS pj 1 2 2 gj been well studied, so it is a goodHamiltonHamilton system dynamics. to dynamics. compare new solution methodsHamilton to, such dynamics. as quantizedenergy gap between the two spin states; andapplicationpj, qj, ωHj of,=µ QHDh ¯jΩIII.,σ and to+ The model ¯hεσg Spin-Bosonj are+ the respectively condensed-phase Model+ µ ω theq with momentum,( aσ bathq + ofq σ harmonic) oscillators.(3) includingx z use of the2µ non-interacting2 j j j − 2blipz approximationj j z 2 and with Quantum two-state systems are frequently encountered in chemistry and physics, for instance j j Hamilton dynamics. coordinate, The time evolution frequency, of quantum reduced expectation mass, and values coupling Tois given obtain strength the dynamicsThe of Spin-Boson the ofjth themodel bath system, consists oscillator. of aequations single two-state of spin motion system coupledwere generated bilinearly to a bathto second of order with a spin- 1 particle or the polarization of light. The behavior of these systems become very complex  3 The time evolution2 of quantum expectation values is giveneq by 1. Ehrenfest’s A third orderharmonicMakarov term theorem, oscillators appeared and through orMakri’s in the the the vibrational equationsnumerical coordinate. of path motion The modelintegrals. for has applicationsσxq jTheand toquantized condensed was decomposed using by Ehrenfest’sThis system theorem, has or been the Heisenberg analyzed2 representation and solved in(eq many 1). ways, including use of the non-interacting II.II. Quantized Quantized Hamilton Hamilton Dynamics DynamicsII. Quantizedwhen interaction Hamilton with their environmentDynamics is considered. Thewhere spin-bosonσx and systemσz are Pauli simulatesphase matrices. nonadiabatic the Theelectron term transfer, Ω low is temperature related toproton tunneling transfer, and probability; macroscopic quantum 2¯hε is the Heisenberg representation (eq 1). 2 eq 2 as follows: coherenceHamilton in SQUIDS. dynamics3 The system scheme is described3 is byapplied the Hamiltonian to this system so that its II. Quantized Hamilton Dynamics coupling of a two-stateblip approximation spin system to anand environment with Makarov of harmonicenergy and gap oscillators. between Makri’s This the numerical two system spin has states; path and integrals.pj, qj, ωj, µjThe, and g quantizedj are respectively the momentum, 2,3 d i results can be compared withp 2exact solutions as well as providing been well studied, so it is a good system to compare new solutioncoordinate, methods frequency, to, such reduced as quantized mass, and coupling strengthj 1 of the2 2 jgthj bath oscillator. Hamilton dynamics schemeA = is[ appliedH,A] to this system(1) so that its resultsH = canh¯Ωσx(1)+ be ¯hεσz compared+ + µjω withj qj ( exactσzqj + qjσz) (3) TheThe time time evolution evolution of of quantum quantum expectation expectationThe time values valuesevolution is is given given of quantum by by Ehrenfest’s Ehrenfest’s expectation theorem, theorem, values or or the is the given byg Ehrenfest’sk σxqjqk ( theorem,σxqj 2 σx orqj the) gk q2µkj +2 σx − 2 gk qjqk + qj gk σxqk Hamilton dynamics. dt  ¯h  This system has been ≈ analyzedan alternate and− solution solved  in to manythisj  problem. ways,  including Additionally, use of this the  non-interactingis the first  solutions as well as providing an alternate solutionk to this problem. Additionally,k  this is thek first k The time evolution of quantumHeisenbergHeisenberg expectation representation representation values is (eq given (eq 1). 1). by Ehrenfest’sHeisenberg theorem, representation or the (eq 1).  2 where σ and σ are Pauli matrices. The term Ω is related to tunneling probability;3  2¯hε is the When treated with Ehrenfest’s theorem, quadratic potentials,blip such approximation as the harmonicand withapplication oscillator,x Makarovz of QHD and yield Makri’s to model numerical the condensed-phase path integrals.2 with2 The a bath quantized of Heisenberg representation (eq 1). d d i i application When treated of QHDwith Ehrenfest’sd to modeli theorem, the condensed-phase quadratic potentials, with aenergy bath gap( betweenofσxq harmonicj the twoσx spinq states;j oscillators.) andgpkj, qjk, ωj+, µjg,j andσxgj (areq respectivelyj qj the) momentum,+ qj gk σxqk (4) A AII.= = Quantized[H,A[H,A] ] Hamilton DynamicsA = [H,A](1)Hamilton(1) dynamics schemecoordinate, is≈ applied frequency, to − reducedthis   system(1) mass, and so coupling that strength its results of the j canth bath be − oscillator.  compared  with exact  d i equations of motion that look exactly classical. However, more complicated potentialsharmonic will oscillators. generate k k dtdt  ¯h¯h  such Toas the obtain harmonic thedynamics oscillator,dt  yield of¯h the equations system,solutions of equations motion as well that as of providing motionThis an system were alternate has generated been analyzed solution2 and to tosolved second this in many problem. orderways, including Additionally, with use of the non-interacting this is the first A = [H,A] an infinite hierarchy of equations(1) of increasing order; the time evolution of q willblip approximation depend To obtain2 onand the withq dynamics Makarov, and Makri’sof the numerical system, path equations integrals.3 The of quantized motion WhenWhen treateddt treated  with¯h with Ehrenfest’s Ehrenfest’s theorem, theorem, quadratic quadraticThe potentials, potentials, time evolutioneq such 1. such A of as third quantum as the the orderharmonic harmonic expectation term oscillator, appeared oscillator, values is yield in given yieldapplication the by equations Ehrenfest’swhere of QHD the of last theorem, to motion modelstepHamilton in or the for eq dynamics the 4condensed-phaseσ makesxq schemej theand is applied approximation was to with this decomposed system a bath so that that of its harmonic resultsqj usingqk can= be oscillators.q comparedj qk for with exactj = k, that is, bath When treated with Ehrenfest’slook2 exactly theorem,classical. However, quadratic3 more potentials, complicated such potentials as the will harmonic  oscillator,1  yield       and the evolutionHeisenberg of representationq will depend (eq 1). on q , etc. A novel method,To obtainoscillators by Prezhdo the dynamics are uncoupled.et.solutionswere of al. the asgeneratedin well system, Similar quantized as providing equationsto decomposition second an alternate of order solution motion andwith to this were approximationseq problem. generated1. A Additionally, third toorder aresecond this is appliedterm the order first ap- to with third order When treated with Ehrenfest’s theorem,equationsequations quadratic of ofmotion motion potentials, that that look look such exactly exactly as the classical.equations classical. harmonic However, of However, oscillator, motion more thatgenerate moreeq yield 2 complicated look as complicated an follows: exactly infinite potentials classical.hierarchy potentials of However,will equations will generate generate moreof increasing complicated order; potentialsapplication of will QHD to generate model the condensed-phase with a bath of harmonic oscillators. Hamilton dynamics (QHD), ends the infinited hierarchyi ateq a2 given1. A thirdterms order. order in the Using term time appearedpeared anderivatives approximation in in the the for equations equationsσxpj , σof ofyq motionj motion, and for forσypjσx.qj and was decomposed using equations of motion that look exactlyanan infinite classical. infinite hierarchy hierarchy However, of ofequations more equations complicated of ofincreasingan increasing infinite potentials order; order;hierarchy will the the generatetimethe of time equationstime evolution evolution evolution of of increasing ofofq qwill willA depend= depend dependorder;[H,A onon the] on q time2 q,, and, evolution the evolution of q willTo obtain depend(1) the dynamics on ofq the2 system,, equations of motion   were generated to second order with dt  ¯h  eq 2 as follows:For simplicityeqh ¯ 1.was A third set order to term unity. appeared The in the set equations of first of motion order for equationsσxqj and was for decomposed the spin using system are 2 2 borrowed3 3 from many-body2 2 theory expectation  3 values1 1 of high  order operators are using broken 1eq 2 down as follows: into    an infinite hierarchy of equationsand ofand increasing the the evolution evolution order; of of theq q timewillwill depend evolution depend onand on ofq theq, etc.will,When evolution etc. A depend treated novelA novel of method,with onofq method, Ehrenfest’sqwill ,willg byk depend σ bydepend Prezhdoxq theorem, Prezhdojqk onon et. quadratic(q et.σ al.,x etc.q etc.j al.in A potentials, Ainnovel quantized2 novelσ quantizedx method,q suchmethod,j ) as by theg Prezhdok by harmonicqk Prezhdo+ et.σ oscillator,x eqet. 2 asg al. follows:k yieldqjinqk quantized+ qj gk σxqk 2 3    products   of lower1 order  operators.   For example≈    a − third order  operator  is decomposed   as,d      equations of motion1 thatk look exactly classical. However, more complicatedk potentials willk generategk σxqjqk ( σxqj 2 σx kqj ) gk qk + σx gk qjqk + qj gk σxqk and the evolution of q will dependHamiltonHamilton on q dynamics, etc.dynamics A novel (QHD), (QHD), method, ends ends by the the Prezhdo infiniteHamilton infiniteet. hierarchy hierarchydynamics al. in quantizedat at (QHD),al.a given a in given quantized order. ends order. theHamilton Using Usinginfinite an dynamics an approximation hierarchy approximation (QHD), at ends a given gthek σ infinitex order.qjqk hi( Using- σxqj an2 σ approximationx qj ≈σz) =g −k  2Ωqk σ+ y σx   gk qjqk + qj  gk σxqk (5)     ≈   −k  2    k   k    k       an infinite hierarchy of equations of increasing order; the time evolutionk of q will depend on q2dt,   k 2   k  k (4) Hamilton dynamics (QHD), endsborrowed theborrowed infinite from from hierarchy many-body many-body at a theory given theory order.expectation expectationborrowed Using values an from values approximation ofmany-body ofhigh high order order theory operators operators expectation are are broken broken values down down of intohigh into order operators are broken down into 2 2 ABCerarchy at a AgivenBC order.+ UsingB( ACσ anxq approximationj + Cσx ABqj )borrowed2 Agk q Bfromk +C gj σx ( q d q(jσxqj)(2) +σx qqj ) gkqkgk+ gσj σxxq(kqj (4)qj ) +qj gk σxqk (4) and the evolution of q2 will depend on q3 , etc. A novel method, by Prezhdo et. al.1 in quantizedj ≈   −      2    −  2       ≈    ≈   −    −    ( σ q  σ −σqx  ) = g 2qε σ+y kg+ 2σ (gqj σyqjq ) + q k g σ q (4) (6) borrowed from many-body theoryproducts expectationproducts of of lower values lower order order of high operators. operators. order operators For Forproducts example example are of abroken third lowera third downorder order ordermany-body intooperator operators. operator theory is decomposedisFor expectation decomposed example  values as, a thirdas, of high order order operator operatorsk is are decomposed x j dt as,x j −k k kj x j j j k x k Hamilton dynamics (QHD), ends the infinite hierarchy at a given order. Using an≈ approximation where the − last  step in eq 4 makes the approximation  j that q q − = q q for j = k, that is, bath   where the last k step in eq 4 makes j k the j approximationk  k that products of lower order operators. For example a third order operator is decomposed as, oscillators are uncoupled. Similar decomposition and approximations are applied to third order Using secondborrowed order from equations, many-bodybroken down theory QHD into expectationproducts can reproduce of lower values order of important high operators. order operators quantum For example are effects broken suchdown into asd zero-point where the last step in eq 4 makes the approximation thattermsq inq the time= derivativesq q for forσxpj ,j σ=yqj ,k and, thatσypj . is, bath ABCABC A ABCBC+ +B BproductsACAC+ of+C lowerCABABC orderAB operators.2 A2 AABBCB ForC C example+ B aAC third+ orderC(2)where operatorAB(2) the is2 last decomposedA stepB inC eq as, 4 makesj k theσ approximationy j =k (2)2Ωσz that+ 2ε qjσqkx that= 2qj is,q gkj bathσforxqjj =. oscillatorsk, that is, bath (7)    ≈ ≈  energy    and  tunneling.   a third These − order − ≈ effects  operator are is decomposed preserved   whileas,  reducing −  the complexity  For simplicity of dt ah ¯quantumwas  set to unity.− The set of first  order − equations  for the spin system are ABC A BC + B AC + C AB 2 A B C oscillators(2) are uncoupled. Similar decompositionoscillators are uncoupled. and approximations Similar decomposition are applied and approximations to third orderj are applied to third order problem to one that is nearly as simple than classical mechanics. are uncoupled.d Similar decomposition and approximations   ≈          −     ABC A BC + B AC + C AB 2 A B C (2) σz = 2Ω σy (5) UsingUsing second second order order equations, equations, QHD QHD canUsing can reproduce reproduce second importantorder important equations,terms quantum quantum in the QHD time effects effects can derivatives reproducesuch such as as zero-point for zero-pointimportantσxptermsj , σ in quantumy theqj , time and (2) derivatives effectsσypj . such for asσxp zero-pointj , σdtyqj  , and  σypj .   ≈          −  The bath terms evolveare applied as  to d third order terms in the time derivatives for  For simplicity h ¯ was set to unity. The setσ of= first2ε orderσ + 2 equationsg σ q for the spin system(6) are Using second order equations, QHDenergyenergy can and reproduce and tunneling. tunneling. important These These quantum effects effects areenergy effectsare preserved preserved and such tunneling. whileas while zero-point reducing reducing These the effects the complexity complexity are preserved of of a quantum a while quantum reducing the complexity of a quantumdt x −  y j y j Using second order equations,For simplicity QHD canh ¯ was reproduce set to important unity. Thequantum set effects of first such order as zero-point equations for the spin systemj are d pj  d q = (8) energy and tunneling. These effectsproblemproblem are to preserved to one one that that while is nearlyis nearly reducing as as simple simple theproblem complexity than than classicalenergy to classical one of and mechanics.that a tunneling. mechanics.quantum is nearly These as effectssimple are than preserved classical1 while mechanics. reducing the complexity of ad quantum σ j = 2Ωσ + 2ε σ 2 g σ q . (7) CMDITRσ = Review 2Ω σdtdt of y Undergraduate −  µzj Research x − Vol.j 2x No.j 1 Summer 2005 39 (5) d z y j problem to one that is nearly as simple than classical mechanics. problem to one that is nearly as simple than classical mechanics. dt     σz = 2Ω σy d (5) The bath terms evolve as 2 dt    d pj = µjωj qj + gj σz . (9) σx = 2εdtσy + 2 −gj σyqj    (6) 1 1 1 dt  −   d pj  d1 qj j =   (8) σx = 2ε σy + 2 gj σyqj dt  µj (6) 1 dt  −    d  d 2 d pj 2= µjωj qj + gj σz . (9) j σy = 2Ω σ2z +dt 2ε σx − 2   gj σxqj . (7)  dt  − qj  =  p −jqj s   (10) d dt  µj   j d 2 2  σy = 2Ω σz + 2ε σx 2 gj σxqj . qj = pjqj s (7) (10) dt  The− bath terms evolve  as −   dt  µj   j 2  2 d pj qj =   (8) The bath terms evolve as dt  µj

d 2 d pj pj = µjωj qj + gj σz . (9) qj =   dt  −     (8) dt  µj

d 2 d 2 2 pj = µjωj qj + gj σz q.j = pjqj s (9) (10) dt  −    dt  µj  

2 d 2 2 qj = pjqj s (10) dt  µj  

2 ω = 17ε/8

ω = 2ε/4

ω = 7ε/4

Coupled Oscillators with ω < 9ε/4 ω = 3ε/2 υ) > ( z σ P<

υ

Coupled Oscillator with ω = 9ε/4 υ) > ( z σ P<

0.25 υ

0.2 0.25

0.15 0.2

0.1 0.15

0.05 0.1

1e+06

0 0.05

1e+06

100000 0

100000

10000

III. The Spin-Boson Model III. The Spin-Boson Model III. The Spin-Boson Model III. The Spin-Boson Model 2 The Spin-Boson model consists of a single two-state spin system coupled bilinearly to a bathd of pj 2 2 The Spin-Boson model consists of a single two-state spin system coupled bilinearly to a bath of pjqj s =   µjωj qj + gj σzqj (11) harmonicThe Spin-Boson oscillators model through consists the ofvibrational a single two-state coordinate. spin The system model coupled has applications bilinearly to to a condensed bathdt of  µj −   10000  harmonic oscillatorsThe Spin-Boson through model the vibrational consists of coordinate. a single two-state The model spin has system applications coupled tobilinearly condensed to a bath of harmonicphase nonadiabatic oscillators through electron the transfer, vibrational low temperature coordinate. The proton model transfer, has applications and macroscopic to1000 condensed quantumd phase nonadiabaticharmonic oscillators electron transfer, through low the temperaturevibrational coordinate. proton transfer, The model and macroscopic has applications quantum to condensedp2 = 2µ ω2 p q + 2g σ p . (12) phasecoherence nonadiabatic in SQUIDS. electron3 The transfer, system low is described temperature by the proton Hamiltonian transfer, and macroscopic quantumdt j  − j j  j js j z j coherencephase in SQUIDS. nonadiabatic3 The systemelectron is transfer, described low by temperature the Hamiltonian proton transfer, and macroscopic quantum 3 3 coherencecoherence in SQUIDS. in SQUIDS.The system3 The systemis described is described by the Hamiltonian by the HamiltonianThe first order spin terms depend on second order terms that mix spin and bath degrees of freedom. coherence in SQUIDS. The system is describedp2 1 by the Hamiltoniang p2 j 2 2 j d 1 H =h ¯Ωσx + ¯hεσz + j 2 1+ 2 µ2 j2ωj qjgj (σzqj + qjσz) (3) H =h ¯Ωσx + ¯hεσz + pj2+µ µ1jpω22 q −g(jσ2zqj + qjσz) (3)σzqj = 2Ω σyqj + σzpj (13) j j pj j 2j 21 2 2 gj dt    µ   H =h ¯ΩHσx=+h ¯ ¯hεσΩσxz + ¯hεσ2zµ+j +2 µjjω+j qj−1µj2ω2q(σ2 zqjg+j (qσjzσqzj)+ qjσz) (3) (3) j H =h ¯Ωσ +j ¯hεσ+ + µ ωj qj (σ q + q σ ) (3) 1000 x  z 2µj 22µj 2−j 2j j − 2 z j j z j  j 2µj 2 − 2  100 d 2 where σx and σz are Pauli matrices. The termj  Ω is related to tunneling probability; 2¯hε σiszp thej = 2Ω σypj µjωj σzqj + gj (14) where σx and σz are Pauli matrices. The term Ω is related to tunneling probability; 2¯hε isdt the    −   whereenergyσwherex gapand betweenσσxz andare Pauliσ thez are two matrices. Pauli spin matrices.states; The and term Thepj, Ωq termj is, ω relatedj, Ωµj is, and related to tunnelinggj are to respectively tunneling probability; probability; the 2¯ momentum,hεdis the 2¯hε is the 1 energy gapwhere betweenσ and theσ twoare spin Pauli states; matrices. and pj, Theqj, ω termj, µj, Ω and is relatedgj are respectively to tunneling the probability; momentum, 2¯hε is the x z σxqj = 2ε σyqj + σxpj energycoordinate, gapenergy between frequency, gap between the two reduced the spin two states;mass, spin and and states; couplingpj, andqj, ωpj strengthj,, µqj, andωj, ofµgjj the,are andjth respectivelygj bathare respectively oscillator. the momentum, the momentum, coordinate,energy frequency, gap between reduced the mass, two andspin coupling states; and strengthpj, qj, ofωj the, µj,jth and bathgj are oscillator. respectively thedt momentum, −   µj   coordinate,Thiscoordinate, system frequency, has frequency, reducedbeen analyzed reduced mass, and and mass, coupling solved and incoupling strength many ways, strength of the includingj ofth the bathj useth oscillator. bath of the oscillator. non-interacting 2 2 This systemcoordinate, hasbeen frequency,2 analyzed reduced and solvedmass, and in many coupling ways, strength including of the usej ofth the bath non-interacting3 oscillator. +2gj σy ( qj qj ) blipThis approximation systemThis system has2 beenand has analyzed been with analyzed Makarov and solved and and in solved Makri’s many in ways, numericalmany including ways, path including use integrals. of3 the use non-interacting ofThe the non-interacting quantized     −   100 blip approximationThis systemand has with been Makarov analyzed and and Makri’s solved numerical in many ways, path includingintegrals. useThe of 10 the quantized non-interacting blipHamilton approximation dynamics2 and scheme with2 is Makarov applied to and this Makri’s system so numerical that its results path integrals. can be compared3 The3 quantized with exact +2 qj gk σyqk Hamiltonblip dynamics approximation scheme is2 appliedand with to thisMakarov system and so thatMakri’s its results numerical can be path compared integrals. with3 The exact quantized     Hamiltonsolutions dynamics as well as scheme providing is applied an alternate to this system solution so to that this its problem. results can Additionally, be compared this with is the exact first k solutionsHamilton as well as dynamics providing scheme an alternate is applied solution to this to system this problem. so that its Additionally, results can thisbe compared is the first with exact  solutionsapplication as well of QHD as providing to model an the alternate condensed-phase solution to with this a problem. bath of harmonic Additionally, oscillators. this is the first +2( σyqj σy qj ) gk qk (15) applicationsolutions of QHD as to well model as providing the condensed-phase an alternate with solution a bath to ofthis harmonic problem. oscillators. Additionally, this is the first   −      applicationToapplication obtain of QHD the of dynamics to QHD model to of the model the condensed-phase system, the condensed-phase equations with of a motion bath with ofa were bath harmonic generated of harmonic oscillators. to second oscillators. order with k To obtainapplication the dynamics of QHD of to the model system, the equations condensed-phase of motion with were a bath generated of harmonic to second oscillators. order with 10 1 d 2 eqTo 1. obtain A thirdTo the obtain order dynamics the term dynamics appeared of the system, of in the the system, equations equations equations of of motion motion of motion were for generatedσx wereqj and generated to was second decomposed to ordersecond withσ usingorderxpj with= 2ε σypj µjω σxqj eq 1. A thirdTo order obtain term the appeared dynamics in of the the equations system, equationsof motion of for motionσxqj wereand was generated decomposed to seconddt using order with−   − j   eqeq 1. 2 A aseq third follows: 1. A order third term order appeared term appeared in the equations in the equations of motion of motion for σxq forj andσxq wasj and decomposed was decomposed using using eq 2 as follows:eq 1. A third order term appeared in the equations of motion for σxqj and was decomposed using+2( σ p σ p ) g q eq 2 aseq follows: 2 as follows:    y j −  y j k k eq 2 as follows: k gk σxqjqk ( σxqj 2 σx qj ) gk qk + σx gk qjqk + qj gk σxqk  1 ω ε ω ε gk σxqjqk ( ≈σxqj 2 −σx  qj)  gk qk + σx   gk qjqk + qj   gk σxqk 0.1  Coupled Oscillators with < 9 /4 Coupled Oscillator with = 9 /4 k k k k +2 pj gk σ1e+06yqk ω = 3ε/2 1e+06 gk σxqjgqk ≈σ q( qσxqj −( σ2 qσx qj2 )σ qgk )qk +g σqx + gσk qjqk g+ qqqj + gqk σxqkg σ q ω ε  k x j k x j Coupled Oscillatorsx with ω j< 9ε/4 k k x k j k Coupled Oscillatorj with ω =k 9ε/4 x k     = 7 /4 k g σ q q ( σ q 2 σk q ) g q k+ σ g q q k + q g σ q ω = 2ε/4  k ≈x j k ≈ − x j −  x j  k k ω xε  k j k  j k x k ω ε k 100000 100000  1e+06  Coupled Oscillators with < 9 /4 1e+062 2 Coupled Oscillator with = 9 /4 ω = 17ε/8 k k   ≈   −  k  k ω = 3ε/2  k   k   k   k   ω ε ω ε ( σxqj σx 1e+06qj ) gkω qkε + gj σx ( q qj ) + qj 1e+06 gk σxCoupledqk Oscillators(4) with < 9 /4  Coupled Oscillator with = 9 /4  k  k = 7 /4  2 kj ω = 3ε/22  k 10000  ( σ q σ q ) g q ω += 2ε/4g σ ( q ωq ε ) + q 1e+06g σ q (4) +2g ω σε ( p q 1e+06p q ) (16) 10000  100000≈x j  −x  j  k k  j x  j 1000002 = − j7 /4  2  j  k x k  j = 3 y/2 j j s j j 100000 k ω = 17ε/8 ω = 2ε/4 2 2 k ω = 7ε/4 ≈ ( σxqj −  σx qj ) gk qk + gj σx ( q −  qj2 ) + qj2 100000 gk σxqk (4)    1000 −    ( σxqj σx qj ) gk qk + gjj σxω = (17ε/8qj qj 100000) + qj gk σxqk (4) ω = 2ε/4 0.1 10000 ( σ q σk q ) g q + g σ ( q q )k +q 0.01 g σ q (4) ω ε 100000 1000 ≈   − x j  x j   k k   j 10000 −x  j j  j dk x k = 17Coupled/8 Coupled Oscillators Oscillators with ω ( 1000 10000 y j 1e+06 1e+06 1e+06z j x j y j 1e+06 1e+06 1e+06 > ( z 10000 ω = 3ε/2ω = 3ε/2 ω = 3ε/2 z 100 σ where the last step in eq 4 makes the approximation k that q q = q q for j = k, thatk is, bath σ 1000  j k 1000 j k dt  −     ωµ = 7ε/4ω = 7ε/4  Coupled Oscillators 10 with ω < 9ε/4 j ω = 7ε/4 Coupled Oscillator with ω = 9ε/4 υ) υ) P< where the last step in eq 4 makes the approximation that q q = q q for j = 1000k 1000, that is, bath 100000 100000 ω = 2ε/4ω = 2ε/4 ω ε P< 100 j k  j  k   100000 = 2 /4 100000 100000 100000 > ( > ( υ) QUANTIZED HAMILTON DYNAMICSυ) APPLIED TO CONDENSED 1e+06 PHASE SPIN-RELAXATION 1000 ω = 17εω/8 = 17ε/8 ω ε 1e+06 z whereoscillators the last are step uncoupled. in eq 4 makes Similar the decomposition approximation 100 and that approximationsq q z = 100 q q arefor appliedj = k, that to third is, bath order 2 2 ω = 3ε/2 = 17 /8 10 σ  σ     υ) j k j k υ) > ( where the last step in eq 4 makes the approximation that qjqk = qj q> ( k for j = k, that is, bath 1 ω = 7ε/4 z 10 z 100 100 10000 10000 2g j10000σx ( qj qj ) P< P< σ σ > ( oscillatorswhere are uncoupled. the last step Similar in eq 4 decomposition makes the approximation and approximations that qjqk are= appliedqj qk tofor 0.001 thirdj = k order, that is, bath > ( 0.01ω = 2ε/4 10000 10000 z           100000 z 100 100000 10000 σ 10 −     − σ  1 P< terms in the time derivatives for σ p , σ q , and σ p . 10 P< ω = 17ε/8 oscillatorsoscillators are uncoupled. are uncoupled. Similar Similar decompositionx j decompositiony j and approximationsy andj approximations  are applied are 10 applied to third to order third order 0.1 P< 1 10 Coupled Oscillators 1000 with 1000 ω < 9ε/4 1000 P< Coupled Oscillator with ω = 9ε/4 terms inoscillators the time derivatives are uncoupled. for σx Similarpj , σyq decompositionj , and σypj . and approximations are applied to third order 10000 1000 1000 1000 1 1e+06 2 qj gk σxqk 1e+06 10 10000 0.1 υ) υ) ω ε 0.01 υ) υ) υ) 1 = 3 /2 υ) termsFor interms the simplicity time in the derivativesh ¯time was0.1 derivatives set for to unity.σxp forj , Theσσypq setj , of andσ firstq σ,y order andpj . σ equationsp . for the spin 1 system are 100 100−  100   > ( > ( x j y j y j ω ε > ( > ( > ( 1 1000 = 7 /4 > ( z z z z z terms in the time derivatives for 0.1 σ p , σ q , and σ p . 100 100 z σ σ For simplicityh ¯ was set to unity. The  setx ofj  firsty orderj equations y j for the spin system are ω εk σ σ 100 σ       100000 = 2 /4 0.001 1 1000 σ 0.01 For simplicity ħ was set to unity. The 0.1 set of first order equa 0.1 - 10 10 ω ε 100000 υ) υ) P< P< 0.01 10= 17/8 P< P< P<       0 0.05 0.1 0.15 0.2 P< 0.25 0 0.05 0.1 0.15 0.2 0.25 For simplicityFor simplicityh ¯ was seth ¯ was to unity. set to The unity. set The of first set orderof first equations order equations for the forspin the system 0.1 spin systemare are 100 > ( 0.01 > ( z z 10 10 For simplicityh ¯ wasd set to unity. The set of first order equations for the 10000 spin system are 0.1 υ 100 10 υ σ 0.001 0.01 0.01 1 1 2( σ x1 qj σx qj 10000) gk qk σ (17) tions for the spin system are 10 P< 0d 0.05σz =0.1 0.001 2Ω 0.15σy 0.2 0.25 0 0.05 0.1 0.01 0.15 0.2 (5) 0.25 −   −      P< 1000 1 1 1 dσdt = 2Ω υσ 0   0.05 0.1 0.15 0.2 0.25 0.001υ 0 0.05 (5) 0.1 0.1 0.15 0.1 0.2 0.1 0.25 0.01 1000 k 10 z d y 0 0.05 0.1 0.15 1 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 υ) d υ υ υ)  dt σz =σ 2Ω =σy 2Ω σ 100 (5)Coupledυ Oscillators(5) with ω > 9ε/4 Coupledυ 0.1to Multiple 0.1 Oscillators > ( z y 0.01 0.01 Coupled Oscillators> ( with ω < 9ε/4 0.1 Coupled Oscillator with ω = 9ε/4 z d d 0.01 z 100 1 σ σ = 2Ω σ 0.1(5) σ dt  dt z   y 1e+06 ω ε 1e+06 2 d σ =  2ε σ + 2 g σ q 10(5) (6) 1e+06 = 19 /8 ω = 3ε/2 1e+06 P< x y j y j σypj = 2Ω σzpωj +ε 2ε σxP< pj µjω σyqj dt    0.001 0.001 0.001 = 10 /4 Coupledj Oscillatorsω = 7ε/4 with ω (18)> 9ε/4 0.01ω ε 0.01 0.01 Coupled to Multiple Oscillatorsω ε 0.01 10 Coupled Oscillators with < 0.19 /4 Coupled Oscillator with = 9 /4 dσdtx  =d 2ε−σy + 2 gj σyqj  100000 dt(6)  0 − 1000000 0.05 0.05 0  0.1 0.050.1  0.15 0.15 0.1 − 0.2100000 0.150.2 0.25 ω = 0.25 20.2ε/4  0.25 0 0 0.05 0.05 0 0.1 0.050.1 0.15 0.15 0.1 0.2 0.150.2 0.25 0.25 0.2 0.25 j 1 1e+06 1e+06 ω = 100000 19ε/8 1e+06 1e+06 dt σ d= − 2ε σ + 2 g σ q (6) υ υ υ ω = 17ε/8 ω = 3ε/2 υ υ υ x σx = ωy ε 2ε σy +j 2 y j gj σyqj (6) ω = 10ε/4ω ε Coupled Oscillators with > 9 /4 j  Coupled to Multiple 10000Oscillators 0.001 1 0.01 = 7 /4 dt  σx− =  2ε σy + 2 gj σyqj (6) 100002( σ p 100000σ p ) 10000g q ω ε 1e+06 dt  −  Coupled Oscillators with ω > 9ε/4 1e+06  0.1 Coupled to Multiple Oscillators 0 0.05x j 0.1 x 100000 0.15j 0.2 k k 0.25 0 10000 = 2 0.05/4 0.1 100000 100000 0.15 0.2 0.25 d ω = 19j ε/8 (6) ω ε ω = 17ε/8 dt 1e+06 − ω ε j   1e+06 Coupled Oscillators with > 9 −/4   −  υ     Coupled to Multiple Oscillators υ = 10 /4 ω = 19ε/8 1000 1000 10000 0.1 d σy = 2Ω σz + 2ε σx j 2 gj σxqj . 1e+06 0.01 (7) 10000 1e+06 k 1000 100000 100000ω = 10ε/4 ω = 19ε/8 1000 10000 10000 υ)  ω ε υ)  υ) dσdt = 1000002Ω σ + 2ε σ 2 g σ q . (7) = 10 /4 υ) y  d − z   x   − j x j  100000 100 100 1000 > ( > ( > ( 0.001 1000 0.01 > ( z 10000 j 100000 z z d 100000 100 z 1000 σ dt σy = − 2Ω σz + 2ε σx − 2 gj σxqj . (7) 0.15 0.2 0.25 σ 100 0.2 0.25 σ σy = 2Ω σz + 2ε σx 100002 gj σxqj . 0 0.05 0.1 (7) 2 pj CoupledCoupled Oscillatorsgk Oscillatorsσx withq kω >with 9ε /4ω > 9ε /40 ω 0.05ε 0.1 0.15σ CoupledCoupled to Multiple to Multiple Oscillators 1000 Oscillators υ) 10000 j  (7) 10 Coupled Oscillators with > 9 /4 υ) Coupled to Multiple Oscillators P< P< υ) σ = 2Ω σ + 2ε σ 2 g σ q . (7) 10 100 υ) P< dt  y−   z  − x   j x j 10000 υ υ P< > ( 1000 dt  −     −   10000 1e+06 1e+06−  1e+06  100 1e+06 1e+06 1e+06 > ( z ω εω ε z > ( j 10000 = 10 19 /8 = 19 /8 ω = 19ε/8 > ( 100 z z σ dt  −     − 1000 j   k 10 σ 100 σ 1000 ω = 10εω/4 = 10ε/4 σ 1 1 Coupled Oscillators 10 with ω > 9ε/4 ω = 10ε/4 Coupled to Multiple Oscillators υ) υ) P< The bath terms evolve as  j 1000 1000  10 P< P<  100000 100000 P< 100 100000 100000 100000 100000 > ( > ( υ) υ) 1e+06 1000 1 1e+06 z The bath terms evolve as 100 z 100  2 σ ( p q p q )ω = 19ε/8 1 (18) 10 10 σ σ 0.1 x j j s j j υ) υ) > ( > ( 0.1 1 ω = 10ε/4 z 10 The bath terms evolve as z 100 100 10000 10000 10000 1 P< P< σ σ > ( The bath terms evolve as −     −> (    10000 10000 z 100000 z The bath terms evolve as 100 0.1 100000 10000 σ 10 σ 1 P< The bath terms evolve as d pj 10 P< 0.01 0.01 0.1 1 10 0.1 0.1 P< 1 d q =p There are eleven 10 equationsCoupled (2 first Oscillators(8) 1000 order, with 1000 ω > 9ε/4 6 mixed 1000 terms, andP< 3 second order)Coupled for to Multiple every Oscillators bath oscillator j j   10000 1000 1000 1000 1 1e+06 0.001 1e+06 10 0.01 10000 0.1 υ) υ) 0.001 ω ε 0.01 υ) υ) 0.01 υ) dqdt =  p µ 1 (8) = 19 /8 υ) 0.1 0.1 j  d j j pj 1 0 0.05 There are0.1 100 eleven 100 0.15 0equations 100 0.2 0.05 (2 0.25 first 0.1 0.01 order, 0.15 6 mixed 0 0.2 terms, 0.05 0.25 and 0.1 0.15 0 0.2 0.05 0.25 0.1 0.15 0.2 0.25 > ( > ( and three for the spin terms that are requiredω ε to describe motion of the system.> ( > ( > ( 1 1000 = 10 /4 > ( z z z z z dt q 0.1d= µ  pj (8) υ 100 υ100 z σ σ  j j υ σ σ 100 υ σ qj =   100000(8) (8) 0.001 1 1000 σ 0.01 0.1 0.1 10 10 0.001 100000 0.01 υ) υ) P< P< 0.01 q = (8) 10 P< P< P< dtd  j µ   3 second order) for every bath oscillator 0 and 0.05 three 0.1 for the 0.15 spin 0.2 P< 0.25 0 0.05 0.1 0.15 0.2 0.25 dt  j 2µj 0.1 100 0 0.05 0.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 > ( 0.01 > ( z z 10 10 d p dt=  µ ω µqj + g σ . 10000 (9) 0.1 υ υ 100 10 υ υ σ 0.001 j 2 j j j j z 0.01 0.01 1 1 1 10000 σ 10 P< terms that are required to describe motion of the system. P< 0 0.05 dpdtj 0.1 0.001=d 0.15 µj−ω 0.2j 2qj + 0.25gj σz .  0 0.05 0.1 0.01 0.15 0.2 (9) 0.25 2 IV. Analysis 1000 of Dynamics 1 1 1 dt pj υ d =0 − µ0.05jω q j0.1 + gj 0.15σz . 0.2 0.25 0.001υ (9) 0 0.05 (9) 0.1 0.1 0.15 0.1 0.2 0.1 0.25 0.01 1000 10 pj =j µjωj2 qj + gj σz . 0 0.05 0.1 0.15(9) 1 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 υ) p = µ ωυ q + g σ . υ (9) υ) dt  dt j−   j j j  j z 100 υ υ 0.1 0.1 > (   −     Figure 2: Power Spectra of 0.01 Asymmetric 0.01 SystemsCoupled Oscillators> ( Coupled with ω > 9ε/4 to Single Oscillators, 0.1 usingCoupled the to sameMultiple Oscillators z 0.01 z 100 1 σ dt  −     0.1 σ 10 1e+06 ω = 19ε/8 1e+06 P< ANALYSIS OF DYNAMICSP< d 2 It has been demonstrated that 0.001 the 0.001 spin-boson 0.001 system can evolveω with = 10ε/4 many 0.01ω differentε 0.01 0.01 patterns 2 conditions as Figure 1. The power 0.01 Figure spectra 2: of Power the asymmetric Spectra 10 ofCoupled systems Asymmetric Oscillators with show > 0.19 /4 two Systems major Coupled oscillations. to Single Oscillators,Coupled to Multiple Oscillators using the same 0 1000000 0.05 0.05 0 0.1 0.050.1 0.15 0.15 0.1 0.2 0.150.2 0.25 0.25 0.2 0.25 0 0 0.05 0.05 0 0.1 0.050.1 0.15 0.15 0.1 0.2 0.150.2 0.25 0.25 0.2 0.25 d 2 qj =2 pjqj s ( 101 ) (10) 1e+06 100000ω ε 1e+06 ranging from coherent oscillations, to incoherent relaxationυ υ andυ complete localization depending = 19 /8 onυ υ υ dqdt = 2pµjqj s  A high frequency oscillation,(10) 0.001 dependentconditions on Ω as and Figureε and 1. Thethe slower power oscillation spectra 0.01 ofω = 10 theisε/4 dependent asymmetric on systemsω. show two major oscillations. Figure 2: Power Spectraj 2 of Asymmetricd 2 j 2 Systems Coupled to Single Oscillators, using It has the been same demonstrated 10000 that the spin-boson 1 system can Figuredt q 2: Powerd= µj  Spectrapjqj 2s of Asymmetric Systems Coupled 0.1 to Single(10) Oscillators, 0 using 0.05 the 0.1 same 100000 0.15 0.2 0.25 0 10000 0.05 0.1 100000 0.15 0.2 0.25 j qj2 = 2 pjqj s (10) υ υ dt q µ = p pjqj s FigureFor 2: an Power oscillator Spectra where ofω Asymmetricis(10) the natural 1000A high Systems frequency frequency Coupled of oscillation, the to spin Single systemdependent Oscillators, more on complicated Ω using and theε and same behavior the slower oscillation is dependent on ω. conditions as Figure 1. Thed  powerdt j  spectraj  jµ j of the2 asymmetric2 systems show 0.01 two majorevolve oscillations. with many different patterns 10000ranging 0.1 from coherent os- conditions asdt Figure  1. Theµj  power spectra of the asymmetric systems show two major oscillations. 1000 10000 υ) pjqj s = µjωj qj + gj σzqj (11) (11) υ) 2 conditionsoccurs. as Figure 1. The power spectra For100 an of the oscillator3 asymmetric where systemsω is the show natural two frequency major oscillations. of the spin system more complicated behavior > ( A high frequency oscillation, dependent on Ω and ε and the slower oscillation 0.001 iscillations, dependent to on incoherentω. relaxation and1000 0.01 complete localization > ( z dt µ z   j −     0.15 0.2 0.25 100 0.2 0.25 σ A high frequency oscillation,2 dependent on Ω and ε andthe 0 Figure slower 0.05Figure oscillation 2: 0.1 2: Power Power Spectrais dependent Spectra of ofAsymmetric on Asymmetricω. 0 Systems Systems0.05 Coupled 0.1 Coupled 0.15σ to to Single Single Oscillators, Oscillators, 1000 using using the the same same υ) 2 Figure 2: 10 Power Spectra of Asymmetric Systems Coupled to Singleυ) Oscillators, using the same P< A high frequency oscillation, dependentυ occurs. on Ω and ε3 and 100 the slower oscillationυ P< is dependent on ω. > ( For an oscillator where ω is the natural2p frequency of the spin system more complicateddepending behavior on various parameters. The particular case studied > ( z d d j 2 2 2 2 10 z 100 σ For an oscillator2 where ωpis2 the2 natural frequency of the spin system more complicated 1 behavior σ pdq =   j µ ω q + g σ q conditionsFigureconditions 2:conditions as(11) Power Figureas Figure Spectra as 1. Figure The1. The of power 1. Asymmetric power Thespectra 10 powerspectra of Systems spectra the of the asymmetric asymmetricof Coupled the asymmetric systems to systems Single show systems Oscillators, show two two show major major using two oscillations. majoroscillations. the same oscillations. P< j pjj s = 2µjωj jpjqjj sj2+ 2gjj Forσzzpjj an. oscillator(12) where 3ω is the(12) natural frequency of the spin system more complicated behaviorP< occurs. p q =   µ ω q + g σvariousq parameters.with TheQHD particular is (11)the asymmetric case studied system, withε > Ω QHD > 0 and is thewatching asymmetric 1 system, ε > Ω > 0 dtdt j j s −µj −  j j j  j z j 0.1 10 occurs. Aconditions highA high frequencyA frequency high as Figure frequency oscillation, oscillation, 1. The oscillation, powerdependent dependent 1 spectra dependent on on Ωof andthe Ω and on asymmetricε and Ωε and the theε slowerand systems slower the oscillation slower oscillationshow two oscillation is major dependent is dependent oscillations. is dependent on onω. ω. on ω. dt  µj −   occurs.  various parameters.3 The particular case studied with QHD is the asymmetric system, ε > Ω > 0 The first order spin terms depend d on second order terms that mix spinand and watching bath degrees a system,a ofsystem, freedom. starting starting withwith 0.01 σx(0) = 1,relax. relax. The The resulting resulting equa equations- 0.1 of the spin-boson 1 2 2 ForAFor high an an oscillator frequencyFor oscillator an oscillator where oscillation, where ω is whereω theis dependent the naturalω 0.1 naturalis the frequencyon natural frequency Ω and frequency ofε and theof the spin the of spin slower system the system spin oscillation more system more complicated complicated more is dependent complicated behavior behavior on ω. behavior 3 pdj 2= 2µjωj pj2qj s + 2gj σFigurezpj . 2: Power Spectra of Asymmetric(12) Systems Coupled to Single Oscillators, using the same various parameters. The The particular first order case spin studied terms depend with on QHD second is order thesystem asymmetric terms were that then system,tions numerically of εthe > spin-bosonΩ > integrating 0.0010and system watchingusing were athen a system, 4th-order numerically starting Runge-Kutta integrating with σx(0) algorithm. 0.01 = 1, relax. The The initial 0.1 resulting equations of the spin-boson dt  p 3=− 2µjω pjqj s + 2gj σzpj . (12) 0 0.05 0.1 0.01 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 variousd parameters.j The particular1 j case studiedconditions with as QHD Figure isoccurs.3For the 1.occurs. The an asymmetric oscillatoroccurs. power spectra system, where ofωε theis > theΩ asymmetric> natural0 frequency systems show of the two spin major system oscillations. moreωj µj complicated behavior dt  −   various parameters. The particular case studied with QHDυ is the asymmetric system, 2ε > Ω > 0 υ and watching a system,mixσ startingzq spinj and= with bath 2Ω σdegreesσyq(0)j + of= freedom. 1,σ relax.zpj The resultingconditions equations for the ofusing bath the a spin-boson4th-order oscillators(13) Runge-Kuttasystem give minimal were algorithm. then 0.001 bath numerically The energy initial andconditions integrating uncertainty: usingp a 4th-order= 0.01and Runge-Kutta algorithm. The initial The first order spin terms depend on second orderx terms that mix spin and bath degrees of freedom. 0 0.05 0.1 0.15 0.2 0.25j 2 0 0.05 0.1 0.15 0.2 0.25 anddt watching  a system, starting µj  with σx(0)A= high 1, relax. frequency The oscillation, resultingoccurs. equations dependent of theon Ω spin-boson and ε and the slower oscillation is dependent  on ω. 2 ωj µj The first order spin terms depend on second order terms thatand mix watching spin2 and1 bath a system, degrees starting of freedom. with σ (0) = 1, relax. The resulting equationsυ of the spin-boson υ system were then numerically integrating using a 4th-order  Runge-Kuttaq = algorithm.. Thefor other the bath The terms oscillators initial areconditions allx give set minimal initially for bath the to energy zero. bath and oscillators uncertainty: give minimal bath energy and uncertainty: pj = 2 and system were then numerically integrating usingj a 4th-order2µj ωj Runge-Kutta algorithm. 3 3  The3 initial d 1 2 systemFor an were oscillator then where numericallyvariousvarious2ω parameters.isvarious theparameters.ω integratingj µj natural parameters.2The frequency usingThe particular1 particular aThe 4th-order ofparticular the case casespin studied Runge-Kuttastudied system case with studied morewith QHD algorithm. QHD complicated with is the is QHD the asymmetric The asymmetricis behaviorthe initial asymmetric system, system,ε system, >εΩ >>Ω 0ε> >0 Ω> 0 conditions for the bath σdz oscillatorspqj = 2Ω giveσypq minimalj + µjωσ bath1zσpzj qj energy+ gj and uncertainty: p = (13)(14)andand q 2 = ωj µ.j The The other other terms terms are all are set all ini- set initially to zero. conditions for the bath oscillatorsj give minimal bath energy(13) andj uncertainty:2 jp3 =2µj ωj and 2 ωj µj 2 1 dt σzqj = 2Ω σyq −j µ+j   σzpj conditionsoccurs. for the bathandvariousand watching oscillatorsFigure watching parameters.and 2:watching a(13) system, give Powera system, minimalj The a Spectrastarting system, starting particular2 bath ofwith starting Asymmetricwith energy caseσx(0)σ withx studied and(0)= 1,σ= Systemsuncertainty:x(0) relax. 1, with relax.= QHD The 1, Coupled relax.The resulting isp resulting the The to= asymmetric Single resultingequations equationsand Oscillators, equations system,of the of the spin-boson usingε spin-bosonof > theΩ the> spin-boson0 same q = . The other2 dt terms1  are all set initially µ toj  zero. tially to zero.        j 2 j 2µj ωj q d = . The other terms1 are all set initially2 A. to1 Baths zero. ofsystem Singlesystem were Oscillatorswere then then numerically numerically integrating integrating using using a 4th-order a 4th-order Runge-Kutta Runge-Kutta algorithm. algorithm. The The initial initial   j d 2µj ωj 2 and watchingconditionssystem a asweresystem, FigureFigure then starting 1. numerically 2: The Power with power Spectraσ integrating spectrax(0) = of of1, Asymmetric therelax. using asymmetric The a 4th-order resulting Systems systems Runge-Kutta equations Coupled show two ofto algorithm.the Single major spin-boson Oscillators,oscillations. The initial using the same   σxqj = 2ε σyqj + σxpj qj = . The other terms are all set initially to zero. ωj µωj j µj σdzpj = 2Ω σypj µjωj σz2qj +gj  2µj ωj 3 (14) A. Baths of Single Oscillators 2 2 2 ωj µj dt σzpj =−  2Ω σyp −j µjµjω σzqj various+ gj parameters.(14) conditionssystemconditionsTheA particular high wereconditions forfrequency for(14) then the the case bath numerically forconditions bath studied oscillation,the oscillators oscillators bath with integrating as oscillators Figuregivedependent QHD give minimal is1. minimal using givethe The on asymmetric bathminimal Ωpower a 4th-orderandbath energy spectraε energyand bath system, Runge-Kutta and the energyof and theslower uncertainty:ε uncertainty: asymmetric> andΩ oscillation> algorithm.uncertainty:0 pj systemsp isj= dependent The=2p showj initial2and=and two on majorω.and oscillations. j For the gas phase (no bath oscillators), the system can be solved analytically, and σz oscillates   2 A. Baths of Singledt Oscillators   2 − 2  2 2 1 BATHS21 1 OF SINGLE OSCILLATORS 2 ωj µj d +2g σ ( q 1 q ) and watching a system,qconditionsj q=j startingFor= q an. foroscillator The= with. the The otherσ bath other.xA(0) The where termshigh oscillators= terms other 1,frequencyω are relax.is areterms all the give all set The natural oscillation,are set initially minimal resulting all initially setfrequency to initially bath dependentequations zero.to zero. energy of to the zero. ofon andspin the Ω and uncertainty:spin-boson system ε and more the complicated slowerpj = oscillation behaviorand is dependent on ω. A. Bathsσd q = of Single2εjσ yq Oscillators+j σ1 jp sinusoidally in time with 2µj ω2 frequency.µj j ωj 2µj ωj For the gas phase (no bath oscillators), the system can be solved  analytically,2 and σz oscillates x σj q = 2yε σj q + −  x σj p A. Baths of Single 2 Oscillatorsoccurs. For the1  gas phase (no bath oscillators), the system can be   dt x j −  y j µj  x j system were(15) then numericallyqj = 2µ ω integrating. The otherFor using terms an oscillator a are 4th-order all set where2 initially Runge-Kuttaω is the to zero. natural algorithm. frequency The of initial the spin system more complicated behavior For the gas phase (nodt bath oscillators), +2 −q  theg system σ qµj  can be solved analytically, and σz joscillatesj sinusoidally in time withΩ frequency. ω µ For the gas phase (noj bathk2 oscillators),y k 2 the system can be solved  analytically, and σz oscillates 2 j j     conditionsFor the gas for phase the bath (nosolved bath oscillators analytically, oscillators), give andoccurs. minimal theω system0 oscillates= 2 bathε can 1sinusoidally energy + be2 solved and in analytically, time uncertainty: with andpj σ =oscillates2and(19) sinusoidally in time with frequency.+2gj σyk( qj 2 qj ) 2 A.A. Baths BathsA. of Baths of Single Single of Oscillators Single Oscillators Oscillatorsε z 2Ω sinusoidally in time with+2 gj frequency.σy (q −   qj ) 2 1 3      j 2 sinusoidallyq = in. The time other withfrequency. termsvariousfrequency. are parameters. all set initiallyThe to particular zero. case studied withω0 QHD= 2ε is1 the + asymmetric2 system, ε > Ω > 0 (19)     − Ω  j This2µ behaviorj ωj changesA. Baths when adding of Single a single Oscillators bath oscillator. If a bath oscillator is addedε with an +2+2(qσj yqj gk σyyqkqj ) gk qk Ω2 (15) 3  +2ωq0 = 2εg 1σ +q ForandFor the thewatching gasFor gas phase(19) the phase a gas (novarious system, (no phase bath bath parameters. starting oscillators),(no oscillators), bathΩ2 with oscillators), theTheσ thex system(0) particular system= the can1, system relax.can be case besolved Thecan solvedstudied be analytically,resulting solved analytically, with analytically,equations QHD and andisσ thez ofσoscillates asymmetricandz theoscillates spin-bosonσz oscillates system, ε > Ω > 0   j −  k y kε2 ω0 = 2oscillationε 1 + frequency not close toThis the natural behavior frequency(19) changes ofwhen the adding spin system a single a new bath oscillation, oscillator. less If a bath oscillator   is added with an  k   k ε2 sinusoidallysinusoidallyFor thesinusoidally in gas timein phase time withω in (no0 with= time frequency. bath2ε frequency. with1 oscillators), + frequency. the system (19) can be solved analytically,(19) and σ oscillates  k  A. Baths of Singlesystem Oscillators were thenand numerically watchingε a2 integrating system, starting using with a 4th-orderσx(0) = Runge-Kutta 1, relax. The algorithm. resultingz equations The initial of the spin-boson This behavior changesd when adding a single bath oscillator. Ifthan a bathω0 oscillator, is added is to added the behavior with anoscillation of σz . The frequency closer ω notj is close to ω0 tothe the stronger2 natural2 the frequency amplitude of the and spin system aωj newµj oscillation, less This behaviorσ p = changes+2(2ε σy whenqpj addingµσyω2qσj a)q singlegk bathqk oscillator. Ifsinusoidally a bathconditions oscillator(15) in time for is the withsystem added bath frequency. with were oscillators an then numerically give minimal integratingΩ bathΩ energyΩ using2 and a 4th-order uncertainty: Runge-Kuttap2 = algorithm.and The initial x j +2( y σjy −qj  j jσy xqj )This gthek behavior qk lower changesthe frequency when addingof the(15) added athan single oscillation.ω0 bath, is added oscillator. This to theω single0 Ifω= behavior a0 2= bathε oscillator 2ωε1 oscillator+=1 of + 2εσ responsez 1. + is The added closer is shown withωj anis in to theω0 the strongerj (19)(19)2 the amplitude(19) and oscillation frequencydt not close to the−  natural  − frequency −    ofk the spin  system a new oscillation,2 1 less 0 2 2 2 2   2 ωj µj oscillation frequency not close to the natural k frequencyFor the gas of phase the spin (no This bathq system = behavior oscillators), a new. changes The oscillation,conditions the other when system terms adding forless can theare a be single all bath solved set bath oscillators initially analytically, oscillaε toΩ- giveε zero. and minimalε σz oscillates bath energy and uncertainty: pj = and oscillationtime-domain frequency of σ notz in closej Figure to2µ thej 1ωj andthe natural the lower frequency frequency the frequency domain of theω of spin in= the Figure2ε system added1 + 2. a oscillation. new oscillation, This singleless oscillator response(19) is shown in2 the than ω0, is added tod the behavior+2( of σσz p. Theσ closerp ω) j isg to qω0 the stronger the amplitude  and 2 1 0 2     than ωσd0,p is added= to2ε theσypj behaviorµ yω2 ofσj qσz .sinusoidally Thek k closer ω inj timeis toThis withωtor.This0thebehavior If frequency. a behavior strongerThisbath oscillator behaviorchanges changes the is amplitudeq whenadded changes when= addingwith adding when andan. Theoscillation a adding single a other single bathfrequency a terms bathsingle oscillator. are oscillator. not bathε all set oscillator. If initially a If bath a bath If oscillator to a oscillator zero. bath oscillator is added is added withis added with an an with an the lower the frequencyx ofj the added oscillation.y j −  j j Thisx2 j single than oscillatorω0, is added response to the is behavior shown in of thetime-domainσz .j The closer2µj ω ofj ωσjz isin to Figureω0 the 1 stronger and the frequency the amplitude domain and in Figure 2. thedt lower σx thepj frequency=− 2ε ofσyp −thej addedµjωj oscillation.σkxqj This(16) singleoscillation oscillatoroscillationoscillation frequency response frequency frequency is not shown not close close notin to the tothe closeΩ the2 natural to natural the frequency natural frequency frequency of theof the spin of spin system the system spin a system new a new oscillation, a oscillation, new oscillation, less less less time-domain of σ indt Figure  1 and the− frequency  − domainthe in Figure lower 2.the frequencyThisclose ofA. behavior to the the Baths addednatural changes oscillation.frequency of Single when of addingthe Oscillators This spin single asystem single oscillator a bathnew oscilla oscillator. response- If is a shown bath oscillatorin the is added with an z ω0 = 2ε 15 + 2 (19)  time-domain of σ+2+2(z pinσj y Figurepj gk σ 1yy andqkpj the) frequencygk qk domain inthan Figureoscillationtion,thanω 0less, 2.ωthan is0 than, added isfrequency addedω0,is tois added added tothe not the to behavior closethe to behavior behavior the to behavior of theε ofofσ naturalz σ.z The of.. The The frequencyσ closer closerz . closer The ωj ω closeris ofj tois the toωω0 spinjωtheis0 the tosystem strongerω stronger0 the a thenew stronger the amplitude oscillation, amplitude the amplitude and less and and  +2( σyp −j  σy  pj )time-domain gk qk of σz in Figure 1 and the frequencyA. Baths domain of Single in Figure Oscillators 2.     k        5  k  −    This behavior changesthethanisthe lowerto whenω lower0,thetheFor the is adding stronger addedthe lowerthe frequency frequencygas a the tothe single phase theamplitude frequency of behaviorbath (no theof the bathand added oscillator. of addedthe of oscillators), the lower oscillation.σz added oscillation.. Ifthe The a frequency bath oscillation. the closer This system oscillator This ofω single j is single can This to is oscillator beω added singleoscillator0 solvedthe strongerwith oscillator response analytically, response an the responseis amplitudeshown is and shownσ is inz shown intheoscillates and the in the  5  k     +2gpj σy ( gpjqσj sq pj qj) 5 time-domainthetime-domain lowersinusoidallytime-domain(16) the of frequency ofσz inσinz time of FigureinForσ ofFigurewith the thein 1 frequency. and Figure addedgas 1 and thephase the1 oscillation. frequency and (no frequency the bath frequency domain Thisoscillators), domain single in domain Figure in oscillatorthe Figure system in 2. Figure 2. response can 2. be solved is shown analytically, in the and σ oscillates j   k y −k   oscillation frequencythe not added close oscillation. to the natural This singlez frequency5 oscillator of response the spin is shown system in a new oscillation, less z +2 pj  gk σyqk      2   d  k   1than ω , is added totime-domain the behavior of ofσzσ sinusoidallyin. Figure The closer 1 and inω time theis to withfrequencyω frequency.the stronger domain inΩ the Figure amplitude 2. and σ q = 2Ω σ q +k 2ε σ q + σ p 0 the time-domain of z in Figure 1 and thej frequency0 domain in y j z j x j y j     ω0 = 2ε 1 + 2 (19) dt  +2− gj σy (pjqj s pj qµjthej) lower the frequencyFigure of 2. the(16) added oscillation. This single oscillator5 5 response5 ε2 is shown in theΩ +2 gjσy ( pjq−j s  pj  qj ) (16)  ω = 2ε 1 + (19)  2   2−    0 2 d 2gj σx ( q qj ) 1time-domain of σz in FigureThis Adding behavior 1 single and the bath changes frequency oscillators when domain at adding frequencies, in a Figure single that5 2.are bath not oscillator. If a bathε oscillator is added with an σdyqj = −2Ω σzqj +j  2 −ε σ  xqj + σ1ypj    dt σyqj =− 2Ω σzqj + 2ε σxqj µ+j  σypj (17) closeoscillation to the spin frequency system’sThis natural behavior not close frequency, changes to the cause natural when relatively adding frequency a single of the bath spin oscillator. system a If new a bath oscillation, oscillator less is added with an dt  2 −qj  gk σxqk   µj   −2g σ ( q2 q 2) simplethan behaviorω0, is in added the oscillationsystem. to the This5 behavior frequencybehavior ofof notσz close.coupled The to closer to the naturalωj is to frequencyω0 the stronger of the spinthe amplitude system a new and oscillation, less j xk j 2 j 2   − 2gjσx (q −j   qj ) the lower the frequencythan ω , of is added the added to the oscillation. behavior of Thisσ single. The oscillator closer ω responseis to ω the is shown stronger in the the amplitude and −     −   a single bath oscillator off of 0 can be described as energy pass- z j 0 2(2 qσj xqj gk σx qkqj ) gk qk (17)   −  2 qj −  gkσxqk   time-domain of theσz lowerin Figure the frequency 1 and the of frequency the added domain oscillation. in Figure This 2. single oscillator response is shown in the −  k   k    k  time-domain of σz in Figure 1 and the frequency domain in Figure 2. 40 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005   d 2( σxqj σx qj ) gk 2qk (17) σypj = −2Ω 2(σzpσjx −q+j  2ε σσxxpj qj )µjωjgkσyqqkj (17) 5 dt  − −    −   −k     k 5 d 2( σxpj σx pj ) gk 2qk σdypj = −2Ω σzpj −+  2ε σxpj µjωj σy2qj dt σ p =− 2Ω σ p + 2ε σp −k µω σ q dt y j −  z j  xj − j j  y j 22(pσj xpj gk σx qkpj ) gk qk −  2( σxp −j  σx  pj )  gk qk −  k  −   k     k  2 pσjx ( pjgqkj sσxqkpj qj ) (18) −  2pj  g−k σ xqk  −  k   There are eleven equations (2 first order, 6 mixed terms,k and 3 second order) for every bath oscillator 2 σx ( pjqj s pj qj ) (18) and three for the spin terms that are required−  2σ tox ( describepjq−j s  motionpj  qj of) the system. (18) There are eleven equations (2 first order, 6 mixed−   terms,  and− 3 second  order) for every bath oscillator There are eleven equations (2 first order, 6 mixed terms, and 3 second order) for every bath oscillator and three for the spin terms that are required to describe motion of the system. IV.and Analysis three for the of spinDynamics terms that are required to describe motion of the system.

IV.It has Analysis been demonstrated of Dynamics that the spin-boson system can evolve with many different patterns rangingIV. from Analysis coherent of oscillations, Dynamics to incoherent relaxation and complete localization depending on It has been demonstrated that the spin-boson system can evolve with many different patterns It has been demonstrated that the spin-boson system can evolve with many different patterns ranging from coherent oscillations, to incoherent relaxation3 and complete localization depending on ranging from coherent oscillations, to incoherent relaxation and complete localization depending on

3 3 EDWARDS

ω = 7ε/4 ω = 9ε/4 1 1 0.95 0.8 0.9 0.6 0.85 0.4 0.2 > > z 0.8 z σ σ 0 < 0.75 < -0.2 0.7 -0.4 0.65 -0.6 0.6 -0.8 0.55 -1 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 t t

ω = 7ε/4 ω = 9ε/4 1 1 ω = 2ε ω = 19ε/8 0.95 0.8 0.9 1 0.6 1 0.4 0.85 0.95 0.2 0.95 > > z z 0.8 σ σ 0 < < 0.75 0.9 0.9 -0.2 0.7 0.85 -0.4 0.85 0.65 -0.6 > >

0.6 z 0.8 -0.8 z 0.8 σ σ

0.55 < -1 < 0 50 100 150 200 250 0.75 300 350 400 0 50 100 150 200 250 300 350 400 0.75 t 0.7 t 0.7 0.65 0.65 0.6 0.6 ω = 2ε ω = 19ε/8 1 0.55 1 0.55 0.95 0 50 100 0.95 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 0.9 0.9 t t 0.85 0.85 > >

z 0.8 z 0.8 σ σ

< 0.75 < 0.75 0.7 0.7 0.65 0.65 0.6 0.6 ω = 17ε/8 ω = 5ε/4 0.55 0.55 0 50 100 150 200 250 300 1 350 400 0 50 100 150 200 250 300 350 400 1 t t 0.95 0.95 0.9 0.9 0.85 ω = 17ε/8 ω = 5ε/4 0.85 1 0.8 1 > > z z 0.8 0.95 0.95 σ σ 0.75 < 0.9 < 0.9 0.75 0.85 0.7 0.85 0.8 0.7 > > z 0.65 z 0.8 σ 0.75 σ < < 0.75 0.7 0.6 0.65 0.7 0.65 0.55 0.6 0.6 0.65 0.55 0.5 0.6 0.55 0.5 0 50 100 0.55 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 t t t t

Figure 1: Asymmetric Systems Coupled toFigure Single 1. Asymmetric Oscillators Systems of frequency Coupledω to. Single In these Oscillators simulations of frequency ω. In these simulations ε = 2Ω = 0.5. Coupling constant g = 9ε ε = 2Ω = 0.5 . Coupling constant g = 0.02. The0.02. fast The oscillation fast oscillation has has a frequencya frequency of ω0 ,while while the slower oscillation is dependent on ω of the bath oscillator. Figure 1: Asymmetric Systems Coupled to≈ Single4 Oscillators of frequency ω. In these simulations the slower oscillation is dependent on ω of the bath oscillator. ε = 2Ω = 0.5 . Coupling constant g = 0.02. The fast oscillation has a frequency of ω 9ε , while 0 ≈ 4 the slower oscillation is dependent on ω of the bath oscillator.CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 41

4

4 QUANTIZED HAMILTON DYNAMICS APPLIED TO CONDENSED PHASE SPIN-RELAXATION

Coupled Oscillators with ω < 9ε/4 Coupled Oscillator with ω = 9ε/4 1e+06 ω = 3ε/2 1e+06 ω = 7ε/4 ω ε 100000 = 2 /4 100000 ω = 17ε/8 10000 10000 1000 1000 υ) 100 Asymmetric System Coupled to 25 Oscillators Power Spectrumυ) of 25 Oscillators Coupled to a Spin System > ( > ( z 1 1e+06 z 100 σ σ 10 Asymmetric System Coupled to 25 Oscillators Power Spectrum of 25 Oscillators Coupled to a Spin System P< 0.8 P< 1 100000 1e+06 10 0.6 1 0.8 10000 100000 0.4 1 0.6 0.1 1000 10000 υ)

> 0.2 z 0.4 > ( z σ 100 0.1

0.01 σ < Asymmetric System Coupled to 25 Oscillators 1000 Power Spectrum of 25 Oscillators Coupled to a Spin System

0 υ) P<

> 0.2 z 1 > ( 1e+06

z 10 σ 0.001 100 0.01 σ < -0.2 0 0 0.05 0.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 0.8 P< 1 100000 -0.4 υ 10 υ -0.2 0.6 10000 -0.6 0.1 1 -0.4 0.4 -0.8 0.01 1000

0.1 υ) -0.6 0 50 100 150 200 0 0.05 0.1 0.15 0.2 0.25

> 0.2 z > ( z σ t 100υ

AsymmetricAsymmetric System System Coupled Coupled to to 25 25 Oscillators Oscillators PowerPower Spectrum Spectrumσ of of 25 25 Oscillators Oscillators Coupled Coupled to to a a Spin Spin System System -0.8 < 0 0.01 0 1 1 50 100 150 200 1e+06 1e+060 0.05 P< 0.1 0.15 0.2 0.25 ω ε 10 Asymmetric System Coupled to 25 OscillatorsCoupledt Oscillators with > 9 /4 Power Spectrum of 25 Oscillators Coupled to a Spinυ System Coupled to Multiple Oscillators 0.8 0.8 -0.2 1 1e+06 1e+06ω ε 100000 100000 1e+06 = 19 /8 1 -0.4 ω = 10ε/4 0.8 0.6 0.6 100000 Asymmetric System Coupled to 25 Oscillators 100000 10000 10000 Power Spectrum of 25 Oscillators Coupled to a Spin System Figure 3: An Asymmetric-0.6 System Coupled to 25 Oscillators. The frequencies 100000 0.1 of the bath oscillators, 0.6 10.4 0.4 3ε 5ε 1e+06 ω, are evenly spaced between to Using the 10000 same Ω, ε, and 1000 1000 coupling as Figs. 1 and 2, the spin Figure 3: 10000 An Asymmetric-0.8 System2 Coupled2 to 25 Oscillators.υ) υ) The frequencies 0.01 of the bath oscillators, > > 0.8 0.2 0.2 100000 10000 z z 0.4 0 50 100 150 > ( > ( 200 0 0.05 0.1 0.15 0.2 0.25 z z σ σ 3ε 5ε 1000 100 100 σ σ system decays< < in the presence of more oscillators.t υ ω, are evenly 1000 0.6 0 0 spaced between to Usingυ) the same Ω, ε, and coupling as Figs. 1 and 2, the spin P< P< > 0.2 2 2 10000 z > ( 1000 z σ 100 10 10 σ < υ) 0 system decays 1000.4-0.2-0.2 in the presence of more oscillators. υ) P< 1000 > ( > ( z z 10 υ) 1 1 100 σ σ

> 0.2-0.4-0.4 z -0.2 Adding single bath oscillators at frequencies, that are> ( not close to the spin system’s natural z σ 10 P< 100 P< σ < -0.6 -0.60 Figure 3: An Asymmetric System 1 Coupled to 0.1 0.1 25 Oscillators. The frequencies of the bath oscillators, -0.4frequency, cause relatively simple behavior in the system.P< This behavior 10 of σz coupled to a single Adding 1 single bath oscillators at frequencies,3ε that5ε are not10 close to the spin system’s natural -0.6 -0.2-0.8-0.8 ω, are evenly spaced between 0.1to Using 0.01 0.01 the same Ω, ε, and coupling as Figs. 1 and 2, the spin bath oscillator 0 0 off of ω0 50 50can bedescribed 100 100 150 150 as energy2 200 200 passing2 from 0 the0 spin 0.05 0.05 system 0.1 0.1 into 0.15 0.15 the harmonic 0.2 0.2 0.25 0.25 frequency, cause relatively simple behavior in the system. This 1 behavior 1 of σz coupled to a single 0.1 υυ -0.8oscillator and-0.4 thensystem returning decays intott in the the spin presence system. 0.01 of more The oscillators. lower frequency oscillation  is the flow of 0 bath oscillator 50 off 100 of ω0 can 150 be described 200 as energy 0 passing 0.05 from 0.1 the spin 0.15 system 0.2 into 0.25 the harmonic -0.6 0.1 0.1 energyoscillator back 0.01 and and then fortht returning from spin into to boson. the spin system. The lower frequencyυ oscillation is the flow of -0.8 0.01 However, 0.001 0 for single 50 oscillators 100 where 150ω is on 200ω cause more 0 complicated 0.05 0.01 0.1 behavior. 0.15 In Figure0.2 0.25 2 energy back 0 and forth 0.05Adding from single spin0.1 to bath boson. 0.15 oscillatorsj 0.2 0 at frequencies,0.25 that 0 are not 0.05 close to 0.1 the spin 0.15 system’s 0.2 natural 0.25 FigureFigure 3: 3: An An Asymmetric Asymmetric9ε System Systemt Coupled Coupled to to 25 25 Oscillators. Oscillators. The The frequencies frequenciesυ of of the the bath bath oscillators, oscillators, the oscillator at ωfrequency,= ω causeshows relativelyυ gives a simple responses behavior over in arange the system. of frequencies This behavior and of of greaterσυ coupled to a single ωωHowever,,, are are evenly evenly for spacedj spacedsingle4 oscillators between betweeno 33ε whereε toto 55εεωUsingUsingj is on the theω0 same samecause Ω, Ω, moreεε,, and and complicated coupling coupling as as behavior. Figs. Figs. 1 1 and and In Figure 2, 2, thez the 2spin spin Figure 3: An Asymmetric System9 Coupledε≈ to22 25 Oscillators.22 The frequencies of the bath oscillators,   magnitude.the oscillator This at behaviorbathω 3=ε oscillator5 needsε ω offshows to ofbeω studied gives0 can a be more responses described in depth. over as energy a range passing of frequencies from the and spin of system greater into the harmonic ω, are evenlysystem spacedsystem betweendecays decaysFigure in inj the the2.to Power4 presence presence UsingSpectrao of of ofthe Asymmetric more more same oscillators. oscillators. Ω,Systemsε, and Coupled coupling to Single Oscillators, as Figs. using 1 and the same 2, the conditions spin as Figure 1. The power spectra oscillator2 2≈ and then returning into the spin system. The lower frequency oscillation is the flow of system decaysmagnitude.Figure in the 3: presence An This Asymmetricof behavior the of asymmetric more needs System oscillators. systems to show Coupledbe studiedtwo major to oscillations. more 25 Oscillators. in depth.A high frequency The frequencies oscillation, dependent of the on bathΩ and ε oscillators, and the slower oscillation B.ω, Larger are evenly Baths spacedenergyis dependent between back and on ω.3ε forth Forto an5ε fromoscillatorUsing spin where the to ω same is boson. the natural Ω, ε, frequency and coupling of the spinas system Figs. more 1 complicated and 2, the behavior spin occurs. Figure 2: PowerAddingAdding single single Spectra bath bath oscillators oscillators of Asymmetric2 at2 at frequencies, frequencies, Systems that that are are Coupled not not close close to to to the the Single spin spin system’s system’s Oscillators, natural natural using the same B.system Larger decays Baths in theHowever, presence for of single more oscillators oscillators. where ωj is on ω0 cause more complicated behavior. In Figure 2 Adding singleAfrequency,frequency, more bath realistic oscillators cause causethecondensed relatively relativelyoscillator at frequencies, simple simpleat phaseω = behavior behavior involves9ε thatω are in manyinshows the thenot system. bathsystem. close gives oscillators. to a This This responsesthe behaviorspin behavior The system’s over of lastof aσσ rangez plotz naturalcoupledcoupled in of Figurefrequencies to to a a single2 single and of greater conditions as Figure 1. The powerj 4 spectra≈ o of the asymmetric systems  show two major oscillations. frequency,shows causebathbathA theing morerelatively oscillator oscillatorfrom power realistic themagnitude. simplespectrum spin off off condensedof ofsystem behaviorωω00 ofcancan into This multiple be be phasethe in behavior described described theharmonic bathinvolves system. needs oscillators,asoscillator as energy energy This many tobe behaviorand passing bathpassing studied coupled then oscillators. fromof from more tomoreσz the the in coupledcomplex spin depth.spin The system systemsystem, last tothan a plota single intosimple into with in the the Figuresum all harmonic harmonic theof the 2 single oscillator power A high frequencyAdding single oscillation, bath oscillators dependent9ε at frequencies, on Ω that and are notε and close the to the slower spin system’s oscillation natural is dependent on ω. bath oscillatorshownshowsoscillatoroscillator off frequenciesreturning the of ω power0 and andcan into exceptthen be spectrumthenthe describedspin returning returning forsystem. ofω =as multiple The into intoenergy4 .lower The the the bath passingfrequency power spin spin oscillators, system. system. spectrum from oscillation the The The coupled spin isis lower morelower systemspectra to complexfrequency frequency the into – spinmost the than oscillation oscillation system,notably harmonic a simple by with the is is sum the thedip all in flow flowof the the of ofspectra just below fre- thefrequency, singleenergyenergy oscillator back back cause and and power relatively forth forth spectra from from simple spin spin –9 mostε behaviorto to boson. boson. notably in the by system. the dip Thisin the behavior spectra of justσz belowcoupled frequencies to a single oscillatorFor an andshown oscillator thenthe frequencies flow returning of whereenergy except into back theω forandis spin ωforth the= system. from4 . natural The spin The to power boson. lower frequency spectrum frequency is more of oscillationquencies the complex spin of is ν thanthe= 0.1. system flow a simple of more sum of complicated behavior bath oscillatorB. off of Largerω can be Baths described as energy passing from the spin system into the harmonic energy backoftheν and= single 0 However,. forthHowever,1. However, oscillator from for for spin for single powersingle single to0 boson. oscillators oscillators spectra oscillators – most where where notablyωωjj isis on on by ωω0 the0 causecause dip more morein the complicated complicated spectra Couplingjust many behavior. behavior. below bath oscillators frequencies In In Figure Figure also changes 2 2 the behavior of occurs. oscillator and then returning99εε into the spin system. The lower frequency oscillation is the flow of However,ofCouplingν forthethe= single 0oscillator oscillator.1. many oscillators at at bathAωωjj more== where oscillators realisticωωωooisshowsshows also on condensedω changes gives givescause a a more phasethe responses responses behavior complicated involves over over of many a aσ range behavior.rangez in bath of ofthe frequencies frequencies oscillators. In time Figure domain 2 and and The into of of last greater greater a plot in Figure 2 more complicated behavior.44 ≈≈ j In Figure0 2 the oscillator at   in the time domain into a more physical result. Finally, is moreenergymagnitude.magnitude. physical back9 result.ε and This Thisshows forth Finally,behavior behavior the from power is needs needsspin the spectrum case to to to boson. be be with studied studied of many multiple more more oscillators. in inbath depth. depth. oscillators, Figure 3 shows coupled the to case the where spin the system, with all the the oscillator atCouplingωj = 4 manyωo bath shows oscillators gives gives a aresponses responses also changes over over a range the a range behaviorof fre of- frequencies oftheσ casez inwith and the many of time greateroscillators. domain Figure into 3 a shows the case where the However,≈ for single oscillators where ω is on 9ωε cause more3ε complicated  5ε behavior. In Figure 2 magnitude.spinmore This system physical behavior is coupled result.shown needs3 Finally, to to frequencies 25be oscillators,studied is the except case more spaced with forinj depth.ω many evenly= oscillators.0. The between power Figure2 spectrumand 32 shows. Theis more the behavior case complex where of thisthan the a simple sum of various parameters.quencies and of greaterThe9ε magnitude. particular This behavior case studied needs4 to be withspin QHD system is is coupled the asymmetricto 25 oscillators, spaced system, evenly betweenε > Ω > 0 systemthe oscillator is obviously attheω more singlej = complex4 oscillatorωo shows than power just gives spectra the a responses addition – most of over notably a new a range3ε by oscillation, the of5 dip frequenciesε in as the evident spectra and ofin just greater the below frequencies spinB.B. systemstudied Larger Larger more is coupled in Baths Baths depth. to 25≈ oscillators, spaced evenly between 2 and 2 . The behavior behavior of of this this system is obviously more and watchingpowermagnitude. spectra a of system, Thisofσ ν behavior(Figure= 0. starting1. 3). needs It decays to with be studied to aσ lowerx(0) more energy= in depth.1, state relax. over time The and resulting the spin stays equations in a of the spin-boson B. Largersystem Baths is obviously z more complex than just the addition of a newcomplex oscillation, than just the as addition evident of in a new the oscillation, as evident in low energyAA moremore state realisticuntil realistic the condensed condensednumerical solution phase phase involves involves eventually many many grows bath bath unstable. oscillators. oscillators. Figure The The 3 greatly last last plot plot resembles in in Figure Figure 2 2 systempower were spectra then of numericallyσzCoupling(Figure many3). It integrating decays bath oscillators to a lower using also energy changes a state 4th-orderthe theover power behavior time spectra and Runge-Kutta ofof the σ spinz (Figurein stays the 3). intime algorithm. It a decays domain to a lower into The aenergy initial   LARGER BATHS 3   A morethelow realisticB. sameshowsshows energy Larger system the the condensed state power powermore thatBaths until spectrum spectrumis phasephysical the presented numerical involves result. of of multiple inmultiple solution Makri’s many Finally, bath bath bath eventually solution, is theoscillators, oscillators, oscillators. casealthough grows with coupled coupled The unstable.many a last more oscillators.to to plot the the Figure careful spin spininFigure 3 Figure system,comparison system,greatly 2 3 resembles showswith with still all all the the the case where2 the ωj µj 99εε state over time and the spin stays in a low energy state until the showsconditions theneeds powershownshown to for spectrum be A frequencies frequencies made. more the realistic of bath multiple except except condensed oscillators for forbathω ωphase= oscillators,= involves.. give The The powermany power coupled minimal bath spectrum spectrum to3oscil the- bath spin is is more more system, energy complex complex with and than than all3ε the uncertainty: a a simple simple5ε sum sum of of pj = 2 and the same systemspin that system is presented is coupled in44 Makri’s to 25 oscillators, solution, although spacednumerical evenly a more between solution careful eventually2 comparisonand 2 grows. The still unstable. behavior Figure of this3 greatly shown2 frequenciesthethe1 Alators. single single exceptmore The oscillator oscillator realistic for last ω plot= powercondensed power in9ε . Figure The spectra spectra 2power phase shows – – spectrum most most the involves power notably notably is many spectrum more by by bath the the complex of dip dip oscillators. in in than the the spectra spectra a simple The lastjust just sum belowplot below of in frequencies frequencies Figure 2 qj =needs2µ ω to. be The made.system other is4 terms obviously are more all complex set initially than just to the zero.resembles addition the of same a new system oscillation, that is presented as evident in Makri’s in thesolution,3 the single oscillatorshowsofofj νmultipleνj== the power 0 0..1.1. powerbath spectra oscillators, spectrum – most coupled of notably multiple to the spin by bath thesystem, dip oscillators, with in the all the spectra coupled just to below the spin frequencies system, with all the V. Conclusionpower spectra of σz 9ε(Figure 3). It decays to a lower energy state over time and the spin stays in a shownshown frequencies frequencies except except forfor ω=  . TheThe power power spectrum spectrum is is morealthough complex a more careful than a comparison simple sum still of needs to be made. of ν = 0.1. CouplingCouplinglow many many energy bath bath state oscillators oscillators until4 the also also numerical changes changes solution the the behavior behavior eventually of of σσz growsz inin the the unstable. time time domain domain Figure into 3 into greatly a a resembles V.the Conclusion single oscillator power spectra – most notably by the dip in the spectra3  just below frequencies A.Coupling BathsThemore manymore model physical physical bath of here Singleoscillatorsthe result. result. has same proven Finally, Finally, alsosystem Oscillators changes to is is that be the the non-trivial is case thecase presented behavior with with with many many in of interesting Makri’s oscillators. oscillators.σz in solution, the behavior. Figure Figure timealthough domain 3 3 shows showsThe approximations into the athe more casea case careful where where the thecomparison still ofspinspinν42= system system 0 CMDITR.1. isis Review coupled coupled of Undergraduate to to 25 25 oscillators, oscillators, Research Vol. 2 spaced spacedNo. 1 Summer evenly evenly 2005  between between 33εε andand 55εε.. The The behavior behavior of of this this more physicalmadeThe result. in QHD model Finally, seem hereneeds isto has the to give proven be case reasonably made. with to be many non-trivial physical oscillators. results with Figure whileinteresting 3 suitably shows behavior.2 the2 simplifying case22 Thewhere the approximations the dynamics. spin system issystemsystem coupledCoupling is is obviouslyto obviously many 25 oscillators, bath more more oscillators complex complex spaced alsoevenly than than changes just just between the the the addition addition3ε behaviorand of of5ε . a aof The new newσz behavior oscillation, oscillation,in the time ofas thisas domain evident evident into in in the the a Formade the in gas QHD phase seem (no to give bath reasonably oscillators), physical results the system2 while2 suitably can be simplifying solved the analytically, dynamics. and σz oscillates system is obviouslymorepowerpower physical spectra spectra more complex result. of of σσzz Finally,(Figure(Figure than just is 3). 3). the the It It case decays decays addition with to to many a aof lower lower a new oscillators. energy energy oscillation, state state Figure over over as 3 evident time time shows and and inthe the the the case spin spin where stays stays the in in a a spin system isV. coupled Conclusion to 25 oscillators, spaced6 evenly between 3ε and 5ε . The behavior of this   powersinusoidally spectralowlow of energy energyσ inz (Figure time state state 3).with until until It the decaysthe frequency. numerical numerical to a lower solution solution energy eventually eventually state over grows grows time unstable. unstable.and2 the spin Figure Figure2 stays 3 3 ingreatly greatly a resembles resembles system  is obviously more complex than just the6 addition33 of a new oscillation, as evident in the low energy statethethe until same same the system system numericalThe that that model is solutionis presented presented here eventually has in in proven Makri’s Makri’s grows to solution, solution, beunstable. non-trivialalthoughalthough Figure withΩ 32 a a greatly interestingmore more careful careful resembles behavior. comparison comparison The still still approximations the same systempowerneedsneeds that spectra to to be isbe presented made. made. of σz (Figure in Makri’s 3). It solution, decays3 toalthough a lower energy a more state careful over comparison time and the still spin stays in a made  in QHD seem to give reasonablyω0 = 2ε physical1 + results2 while suitably simplifying the dynamics. (19) needs to be made.low energy state until the numerical solution eventually grows unstable.ε Figure 3 greatly resembles the same system that is presented in Makri’s solution,3although a more careful comparison still V.V. Conclusion Conclusion This behaviorneeds to be changes made. when adding a single bath oscillator.6 If a bath oscillator is added with an V. Conclusion oscillation frequencyTheThe model model here here not has has close proven proven to to to bethe be non-trivial non-trivial natural with with frequency interesting interesting of behavior. behavior. the spin The The approximations systemapproximations a new oscillation, less thanThe modelω0V.,mademade is here added Conclusion in in has QHD QHD proven to seem seem the to to to be behavior give give non-trivial reasonably reasonably of with physical physicalσ interestingz . The results results behavior. closer while while suitably suitablyω Thej is approximations to simplifying simplifyingω0 the the the stronger dynamics. dynamics. the amplitude and made in QHD seem to give reasonably physical results while suitably simplifying the dynamics. the lower theThe model frequency here has of proven the to added be non-trivial oscillation. with interesting This behavior. single oscillator The approximations response is shown in the 66 time-domainmade in of QHDσz seemin to Figure give reasonably 1 and thephysical frequency results while domain suitably in simplifying Figure the 2. dynamics.   6

6 5 EDWARDS

Asymmetric System Coupled to 25 Oscillators Power Spectrum of 25 Oscillators Coupled to a Spin System 1 1e+06

0.8 100000

0.6 10000 Asymmetric System Coupled to 25 Oscillators Power Spectrum of 25 Oscillators Coupled to a Spin System 1 0.4 1e+06 1000

0.8 100000 υ)

> 0.2 z > (

0.6 z σ 10000 100 σ < 0.4 0 1000 P< υ) 10 > 0.2 z > ( z σ -0.2 100 σ < 0 P< 10 1 -0.2 -0.4 1 -0.4 -0.6 0.1 -0.6 0.1

-0.8 -0.8 0.01 0.01 0 50 0 100 50 150 200100 150 0 0.05 2000.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 t υ t υ

Figure 3: An AsymmetricFigure System 3. An Coupled Asymmetric to 25 System Oscillators. Coupled The to 25 frequencies Oscillators. The of the frequencies bath oscillators, of the bath oscillators, ω, are evenly spaced between 3” ω, are evenly spaced between 3ε to 5ε Using the same Ω, ε, and coupling as Figs. 1 and 2, the spin 2 2 . Using the same Ω, ε, and coupling as Figs. 1 and 2, the spin system decays in the presence of more oscillators. systemFigure decays in 3: the An presence Asymmetric of more oscillators. System Coupled to 25 Oscillators. The frequencies of the bath oscillators, 3ε 5ε ω, are evenly spaced between 2 to 2 Using the same Ω, ε, and coupling as Figs. 1 and 2, the spin Adding single bath oscillatorsCONCLUSION at frequencies, that are not close to the spin system’s natural ACKNOWLEDGEMENTS frequency,system cause decays relatively simple in the behavior presence in the system. of more This behavior oscillators. of σ coupled to a single  z bath oscillator The off of modelω0 can here be describedhas proven as to energy be non-trivial passing from with the interesting spin system into the Research harmonic support is gratefully acknowledged from the Na- oscillatorbehavior. and then returningThe approximations into the spin made system. in QHD The lower seem frequency to give rea oscillation- tional is the Science flow of Foundation Center on Materials and Devices for energy back and forth from spin to boson. sonablyAdding physical single results bath while oscillatorssuitably simplifying at frequencies, the dynamics. thatInformation are not Technology close Research to the (CMDITR), spin system’s DMR-0120967. natural However, for single oscillators where ωj is on ω0 cause more complicated behavior. In Figure 2 frequency,However, there cause9ε is still relatively much to explore simple and verify behavior with this in mod the- system. This behavior of σ coupled to a single the oscillator at ωj = ωo shows gives a responses over a range of frequencies and of greater z However, there4 ≈ is still much to explore and verify with this model. The frequency of the new spin- However, there is still much to exploremagnitude. andbath verifyel. This oscillatorThe withbehaviorfrequency this needs off of model. the of to new beω0 studiedThe spinoscillationcan frequency bemore described in depth. induced of the by new as coupling energy spin- passing from the spin system into the harmonic oscillation induced by coupling to a harmonicoscillation oscillator induced should by be coupling determined to a harmonic as a function oscillator of ω should be determined as a function of ω However, there is still much to explore and verify with this model.oscillatorto a harmonic The and frequency oscillator then of returningshould the new be determined spin- intothe as a function spin system. of j The lower frequency oscillationj is the flow of and perhaps g . Also, the frequencyB. response Largerand from perhaps Baths addingg . oscillators Also, the frequency near ω should response be determined. from adding oscillators near ω should be determined. oscillation induced by coupling toj a harmonic oscillator shouldenergyand be perhaps determined back andj Also, as forth athe function frequency from of spin 0responseωj to from boson. adding oscil- 0 In addition to frequency responses, the differenceIn addition between to frequency treating responses, the system the as difference closed with between treating the system as closed with and perhaps gj. Also, the frequency response from adding oscillatorsA morelators realistic nearnear condensedω0 should phase be be determined. determined. involves many bath oscillators. The last plot in Figure 2 In addition to frequencya finite set responses, of oscillators the difference at discrete betweenshows frequencies thea treating finiteHowever, power In or setaddition spectrum as the of an for systemoscillators to openof frequency single multiple system,as closed at oscillatorsbath responses, discrete oscillators,like with a the frequenciesLangevin where coupleddifference toω equation, j orthebetweenis as spin on an system,ω open0 cause with system, all more the like complicated a Langevin equation, behavior. In Figure 2 shown frequencies except for ω = 9ε . The9ε power spectrum is more complex than a simple sum of a finite set of oscillatorswhere at the discrete bath has frequencies continuous or as spectral an openthe densitywheretreating system, oscillator the should the like bath system at abe has Langevin asω addressed. closed continuous4= with equation, a There ωfinite spectralshows set could of densityoscillators begives a wayshould aat dis responses to- be addressed. over a There range could of frequencies be a way to and of greater the single oscillator power spectraj – most4 notably≈ o by the dip in the spectra just below frequencies where the bath hasincorporate continuous the spectral QHD density approximation shouldof be intoν =magnitude. addressed. 0incorporate. a1.crete Langevin-type frequencies There the This QHD or could formalism, behavior as an approximation be open a waymaking system, needs to like usetointo abe of aLangevin theLangevin-type studied tools equa ofmore- formalism, in depth. making use of the tools of incorporate the QHDstatistical approximation mechanics. into a Langevin-typeCoupling formalism,statisticaltion, many where making bath mechanics. the oscillators bath use has of alsocontinuous the changes tools spectral the of behavior density of shouldσz in thebe time domain into a   statistical mechanics. In all, the QHD spin-boson modelmore successfully physicaladdressed.In result. all, givestheThere Finally, QHD could a is physical the spin-bosonbe case a way with solution to manyincorporate model oscillators. for successfully the the QHD asymmetric Figure approxi 3 givesshows- the a physical case where solution the for the asymmetric spin system is coupled to 25 oscillators, spaced evenly between 3ε and 5ε . The behavior of this 3 system. The decay of spin is qualitativelyB.system. the Larger same The as decay the Baths data of spin presented is qualitatively by Makri. the3 There2 same2 areas the data presented by Makri. There are In all, the QHD spin-boson model successfully givessystem a physical ismation obviously solutioninto more a Langevin-type complex for the than asymmetric formalism, just the addition making of use a new of the oscillation, tools as evident in the 3 system. The decaystill of spin many is qualitatively more aspects the of same this model aspower the to data spectrastill explore,of presentedstatistical many of σ such more(Figure mechanics. by as aspectsMakri. 3). symmetric It decays ofThere thisto systems,a lower model are energy and to state explore, asymmetric over time such and as the symmetric spin stays in a systems, and asymmetric  z still many more aspectssystems of thiswith model Ω > ε toas explore, well as decipheringsuchlow as energy symmetricsystems A thestatemore Inresponse until all,with systems, the the realistic Ω numerical QHD of> εthese and as spin-boson solutioncondensed well asymmetric systems as eventually model deciphering to phase various successfully grows unstable. involvesthe baths response gives Figurewith a many 3 of greatly these bath resembles systems oscillators. to various The baths last with plot in Figure 2 the same system that is presented in Makri’s solution,3 although a more careful comparison still systems with Ω > εdifferentas well spectral as deciphering densities. the response ofshows thesedifferentphysical systems the spectral solution power to variousdensities. for spectrum the asymmetricbaths of with multiple system. The bath decay oscillators, of coupled to the spin system, with all the needs to be made. different spectral densities. 9ε 3 shownspin is frequencies qualitatively the except same as for the ωdata= presented4 . The by powerMakri. spectrum is more complex than a simple sum of V.the ConclusionThere single are oscillator still many more power aspects spectra of this – model most to notably explore, by the dip in the spectra just below frequencies of νsuch= 0 as.1. symmetric systems, and asymmetric systems with The modelΩ > ε here as well has provenas deciphering to be non-trivial the response with interestingof these systems behavior. to The approximations made in QHDCoupling seem to give many reasonably bath physical oscillators results while also suitably changes simplifying the the behavior dynamics. of σ in the time domain into a various baths with different spectral densities.  z more physical result. Finally, is the case with many oscillators. Figure 3 shows the case where the 6 3ε 5ε spin system is coupledREFERENCES to 25 oscillators, spaced evenly between 2 and 2 . The behavior of this system1O. V. isPrezhdo obviously and Y. V. more Pereverzev, complex J. Chem. than Phys. just113, 6557 the addition of a new oscillation, as evident in the power(2000). spectra of σ (Figure 3). It decays to a lower energy state over time and the spin stays in a  z low2A. energy J. Leggett state et al., until Rev. Mod. the Phys. numerical 59, 1 (1987). solution eventually grows unstable. Figure 3 greatly resembles the3N. same Makri, system J. Math. Phys. that 36, is 2430 presented (1995). in Makri’s solution,3 although a more careful comparison still needs to be made.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 43 V. Conclusion

The model here has proven to be non-trivial with interesting behavior. The approximations made in QHD seem to give reasonably physical results while suitably simplifying the dynamics.

6

7 7 7 44 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Investigating New Cladding and Core Materials for Hybrid Electro-optic Modulators

Parissa Fathalipour Robert A. Norwood, Chris DeRose and Amir Fardad University of Arizona Norwood Lab, Optical Science Center University of Arizona

INTRODUCTION ments were taken at four different wavelengths 632.8nm, 830nm, 1300nm, and 1550nm to ensure thorough results. The sample Detailed knowledge of the optical properties of materials that films involved were formulated and fabricated using a technique are used for electro-optic devices will help lead to the design and called spin coating which is commonly used in the semiconduc- formulation of improved waveguide cladding materials. This in tor industry. The films to be investigated were initially of known turn will also lead to the improvement in design and performance materials such as silicon dioxide on a silicon wafer, hydrogen of various optical devices which use these optical waveguide ma- based sol-gels, and fluorinated sol-gels. For all samples the TE terials. index and thickness was to be found at each wavelength as well as the TM index and thickness also at each wavelength. The TE OBJECTIVE AND MOTIVATION index is the refractive index for waveguide modes polarized par- The purpose of this research is to investigate and character- allel to the surface of the film, while TM modes are polarized ize a variety of electro-optic materials, sol gels, and nanoparticle normal to the surface of the film. doped sol-gel thin films by use of a prism coupler. Detailed mea- Experience was used to set a standard as to what was an ac- surements of the dependence of the refractive index on wave- ceptable standard of deviation for the measurements, with the length and polarization were taken, as well as measurements of requirement that the error in the index of refraction was to be no the uniformity of the refractive index over the area of the prepared greater than 0.002% and that in the thickness had to be less than film. The data collected gave more insight into the make-up and 1.0%. The initial measurements were taken of the silicon dioxide structure of thin films being used in waveguides, provided better on a silicon wafer. The TE index, TM index, and thickness were input data for modeling and simulations, assisted in the develop- found, and the measurements were repeated until standard devia- ment of improved devices, and may eventually lead to new or tions were within the acceptable standard range. The data was improved cladding materials. In particular, the observation large then analyzed and fit using a plotting program called Axum. The birefringence in some of our standard sol-gel materials provides data was to be fitted to the Cauchy dispersion formula given in a potential explanation for the anomalously high losses that have Equation (1) given below: been observed in these materials. (1) APPROACH Optical characterization of thin films can be easily and ac- which is generally applicable to materials far from any significant curately done by use of a prism coupler, specifically a Metricon absorption bands. Using this equation coefficient A, B, and C Model 2010 prism coupler. The prism coupler is able to make its were calculated and then expected values were found and com- highly accurate measurements (±.0001) by directing a laser of a pared to the measured values. With these values the dispersion specific wavelength to a prism of a high refractive index, which curves were made, which give the index of refraction vs. wave- is coupled with a sample with a pressure ranging from 20psi to length. This was repeated also for the TM data. In order to get about 40psi. The prism, sample, and photo detector are mounted an understanding of the birefringence of the sample the index on a stage which rotate and varies the angle at which the laser difference had to be found, TE index minus TM index. This was enters the prism thuss allowing the waveguide modes of the film also graphed versus wavelength and gave us our birefringence to be found. When three or more modes are found the prism graph. Once the measured data was accurate and concurred with coupler is able to calculate the index of refraction and thickness the previous data the focus was moved to three hydrogen based of the substrate film; it also is able to give the standard deviation films Sample 6 spin speed of 500r/30s, Sample 7 1500r/30s, and of the measurements of the thickness and index. The measure- Sample 8 3000r/30s, which were supplied by Chris DeRose. The same procedure was followed as with the silicon dioxide. One

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 45 INVESTIGATING NEW CLADDING AND CORE MATERIALS FOR HYBRID ELECTRO-OPTIC MODULATORS difficulty that was found with these samples was to attain a good RESULTS coupling point. The higher spin speed proved to be the most dif- The results of this research became most interesting when ficult, but after multiple runs accurate data was found. The re- the hydrogen based sol-gels were analyzed. All other samples search then became more detailed by investigating the cladding acted as they should, demonstrating little birefringence and low and core films wavelength dependent refractive index and the dispersion. As can be seen from Figures 1 and 2 the dispersion effects of TE and TM polarization. These were the fluorinated for SiO2 is low. These figures also show the general shape of the sol-gel samples provided by Amir Fardad. The same steps were curves this should not vary much from sample to sample. Also taken and as before these samples proved to be more challenging seen below is Figure 3, which shows the low birefringence of the because of their nonuniformities and rough surfaces. Coupling silicon dioxide film. spots were hard to find because the surface of the samples were nonuniform and also the silicon wafers used were much thinner than with the previous samples. This resulted in the need to ad- just measurement pressure, which had to be lowered and varied and the samples had to be moved to multiple locations until ac- curate data could be collected. All plots were then analyzed to find characteristics of the films such as high or low dispersion and birefringence.

Figure 3. Birefringence of SiO2 film

The data below is for the fluorinated core sol-gel film, which also exhibits low dispersion and birefringence. The only main difference is that the index of refraction is much lower than the

SiO2 and the hydrogen based film and this is due to the substi- tution of hydrogen with fluorine, which reduces the molecular polarizability of the sol-gel. Figure 4 shows the TE dispersion Figure 1. TE dispersion curve of SiO film 2 curve and Figure 5 shows the birefringence.

Figure 2. TM dispersion curve of SiO2 film Figure 4. Fluorinated TE dispersion curve

46 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 FATHALIPOUR

Figure 5. Fluorinated sol-gel birefringence Figure 7. Birefringence of Hydrogen based film

When the hydrogen based sol-gels were analyzed the plots CONCLUSIONS show a low amount of dispersion in both TE and TM direction. The data shown is for the spin speed of 1500/30 rps, which dem- The hydrogen based sol-gel samples had the largest birefrin- onstrates the low dispersion that all the hydrogen based sol-gels gence this finding is important because it means that there is a exhibit as seen in Figure 6. large change in index depending if its TE or TM. The large dif- ference means that the sol-gel is polarization dependent, which could have a large impact in optical systems. The large birefrin- gence also means that there is potentially birefringence induced waveguide loss in the hydrogen-based sol-gels. The loss is due to the microdomain scattering that happens within the waveguides.

FUTURE WORK In future research poled and un-poled films will also be looked at, the main focus will be on poled films and how the in- dex of refraction varies across the film, as poling nonuniformities can lead to significant problems in fabricating high performance devices. Nanoparticle doped sol-gels will also be investigated in future research. The data will then be used to investigate new or improved cladding materials for a target core material. A key goal of this next phase will be to develop a cladding material with Figure 6. TE dispersion curve a refractive index much closer to that of the electro-optic polymer material, which has a refractive index of approximately 1.62 at However, the birefringence for these samples was exception- 1550mn. Two potential approaches are considered. One is high ally high compared to the SiO2 and fluorinated samples. Looking index nanoparticle doping of sol-gels; silicon nanoparticles with at the scale it can be seen that the magnitude of the birefringence indices of 3.5 have been obtained and will be dispersed in pho- for the hydrogen based sol-gel is ten times larger than that of topatternable sol-gels that have been previously developed. The the SiO2 and fluorinated films that is why it is such an interest- dependence of the refractive index on nanoparticle loading will ing result. This high birefringence that these samples exhibited be studied, assuming that good dispersions can be made. Another is demonstrated in Figure 7. High birefringence can lead to lo- approach is to investigate the use of selective photobleaching of cal variations in the refractive index that can subsequently result the electro-optic polymer layer to alter the refractive index, there- in scattering. It is hypothesized that the high birefringence of fore allowing it to be a cladding layer for itself. the hydrogen-based sol-gels could be a major contributor to the anomalously high waveguide loss that has been observed in these films.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 47 INVESTIGATING NEW CLADDING AND CORE MATERIALS FOR HYBRID ELECTRO-OPTIC MODULATORS

ACKNOWLEDGEMENTS Dr. Robert Norwood Chris DeRose Amir Fardad University of Arizona College of Optical Sciences

Funding for this research provided by the Center on Materials and Devices for Information Technology Research (CMDITR), an NSF Science and Technology Center No. DMR 0120967

Parissa would like to thank Dr. Norwood for giving me an opportunity to gain invaluable experience this summer.

48 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis of TPD-Based Compounds for Use in Modification of ITO Surfaces and Metal Nanoparticles

Aaron D. Finke Michal Malicki and Seth R. Marder University of Arizona Marder Lab Georgia Institute of Technology

INTRODUCTION mon resonance, which is thought of as collective oscillations of surface electrons. The induced oscillating electric field in close Metal oxides, in particular the transparent semiconductor proximity to the surface of the nanoparticle shows much higher indium tin oxide (ITO), have been shown to be effective charge amplitude than the amplitude of the electric field of the incoming transport materials in organic photovoltaic cells. However, the light. TPD compounds show two-photon absorption in roughly relative ease in which ITO surfaces can lose their charge trans- the same region as the surface plasmon resonance. Therefore, at- fer properties, thus creating “dead spots” on their surfaces, has taching TPD to a gold nanoparticle via a thiol may increase the raised questions recently about the relative efficiency of these rate of two-photon absorption of TPD 4,5. metal oxides to transfer charge. Therefore, the synthesis of a TPD-based compound which The surface of ITO is highly reactive in ambient conditions, can allow for a variety of different functionalities to be synthe- making the surface problematic for practical use. Upon con- sized will be important in showing the effects of TPD in various tact with water, ITO hydrolyses to form surface metal hydroxyl systems. Synthesis of TPD with a phenolic functionality can give groups and, in many cases, complete hydrolysis of the surface TPD such versatility and should allow for several different TPD- metal to form metal hydroxides, which dissociate form the sur- based compounds to be made from a single source compound. face and are physiosorbed onto the metal surface. In conditions where oxygen is absent (such as those used in vapor deposition), METHOD oxygen deficiencies can form on the surface, leading to positive charge formation on the surface. Physiosorbed metal hydroxides The proposed synthetic scheme for phenol-functionalized are especially poor at transferring charge, are insoluble in most TPD is shown in Figure 1, which is partly based on previous solvents, and bind tightly to ITO. Their presence can result in re- fuctionalized syntheses6 . The major route for synthesis involves duced device efficiency and lifetime, as a result of thermal break- multiple steps utilizing Buchwald-Hartwig palladium-catalyzed down that can occur due to charge buildup on the surface1. aminations 7. A tert-butyldiphenylsilyl (TBDPS) group is used as Chemisorption (chemical bonding) of molecules onto ITO a protecting group for a phenol to be revealed after the coupling surface defects which can facilitate charge transfer has been reactions; this is necessary as the acidic phenol hydrogens are shown to be an effective solution to this problem. Several moi- incompatible with the reaction conditions of Buchwald-Hartwig eties have been shown to aid in charge transfer when chemically amination. This phenol can be used as a starting material for a bonded to ITO: recently, ferrocenedicarboxylic acid and 3-thio- variety of different reactions, allowing for several different func- pheneacetic acid have both been shown to greatly aid in hole tionalities to be attached easily onto it. transfer across the ITO surface in OLEDs. A recent question has been asked in light of these results: will directly binding the hole RESULTS AND DISCUSSION transport layer commonly used in OLEDs, a class of bis(triaryl) Synthesis of TPD was performed in good yield for all re- 2,3 amines called TPD , aid in hole transport as well? Furthermore, actions, except for one step. The amination of 4-bromo-4’-io- it has been shown that while carboxylic acid moieties bind well to dobiphenyl with phenyl-m-tolyl amine was attempted several surface defects on ITO, phosphonic acid moieties bind even more times with two different conditions. Initially, the amination was tightly, and therefore are of interest. Synthesis of TPD with an attempted through copper(I)-catalyzed Ullmann condensations. alkyl phosphonic acid functionality would be used to test the ef- However, the reaction never went to completion, even after sev- fects of direct attachment of the hole-transport layer via a strong eral days of heating. After workup, the product was isolated in binding functionality. 16% yield. Another pathway was chosen, through palladium- TPD may also have other applications in the functionaliza- catalyzed amination. While the reaction progressed much more tion of nanoparticles. Gold nanoparticles exhibit surface plas- quickly (~24h), these conditions aminate both aryl bromides and

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 49 SYNTHESIS OF TPD-BASED COMPOUNDS FOR USE IN MODIFICATION OF ITO SURFACES AND METAL NANOPARTICLES

Figure 1. Synthesis of TPD with one phenol group.

iodides, thus making diamination a major byproduct of the re- room temperature under nitrogen atmosphere overnight. A white action. With these conditions, the product was isolated in 26% precipitate formed over this time. The reaction was added to 50 yield. mL CH2Cl2 to dissolve the precipitate, then poured into 50 mL

The TPD phenol was isolated and, as it is a compound not water. The mixture was extracted with CH2Cl2 three times, the ex- previously described in literature, it will need to be fully charac- tracts of which were combined and washed with water four times terized before proceeding with alkylation, using NMR, MS, and (50 mL each) and brine three times (50 mL each), then dried elemental analysis. over anhydrous MgSO4. The solvent was removed under reduced

pressure. Column chromatography in 4:1 hexanes:CH2Cl2 gave EXPERIMENTAL a colorless oil which crystallized to form a white solid overnight (24.1g, 97% yield). 1H NMR (500 MHz, , CD Cl , ∂): 7.70 (dd, Synthesis of 4-bromo-(tert-butyldiphenylsilyl) 2 2 4H), 7.45 (tt, 2H), 7.38 (t, 4H), 7.19 (dt, 2H), 6.65 (dt, 2H), 1.05 oxybenzene (s, 9H). To a 500 mL round-bottom flask was added 4-bromophenol (10.0 g, 0.057 mol), imidazole (4.77 g, 0.07 mol), and DMF (50 mL). Tert-butylchlorodiphenylsilane (19.0g, 0.069 mol) was added with to the mixture while stirring. The reaction was stirred at

50 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 FINKE

Synthesis of 4-bromo-4’-iodobiphenyl to the flask via syringe. The following reagents were added to To an oven-dried 500 mL round-bottom flask was added 4- the flask in order: phenyl m-tolyl amine (7.66 g, 0.0418 mol), 4-bromo-4’-iodobiphenyl (14.3 g, 0.0402 mol), 1,10-phenanthro- bromobiphenyl (44.6 g, 0.191 mol), KIO4 (22.0 g, 0.095 mol), and iodine (24.6 g, 0.097 mol). Acetic acid (132 mL, 2.30 mol) was line (0.290 g, 1.61 mmol), CuCl (0.161 g, 1.62 mmol), and KOH added to the flask, followed by water (11 mL) and concentrated (17.7 g, 0.316 mol). A Dean-Stark trap and reflux condenser were sulfuric acid (5.6 mL). The reaction was heated with stirring at 70 attached to the flask and the reaction was heated to reflux for one °C overnight. The reaction mixture was cooled to room tempera- week. The reaction was taken off heat and the toluene evapo- ture and water (250 mL) was added to the flask. The mixture was rated under reduced pressure. The resulting black sludge was dis- filtered and washed with water to give a pinkish-white solid. The solved in 500 mL dichloromethane and washed with water three precipitate was recrystallized in toluene to give colorless, needle- times, dried over anhydrous MgSO4 and the solvent evaporated. 1 Column chromatography in 9:1 hexanes:dichloromethane gave a like crystals (64.9 g, 94% yield). H NMR (500 MHz, CD2Cl2, ∂): 7.78 (d, J = 5.4 Hz, 2H), 7.58 (d, J = 5.4 Hz, 2H), 7.45 (d, J = 5.4 white solid (3.5 g). The solid was recrystallized in ethanol to give Hz, 2H), 7.32 (d, J = 5.4 Hz, 2H). white crystals (2.77 g, 16% yield). 1H NMR: 7.52 (dt, 2H), 7.41 (m, 4H), 7.24 (m, 2H), 7.11 (m, 5 H), 7.02 (t, 1H), 6.90 (m, 3H), 2.24 (s, 3H).

Synthesis of 4-(tert-butyldiphenylsilyl)oxyphenyl m-tolyl amine. An oven-dried 250 mL three-neck flask was sealed with septa and was purged with nitrogen gas for ten minutes. 100 mL anhydrous toluene was added to the flask. Tris(dibenzylideneace tone)dipalladium (0) (Pd2(dba)3) (0.169 g, 0.184 mmol) and di- phenylphosphinoferrocene (DPPF) (0.212 g, 0.382 mmol) was added to the flask and the mixture was stirred for ten minutes. Buchwald-Hartwig amination of (4’-bromo- 4-bromo-(tert-butyldiphenylsilyl)oxybenzene (6.06 g, 0.0147 biphenyl-4-yl)-phenyl-m-tolyl-amine. mol), m-toluidine (2.0 mL, 0.0184 mol) and sodium tert-butoxide To an oven-dried large Schlenk tube there was added

(1.71 g, 0.0184 mol) were added to the flask. A reflux condenser Pd2(dba)3 (0.626 g, 0.884 mmol), DPPF (0.889 g, 1.60 mmol), was added to the flask and the reaction was heated to reflux under sodium tert-butoxide (4.22 g, 44.0 mmol), and 4-bromo-4’-iodo- nitrogen atmosphere for ~36 h. Upon completion of the reaction biphenyl (14.9 g, 42.0 mmol). The tube was sealed with a sep- as monitored by TLC, the solvent was evaporated under reduced tum and evacuated under vacuum for ten minutes, then purged pressure and the resulting brown oil was dissolved in 100 mL with argon gas for ten minutes. Phenyl-m-tolylamine (6.41g, 34.9 dichloromethane and washed with water three times, dried over mmol) was added via syringe and anhydrous toluene (70 mL) was added via cannula. The mixture was heated with stirring to MgSO4 and the solvent evaporated. Column chromatography in 7:3 hexanes:dichloromethane gave a yellow, viscous oil (4.50g, 100°C for 24h. The solvent was evaporated under reduced pres- 1 sure, and the resulting oil was dissolved in 50 mL dichlorometh- 70% yield). H NMR (300 MHz, CDCl3, ∂): 7.71 (m, 4H), 7.39 (m, 6H), 7.08 (m, 1H), 6.86 (d, 2H), 6.70 (m, 5H), 2.24 (s, 3H), ane, which was filtered through silica to give a red oil. Column

1.12 (s, 9H). chromatography on SiO2 using 9:1 hexanes:dichloromethane as eluent gave a white solid (4.50 g). The solid was recrystallized in ethanol to give colorless, needle-like crystals (3.79g, 26% yield). 1 H NMR (300 MHz, CD2Cl2, ∂): 7.52 (dt, 2H), 7.41 (m, 4H), 7.24 (m, 2H), 7.11 (m, 5 H), 7.02 (t, 1H), 6.90 (m, 3H), 2.24 (s, 3H).

Synthesis of N4’-[4-(tert-Butyl-diphenyl-silanyloxy)- phenyl]-N4-phenyl-N4,N4’-di-m-tolyl-biphenyl-4,4’- Ullmann condensation of (4’-Bromo-biphenyl-4-yl)- diamine. phenyl-m-tolyl-amine. To an oven-dried Schlenk tube there was added (4’-Bromo-bi- An oven-dried 250 mL round-bottom flask was purged with ni- phenyl-4-yl)-phenyl-m-tolyl-amine (2.02 g, 4.86 mmol), 4-(tert- trogen gas for ten minutes. 70 mL anhydrous toluene was added butyldiphenylsilyl)oxyphenyl m-tolyl amine (2.50 g, 5.70 mmol),

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 51 SYNTHESIS OF TPD-BASED COMPOUNDS FOR USE IN MODIFICATION OF ITO SURFACES AND METAL NANOPARTICLES

Pd2(dba)3 (88.6 mg, 0.0968 mmol), DPPF (110 mg, 0.198 mmol), and sodium tert-butoxide (0.584 g, 6.08 mmol). The tube was sealed with a septum, evacuated under vacuum for ten minutes and purged with argon gas for ten minutes. Anhydrous toluene (20 mL) was added via cannula. The mixture was heated with stirring to 100°C for 24h. Toluene was evaporated under reduced pressure, and the resulting brown oil was dissolved in 20 mL di- chloromethane, then filtered through silica gel, and the solvent Aaron Finke is a senior majoring in chemistry at the University of Arizona. He will graduate in May 2006 and plans on attending evaporated. Column chromatography on SiO2 using 4:1 hexanes: graduate school in organic or inorganic chemistry. toluene as eluent gave a glassy, yellow solid (3.16g, 84 % yield). 1 H NMR (300 MHz, CD2Cl2, ∂): 7.79 (d, 4H), 7.42 (m, 10H), 7.28 (t, 2H), 7.12 (m, 6H), 7.01 (m, 4H), 6.90 (m, 9H), 2.28 (s, 3H), 2.26 (s, 3H), 1.17 (s, 9H).

REFERENCES (1) Brumbach, M. Doctoral Thesis; University of Arizona: Tuc- son, 2003. (2) Okumoto, K.; Wayaku, K.; Noda, T.; Kageyama, H.; Shirota, Y. Synthetic Mealst 2000, 111, 473-476. (3) Shirota, Y.; Okumoto, K.; Inada, H. Synthetic Mealst 2000, 111, 387-391. (4) Stellacci, F.; Bauer, C. A.; Meyer-Friedrichsen, T.; Wendel- eers, W.; Marder, S. R.; Perry, J. W. J. Am. Chem. Soc. 2003, 125, 328-329. (5) Wendeleers, W.; Stellacci, F.; Meyer-Friedrichsen, T.; Man- gel, T.; Bauer, C. A.; Pond, S. J. K.; Marder, S. R.; Perry, J. W. J. Phys. Chem. B 2002, 106, 6853-6863. (6) Hreha, R. D.; Zhang, Y. D.; Domercq, B.; Larribeau, N.; Haddock, J. N.; Kippelen, B.; Marder, S. R. Synthesis-Stuttgart 2002, 1201-1212. (7) Hartwig, J. F. Angew. Chem. Int. Ed. 1998, 37, 2046-2067.

ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

52 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Optimizing Hybrid Waveguides

Andrew Gardner Kjersti Kleven and Scott Dunham Highline Community College Department of Electrical Engineering University of Washington

This paper outlines research into how much the propagation constant of hybrid waveguides changes when the index of refraction of the cladding layer is altered. Various indexes of the cladding layer, heights of the silicon block, widths of the silicon blocks, and spaces between the two silicon blocks were modeled to find optimal dimensions to produce the maximum change in the propagation constant.

INTRODUCTION Waveguides have been used in optical modulators for more than ten years. Recently, the concept of hybrid waveguides has been a focus of research. Hybrid waveguide can perform the same function as traditional waveguides, although they designed and operate much differently. Figure 1. Right: The propagation constant vector relationship. Waveguides act much like wires for light. While electric wires Left: A waveguide with two guided modes. conduct or guide electricity, waveguides propagate or guide light. Traditional waveguides do this by guiding the light inside a high in- The power flow diagram of a 2-D cross-section of a wave- dex material surrounded by lower index cladding layer. However, guide (see Fig. 2) shows how light waves propagate through a hybrid waveguides are unique in that they actually guide the light in 3-D waveguide. The red area indicates where the majority of the a low index region between two blocks of high index material.1 light is concentrated. Fig. 2 demonstrates the difference between Light propagates through a waveguide the as a mode. Modes a hybrid waveguide and a traditional waveguide. In the hybrid are a spatial distribution of optical energy or an electromagnetic waveguide the mode is much more confined in the narrow low wave that is a solution of Maxwell’s wave equations. Two impor- index region rather than in the high index blocks. This happens tant types of modes for this application are guided modes and sub- because of a large discontinuity in the transverse E-field at the 2 strate radiation modes. The substrate radiation mode is an unguided high-index material–cladding interface. mode which means the wave will disperse into the substrate layers. A Another advantage of a hybrid waveguide is that the modes guided mode is a wave that experiences total internal refraction and of the waveguide depend more on the index of refraction (index) remains well confined within the waveguide. There can also be more of the cladding layer than traditional waveguides because the that one guided mode inside a waveguide. light propagates through the cladding material. If an active mate- For every mode of a waveguide, there is an associated propaga- rial is used for the cladding layer, the index of this material could tion constant. The propagation constant describes how fast the wave be changed, which would affect the propagation of the mode. moves along the waveguide. The triangle in Fig. 1 shows a vector Recent advances in electro-optic (EO) polymers allow them to relationship between how fast the wave moves through the material experience a very large change in their index of refraction when 3 and the propagation constant, beta(β). Theta(θ) is the actual angle an electric field is applied across them. This effect can be used of wave movement inside the waveguide. The planar waveguide in with hybrid waveguides to change the propagation constant of the Fig. 1 demonstrates how a waveguide can have two modes. Notice modes of a hybrid waveguide enough to allow the waveguides that both modes experience total internal refraction, but they bounce to be reduced in size for modulator applications and still remain off the surfaces at different angles. This shows why the two modes effective. For an example, in a Mach-Zehnder modulator a tra- are different. Since the waves travel at the same speed inside the ditional waveguide arm would have to be about 2cm in length material but at different angles, they travel along the X axis at dif- while a hybrid waveguide arm would only have to be .5 to 1mm ferent speeds. The same principle applies to all guided waves in a in length to be equally effective. waveguide.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 53 OPTIMIZING HYBRID WAVEGUIDES

propagation constants; one when the index of the cladding layer is unshifted or without an applied voltage (for example 1.202), the other is when the index of the cladding layer is shifted due an applied voltage (for example 1.200).

Index of Refraction Looking at the effect the shift has on the propagation con- stant when the base index of refraction of the cladding layer is varied, it is found that over the range of 1 to 1.5 the changes is relatively constant. With dimensions of height 200nm, width 180nm and space 50nm, the change only varies by 6m-1 over an index change of 0.03, as illustrated in Fig. 3. A larger range of indexes was sampled and similar results were found.

Figure 2. Top: A power flow diagram of a hybrid waveguide’s guided mode. Bottom: Power flow diagram of a traditional waveguide guided mode.

The goal for this summer’s research is to discover how changing the dimensions of hybrid waveguides will affect the waveguide’s propagation constant and also what dimensions will produce the greatest change in the propagation constant due to a constant change in the index of refraction without allowing light Figure 3. The amount the propagation constant changes with a - .002 change of the index of the cladding layer. Data observed to form a guided mode in the silicon blocks. at height: 200nm, width: 180nm, and space: 50nm. To accomplish this task, a modeling program called FEM- LAB 3 was used for simulation of the structures. The dimen- It is important to note that the larger the index of cladding sions that can be changed are: the index of the cladding layer, the layer will result in a larger change in the index of refraction, ac- height and width of the blocks, and the space between the two cording to Eq. (1). This was not taken in to account in the above blocks (space). analysis.

RESULTS Height of the Silicon Blocks An EO polymer’s ability to alter the index of refraction can Adjusting the height of the silicon blocks has a much differ- be modeled by a simple equation: 3 ent effect than changing the index of the cladding. To determine the maximum change in propagation constant, a range of heights from 200nm to 390nm were sampled with a width of 200nm, a (1) space of 50nm, and a index of the cladding layer of 1.002 (see Fig. 4). A maximum change was found at 370nm. However, with

Where n is the original index of refraction, r33 is an elec- silicon blocks this tall, a second guided mode exists within the tro-optic coefficient, V is voltage applied, and d is the distance blocks themselves. The shorter the silicon blocks are, the more between electrodes. Using appropriate values, the typical Δn is the undesired guided mode becomes a substrate radiation mode. around -.002 to -.003. In this paper all shifts of index of refrac- The question then becomes, when does the guided mode in tion of the cladding layer will be -.002, and will be referred to the silicon blocks become primarily a substrate radiation mode as “the shift”, and “the change” or “change” will refer to how (see Fig. 5)? It seems that a useful range for heights will be be- much the propagation constant changes due to a -.002 change in tween 160nm to 270nm depending on the width of the silicon the index of refraction of the cladding layer. All the graphs will blocks. In this range the power going through the substrate will show the absolute change of the propagation constant. Also note be around half as much as the power going through the space. It that the graphed data is derived from finding the difference of two should also be noted that since the range of effective heights is far

54 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 GARDNER away from the height that has the maximum change, variations of the It is possible to get a second guided mode within a hybrid height will cause a noticeable difference in the propagation constant waveguide by increasing the width of the silicon blocks, but the and also how much it changes due to the shift, as shown in Fig. 6. blocks would have to be much wider than where the optimal di- mensions illustrated in Fig. 7. Also if the blocks are not tall enough to have a guided mode increasing the width will have little affect on how well guided the second, undesired mode will be.

Figure 4. The amount the propagation constant changes due to a - .002 change of the index of the cladding layer. Data observed at width: 200nm, space: 50nm, and index of cladding: 1 to 1.002

Figure 7. The amount the propagation constant changes due to a - Width of Silicon Blocks .002 change of the index of the cladding layer. Data observed at When the width of the silicon blocks was model at various di- height: 200nm, space: 50nm, and index of refraction: 1.2 to 1.202 mensions, the results were more pleasing than when the height was varied. Adjusting the width can produce a maximum change while Space between the Silicon Blocks the second undesired mode is remains unguided. This is convenient Looking at what happens when the space between the two because variations in the width due to manufacturing process will silicon blocks is the varied, a maximum change can be observed, not alter the change significantly (see Fig. 7). However, there is as shown in Fig. 8. Unlike height and width, the ideal space is not a universal ideal width. The ideal width depends on the height, at a currently hard to manufacture dimensions (typically around space, and the index of the cladding layer, but it was observed that 50nm). At currently manufacturable dimensions the change de- the range of ideal widths is between 170nm to 250nm and it is al- creases near linearly as space increases (see Fig. 8). ways 10 to 40nm less than the height of the silicon blocks.

Figure 8. The amount the propagation constant changes due to a - Figure 6. The amount the propagation constant changes due to a -.002 change .002 change of the index of the cladding layer. Data observed at of the index of the cladding layer over an appropriate range. Data observed height: 210nm, width: 200nm, and index of refraction: 1 to 1.002 at width: 180nm, space: 50nm, and index of refraction: 1.2 to 1.202

Figure 5. This shows how decreasing the height (370 nm, 250 nm, 230 nm, 200 nm, 170 nm) can direct the light in to the substrate.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 55 OPTIMIZING HYBRID WAVEGUIDES

At closer spaces the second undesired guided mode becomes more guided, and less power flows though the space. Wider spac- es will result in a less guided undesired mode for the modeling index of refraction and dimensions of the silicon blocks.

CONCLUSIONS As the index of refraction of the cladding layer increases, the change in propagation constant due to a 0.002 decrease in the index of the cladding remains largely constant. When the height is increased, a maximum change of propa- gation constant is observed due to a constant change in the index of refraction. However, as the silicon blocks get taller a second guided mode is also formed. A greater silicon width also results in a maximum in the change in propagation constant with a constant variation of the index of the cladding. Again, wider silicon blocks more strongly support a second, undesired guided mode. When the space between the two silicon blocks is increased, the change in propagation constant experiences a maximum at cur- rently difficult to manufacture dimensions. There is a tradeoff between how strongly guided the unde- sired mode is and how much the propagation constant can change. The larger the propagation constant change, the more strongly guided a mode is within the silicon blocks.

REFERENCES 1. Qianfan Xu, Vilson R. Almeida, Roberto R. Panepucci, and Micheal Lipson. Opt. Lett. Vol. 29, No. 14, July 15, 2004

2. Vilson R. Almeida, QianFan Xu, Carlos A. Barrios, and Michal Lipson. Opt. Lett. Vol. 29, No. 11, June 1, 2004.

3. G. L. Lee and P. K. L. Yu. Journal of Lightwave Technology, vol 21, no. 9 Sept 2003.

ACKNOWLEDGEMENTS Prof. Scott Dunham, Kjersti Kleven, (mentor), University of Washington Electrical Engineering Dept., Hooked on Photonics program, National Science Foundation, STC-MDITR

Andrew Gardner recently earned an AS in Engineering from Highline Community College and is currently in the Electrical Engineering Department at the University of Washington. He intends to graduate with a BSEE in spring 2007.

56 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis and Analysis of Thiol-Stabilized Nanoparticles

Eddie Howell Joe Perry and Wojtek Haske Norfolk State University Perry Lab, School of Chemistry and Biochemistry Georgia Institute of Technology

Dr. Carl Bonner Center for Materials Research Norfolk State University

INTRODUCTION ice bath and stirred for 15 minutes. A solution of sodium borohy- dride dissolved in ethanol was then added dropwise to the silver Metal nanoparticles have gained much interest in the sci- nitrate solution and allowed to reflux for another 15 minutes. The entific community in recent years due to the unique optical and resulting solution was then placed in a freezer and allowed to electronic properties that these particles possess. Potential appli- sit overnight, causing the thiol-coated nanoparticles to precipitate cations for metal nanoparticles include biological imaging, three- out of the suspension and settle to the bottom of the flask. The dimensional microfabrication, and optical data storage.1 nanoparticles were then filtered and washed sequentially with The primary obstacle scientists are currently trying to over- ethanol, distilled water, and acetone to remove any excess sodium come is identifying ways to manipulate the size, shape, and dis- borohydride. The particles were dried in a vacuum oven at room tributions of these nanoparticles. A characteristic of all metal temperature overnight. nanoparticles is that delocalized electrons exhibit collective oscil- lations in an electric field. These collective oscillations are com- monly referred to as plasmons. The particles can absorb energy from electromagnetic waves when the frequency is in resonance with the plasmons. Large local electric fields are present near the nanoparticle surface. This effect is called near-field enhancement of the electric field. It has been shown in some cases that this near field enhancement can enhance the nonlinear optical properties of molecules near the surface of the nanoparticle. The first four weeks of this research project were spent at Norfolk State University working with Dr. Carl Bonner to find suitable methods and reagents for synthesizing the silver nanopar- ticles. During the following five weeks in Dr. Perry’s research lab at Georgia Tech the research project has involved the synthesis of silver nanoparticles with different thiol ligands on their sur- face. Polymer nanocomposite films were prepared using these nanoparticles and their optical properties were investigated

METHODS Synthesis of Silver Nanoparticles A single-phase method was used for synthesizing the silver nanoparticles which involves reducing silver ions with sodium borohydride in the presence of thiol ligands (see Fig. 1 for the li- gands used). The process begins with making a solution of silver nitrate dissolved in ethanol. The desired amounts of thiol ligands were then added to the solution. The ratio or type of ligands for Figure 1. Thiol ligands used to coat the silver nanoparticles. each batch was varied, but the silver to total thiol ratio was held constant at 3:1 for all batches. The solution was then placed in an

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 57 SYNTHESIS AND ANALYSIS OF THIOL-STABILIZED NANOPARTICLES

(a) (b)

Figure 2. Absorption spectra of silver nanoparticles in polystyrene thin films. The particles were coated with octanethiol to dodecanethiol ratios of (a) 1:1 and (b) 3:1

Nanocomposite Films fraction measurements, but the particles were not very well dis- The silver nanoparticles were dissolved in a polystyrene persed throughout the film. This was likely due to the particles matrix and made into thin films to study their optical properties. aggregating overnight as the solvent evaporated. The first method used to make these films was spin coating. The To try and prevent the aggregation of particles the solutions nanoparticles and polystyrene were first dissolved in a solvent were filtered in the same way but the films were prepared at a such as toluene and placed in a small vial. A small amount of higher temperature. The temperature had to be lower than the solution was then drawn from the vial with a syringe. A glass boiling point of the solvent to keep the film from bubbling, so substrate was placed onto the spin coater and the solution was a solvent with a high boiling point was used (dichlorobenzene). filtered and dispensed onto the substrate evenly. The substrate The modified slides were placed on a hot plate and first heated was allowed to spin for two minutes at 1000 rpm. UV-vis ab- to 120°C. The nanocomposite solution was then filtered onto the sorption spectra were measured for these films with varying li- slides and allowed to sit for 1 hour. gand ratios and nanoparticle concentrations. Films made in this manner, however, were too thin to be used for x-ray diffraction RESULTS studies, so a different method, (solvent-evaporation method), was Absorption spectra were obtained for the nanocomposite employed. The glass substrates also had to be modified with an films made by spin coating (Fig. 2). The nanoparticles were syn- adhesion promoter to keep the polymer from peeling off of the thesized with octanethiol to dodecanethiol ligand ratios of 1:1 and glass slides. 3:1. The concentration of nanoparticles for each successive film Three drops of the adhesion promoter solution were put onto in each graph was decreased by one-half. The absorption spectra the slide and the slide was then rotated at 3000 rpm for 3 sec- for the 1:1 films exhibit a shoulder at about 400 nm, which as of onds on the spin coater. The slide was then heated at 90°C on a now cannot be fully explained. The absorption peaks of the 1:1 hot plate for 30 seconds. This process was repeated 3 times. Ini- films are narrower than those of the 3:1 films, which may corre- tially, to make the nanocomposite films, a microscope slide was late with a change in the shape and distribution of silver nanopar- covered with the nanoparticle/polymer solution and placed in an ticles within the polymer matrix. A blue shift in the position of the evaporation chamber overnight to allow the solvent to slowly plasmon resonance band is observed when the concentration of evaporate. The films obtained were thick enough for x-ray dif- nanoparticles in the film decreases.

58 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 HOWELL

CONCLUSIONS Changing the thiol ligands bonded to the silver nanoparticles has shown to alter the optical properties of the nanocomposites. This is likely due to the fact that different ligands or different ligand ratios in the case of mixed coatings, can affect the interac- tion between nanoparticles or between the nanoparticles and the polymer, and the arrangement in the solid matrices. For weak binding, as the concentration of nanoparticles decreases, the space between individual particles increases, and the interactions between them could become weaker. More studies are needed to determine why this causes a blue shift in the absorption peaks.

FUTURE WORK In the future, properties of chromophore thiol-coated metal nanoparticles will be investigated to determine how different ligands affect nonlinear optical properties of the nanoparticle- chromophore system. By using solvent evaporation methods, the films can be made with a thickness that is sufficient for use in x-ray diffraction studies. X-ray diffraction will be used to further study the spatial arrangements of the silver nanoparticles in poly- mer matrices.

ACKNOWLEDGEMENTS Dr. Joe Perry, Faculty Advisor, Georgia Institute of Technology Wojtek Haske, Graduate Student, Georgia Institute of Technology Perry Group, Georgia Institute of Technology Dr. Carl Bonner, Faculty Advisor, Norfolk State University Center for Materials Research, Norfolk State University

REFERENCES 1. Stellacci, F.; et al. J. Am. Chem. Soc. 2003, 125, 328

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 59 60 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Quinoxaline-Containing Polyfluorenes: Enhanced Blue Electroluminescence by Addition of a Hole Blocking Layer

Kelli A. Irvin Jessica M. Hancock and Samson A. Jenekhe Montana State University Deptartments of Chemical and Chemical Engineering University of Washington

INTRODUCTION sive layer in OLEDs to obtain stable blue electroluminescence. The motivation for incorporating the quinoxaline moiety was to The commercial market today is saturated with products us- improve the electron transport of commercially produced PFO. ing flat panel displays from cell phones to computer screens and Copolymerizing a p-type and n-type material could eventually personal data organizers to billboards. Currently, liquid crystal lead to the realization of a material capable of efficient electron displays (LCDs) dominate the market with high resolution and and hole transport. The configuration of the devices incorporated affordability1. The only challenger to this domination may per- poly(N-carbazole) (PVK) to enhance the injection of holes into haps be the organic light-emitting diode (OLED). OLEDs show the device. The work outlined in the proceeding sections intro- promise to consume less energy, lower costs due to ease of manu- duced a hole blocking layer into the architecture of the device in facturing of conjugated polymers, and be brighter than LCDs. In an effort to enhance brightness and efficiency. Ultimately, the addition, conjugated polymers have properties such as ease of true ability of QXF as an emitter will be shown or the realization processing using spin casting, low drive voltages due to the ex- of QXF as a bifunctional material will be demonstrated. cellent charge mobility, and novel properties such as flexibility.2 OLEDs don’t have the need for backlighting and the overall prod- uct would have a slimmer profile due to the nanoscale thickness of the layers used in processing. 3 Conjugated polymers were first recorded to be electrolumi- nescent in the 1990s by J.H. Burroughes, et al., using poly(p- phenylene vinylene), or PPV. 4 Since their initial discovery, great advancements have been made included tunable color, stable color across a voltage range, and more efficient devices using multiple layers. Figure 1. Chemical structures of (left) QXF and (right) TPBI. Challenges that need to be overcome for OLEDs include display lifetimes and stable true blue color. Poly(9,9’-dioctyl- Fabrication and Characterization of LEDs fluorenes), or PFO, has been widely researched as an emissive Three configurations of devices were fabricated using three se- 5 layer yielding stable blue color. PFO is a p-type material that lected molar ratios based on the quinoxaline content as the emissive transport holes and blocks electrons with high mobility through- layer. The three molar ratios will be noted as QXF2, QXF5, and out the delocalized conjugated chain. Although PFO is capable QXF15 (Figure 1a). All devices were fabricated with the same an- of efficiently transporting holes, it lacks the ability to efficiently ode and cathode using indium-tin oxide (ITO) and LiF/Al, respec- transport electrons and block holes. Copolymerization with an tively. Poly(ethylenedioxythiophene) doped with poly(styrene- n-type material shows promise to enhance the electron transport sulfonate) (PEDOT) was spin cast for improved injection of holes. properties of the polymer while maintaining the blue emission. The first device as shown in Figure 2a was a single layer LED with Therefore, more research is needed to find a material that exhibits the copolymer QXF as the emissive layer. PVK was incorporated stable blue color and has the ability to be a bifunctional emissive into the architecture of the second configuration (Figure 2b). The layer. third device (Figure 2c) included both the PVK layer and incorpo- rated 1,3,5-tris(N-phenylbenzimidizole-2-yl)benzene (TPBI) (see BACKGROUND Figure1b) as a hole blocking layer. The three architectures based on Previously, Kulkarni, et al.6, completed work using six molar ra- QXF as the emissive layer were ITO/PEDOT/QXF/LiF/Al (Diode tios of the copolymer 2,3-bis-(p-phenylene)quinoxaline-contain- I), ITO/PEDOT/PVK/QXF/LiF/Al (Diode II), and ITO/PEDOT/ ing poly(9,9’-dioctyl-fluorenes) (QXF) (Figure 1a) as the emis- PVK/QXF/TPBI/LiF/Al (Diode III).

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 61 photoluminescent (PL) emission spectra of each and QXF15 in (a) thin film and (b) solution molar ratio. Thin film data was collected from state. 50 nm thin films cast from 1% wt solutions in toluene. A Lambda 900 UV/Vis/Near-IR material. The QXF2, QXF5, and QXF15 spectrophotometer was used to record the copolymers showed maximum absorbance absorption spectrum. The PL emission was wavelengths of 385, 382, and 377 nm obtained using a Photon Technology respectively in solution state. As the molar ratio International spectrofluorimeter with an of quinoxaline increased, a blue shift of 8 nm excitation wavelength of 380 nm for all devices. was observed in the solution state absorption Table 1 shows all results for absorption and PL spectra. Likewise, the thin film exhibited a emission spectra. slight shift of 2 nm in the maximum absorption peaks of QXF15 (381 nm) to QXF5 (383 nm). Table 1. Photophysical Properties It is understood that increased conjugation length a b a b leads to red shifted absorption. Therefore, the QUINOXALINE-CONTAINING POLYFLUORENES: ENHANCED BLUE ELECTROLUMINESCENCE BY ADDITIONmax OFmax A HOLEmax BLOCKINGmax LAYER sol’n sol’n thin thin higher energy absorption band of the QXF15 relative to QXF2 is indicative of decreasing UV/Vis/Near-IRpolymer spectrophotometer (nm) (nm) was usedfilm to recordfilm the ab- (nm) (nm) conjugation length due to the increasing sorption spectrum. The PL emission was obtained using a Photon quinoxaline content. The quinoxaline content TechnologyQXF2 International 385 spectrofluorimeter 415 383 with an 438 excitation interrupts the conjugation along the carbon wavelengthQXF5 of 380 nm for 382 all devices. 415 Table 382 1 shows 437 all results backbone of the polymer by introducing kinks QXF15 377 415 381 428 and bends which disturb the delocalized electron for absorptiona and PL emission spectra. The absorption maximum in toluene solution or in cloud. a thin film. a b a b λmax λmax λmax λmax b Thin Film PL Solution State PL Emission The PL emissionsol’n maximum sol’n in toluenethin solutionthin or 1.2 E i i Figure 2. Device Configurations of (top left) ITO/PEDOT/QXF/LiF/ in polymera thin film. (nm) (nm) film film QXF2 QXF2 Al (Diode I), (top right) ITO/PEDOT/PVK/QXF/LiF/Al (Diode II), 1 QXF5 (nm) (nm) b QXF5 a QXF15 and (bottom) ITO/PEDOT/PVK/QXF/TPBI/LiF/Al (Diode III). QXF15 PhotophysicalQXF2 385 properties415 were383 measured438 in QXF5 382 415 382 437 0.8 solidQXF15 state and377 a solution415 state.381 The solution428 Glass coated with indium-tin oxide (ITO) was used as the state environment afforded a study of a single 0.6 a The absorption maximum in toluene solution or in a thin film. substrate and anode. The ITO glass was cleaned by a process b Thepolymer PL emission chain. maximum Conversely, in toluene solution the thin or in film a thin state film. 0.4 Normalized Intensity of twenty minute cycles of sonication in deionized (DI) water, a allows insight into the how the polymer chains 0.2 50:50 by volume mixture of DI water and isopropyl alcohol, tolu- interact withTable one1. Photophysical another. Properties Therefore, the thin film, or solid state, expresses the properties of a 0 ene, DI water, and acetone, respectively. The substrate was dried 400 450 500 550 600 450 500 550 600 system of polymer chains and how their overnight under vacuum. PEDOT was spin cast from a filtered Photophysical properties were measured in solid state and a Wavelength Wavelength interactions effect emission and absorption of () () solution state. The solution state environment afforded a study 1:3 dispersion of PEDOT and DI water. The PEDOT layer was wavelengths of light. of a single polymer chain. Conversely, the thin film state allows Figure 4. Photoluminescent Emission Spectra of spin cast onto the ITO to achieve on average a thickness of 55 The absorption spectra of both the thin film QXF2, QXF5, and QXF15 in (a) thin film and (b) nm. The PEDOT layer was heat treated at 180°C for ~15 minutes insight intoand the solution how the state polymer are shown chains in interact Figure with3a and one 3b. another. solution. under vacuum. Where applicable, a 10-15 nm layer of PVK was Therefore,The the absorption thin film, or spectra solid showstate, expresses the absorption the properties of The PL emission spectra for both thin film energy as electrons are promoted to various cast onto the PEDOT layer and dried four hours at 50°C under of a system of polymer chains and how their interactions effect and solution state are shown in Figure 4 with emission vibronicand absorption levels of withinwavelengths the excitedof light. state and maximum emission at 383 nm (QXF2) and 438 vacuum. The copolymer [QXF(2,5,15)] was cast subsequently allows for the study of the ground state of the nm (QXF2), respectively. The PL emission is a from a filtered 1% weight solution in toluene to achieve a 50-60 Thin Film Absorption Solution State Absorption snapshot excited state of the copolymer and 1.2 1.2 nm layer and was dried overnight under vacuum at 50°C. When emission is a result of the relaxation of excited QXF2 1 1 QXF2 applicable, a 20-25 nm layer of TPBI was deposited via thermal a QXF5 b QXF5 electrons from the lowest vibration level of the QXF15 -6 QXF1 5 excited state to the ground state. The solution evaporation at pressures of less than 3 x 10 torr and a deposition 0.8 0.8 rate of 0.2-0.3 nm/s.. The electrodes were also deposited using state PL emission spectrum (Figure 4b) is red 0.6 0.6 shifted ~34 nm from the solution state thermal evaporation. 2 nm of LiF was evaporated followed by 0.4 0.4 absorption spectrum, and the thin film PL NormalizedAbsorbance 145-150 nm of Al deposited without breaking vacuum at pres- NormalizedAbsorbance spectrum (Figure 4a) is red shifted ~50 from the sures <2 x 10-6 torr. The resulting diodes exhibited an area of 0.2 0.2 0.2 thin film absorption spectrum. The thin film PL 2 cm . 0 0 emission spectra shows a ~15 nm red shift 300 350 400 450 500300 350 400 450 500 relative to the solution state emission. The red Characterization of the devices was performed directly after Wavelength Wavelength () () shift in the PL spectra suggests increasing the deposition of the electrodes to obtain maximum results. Lu- Figure 3. Absorption Spectra of QXF2, QXF5, Figure 3. Absorption Spectra of QXF2, QXF5, and minance-current-voltage (L-I-V) characteristics were measured QXF15 in (a) thin film and (b) solution state. simultaneously using an optometer with a luminance sensor head to detect photons and a semiconductor parameter analyzer used to The absorption spectra of both the thin film and solution measure current density as a function of voltage. state are shown in Figure 3a and 3b. The absorption spectra show the absorption of energy as electrons are promoted to various vi- RESULTS AND DISCUSSION bronic levels within the excited state and allows for the study of Synthesis and Characterization the ground state of the material. The QXF2, QXF5, and QXF15 The synthesis and characterization of the varying molar ra- copolymers showed maximum absorbance wavelengths of 385, tios of the QXF copolymer were completed by Kulkarni, et al., 382, and 377 nm respectively in solution state. As the molar ra- and is discussed in detail in that work. 6 tio of quinoxaline increased, a blue shift of 8 nm was observed in the solution state absorption spectra. Likewise, the thin film exhibited a slight shift of 2 nm in the maximum absorption peaks Photophysical Properties of QXF15 (381 nm) to QXF5 (383 nm). It is understood that Dilute solutions (10-7 M) in toluene were used to investigate increased conjugation length leads to red shifted absorption. the optical absorption and photoluminescent (PL) emission spec- Therefore, the higher energy absorption band of the QXF15 rela- tra of each molar ratio. Thin film data was collected from 50 nm tive to QXF2 is indicative of decreasing conjugation length due thin films cast from 1% wt solutions in toluene. A Lambda 900

62 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 photoluminescent (PL) emission spectra of each and QXF15 in (a) thin film and (b) solution molar ratio. Thin film data was collected from state. 50 nm thin films cast from 1% wt solutions in toluene. A Lambda 900 UV/Vis/Near-IR material. The QXF2, QXF5, and QXF15 spectrophotometer was used to record the copolymers showed maximum absorbance absorption spectrum. The PL emission was wavelengths of 385, 382, and 377 nm obtained using a Photon Technology respectively in solution state. As the molar ratio International spectrofluorimeter with an of quinoxaline increased, a blue shift of 8 nm excitation wavelength of 380 nm for all devices. was observed in the solution state absorption Table 1 shows all results for absorption and PL spectra. Likewise, the thin film exhibited a emission spectra. slight shift of 2 nm in the maximum absorption peaks of QXF15 (381 nm) to QXF5 (383 nm). Table 1. Photophysical Properties It is understood that increased conjugation length a b a b leads to red shifted absorption. Therefore, the max max max max sol’n sol’n thin thin higher energy absorption band of the QXF15 polymer (nm) (nm) film film relative to QXF2 is indicative of decreasing IRVIN (nm) (nm) conjugation length due to the increasing to the increasingquinoxaline quinoxaline content. content. The quinoxaline The quinoxaline content content Electroluminescent Properties QXF2 385 415 383 438 interruptsinterrupts the conjugation the conjugation along the carbon along backbone the carbon of the poly- Table 2 shows the electroluminescent properties of Diodes I, QXF5 382 415 382 437 mer by introducingbackbone ofkinks the and polymer bends which by introducing disturb the kinks delocalized II, and III (Figure 2) based on QXF2, QXF5, and QXF15 as the QXF15 377 415 381 428 and bends which disturb the delocalized electron a electron cloud. emissive layers. The results for the single layer (Diode I) devices The absorption maximum in toluene solution or in cloud. a thin film. will not be shown graphically or in tabular form. The electronic b The PL emission maximum in toluene solution or Thin Film PL Solution State PL Emission 1.2 E i i structures for Diodes I, II, and II are shown in Figure 5. The val- in a thin film. QXF2 QXF2 1 QXF5 ues are of the electron affinity (EA) and ionization potential (IP) b QXF5 a QXF15 QXF15 of each material. The electronic structures of each diode show Photophysical properties were measured in 0.8 solid state and a solution state. The solution the extent of barrier to charge injection between adjacent layers state environment afforded a study of a single 0.6 and demonstrate the ease of charge mobility through the diode. polymer chain. Conversely, the thin film state 0.4 QXF has a low-lying IP of 5.8 eV relative to vacuum level (0 eV). Normalized Intensity allows insight into the how the polymer chains The addition of PEDOT as a hole injection layer between the an- interact with one another. Therefore, the thin 0.2 ode and the emissive layer minimized the barrier to hole injection, film, or solid state, expresses the properties of a 0 400 450 500 550 600 450 500 550 600 but a significant difference of 0.6 eV still existed between the IPs system of polymer chains and how their Wavelength Wavelength interactions effect emission and absorption of () () of PEDOT and QXF. PVK has an ionization potential of 5.8 eV wavelengths of light. FigureFigure 4. 4. PhotoluminescentPhotoluminescent Emission Emission Spectra of Spectra QXF2, of that approaches ohmic contact between the IP of QXF and the IP The absorption spectra of both the thin film QXF2,QXF5, QXF5, and QXF15 and QXF15 in (a) thin in film (a) and thin (b) filmsolution. and (b) of PVK implying that no barrier for the injection of holes into and solution state are shown in Figure 3a and 3b. solution. the emissive layer exists. Diode I was fabricated to insure that The absorption spectra show the absorption of The PL emission spectra for both thin film The PL emission spectra for both thin film and solution state Diode II was demonstrating enhanced device performance for all energy as electrons are promoted to various and solution state are shown in Figure 4 with vibronic levels within the excited state and are shownmaximum in Figure emission4 with maximum at 383 nmemission (QXF2) at 383 and nm 438 (QXF2) molar ratios due to the minimization of the barrier to charge in- allows for the study of the ground state of the and 438 nm (QXF2), respectively.respectively. The The PL PL emission emission is isa snapshota jection. Kulkarni, et al., completed the cyclic voltammetry of the Thin Film Absorption Solution State Absorption excited statesnapshot of the excitedcopolymer state and of emission the copolymer is a result andof the re- copolymer and found no significant differences in the ionization 1.2 1.2 laxation emissionof excited is electrons a result offrom the the relaxation lowest vibration of excited level of potential and electron affinity of the copolymer relative to PFO.6 QXF2 1 1 QXF2 a QXF5 b QXF5 electrons from the lowest vibration level of the QXF15 the excited state to the ground state. The solution state PL emis- However, it was expected that the addition of the n-type quinoxa- QXF1 5 excited state to the ground state. The solution 0.8 0.8 sion spectrum (Figure 4b) is red shifted ~34 nm from the solution line moiety would improve the electron transport properties of state PL emission spectrum (Figure 4b) is red 0.6 0.6 state absorptionshifted spectrum, ~34 nm and from the thin the film solution PL spectrum state (Figure the materials leading to a higher electron affinity. TPBI has a 4a) is red shifted ~50 from the thin film absorption spectrum. The low- lying ionization potential at 6.7 eV that blocks holes and 0.4 0.4 absorption spectrum, and the thin film PL NormalizedAbsorbance NormalizedAbsorbance thin filmspectrum PL emission (Figure spectra 4a) isshows red shifteda ~15 nm~50 red from shift the relative aides in confining charge recombination to the emissive layer. 0.2 0.2 to the solutionthin film state absorption emission. spectrum. The red The shift thin in filmthe PL PL spectra Therefore, the addition of the TPBI hole blocking layer could im- 0 0 emission spectra shows a ~15 nm red shift 300 350 400 450 500300 350 400 450 suggests500 increasing conjugation length as a result of interchain prove performance and efficiency of the devices by minimizing relative to the solution state emission. The red Wavelength Wavelength interactions between neighboring polymer and increased delocal- quenching of the charges at the interfaces. () () shift in the PL spectra suggests increasing Figure 3. Absorption Spectra of QXF2, QXF5, ization of electrons that resulted in a lower energy emission.

L J Va V λ EL EQEb LEc L max L max on max @ LE CIE 1931 (cd/m2) (mA/cm2) (V) (V) (nm) % (cd/A) (cd/m2) PVK Device QXF2 305 170 11 ~5 422 0.81 0.20 235 (0.16, 0.04) QXF5 1180 470 15 ~5 421 0.33 0.25 1180 (0.16, 0.04) QXF15 176 221 11 ~6 421 0.34 0.092 110 (0.16, 0.05) TPBI Device QXF2 1620 185 13 ~5 424 1.52 0.91 1530 (0.19, 0.09) QXF5 1510 265 12 ~5 437 1.77 0.71 115 (0.16, 0.05) QXF15 425 137 13 ~6 424 0.81 0.40 165 (0.15, 0.08) PFO 700 170 8 ~4 421 3.2 0.91 108 (0.16, 0.11)

a Drive Voltage, b External Quantum Efficiency is defined as the ratio of the number of photons emitted per number of electrons injected c Luminous Efficiency is a calculation made by dividing the maximum

luminance (Lmax) but the current density at the Lmax.

Table 2. Electroluminescent Characteristics

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 63 QUINOXALINE-CONTAINING POLYFLUORENES: ENHANCED BLUE ELECTROLUMINESCENCE BY ADDITION OF A HOLE BLOCKING LAYER

Figure 5. Electronic Structures of (a) Diode I, (b) Diode II, and (c) Diode III.

The EL spectra of Diode II and Diode II of each molar ratio are shown in Figure 6. No significant shifts in the peaks or line shape were found for any diodes or molar ratios over the applied voltage range. All diodes were blue emitting with an average CIE of ~(0.16, 0.04), which is similar to the true blue color emitted from the PFO homopolymer. On average, a blue shift of ~15 nm was shown for the EL emission relative to the thin film PL emis- sion. The current density-voltage and luminance-voltage graphs of Diodes II and III are shown for each molar ratio in Figures 7 and 8, respectively. The turn-on voltages for the QXF2 and QXF5 were ~5 V for all device configurations, while the QXF15 had consistently higher turn-on voltages of ~6 V. The quinoxa- line moiety improves electron injection, but the higher turn-on voltages for QXF15 suggest that the improved electron injection lead to a charge imbalance. Maximum luminances for Diode I were 230, 160, and 24 cd/m2 with luminous efficiencies (LE) of 0.05, 0.03, and 0.01 cd/ A for QXF2, QXF5, and QXF15, respectively. The EQE val- ues were very low at 0.19, 0.14, and 0.032% for QXF2, QXF5, and QXF15. The low brightnesses and efficiencies for the device Figure 7. (left column) Luminance-Voltage Characteristics were expected due to the large barrier to hole injection of 0.6 V of (top) QXF2, (middle) QXF5, and (bottom) QXF15. Red is Diode III, Blue is Diode II unless otherwise noted. between the PEDOT and QXF layers. Highest brightness for Di- ode II was exhibited by QXF5 with a luminance of 1180 cd/m2, Figure 8. (right column) Current Density-Voltage Characteristics of (top) QXF2, (middle) QXF5, and (bottom) QXF15. Red is Diode II, Blue is Diode III. LE of 0.25 cd/A, and EQE of 0.33%. For Diode III, the QXF2 copolymer exhibited the highest brightness, relative to QXF5 The addition of the TPBI hole blocking layer in Diode III for and QXF15, of 1623 cd/m2 with a LE of 0.91 cd/A and EQE of all molar ratios improved luminance and efficiency. For QXF2, 1.52%. Diode II had a maximum luminance of 310 cd/m2, LE of 0.20 cd/ A, and EQE of 0.81%. The addition of the TPBI layer in Diode III for QXF2 showed a brightness of 1620 cd/m2, LE of 0.91 cd/ A, and EQE of 1.52%. The brightness improved by a factor of 5 and the EQE showed an increase of 0.7% in Diode III relative to Diode II for QXF2. Similar improvements between Diode II and Diode III were shown for QXF5 and QXF15. These improve- ments in luminance, LE, and EQE between Diode II and Diode III for all molar ratios are due to the low lying IP of TPBI which aides in the containment of charge recombination to the emissive layer. Figure 6. EL Spectra of QXF2, QXF5, and QXF15 All molar ratios exhibited an increase in efficiency and bright- for (a) Diode II and (b) Diode III. ness when TPBI was added to the device architecture. However,

64 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 IRVIN

QXF2 and QXF5 displayed higher brightnesses for all diodes 5. Leclerc, M. Polyfluorenes: Twenty Year of Progress. Journal than QXF15. The increased quinoxaline content in QXF15 lead of Polymer Science: Part A: Polymer Chemistry 2001, Vol 39, to an imbalance in charge injection that resulted in poor device 2867-2873. results for all diodes relative to the results for QXF5 and QXF2. 6. Kulkarni, A.P.; Zhu, Y.; Jenekhe, S.A. Quinoxaline-Containing Kulkarni et al. observed a similar trend for molar ratios up to 50 Polyfluorenes: Synthesis, Photophysics, and Stable Blue E Elec- mol % quinoxaline (QXF50).6 troluminescence. Macromolecules 2005, 38, 1553-1563. For the Diode III architecture, QXF2 demonstrated a higher brightness of 1620 cd/m2 than the same device based on QXF5 ACKNOWLEDGEMENTS (L of 1510 cd/m2). However, the EQE value for Diode III based max Prof. Jenekhe, Jessica Hancock (mentor), Deptartments of on QXF5 was 1.77% which is higher than the EQE of the same Chemistry and of Chemical Engineering at the University of diode based on QXF2 (1.52%). For Diode II, a similar trend of Washington, Hooked on Photonics program, National Science higher brightness and lower EQE for QXF5 relative to QXF2 was Foundation, STC-MDITR exhibited. Therefore, better balance in charge injection cannot be conclusively attributed to either QXF2 or QXF5. A device with configuration of Diode III was constructed based on the homopolymer PFO as the emissive layer. A peak brightness of only 700 cd/m2 was shown for this diode. The brightness of 700 cd/m2 is significantly lower than peak brightness achieve for similar diodes based on QXF2 and QXF5. QXF2 and QXF5 demonstrated peak brighnesses >1510cd/m2. The addition of the n-type material improved brightness relative to PFO and Kelli Irvin is currently studying chemical engineering at Montana this improvement can be attributed to better electron transport in State University. She will be transferring to the University of the QXF2 and QXF5 copolymers. Washington in Winter of 2006 to complete her degree. Kelli intends to obtain an engineering position in industry upon graduation.

CONCLUSIONS QXF2 and QXF5 show promise as copolymers capable of stable blue electroluminescence over a wide voltage range. Di- ode III for QXF2 and QXF5 exhibited improved brightness from a similar diode using PFO as the emissive layer. The improve- ment in brightness suggests less quenching of luminance at the interface for devices based on a copolymer of n-type and p-type materials compared to a device using a p-type material as the emissive layer. The incorporation of TPBI as a hole blocking layer into a device with PVK as a hole injection layer exhibited enhanced brightness and efficiency when compared to a device without the TPBI layer. The power of n-type and p-type materials compolymerized as an emissive layer with the incorporation of a hole blocking layer into a device was demonstrated.

REFERENCES 1. Kelly, S.M.; Flat Panel Displays. In Flat Panel Displays: Ad- vanced Organic Materials; Connor, J.A.; RSC Materials Mono- graphs Series; R. Soc. Chem: Cambridge, UK 2000; pp 5-7. 2. Friend, R.; Burroughes, J.; Shimoda, T.; Polymer diodes. Phys- ics World 1999, pp. 35-36. 3. Howard, Webster E. Better Displays with Organic Films. Sci- entific American, Jan 12, 2004, pp 1-4. 4. Burroughes, J.H., et al. Light-emitting diodes based on conju- gated polymers. Nature 1990, 347, 539-541.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 65 66 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis of Dendron-Functionalized Chromophores: An Approach to Supramolecular Assembly for Electro-optic Applications

Zerubba Levi Sei-Hum Jang and Alex Jen Gonzaga University Department of Materials Science and Engineering University of Washington

INTRODUCTION Recent efforts in the Jen group have provided solid state evidence that arene-fluoroarene dendrons can produce supra- Organic nonlinear optical (NLO) chromophores present a molecular structures. The results provided by the Jen group il- number of applications in optical communication systems and lustrate chromophores capable of the arene-fluoroarene (Ar-ArF) storage devices. Basic research efforts are aimed at pushing the interactions produced enhanced r coefficients, T values, and bandwidths of optical networks to 80, 100 and even 160 GHz 33 g thermal stabilities relative to those chromophores that could only with the use of novel organic chromophores. The application of exhibit Ar-Ar or ArF-ArF interactions. This suggests the Ar-ArF organic electronics to commercial devices demands the thermal, interactions between monolithic dendrons ameliorate the poling chemical, and photo-stability of the materials. Balancing the op- process through π-π attractive forces that counteract the dipole- tical transparency while maximizing the nonlinear response and dipole forces among the chromophores. Despite the significant ensuring accentric geometry of the chromophores in bulk materi- improvements of the r coefficients of these systems over other als is also of critical importance. 33 materials, work still needs to be made to improve the thermal Recent work in the synthesis of NLO materials has exten- properties of the materials. sively used dendrimers to reduce dipole-dipole interactions be- tween chromophores through the site-isolation effect. As a result, much higher poling efficiencies, temporal alignment stabilities and nonlinear responses have been observed. In this project a series of NLO chromophores were synthesized with the careful selection of monolithic and binary dendron structures to produce pre-aligned systems with the capacity for supramolecular assem- bly upon electric field poling. Benzene and hexafluorobenzene (HFB) are known to cocrys- talize in a face-to-face manner through several noncovalent asso- ciative interactions such as electrostatic quadrupolar interactions (see Figure 1).1

Figure 2. Chromophore functionalized with dendrons

Studies using ab initio calculations performed with the DFT

method indicate the cohesive energy of heterodimers (Ec) in- creases as the electron density of the π-donating system increases (see Table 1).

Heterodimer EC (kcal/mol) Benzene/HFB 4-6 Anthracene/HFB 7.7 Pyrene/HFB 8.3 Triphenylene/HFB 8.6

2,3 Figure 1. Stacking arrangement of Benzene/HFB complex. Table 1. Cohesive Energies (EC) for Heterodimers

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 67 SYNTHESIS OF DENDRON-FUNCTIONALIZED CHROMOPHORES: AN APPROACH TO UPRAMOLECULAR ASSEMBLY FOR ELECTRO-OPTIC APPLICATIONS

Therefore, a monolithic anthracene-based dendron was syn- static quadrupole-quadrupole interactions between dendrons still thesized and covalently attached to a chromophore. It was hy- exists, but it is unclear whether the binary dendron systems offer pothesized that replacement of a phenyl group with an anthrace- any energetic or geometric advantages for supramolecular as- nyl group would lead to an increased cohesive energy between sembly. With this uncertainty present binary dendrons were also the dendrons (An-ArF) thereby strengthening the self-aligning synthesized with arene/fluoroarene moieties. properties of the chromophore. Furthermore, it was expected that the increase in noncovalent attraction between dendrons would EXPERIMENTAL result in a higher T and an increased thermal stability upon elec- g Synthesis of Dendron-Functionalized Chromophores tric field poling. The reaction scheme above illustrates an important synthetic While previous work has only been concerned with mono- simplification of the binary dendron systems (dendrons 1 and 2) lithic dendrons, it is important to inquire about the nature of bi- relative to the monolithic dendron systems (dendrons 3 and 4). nary dendron systems. The potential for noncovalent, electro-

λ in thin Materiala T (oC)b λ in soln. (nm)c max Poling Field (MV/cm) r (pm/V)d g max film (nm) 33 1 57 671 719 0.75 52 2 76 655 689 0.75 51 3 75 667 703 1 108 4 153 678 771 ------I. 122 670 720 0.75 95 II. 76 660 700 0.73 102 III. 122 670 720 0.75 51

a. Materials 1-4 synthesized previously by Jen Group o b. Measured by DSC (10 C/min in N2) c. Solution in 1,4-dioxane d. E-O coefficient measured by simple reflection at the wavelength of 1300 nm

Table 2. Results and Comparisons

68 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 LEVI

The binary dendron system requires a single deprotection step followed by a single catalytic esterification. Employing mono- lithic dendrons necessitate two separate deprotection reactions and two esterifications. Furthermore, r33 measurements (see Table 2 below) illustrate the binary dendron systems offer signifi- cant improvements in EO activity while maintaining the thermal properties of the monolithic dendron systems. Chromophore I. displays 100 % stability in the EO activity Zerubba Levi will be graduating from Gonzaga University in the spring of 2006 with a B.S. in Chemistry. Following graduation after 24 hours at 75OC; similar temporal stability is absent in ma- he intends to pursue a Ph.D. degree in Organic Chemistry. terials 1-4.

CONCLUSIONS In summary, a series of monolithic and binary dendrons were synthesized, and covalently attached to highly efficient NLO chromophores. The primary motivation of this work is fine tun- ing the strengths of arene-fluoroarene interactions between den- drons to produce pre-aligned supramolecular self-assembly to improve material stability and poling efficiencies. Anthracene was incorporated into the dendrons in an effort to increase the noncovalent, arene-fluoroarene interactions between dendrons. While the presence of anthracene increased the thermal stability of the materials, it is unclear for now whether the replacement of the phenyl ring with anthracene leads to stable pre-alignment or improved poling efficiency. However, the binary dendron sys- tems show significantly improved EO activities relative to pre- vious monolithic dendron systems studied. The basis for this improvement will be further studied. Full characterization of structures and material properties are ongoing.

REFERENCES 1. Patrick, C. R.; Prosser, G. S. Nature 1960, 187, 1021. 2. Collings, J. C.; Roscoe, K. P.; Robins, E. G.; Batsanov, A. S.; Stimson, L. M.; Howard, J. A.; Clark, S. J.; Marder, T. B. New J. Chem. 2002, 26, 1740-1746. 3. Castellano, R. K.; Diederich, F.; Meyer, E. A. Angew. Chem. Int. Ed. 2003, 42, 1210-1250.

ACKNOWLEDGEMENTS Many thanks are given to Tae-Dong Kim, and Zhengwei Shi for helpful discussions and important synthetic intermediates,

Steve Hau for sample poling and r33 measurements, and the NSF STC-MDITR 2005 Summer REU Program for funding.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 69 70 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Building an Optical Oximeter to Measure the Oxygen Content of Blood Non-invasively

JoAnn Lin Dr. Antao Chen University of Washington Applied Physics Laboratory University of Washington

INTRODUCTION photons collected have on average penetrated a shallower depth of tissue compared to photons collected from a more distant Light of certain wavelengths is able to penetrate deeply into source. the skin and may be absorbed or reflected by the blood. The An optical oximeter uses near infrared spectroscopy to mea- particular wavelength of light and the color of the blood deter- sure the level of oxygen in the blood. One application of optical mine whether the light is reflected or absorbed (Ma et al., 1999). oximetry is the detection of hemorrhages in the brain. Blood that The hemoglobin in red blood cells changes color when bound to is not circulating will appear darker due to lower oxygen concen- oxygen, causing oxygenated blood to appear bright red. The red trations. The oximeter uses low intensity laser light for measure- color is visible because oxygenated blood reflects red light and ment and is non-invasive. absorbs other wavelengths. Deoxygenated blood appears dark red because deoxyhemoglobin does not reflect much red light. Deoxyhemoglobin and oxyhemoglobin have significant differ- The Oximeter ences in absorption of light in the near infrared region (approxi- An optical oximeter was built to measure the oxygen content of mately 700 nm-900 nm, figure 1). blood, with the goal of focusing on measuring the oxygenation of blood in the brain. The oximeter consists of a probe that is placed over the skin on the head, circuits to process signal information, and a computerized system to record data. The components were tested separately to ensure they worked as expected. Wavelengths of 635 nm (red light) and 800 nm (near infrared light) are produced by laser diodes which are configured to oscil- late in brightness and are flashed on the skin. Only one wave- length is flashed at a time. The light penetrates the skin and blood vessels, and some of the light is reflected from the blood. The reflected light is collected by a photomultiplier tube (PMT) which converts the light signal into an electrical signal.

Signal Reference Photomultiplier Input Lock-In out: 1 kHz Tube (PMT) Amplifier

Voltage Controlled Figure 1. Taken from p. 373 of “Near-infrared spiroximetry: noninvasive Crystal Oscillator measurements of venous saturation in piglets and human subjects” (Franceschini et al., 2002). Near-infrared absorption spectra of 100 μM hemoglobin Phase Locked Loop Input 1 concentration for different values of oxygen saturation (SO2) in the range of Input 2 0-100%. The curve for 0% saturation corresponds with deoxyhemoglobin. Oximeter probe (held in place by elastic The curve for 100% saturation corresponds with oxyhemoglobin. band) Control voltage

Voltage Controlled Optical fiber Laser diode Crystal Oscillator Light entering the tissue is scattered in many different di- rections. A detector placed at a given location collects photons Target area – dashed line represents average path of that have on average traveled through the tissue along a banana- photons DC Bias shaped path. The depth of light penetration can be controlled Figure 2 by the choosing the distance between the light source and the . DiagramFigure of one 2. channelDiagram of the of oximeter. one channel 45 mA is usedof thefor the oximeter. DC bias. Circles 45 representmA is theused frequency mixers. The laser light for is intensity the DC modulated bias. Circles to allow therepresent al., the 1999). frequency The modulation mixers. for the oximeter was detector. If the detector and source are placed close together, the measurement of intensity as well as phase shift. chosen to be 54 MHz because the laser diodes had From this information the absorption and scattering higher intensities at this frequency than at the other may be determined (Hueber et al., 2001). Typically frequency choice of 100 MHz (figure 4). The setup in frequencies around 100 MHz or higher have been used figure 3 was used to test the frequency range in otherCMDITR oximeters Review (Franceschini of Undergraduate et al., 2002; Ma et Research available. Vol. 2 No. 1 Summer 2005 71

Oscilloscope Ch2 Ch1 Figure 3. Setup for the test of the laser diode and PMT signal. The oscilloscope (Tektronix model TDS Signal Generator 3054) is set to acquire at 64 PMT averages.

Laser Near infrared laser diode Supply Figure 4. Frequency range test for the red and near infrared laser diodes. The amplitude% PM ofT theAm signalp vs. Freque

from the PMT is greater around 50 MHz than at 100 MHz. In the oximeter, the light collected) will be a small fraction of the 2.5 %

light emitted by the laser diode sources, so it is important that the laser diodes emit at as high( an intensity as practical.

p 2.0 m Experiment tissue in preliminary testingA (figure 5). A 1/4” hole f

e 1.5 A white phantom was used to simulate biological in the phantom allows R a cylinder of black plastic or / p

red plastic to be inserted.m 1.0 The black plastic is expected to absorb light A and the red plastic is expected T

1-2 cm M 0.5

to reflect red light. TheP red laser diode was used in Source (laser Detector (optical diode, 635 nm) fiber to PMT) 0.0 10 20 30 40 50 60 70 80 90 100 Frequency (MH

Phantom BUILDING AN OPTICAL OXIMETER TO MEASURE THE OXYGEN CONTENT OF BLOOD NON-INVASIVELY

A 1/8” diameter optical fiber collects and directs light to the PMT. Voltage controlled crystal oscillators (VCXO) are used to provide the AC signal and a laser current supply is used to Signal Reference provide the DC power.Photomultiplier The laser diode is poweredInput to its Lock-Inlasing out: 1 kHz Tube (PMT) Amplifier threshold, so that the AC signal appears as a change in intensity. Changes in the oxygenation levels of the blood are expected to Voltage Controlled correspond with changesCrystal in the Oscillator signal amplitude. A phase locked loop (PLL) locks the phase between the VCXO signals which Phase Locked Loop prevents drift in the frequency. Frequency mixers Inputoutput 1 the dif- Input 2 Oximeter probe (held ference betweenin place by elastic two input frequencies (figure 2). Theband) laser light is intensity modulated to allow the measureControl- voltage Voltage Controlled ment of intensityOptical fiber as well asLaser phase diode shift. From this informationCrystal Oscillator the absorption and scattering may be determined (Hueber et al., Target area – dashed line 2001). Typicallyrepresents frequencies average path of around 100 MHz or higher have photons been used in other oximeters (FranceschiniDC Bias et al., 2002; Ma et al., 1999). The modulation for the oximeter was chosen to be Figure 2. Diagram of one channel of the oximeter. 45 mA is used for the DC bias. Circles represent the frequency mixers. 54 MHz because the laser diodes had higher intensities at this The laser light is intensity modulated to allow the al., 1999). The modulation for the oximeter was frequencymeasurement than at of the intensity other as wellfrequency as phase shift.choice ofchosen 100 to MHz be 54 MHz(figure because the laser diodes had From this information the absorption and scattering higher intensities at this frequency than at the other 4). Themay setup be determined in figure (Hueber 3 et al.,was 2001). used Typically to test thefrequency frequency choice of 100range MHz (figure 4). The setup in frequencies around 100 MHz or higher have been used figure 3 was used to test the frequency range available.in other oximeters (Franceschini et al., 2002; Ma et available.

Oscilloscope Ch2 Ch1 Figure 3. Setup for the test of the laser diode and PMT signal. The oscilloscope (Tektronix model TDS Signal Generator 3054) is set to acquire at 64 PMT averages.

Figure 4. Frequency range test for the red and near infrared laser Laser Near infrared laser diode Supply Figure 4. Frequency range test for the red and near infrareddiodes. laser The amplitude of the signal from the PMT is greater around diodes. The amplitude% PM ofT theA50m signalp MHzvs. Fre thanque at 100 MHz. In the oximeter, the light collected will be

from the PMT is greater around 50 MHz than at 100 MHz. In the oximeter, the light collected) will be a small fraction of the 2.5 a small fraction of the light emitted by the laser diode sources, so it is Figure 3. Setup for the test of the laser diode and PMT signal. The % light emitted by the laser diode sources, so it is important that the laser diodes emit at as high( an intensity as practical.

p 2.0 important that the laser diodes emit at as high an intensity as practical. oscilloscope (Tektronix model TDS 3054) is set to acquire at 64 averages.m Experiment tissue in preliminary testingA (figure 5). A 1/4” hole f

e 1.5 A white phantom was used to simulate biological in the phantom allows R a cylinder of black plastic or / p

red plastic to be inserted.m 1.0 The black plastic is expected to absorb lightA and the red plastic is expected T

1-2EXPERIMENT cm M 0.5 to reflect red light. TheP red laser diode was used in Source (laser Detector (optical diode, 635 nm) fiber to PMT) 0.0 A white phantom was used to simulate biological tissue in10 20 30 40 50 60 70 80 90 100 preliminary testing (figure 5). A 1/4” hole in the phantom allows Frequency (MH a cylinder of black plastic or red plastic to be inserted. The black Phantom plastic is expected to absorb light and the red plastic is expected to reflect red light. The red laser diode was used in the test. The sig- nal was collected through a BNC board using the NI-DAQ system with LabView 6.0. The oximeter was tested with a source-detector separation of 1 cm (figure 6) and 2 cm (figure 7).

Figure 6. Oximeter tested using the red laser source (635 nm) with 1 cm between the source and detector. Left: Signal amplitude when the red and black plastic are alternated. Right: Signal amplitude with only the Figure 5. Phantom test for the oximeter. The phantom is placed so the hole red or black plastic in the phantom. The amplitude for the red plastic is equidistant from the centers of the source and detector. The dashed line is larger relative to the amplitude for the black plastic. Differences in represents the average path traveled by photons collected by the detector. the red/black values between tests are most likely due to irregularities in the testing procedure (e.g. phantom was not well secured).

72 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 LIN

ACKNOWLEDGEMENTS This research funded by the National Science Foundation STC-MDITR and Research Experiences for Undergraduates pro- gram. Dr. Antao Chen of the Applied Physics Laboratory at the University of Washington mentored the research.

JoAnn Lin plans to continue research work and is currently pursuing a bachelor’s degree in Bioengineering at the University of Washington.

Figure 7. Oximeter tested using the red laser source (635 nm) with 2 cm between the source and detector. Left: Red and black plastic are alternated. Right: Only the red or black plastic in the phantom. The amplitude for the red plastic is larger relative to the amplitude for the black plastic. Differences in the red/black values between tests are most likely due to irregularities in the testing procedure. At 2 cm separation, the light collected by the PMT is more scattered and has lower intensity than at 1 cm. The noise in the signal is significant at 2 cm separation.

FUTURE WORK One channel of the oximeter was successfully built. Test- ing is ongoing, and will include the phantom test with the near infrared laser diode and a greater variation of source-detector separation. The spacing of the optical fiber and laser diode on the oximeter probe will be adjusted according to the values for maximum and optimum source-detector separation. Future work includes building another channel for the sec- ond laser diode and reducing noise in the signal. The oximeter must also be made more compact for practical use. In the future the oximeter may be tested on live subjects (pending approval).

REFERENCES 1. Franceschini, M.A., et. al. J. Appl. Physiol. 2002, 92, 372-384. 2. Hueber, D.M., et. al. Phys. Med. Biol. 2001, 46, 41-62. 3. Ma, H.Y., et. al. Proc. SPIE 1999, 3597, 642-9.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 73 74 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Moisture Uptake of Thin Polymer Films

Ephriam Lucas Dr. Sue-Ann Bidstrup Allen and Dr. Paul Kohl Georgia Institute of Technology Georgia Institute of Technology

Ate He Fort Valley State University

Thin films are layers of either conductive or non-conductive ACKNOWLEDGEMENTS films that are added to the top surface of wafers. These films Research support is gratefully acknowledged from the Na- may be used to form interconnects between devices or insulators tional Science Foundation Center on Materials and Devices for between interconnects layers. A variety of materials can be de- Information Technology Research (CMDITR), DMR-0120967. posited on a wafer. Three main categories of thin films materials are conductors, insulators, and semiconductors. In this research a polymer “insulator” is used as the thin film. This polymer film is used and tested to see how much moisture it absorbs under different circumstances. These films can later be used to coat microchips for computer, cell phones etc. A slide of quartz crystal with a gold electrode on each side is used in this project. After coating the quartz with a thin poly- mer film on one side, the quartz crystal is attached in a sealed glass cell. The cell is connected to nitrogen to test the moisture absorption. The nitrogen enters into the cell in two forms: wet nitrogen and dry nitrogen. The nitrogen that has been condensed with water is called wet nitrogen, and the direct nitrogen flow is dry nitrogen. While using dry nitrogen, the relative humidity is zero, which means that there is no water at all in the cell. When a voltage is applied to the quartz, it vibrates and the frequency is proportional to the mass of the film. Once the mass of the film is changed by moisture uptake, the frequency reflects that change. The more moisture the film absorbs the heavier the film will be causing the frequency to be lower. The theory behind this is the higher the condenser temperature, the higher the moisture content inside the cell, which increasing the relative humidity. The higher the relative humidity, the more moisture the polymer film will uptake, which increases the mass of the thin polymer film. There has not been much progress in this research as of yet due to inaccurate data. The data recorded show that there was a 2.5% increase in weight when changing from dry nitrogen to wet nitrogen and a 2.21% decrease in weight when changing back from wet to dry. This tells us that all the moisture is not being reabsorbs. We believe the frequency values are inaccurate due to the fact that the nitrogen flow is not constant. The nitrogen flow rate constantly fluctuates when changing from wet to dry and dry back to wet. A way to stabilize the nitrogen flow at a slow steady rate is needed before this research can be continued.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 75 76 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Toward Molecular Resolution c-AFM with Carbon Nanotube Tips: Development of Carbon Nanotube Growth Techniques

Austin McLeod Neal R. Armstrong, Paul Lee and Ken Nebesny Northern Arizona University Armstrong Lab, Department of Chemistry The University of Arizona

Carbon nanotube tips (CNTs) will be formed for use in nanometer scale imaging in atomic force microscopy. Optimized versions of these tips promise to provide for very high-resolu- tion imaging, minimization of sample damage due to the weak interactions anticipated be- tween the tip and soft materials (organic thin films), and the prospect for measuring electrical properties of organic thin film, owing to the electrically conductive nature of the CNT.

OBJECTIVES / THESIS be to grow carbon nanotubes on silicon nitride surfaces because that is what AFM tips are made from, growing nanotubes directly AFM nanometer-scale imaging is hindered by standard AFM onto the AFM tip at the correct orientation might be difficult, so a tips because they have poor aspect ratios, which can cause sample benchmark experiment was preformed. damage during the investigation of the surfaces of organic films. In our first experiment we tried to grow carbon nanotubes on AFM tips modified with carbon nanotubes (CNTs) should help a silicon nitride wafer using a ferric nitrate, Fe(NO ) • 9H O, greatly to increase the resolution, owing to the small diameter 3 3 2 solution to coat the silicon surface with iron particles. Chemi- of the CNT and excellent aspect ratios. The Dai group at Stan- cal vapor deposition (CVD) was utilized to grow the SWNT in a ford University has developed a process to grow carbon nano- quartz tube using acetylene as the carbon source and ultra-high tubes on silicon wafer tips in mass quantities.1 Our approach is purity (UHP) nitrogen gas as the carrier gas. AFM was used to to grow nanotubes in carpet like fashion and pick them up using characterize the silicon nitride surface. AFM showed that the the method of dielectrophoresis developed by Hyung Woo Lee.2 ferric nitrate iron source yielded, what looked like, “megatubes” Ultimately, we anticipate growing carbon nanotubes directly onto (Figure 1). individual AFM tips. Most carbon nanotubes are about 10-50 nm in diameter and between 100 – 1000 nm long, some have been grown to about 2 mm in length, but that length will not be necessary for our work. Carbon nanotubes are typically grown from a catalytic nanopar- ticle surface (the iron particle), which becomes the nucleation site of the nanotube’s growth and a carbon source, which becomes the structure of the nanotube. Nanotubes with a 10:1 aspect ratio would be preferred, meaning about 100-200 nm long and 10-20 nm in diameter.

RESEARCH METHODS / RESULTS Carbon nanotubes are typically grown via chemical vapor de- position at low pressures, using catalytic nanoparticles to initiate tube growth and location of the tube. Usually, carbon nanotubes attach very tightly to the nanoparticles, making separation dif- Figure 1. AFM image of CVD with Ferric Nitrate Catalyst source ficult. Carbon nanotubes can also be attached directly to con- ventional AFM tips. A group from the Korea Advanced Institute The next technique involved using an iron source that was al- of Science and Technology has developed a technique to attach ready on the nanometer scale. Ferrosound EMG 1111 nanoparti- 2 the nanotubes directly to the AFM tip. Our first attempts will cles made by Ferrotec, Inc were cleaned and separated in ethanol,

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 77 TOWARD MOLECULAR RESOLUTION C-AFM WITH CARBON NANOTUBE TIPS: DEVELOPMENT OF CARBON NANOTUBE GROWTH TECHNIQUES and they were spun-coated on the surface of the silicon nitride. AFM images show the nanoparticles had an even distribution across the surface (Figure 2). The carbon nanotubes grown on these surfaces were very difficult to image with AFM because some soot was forming on the surface.

Figure 4: FESEM image of a carbon nanotube

In Figure 5, the nanotubes shown in the FESEM image show bright spots, which are believed to be iron particles, the nucle-

Figure 2. Iron nanoparticles on silicon nitride surface ation site of the nanotubes. These nanotubes appear to be about 20 nm in diameter and about 200 nm between each bright spot. The last technique utilized ferrocene as the iron source, and Recent findings show that CNTs grow for a certain amount of xylenes as the carbon source.3 The CVD was preformed in a time (length) before they start to close up, but before they close reducing atmosphere of hydrogen gas, and in the presence of a up another iron particle inserts itself and keeps the tube open, and 4 carrier gas, argon. The wafers were cleaned in “piranha” for 30 that particle becomes the spot for the next CNT to grow from. minutes (4:1 ratio of sulfuric acid and 30% hydrogen peroxide), rinsed with nanopure water, and dried with compressed nitrogen gas before they were placed inside the quartz tube. The wafers were placed in the quartz tube using a measuring tape into the “hot zone.” The “hot zone” from prior experiments was deter- mined to be in the middle of the furnace, the hottest part of the center of the furnace. (Figure 3).

Figure 3. The “hot zone” is circled in red in figure 3 and indicated by the blackened nanotubes area.

Field Emission Scanning Electron Microscopy (FESEM) Figure 5: FESEM image of a carbon nanotube characterized the silicon nitride wafers with the carbon nano- tubes. FESEM was used to characterize the surface because the Electron Dispersion X-ray Spectroscopy (EDX or EDS) was resolution is much higher compared to standard SEM. In Fig- performed on one of these bright spots (Figure 5), but since these ure 4, the carbon nanotubes is of excellent aspect ratio (about 10 are on silicon substrate, the silicon signal is much too high to see length to 1 diameter). This SWNT is about 20 nm in diameter if there is indeed iron within the nanotubes (Figure 6). and about 200 nm long.

78 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 MCLEOD

Figure 6. EDX of Substrate

FUTURE WORK An experiment to perform, would be to grow the nanotubes on gold foil, and then perform EDX on the gold foil to see if a carbon and iron peak show up on the gold because the gold peak is shifted over from the silicon peak. Carbon nanotubes can be grown in carpet like fashion as seen by Maldonado et al. (Figure 7). Dielectrophoresis can be preformed on the nanotubes to attach them directly to the AFM tip (Figure 8).2 Although, it might be possible to grow a carbon nanotube directly onto an AFM tip, the alignment of the tip and Figure 9: AFM tip with CNT attached the carbon nanotube is crucial to obtaining an excellent AFM im- 5 age see figure 9. Therefore growing “nanotube carpets” followed CONCLUSIONS by dielectric attachment of the nanotube enables more control of the alignment of the AFM tip with respect to the CNT. Future When carbon nanotubes can be attached, repeatedly, to AFM work needs to be done to attach the CNTs to the AFM tip.5 tips, they will become very important for performing conductive atomic force microscopy (c-AFM) measurements and increasing the resolution of the AFM image. In addition to decreasing the aspect ratio of AFM tips using CNT they also have interesting properties that also can affect surface measurements. Carbon nanotubes are almost entirely composed of carbon, and carbon will have much lower Van der Waals interaction with the organic thin film surface, which means that the surface damage will be minimized. Most AFM tips are made of silicon, and these types of AFM tips do cause sample damage to soft materials. AFM tips can also be made of gold, platinum, or boron doped diamond tips. These conductive tips can cause sample damage, but these carbon nanotube tips will enable scientists to perform c-AFM ex- Figure 7: “Nanotube Carpets” periments to soft samples without sample damage.

REFERENCES 1 Erhan Yenilmez, Qian Wang, Robert J. Chen, Dunwei Wang, and Hongjie Dai. Wafer scale production of carbon nanotube scanning probe tips for atomic force microscopy. Applied Phys- ics Letters, 2002, 80, 2225-2227. 2 Hyung Woo Lee, Soo Hyun Kim, and Yoon Keun Kwak, Chang Soo Han. Nanoscale fabrication of a single multiwalled carbon nanotubes attached atomic force microscope tip using an electric field. Review of Scientific Instruments, 2005, 76, 046108-1. 3 Figure 8: Dielectric Attachment of CNT to AFM tip Stephen Maldonado and Keith J. Stevenson *. Influence of Ni-

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 79 TOWARD MOLECULAR RESOLUTION C-AFM WITH CARBON NANOTUBE TIPS: DEVELOPMENT OF CARBON NANOTUBE GROWTH TECHNIQUES trogen Doping on Oxygen Reduction Electrocatalysis at Carbon Nanofiber Electrodes. J. Phys. Chem. B 2005, 109, 4707-4716. 3 Stephen Maldonado and Keith J. Stevenson *. Direct Prepa- ration of Carbon Nanofiber Electrodes via Pyrolysis of Iron(II) Phthalocyanine: Electrocatalytic Aspects for Oxygen Reduction. J. Phys. Chem. B 2004, 108, 11375-11383. 4 Christian P. Deck, Kenneth Vecchio *. Growth mechanism of vapor phase CVD-grown multi-walled carbon nanotubes. Car- bon 43, 2005, 2608–2617. 5 R. Schlaf* et al. Using carbon nanotubes catilevers in scanning probe metrology. In press.

ACKNOWLEDGEMENTS -Dr. Neal R. Armstrong, Professor of Chemistry and Optical Sciences, University of Arizona -Armstrong Research Group -Dr. Ken Nebesny, associate staff scientist, LESSA facility, Uni- versity of Arizona -Paul Lee, assistant staff scientist, LESSA Facility, University of Arizona -Clayton Shallcross, Graduate Student, University of Arizona -Alex Veneman, Graduate Student, University of Arizona -Margo Ellis – FESEM operator -Dr. Timothy Vail, Dept. of Chemistry, Northern Arizona Uni- versity

Funding provided by the Center on Materials and Devices for Information Technology Research (CMDITR), an NSF Sci- ence and Technology Center No. DMR 0120967

I’d like to thank Dr. Neal R. Armstrong, and his group for mentoring me this summer at the University of Arizona.

80 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis and Characterization of Extended Squaraine Compounds

Tehetena Mesganaw Shijun Zheng and Seth Marder Georgia Institute of Technology Marder Lab, School of Chemistry and Biochemistry Georgia Institute of Technology

INTRODUCTION Organic dyes encompass a broad range of applications that are being investigated in the quest for new advancements in vari- ous technological areas. Squaraines are a novel class of organic Scheme 2. A revised synthetic scheme to successfully isolate (2). dyes that have delocalized π-electron systems that absorb at long (near infrared) wavelengths. These extended squaraines with al- Synthesis of (3) kyl side chains can easily form good optical quality, thin films. 1H-Pyrrole-2-carbaldehyde was reacted with 3-bromometh- These molecules are nonlinear optical materials and as such can yl-heptane with excess of NaH and DMF under N2. After 5 days potentially be used for optical switching applications. In a nonlin- at room temperature, the solution was washed with ice cold water ear optical material, when the light intensity shined on the mate- and extracted with ethyl ether (×6). It was washed with deionized rial increases, the electron cloud distorts, and the refractive index water (×4) and dried over magnesium sulfate. Column Chroma- changes. As a result, the way light propagates through the mate- tography, eluent hexanes in silica, was used for purification to rial changes. Therefore, squaraine compounds are on the rise and give (3), 1-(2-ethyl-hexyl)-1H-pyrrole-2-carbaldehyde. demands for these dyes will increase rapidly in years to come.

RESEARCH METHOD Synthesis of (1) The synthesis of (1) provides one of the building blocks for Scheme 3. Synthesis of pyrrole (3) that will be the final squaraine compound. Excess pyrrole was reacted with used for the condensation with squaric acid. 11-bromomethyl-tricosane in the presence of NaH and DMF. The reaction was done under N2 at 60 ºC and left for 3 days. The de- Synthesis of (4) 1 sired product (1) resulted in a yield of <50%. By H NMR it was Dibutyl-phenyl-amine was reacted with POCl3, in a Vilsmeir concluded that an elimination product (E2) formed because of the formylation to give (4)1. 4-Dibutylamino-benzaldehyde, (4), was heat added to the reaction, and the fact that pyrrole anion is very reduced to an alcohol by sodium borohydride in ethanol. After 45 basic also aided in the formation of the elimination product. minutes at room temperature, the solution was concentrated and aqueous ammonium chloride was added. Fresh 100 mL of dichlo- romethane was added and the organic layer was extracted (×2), dried over magnesium sulfate and concentrated to give a yellow oil (5). (4-Dibutylamino-phenyl)-methanol was treated with triethyl

phosphite in iodine at 0ºC and left overnight under N2. Vacuum distillation was performed at 60 ºC and remaining liquid was purified by column chromatography, eluent 1:1 hexanes to ethyl acetate to give rise to (6). By a Horner-Emmons condensation, Scheme 1. The attempted synthesis of the pyrrole that will compound (6) (4-Dibutylamino-benzyl)-phosphonic acid diethyl be used for squaraine resulted in a low yield of (1). ester in dry THF was reacted with 3,4-Dibutoxy-thiophene-2- carbaldehyde in a solution of potassium tert-butoxide in dry THF Synthesis of (2) under N . After an hour, the concentrated solution was purified by 1H-pyrrole-2-carbaldehyde was reacted with excess 11-bro- 2 column chromatography, eluent 10:1 hexanes to ethyl acetate in momethyl-tricosane in NaH and DMF under N to yield (2). 2 silica to give the compound (7).

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 81 SYNTHESIS AND CHARACTERIZATION OF EXTENDED SQUARAINE COMPOUNDS

Scheme 4. Synthesis of the desired product, squaraine (11) using (3).

Dibutyl-{4-[2-(3,4-dibutoxy-thiophen-2-yl)-vinyl]-phenyl}- Many of these reactions caused side products which required amine, (7) was then treated with butyl lithium in dry THF at 0 ºC, careful purification techniques that took longer than expected. after two hours DMF was added and left overnight all under N2. Therefore, the desired product squaraine is one reaction away The red solution was washed with water and extracted with ethyl from completion. After pure squaraine is obtained, tests will be acetate (3×). The organic phase is collected and dried over mag- performed on the organic dye to determine its efficacy in produc- nesium sulfate. The concentrated solution was purified by column ing optical quality, thin films. chromatography, eluent 10:1 hexanes to ethyl acetate in silica to give (8). 3,4-dibutoxy-5-[2-(4-dibutylamino-phenyl)-vinyl]-thio- REFERENCES phene-2-carbaldehyde, (8), was reduced by sodium borohydride M.J. Plater, T. Jackson. Tetrahedron 59 (2003). pages 4673-4685 in ethanol, similar to the procedure for compound (5), which was (page 4679). then reacted with tri-ethyl phosphate in iodine at 0 ºC and left overnight under N . Vacuum distillation was performed on the 2 ACKNOWLEDGEMENTS solution for two hours and purified by column chromatography, eluent 4:1 hexanes to ethyl acetate to yield (9). Research support is gratefully acknowledged from the Na- Compound (3), 1-(2-ethyl-hexyl)-1H-pyrrole-2-carbalde- tional Science Foundation Center on Materials and Devices for hyde was reacted with {3,4-Dibutoxy-5-[2-(4-dibutylamino-phe- Information Technology Research (CMDITR), DMR-0120967. nyl)-vinyl]-thiophen-2-ylmethyl}-phosphonic acid diethyl ester, compound (9), in dry THF and a solution of potassium tert-bu- toxide in dry THF by a Horner-Emmons Condensation, under

N2, for an hour and a half. The concentrated oil was purified by column chromatography, eluent 10:1 hexanes to ethyl acetate in silica to give rise to compound (10). Dibutyl-{4-[2-(3,4-dibu- toxy-5-{2-[1-(2-ethyl-hexyl)-1H-pyrrol-2-yl]-vinyl}-thiophen- 2-yl)-vinyl]-phenyl}-amine, (10) is treated with squaric acid to Tehetena Mesganaw is currently attending the Georgia Institute of give the desired compound (11) Squaraine. Technology and majoring in Chemistry. After graduation im Fall 2006, she plans to attend graduate school and obtain her Ph.D in Organic Chemistry. From there, she plans to do research on the AIDS epidemic. CONCLUSION After several failed reactions, it was determined that an al- kylated pyrrole carbaldehyde can be synthesized in high yield.

82 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Enhanced Heat Dissipation Substrates for Organic Semiconductor Devices

Aaron Montgomery Dr. Samuel Graham University of Virginia Georgia Institute of Technology Ashante Allen, Erik Sunden and Adam Christensen

`Many electronic devices are susceptible to overheating be accomplished at 460 degrees Celsius for 5 minutes and is used which can result in failures of the device. This can be exempli- to for creating microfluidic channels for active cooling. Con- fied through electronic chips located inside a computer. To ad- cerning polymers, a hydraulic press successfully bonded PMMA dress these failures, much research has been performed to de- with a carbon nanofiber interlayer at 250 degrees Fahrenheit velop thermal management solutions for microelectronic devices. while maintaining 400lbs of force for 5 minutes (Figure 2). This These solutions have generally been geared towards Si-based laminated structure was used to create a flexible heat spreader microelectronics and have achieved cooling capabilities of 100 in polymer substrates for improve heat dissipation. Future work W/cm2 or greater. While such power densities are not expected in will involve finding a feasible bonding method for PET which is organic semiconductors, these devices have unique thermal man- more commonly used for flexible organic semiconductors and to agement challenges arising from their inherent low thermal con- continue to produce CNT growth onto metal substrates with the ductivity (results in high thermal resistance), the use of thermally least amount of defects. resistive substrates for flexible electronics, and the need to have transparent materials for photon transfer. Thus, new concepts for both active and passive thermal management of organic semicon- ductor devices (OSD) must be explored. In this work, I will primarily concentrate on developing new schemes for the removal of thermal energy through both active and passive mechanisms by “thermally connecting” the OSD to high thermal conductivity substrates. The thermal connection will be based on carbon nanotubes (CNTs) which will act as a thermal interface material (TIM) with superior properties to convention- al TIMs. These CNTs possess a very high thermal conductiv- ity (900-10,000w/mk). I will investigate the use of multilayer Figure 1. Growth of carbon nanotubes on a copper substrate (left) and a scanning electron microscope image showing the details of the growth (right). catalysts to produce highly aligned CNTs on metal substrates and the creation of actively cooled PMMA and Si (gold coated) sub- strates using various bonding techniques. There is a severe lack of attention on the thermal characterization and heat dissipation in OSD. Much of what I do here will be new and provide much needed contributions to the challenges of thermal management in OSDs. In addition, with overheating of chips stalls the advance- ment of better electronic devices. The multilayer catalysts deposition onto metal substrates has proven successful in producing some CNT growth (Figure Figure 2. Growth of carbon nanofibers (left) which were laminated between sheets of PMMA to create a flexible heat 1). Some parameters that contribute to this growth are the type spreader (right) integrated into a polymer substrate. of catalyst deposited onto the metal, the maximum temperature during the procedure, and the length that the sample is exposed to ACKNOWLEDGEMENTS the gases. By changing these parameters I am hoping to produce Research support is gratefully acknowledged from the Na- a recipe that will generate better quality CNT growth with fewer tional Science Foundation Center on Materials and Devices for defects. Also, success was achieved at bonding Si (gold coated) Information Technology Research (CMDITR), DMR-0120967. to Si and PMMA to PMMA. Using a furnace, Si bonding can

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 83 My name is Aaron Montgomery and I am currently a third year student at the University of Virginia. My current major is Mechanical Engineering with an Engineering and Business Minor. My expected graduate date is May 2007. My plan after obtaining my degree is to pursue graduate school.

84 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 1,1-Diphenyl-2,3,4,5-tetrakis(9,9-dimethylfluoren-2-yl)silole Properties in Organic Light-Emitting Diodes and Organic-Field Effect Transistors Sarah Montgomery Bernard Kippelen, Benoit Domercq Purdue University School of Electrical and Computer Engineering Georgia Institute of Technology

Organic light-emitting diodes (OLED) and organic field-effect transistors (OFET) have been central to current research because displays and integrated circuits made with organic materials can be fabricated on plastic, making them flexible and lightweight and potentially reducing production costs. OLEDs and OFETs were fabricated out of a new family of ma- terials based on silole derivatives, specifically 1,1-Diphenyl-2,3,4,5-tetrakis(9,9-dimethyl- fluoren-2-yl)silole (XZ-III-20), shown in Scheme 1, using vapor deposition techniques. These devices were then tested to determine their physical characteristics and properties using lab equipment and the software program LabView.

INTRODUCTION Siloles have been found to have close to 100% photolumi- nescence efficiency, in turn allowing for a high external quantum The extent of research currently on OLEDs is extensive. One efficiency of 8% [2]. This photoluminescence efficiency is over of the advantages of an OLED display to that of a liquid crystal three times higher than that of tris-(8-hydroxyquinoline) alumi- display (LCD) is it does not require a backlight which can ulti- num (AlQ ), one of the leading emitting electron transport lay- mately reduce its size, power consumption and production cost. 3 ers used in OLEDs today. These silole materials are of interest OLEDs consist of subsequent organic layers sandwiched between because they are primarily singlet emitters but it has been pro- electrodes that inject holes and electrons into the organic layers. posed that there may be phosphorescent emissions as well [2]. The organic layers must transport holes, electrons and emit light. XZ-III-20 has been the focus of this research. From its expected Holes are injected from the anode, a high work function metal electron affinity and ionization potential which are comparable while electrons are simultaneously injected from the cathode, a to N-N’-diphenyl-N-N’bis(3-methylphenyl)-[1-1’-biphen- yl]-4- low work function metal. The differing electrodes call for differ- 4’-diamine (TPD) and AlQ it could be ambipolar or work ef- ent organic materials, a hole transport layer (HTL) that has a low 3; fectively as both an ETL and a HTL. XZ-III-20 has four fluorene ionization potential and an electron transport layer (ETL) that has groups attached to the silole. These fluorene groups decrease the a high electron affinity [1]. Holes are transported in the HTL ionization potential of the silole material which could lead to fa- while electrons travel in the opposite direction through the ETL; vorable hole-transport properties. the charges will either find an energy barrier between the HTL and ETL or at the emission layer and will ultimately recombine, creating an exciton which emits a photon when it relaxes to the ground state.

Scheme 2. Structure of a hole transport material: N-N’-diphenyl-N- N’bis(3-methylphenyl)-[1-1’-biphenyl]-4-4’-diamine (TPD).

OFETs can be used in integrated circuits as a switching de- vice in technology such as computers and displays [3,4]. Like inorganic field-effect transistors, OFETs are three terminal de- vices consisting of a gate, drain and source. In an OFET the gate Scheme 1. 1,1-Diphenyl-2,3,4,5-tetrakis(9,9- dimethylfluoren-2-yl)silole (XZ-III-20). voltage controls the S-D current or ID. OFETs are being used to determine the dominant charge carriers present in XZ-III-20 and

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 85 1,1-DIPHENYL-2,3,4,5-TETRAKIS(9,9-DIMETHYLFLUOREN-2-YL)SILOLE PROPERTIES IN ORGANIC LIGHT-EMITTING DIODES AND ORGANIC-FIELD EFFECT TRANSISTORS its charge carrier mobility, μ, which is the velocity of the charge OFETs were fabricated using a bottom contact geometry carriers under a given electric field. This can be done by applying with the organic semiconductor, XZ-III-20 (40 nm), deposited a voltage of the same polarity to both the gate and drain. If there on top of a heavily doped silicon wafer coated with a thermally is a small current present, then the charge carrier has a charge op- grown oxide and gold electrodes defining the source and drain. posite of the applied voltage. The organic semiconductor, XZ-III-20, was deposited on top of the gate insulating layer and gate electrode. The use of gold elec- trodes favors a p-type response due to its work function. The device was tested as n-type and p-type, however only the p-type structure produced an output. For a channel width of 500 nm and length of 50 nm, a drain sweep was conducted from 0 to -80 V with gate voltages ranging from 0 to -80 V in step sizes of 10 V.

The transfer curve was obtained by sweeping the gate holding VD Scheme 3. Structure of an electron transport and emitting constant at - 80 V. As the channel length decreased and the width material: tris-(8-hydroxyquinoline) aluminum (AlQ3). increased, the voltage sweep had to be increased to -100 V. The output characteristic and transfer curve are shown in Figure 2 and DEVICE FABRICATION AND TESTING Figure 3 respectively. OLEDs were fabricated on glass substrates coated with in- dium tin oxide (ITO) as the anode. The subsequent organic layers were deposited using high vacuum thermal evaporation. The metal cathode is composed of a very thin layer of Lithium Fluoride and 300 nm of Aluminum. XZ-III-20 was tested by fabricating OLEDs using various device geometries: as an ETL (1) ITO/ TPD (40 nm)/ XZ-III-20 (40 nm)/ LiF (1nm)/ Al (300 nm), as a HTL (2) ITO/

XZ-III-20 (40 nm)/ AlQ3 (40 nm)/ LiF (1nm)/ Al (300 nm), as an emission layer (3) ITO/ TPD (40 nm)/ XZ-III-20 (40 nm)/ AlQ3 (40 nm)/ LiF (1nm)/ Al (300 nm) and as a single layer of varying thickness (4) ITO/ XZ-III-20/ LiF (1nm)/ Al (300 nm). Each device structure was fabricated and tested in a nitrogen environment and current-forward light output measurements, as Figure 2 I-V characteristics at several values of the gate voltage (VGS) for an a function of the applied voltage, were acquired using a Keithley OFET using XZ-III-20, having a channel width of 500 nm and length of 50 nm. 2400 sourcemeter and a silicon photodiode interfaced using Lab- View software. The electroluminescent (EL) spectrum was taken for each device structure at different voltages to ensure that there was not a shift in the emission peak when increasing the applied voltage. The EL spectrum of each device geometry is shown, along with a device fabricated with AlQ3, a widely used material in OLED (Figure 1). The geometry of the AlQ3 device shown in

Figure 1 is ITO/ TPD / AlQ3 / LiF/ Al.

Figure 3. Transfer curve with VD = -80 V for an OFET using XZ- III-20, having a channel width of 500 nm and length of 50 nm.

RESULTS AND DISCUSSION From the OLED testing, current density (J), luminance and external quantum efficiency were calculated. Luminance takes into account photopic response (how the human eye responds to

Figure 1. EL spectrum of devices fabricated with XZ-III-20 in various the light emitted), the sensitivity response of the photodetector

device geometries. EL spectrum of a device using AlQ3 as an ETL. and the physical geometry of the measurement set-up. External

86 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 MONTGOMERY quantum efficiency is the measurement of photons emitted from Single layer devices using XZ-III-20 were fabricated with the device in the forward direction divided by the amount of elec- varying thicknesses: 50 nm, 80 nm, 100 nm, 120 nm, and 150 trons injected into the device [5]. Figure 4 (top) shows a com- nm. Figure 5 (top) and Figure 5 (bottom) show respectively the parison of current density for different device geometries (1), (2), current density and luminance, external quantum efficiency as a and (3), while Figure 4 (bottom) compares the luminance and function of the applied voltage for each of the thicknesses. It is external quantum efficiency. It is seen from the plot of current seen that with increasing thickness, the current density decreases density, when XZ-III-20 acts as an ETL and HTL, the responses as expected. The luminance decreases while the external quan- are comparable, however the luminance for the ETL structure is tum efficiency increases, however the turn-on voltage is signifi- lower, leading to a lower external quantum efficiency. The cur- cantly increased from 3.6 V for 50 nm to 10.4 V for 150 nm, and rent density for (3) is lower than that of (1) or (2) primarily be- to 11.6 V for 120 nm thick. With a thickness of 50 nm, a lumi- cause it is 40 nm thicker. The small jagged bumps in the J-V plot nance of 300 cd/m2 is reached at 6.5 V with an external quantum represent leakage current; there is current flowing through the efficiency of 0.08%; while with a thickness of 150 nm, the same device that is not producing photons. In device structure (3) the luminance is achieved at 16.9 V with an external quantum effi- holes and electrons are transported by TPD and AlQ3 respectively ciency of 0.09%. and trapped in the emission layer resulting in such high external quantum efficiency. A standard LCD monitor has a luminance of about 300 cd/ m2. For XZ-III-20 as an ETL, it reaches a brightness of 300 cd/ m2 at 6.3 V with an external quantum efficiency of 0.55%. As an HTL it reaches the same brightness at 5.8 V with an external quantum efficiency of 0.6%. When XZ-III-20 acts as an emission layer it has a much higher external quantum efficiency value of 0.95% at 7.6 V when the luminance is 300 cd/m2.

Figure 5. Current density (top) and luminance and external quantum efficiency (bottom) as a function of the applied voltage of devices using XZ-III-20 as a single layer of varying thickness.

From the OFET output characteristic, it is evident that there

is a very small ID current flowing, on the order of nA, and the signal is noisy. The on/off ratio, threshold voltage and mobility are extracted from the transfer curve. For a channel width of 500 nm and length of 50 nm, the mobility is 4.0 x 10-6 cm2/Vs, The threshold voltage, the voltage required to switch the device on, is very high, -56 V. The on/off ratio, the ratio of the current flowing through the device when it is on, divided by the current flowing through the device when there is 0V applied at the gate, is 10. Figure 4. Current density (top), luminance and external quantum These values are very low compared to pentacene, a common efficiency (bottom) as a function of the applied voltage of devices organic hole-transport material used in OFETs, having a hole mo- using XZ-III-20 as an ETL, HTL, and emission layer. bility of μ = 2.4 cm2/V-s and an on/off ratio of 108 [3].

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 87 1,1-DIPHENYL-2,3,4,5-TETRAKIS(9,9-DIMETHYLFLUOREN-2-YL)SILOLE PROPERTIES IN ORGANIC LIGHT-EMITTING DIODES AND ORGANIC-FIELD EFFECT TRANSISTORS

CONCLUSION From the information learned about XZ-III-20 in the OLED device structures of (1), (2), (3) and (4), it can be concluded that XZ-III-20 works both as an ETL and/or a HTL; however it has the highest external quantum efficiency, 0.95%, at 300 cd/m2, in the emission layer geometry. In the future, the thicknesses of TPD and AlQ3 as injection layers should be optimized. As a single layer, there was a high current density that leaked produc- ing low external quantum efficiency, as well as high turn-on volt- ages for the thicker devices. Lifetime measurements at constant current have been taken for most of the devices but need to be completed. For XZ-III-20 in the bottom contact, p-type OFET geometry, the mobility was very low with a high threshold volt- age and small on/off ratio. In the future, a top contact OFET should be fabricated, using a different metal for electrodes that would promote n-type behavior to test if XZ-III-20 also has elec- trons as charge carriers.

ACKNOWLEDGEMENTS This research was supported by MDITR, a Science and Technology Center of NSF under Agreement Number DMR- 0120967. Special thanks go to Andreas Haldi, Joshua Haddock, Xiaowei Zhan, Benoit Domercq, Seth Marder, Bernard Kippelen and the Kippelen Research Group.

REFERENCES [1] Dini, D. Chem. Mater. 2005, 17, 1933-1945. [2] Chen, H. Y.; et al. Appl. Phys. Lett. 2002, 81, (4), 574-576. [3] Dimitrakopoulos, C. D.; Malenfant, P. R. L. Adv. Mater. 2002, 14, (2), 99-117. [4] Newman, C. R.; et al. Chem. Mater. 2004, 16, 4436-4451. [5] Okamoto, S.; et al. Jpn. J. Appl. Phys. 2001, 40, (2), 783-784.

Sarah Montgomery is currently studying Electrical Engineering at Purdue University. She plans to pursue a Masters Degree after her graduation in May.

88 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Effects of Surface Chemistry on Cadmium Selenide Nanocrystal Fluorescence

Marsha S. Ng Andrea M. Munro and David S. Ginger University of Hawaii Department of Chemistry University of Washington

INTRODUCTION reached 60°C, the NCs were precipitated with acetone and cen- trifuged. The synthesis yielded monodisperse ~4 nm diameter CdSe semiconductor nanocrystals (NCs) possess size-tunable CdSe NCs with maximum absorbance of 586 nm, full width at optical and electronic properties that make them promising can- half max at ~30 nm. NC size was determined using the absor- didates for applications such as light emitting diodes, solar cells, bance peak8, and the initial quantum yield (QY) was found to be and biological labels. To be useful in these applications, the NCs 13%. must be highly fluorescent. These potential uses have propelled It was observed that NCs in different regions of the cen- many studies exploring various means of increasing the quantum trifuge tube had differing fluorescence intensities. The NCs on yield (QY) of CdSe NCs. The NCs have a large surface to volume the top layer of the centrifuge tube were more fluorescent than ratio, causing their surface chemistry to be a major factor in their the middle layer, and the middle layer more fluorescent than the fluorescence. Dangling bonds at the surface cause trap states that bottom layer. To eliminate this source of variation, we mixed lower the chances for radiative recombination of an electron hole all nanocrystals thoroughly after centrifugation unless otherwise pair by creating non-radiative decay pathways. One solution to noted. this problem is to grow an inorganic shell (e.g. CdS or ZnS), an- For the ligand exchange experiments, 7.6E-7 mmol of CdSe other method is to passivate the surface with organic ligands. The NCs were mixed with ligand to make 3 mL solutions. The BA functionalization of the ligand bound to the NC surface is impor- was purged with nitrogen for six hours. ODA and ODT stock tant (e.g. hydrophilic functional group will allow NC to be soluble solutions were prepared by dissolving the ODA or ODT solid in water) and the effects of the ligands on fluorescence are still in chloroform. ODA stock solutions were prepared at ~0.10 M, controversial1-6. To gain a greater understanding of the effects of its estimated saturation point in chloroform. ODT stock solution organic ligands on NC fluorescence, we study how the concentra- concentrations were generally prepared at ~1.8 M. tion of 1-octadecanethiol (ODT), 1-butanethiol (BT), 1-octadecyl- A Perkin Elmer LS-50B Fluorimeter was used to measure amine (ODA), and 1-butylamine (BA) alter the photoluminescence the emission intensity of the test solutions. The fluorimeter re- (PL) of CdSe nanoparticles in chloroform solution. sponse was calibrated using an external light source with known light intensity and Rhodamine-101 (R101) with a 100% quantum EXPERIMENTAL efficiency was used as a fluorescence standard. During emission We synthesize CdSe NCs by a derivative of the method of spectra runs, the emission spectrum of 1E-6 M R101 in HCl/ 7 Peng et al. The synthesis is conducted under nitrogen gas in EtOH solution was monitored at fixed intervals to allow the data minimal light conditions. A thermocouple and heating mantle are to be corrected for fluctuations in lamp intensity. used to regulate reaction temperature. Cadmium stearate stock To investigate if H-NMR can be used to determine the rela- solution is formed by heating 0.514g cadmium oxide and 4.552g tive amount of free and bound ligand in solution, solution state stearic acid at 220°C for about 2 hours. Tri-n-butylphosphine se- 1H-NMR data was taken on a 300 MHz Bruker AV300. All CdSe lenium (TBP=Se) solution is formed by mixing 1.421g selenium NCs used in the NMR experiments were washed two times with powder, 3.843g TBP, 12.33g octadecence (90%). In a 25-mL acetone and dried prior to use. Deuterated chloroform (CDCl3) 3-neck flask, 0.539g cadmium stearate powder, 2.0g octadecene was used as the lock solvent. The free ODA test solution was (95%), 1.0g trioctylphosphine oxide (90%), 3.0g octadecylam- prepared with 0.019g ODA in 0.6 mL CDCl3. The free ODT test ine (97%) are purged for 10 minutes and heated to an injection solution was made by adding 0.1g ODT to 0.6 mL. The CdSe- temperature of 270°C. At the injection temperature, 2.25g of the ODA test solution was prepared by adding 0.019g ODA to 1E-5 TBP=Se solution is swiftly injected and the temperature reset to mmol CdSe in 0.6 mL CDCl3. The CdSe-ODT test solution was 250°C for growth. When the NCs reach their desired size the prepared by adding 0.02g ODT to 1E-5 mmol CdSe in 0.6 mL reaction is stopped by removing the heat. When the solution CDCl3.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 89 EFFECTS OF SURFACE CHEMISTRY ON CADMIUM SELENIDE NANOCRYSTAL FLUORESCENCE

RESULTS Addition of degassed BA at low concentrations increased the QY, but quenching occurred at concentrations above 4E-3 M. Figures 1 and 2 show that addition of both ODT and BT Mixing BA and chloroform is exothermic, and the temperature quench the fluorescence of CdSe NCs. Figures 3 and 4 show that rise for BA concentrations above ~0.6 M is noticeable. Figure at low concentrations (below ~0.01 M) of ODA and BA, the NP 5 shows the blue-shift phenomena of the maximum absorbance fluorescence is enhanced. This quenching can be fit with a single peak at high BA concentrations. component Langmuir isotherm. Due to solubility issues, solu- tions with ODA concentrations above ~0.12 M were not tested. Solutions with 0 M ligand concentrations are plotted as 1E-15 M for the log scale in figures below.

Figure 4. Quantum yield vs. 1-butylamine concentration [M] for 4 nm diameter CdSe NCs in chloroform.

Figure 1. Quantum yeild vs 1-octadecanethiol concentration [M] for 4 nm diameter CdSe NCs in chloroform. (Core vials not mixed so absolute initial quantum yeild has err or).

Figure 5. The maximum absorbance peak vs 1-butylamine concentrations [M].

Figure 2. Quantum yeild vs 1-butanethiol concentration [M] for 4 nm diameter CdSe NCs in chloroform. (Core vials not mixed so absolute initial quantum yeild has error). Figure 6. The quantum yield vs ODA concentration [M] with constant [ODT] = 1.2E-5 M present in solution.

Figure 3. Quantum yeild vs 1-octadecylamine concentration Figure 7. The quantum yield vs ODA concentration [M] [M] for 4 nm diameter CdSe NCs in chloroform. with constant [ODT] = 6.2E-6 M present in solution.

90 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 NG

The QY was enhanced more by the addition of ODA than BA. In 1H-NMR studies of ODT, the proton on the sulfur atom To investigate the reversibility of ligand binding, ODA was (circled region in Figure 10) is presumed to have a chemical shift added to CdSe NCs already partially exchanged with ODT. The QY near the methylene peak at ~1.2 ppm making the peak difficult to of CdSe NCs in 1.2E-5 M ODT was ~6%. Figure 6 shows that upon resolve. addition of ~0.1M ODA, the QY was increased to ~21%. The QY of CdSe NCs in 6.2E-6 M ODT was ~10% and figure 7 shows that DISCUSSION the addition of ~0.04 M ODA increased QY up to ~35%. We hypothesized that the ligand binding could be described 9,10 Figure 8, 9, and 10 show the H-NMR spectra that was by a Langmuir isotherm (eqn 1). The Langmuir isotherm relates performed to attempt to determine the relative amounts of each the fractional surface coverage, θ, of ligand on surface sites to the Figure 8, 9, and 10 show the H-NMR9,10 spectra that was performed to attempt to determine the ligand in solution that is free or bound to a CdSe NC. H-NMR concentration of ligand in solution and can be derived from an spectrarelative of amounts 1E-5 mmol of each washed ligand CdSe in solution NCs, ODA, that is ODA free orbound bound to to a CdSe NC. H-NMR spectra of 1E-5 mmol washed CdSe NCs, ODA, ODA bound to CdSe, ODT,equilibrium and ODT boundreaction to givenCdSe thein equilibrium binding constant, K, CdSe,deuterated ODT, chloroform and ODT boundwere conducted. to CdSe in deuterated chloroform and ligand concentration, [ligand]. were conducted. FigureFigure 8 8, 9, and 10 show the H-NMR9,10 spectra that was performed to attempt to determine the relative amounts of each ligand in solution that is free or bound to a CdSe NC. H-NMR spectra of 1E-5 mmol washed CdSe NCs, ODA, ODA bound to CdSe, ODT,Figure and 8 ODT shows bound the 300to CdSe MHz in deuterated chloroform were conducted. solution state H-NMR spectra of 1E-5 mmol CdSe NCs in deuterated Figure 8 In order to use the isotherm, four assumptions were made: (1) the reaction is reversible, (2) only one monolayer can be formed, Figure 9 Figure 8 shows the 300 MHz (a) Figure 8. shows the 300 MHz solution state H-NMR spectra solution(3) binding state H-NMR of a ligandspectra toof a1E-5 surface site does not affect the binding of 1E-5 mmol CdSe NCs in deuterated chloroform. b d mmolat any CdSe other NCs sites, in and deuterated (4) the QY is linearly dependent on the e a c fractional surface coverage, θ. We also assumed that the NCs H 2 H2 H2 H 2 H2 H2 H C C C C 2 H2 H2 C C C C H2N C C C C H C C C 2 H2 H2 H C C CH 2 H2 H2 H 3 2 H2 f Figure 9 started with all empty surface sites, which is not true because (a) there are stearic acid, TOPO, ODA bound to the surface. b d Figure 9. shows the structure of 1-octadecylamine (a) and e A more accurate model would use a multiple component a c b d,e

H2 H2 H 2 H2 H2 H C C 2 H2 H2 H C C C C 2 H N C C C C 2 C C C C H2 H C C C corresponding peak assignments (b) on the colution state H-NMR 2 H2 H CH 2 H2 H2 H 3 2 H2 f FigureLangmuir 9 showsf isotherm the structure in which of 1- the ligand already bound to the sur- (b) spectra of ODA. Figure 9 (c) shows the H-NMR spectraa when octadecylamineface is taken (a) into and account. corresponding The fluorescence quenching that occurs ODA is bound to IE-5 mmol CdSe deuteratedc chloroform. peak assignments (b) on the solution stateat H-NMRhigh amine spectra concentrations of ODA. does not agree with the assumption Figure 9 (c) shows the H-NMR b d,e (4) that the QY increases linearly as the surface is passivated with * o f spectraFiguremore when 9 amine. shows ODA the For is structurebound the amine to of 1E-5 1-curves the fits were made with data at (b) o a mmoloctadecylamine CdSe in deuterated (a) and correspondingchloroform. (c) c peakamine assignments concentrations (b) on the below solution the solution saturation point of ODA state(~0.01 H-NMR M). spectra of ODA. In the case of ODA binding to CdSe, the triplet peak (b) at ~2.66 ppm corresponding to the alpha- Figure 9 To (c) improve shows the the model, H-NMR fits with the Hill equation (eqn 2) protons (circled in Figure 9) is broadened. The pentet at ~3.19 ppmspectra leans when towards ODA the is boundupfield to peaks 1E-5 * o o indicating that it is interacting with the ODA.o *Acetone and stearicmmol wereacid CdSe impurities.performed. in deuterated The chloroform.Peaks Hill are equation a is widely used as an empirical result(c) of amine bound to NC. fitting function when some degree of deviation from the Lang- In the case of ODA binding to CdSe, the triplet peak (b) at ~2.66 ppmmuir corresponding isotherm assumptions to the alpha- is expected. protons (circled in Figure 9) is broadened. The pentet at ~3.19 ppm leans towards the upfield peaks indicating that it is interacting with the ODA. *Acetone and stearic acid impurities. oPeaks are a result ofFigure amine 10. bound shows tothe NC. structure of 1-octadecanethiol (a) and corresponding peak assignments (b) on the solution state H-NMR spectra of ODT. Figure 10 (c) shows the H-NMR spectra when ODT is bound to 1E-5 mmol CdSe in deuterated chloroform. Table 1 Langmuir isotherm- K Hill equation- K Hill equation- n In the case of ODA binding to CdSe, the triplet peak (b) at CdSe + ODT 106980 +/- 24500 94877 2 ~2.66 ppm corresponding to the alpha-protons (circled in Figure 9) is broadened. The pentet at ~3.19 ppm leans towards the upfield CdSe + BT 101800+/- 4300 186470 2 peaks indicating that it is interacting with the ODA. *Acetone and CdSe + ODA 2001.7 +/- 522 4180 1 o stearicFigure 10acid impurities. Peaks are a result of amine bound to NC. CdSe + BA 1787.3 +/- 769 2000 1 Figure2 10 shows the structure4 of 1-octadecanethiol (a) and 5 Table 1. shows the equilibrium binding constant (K) values 1 3 6 corresponding peak assignments (b) on the solution state H-NMR found for least-squares fits of the Langmuir isotherm spectra of ODT. Figure 10 (c) shows the H-NMR spectra when and Hill equations to the experimental data. ODTFigure is 10 bound to 1E-5 mmol CdSe in deuterated chloroform. 2 4 5 4 1 3 CMDITR6 Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 91

4 EFFECTS OF SURFACE CHEMISTRY ON CADMIUM SELENIDE NANOCRYSTAL FLUORESCENCE

The similarity between the K-values of the amine and thiol CONCLUSION pairs suggest that the alkyl chain effects are secondary to the ef- Our studies have shown that thiols quench the fluorescence fects of the head group. The larger K-values for the thiols shows of 4 nm diameter CdSe NCs, whereas at low concentrations (be- that the sulfur binds better than nitrogen, which is consistent with low ~0.01 M) amines enhance PL and at high amine concentra- the Hard-Soft Acid-Base Principal (HSAB). HSAB states that tions quench PL. At low concentrations, the thiol and amine can soft acids such as cadmium preferentially bind to soft bases such be reversibly exchanged. Reversibility tests indicate that the re- as sulfur; nitrogen is a borderline soft base and is less likely to action with high concentrations of amine is not completely re- coordinate to the cadmium. versible due to surface etching. Fluorescence measurements can Quenching occurs in the presence of thiol due to increased allow us to calculate the amount of ligand bound to the core if we hole traps11. The thiol’s sulfur atom is less electronegative than the assume that PL is proportional to surface coverage. In the future, amine’s nitrogen and therefore would donate electrons more readi- H-NMR could be used to quantify surface coverage to comple- ly. When the sulfur atom binds to the cadmium, the cadmium atom ment fluorescence data. becomes more negatively charged. The increased partial negative charge on the cadmium attracts holes in the semiconductor valence ACKNOWLEDGEMENTS band resulting in “hole trapping.” The trapping of holes hinders recombination of an electron-hole pair resulting in fewer radiative Many thanks to my mentors, Andrea Munro and David Gin- decay events and thus a lowered emission intensity. ger, for their help and guidance. Thank you to the NSF Science The QY enhancement occurring at amine concentrations be- & Technology Center for Materials and Devices for Information low ~0.01 M can be explained by an initial decrease in electron Technology Research for funding the Summer 2005 Hooked on traps when the nitrogen binds to cadmium sites. After all the elec- Photonics REU program and allowing me to enrich my summer tron traps have been passivated, the amine will then bind creating with research. Thanks to Loren Kruse and Rajan Paranji of the hole trap sites and quenches the florescence. This could suggest UW NMR facilities and Brenden Carlson and Kolby Allen of the that the amine preferentially binds to electron trap surface sites. Dalton Lab for their help in NMR spectroscopy. Also thanks to The quenching by amines could result from several possibili- Yeechi Chen for her support and guidance. ties: (1) impurities in BA, (2) etching of CdSe NCs at high amine concentrations, and (3) temperature changes. The molar ratio of REFERENCES CdSe NCs to water is ~10 in our test solutions. However, purg- 1. Dannhauser, T.; O’Neil, M.; Johansson, K.; Whitten, D.; ing the BA with N2 (g) did not appear to change the amount of McLendon, G. “Photophysics of Quantized Colloidal Semi- quenching. The quenching and blue-shifting of the absorbance conductors Dramatic Luminescence Enhancement by Binding and PL peaks at high amine concentrations implies that the amine of Simple Amines” J. Phys. Chem 1986, Vol 90, No. 23, 6074- 12 is etching the NC surface as demonstrated by Li et al. Adding 6076. BA to chloroform caused the solution to heat up 10°C. This sug- 2. Landes, C.; Burda, C.; Braun, M.; El-Sayed, M. A. “Photo- gests that the heat released could be the heat of solvation of BA luminescence of CdSe Nanoparticles in the presence of a Hole- in chloroform solvent. Acceptor: n-butylamine” J. Phys. Chem. B 2001, Vol 105, No 15, To check the reversibility of the BA quenching, the BA from 2981-2986. a test solution with 6.75 M BA concentration was pumped off 3. Landes, C.; Braun, M.; El-Sayed, M. A.;“On the Nanopar- and chloroform was added back. Before pumping, the CdSe NCs ticle to Molecular Size Transition: Fluorescence Quenching Stud- had a maximum absorbance peak at 563 nm and quantum yield ies” J. Phys. Chem. B 2001, Vol 105, No 43, 10554-10558. of 0.5%. After pumping off the BA, QY was 0.9% and the ab- 4. Hohng, S.; Ha, T. “Near-Complete Suppression of Quantum sorbance spectra appeared more broadened. However, the large Dot Blinking in Ambient Conditions” J. Am. Chem. Soc. 2004, blue shift of ~22 nm in the absorbance spectra was not reversible Vol 126, No 5, 1324-1325. suggesting that indeed the nanoparticles were etched. 5. Kalyuzhny, G.; Murray, R. “Ligand Effects on Optical Prop- The restoration of the quantum yield upon addition of ODA erties of CdSe nanocrystals.” J. Phys. Chem. B 2005, Vol 109, No to ODT-quenched CdSe NCs indicates that the ligand binding 15, 7012-7021. may be reversible at low concentrations. However, the initial 6. Jeong, S.; Achermann, M.; Nanda, J.; Ivanov, S.; Klimov, 13% QY however was not achieved with the addition of ODA. V.; Hollingsworth, J. “Effect of Thiol–Thiolate Equilibrium on At the high ODA concentrations, a drop in QY occurs. This drop the Photophysical Properties of Aqueous CdSe/ZnS Nanocrystal is observed at the same ODA concentration as in the earlier tests Quantum Dots” J. Am. Chem. Soc. 2005, Vol 127, No 29, 10126- of CdSe NCs and ODA. 10127.

92 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 NG

7. Peng, Z.; Peng, X. “Nearly Monodisperse and Shape-Con- trolled CdSe Nanocrystals via Alternative Route: Nucleation and Growth” J. Am. Chem. Soc. 2002, Vol 124, No 13, 3343-3353. 8. Yu, W.; Qu, L.; Guo, W.; Peng, X. “Experimental Determina- tion of the Extinction Coefficient of CdTe, CdSe, and CdS Nano- crystals.” Chem. Mater. 2003, Vol 15, No14, 2854-2860. 9. Sachleben, J.; Wooten, E.; Emsley, L.; Pines, A.; Colvin, V.; Alivisatos, A. “NMR studies of the surface structure and dynam- ics of semiconductor nanocrystals” Chem. Phys. Lett, 1992, Vol 198, No 5, 431-436. 10. Berrettini, M.; Braun, G.; Hu, J.; Strouse, G. “NMR Analy- sis of Surfaces and Interfaces in 2-nm CdSe” J. Am. Chem. Soc. 2004, Vol 126, No 22, 7063-7070. 11. Wuister, S.; de Mello Donega, C.; Meijerink, A. “Influence of Thiol Capping on the Exciton Luminescence & Decay Kinet- ics of CdTe and CdSe Quantum Dots.” J. Phys. Chem. B 2004, Vol 108, No 45, 17393-17397. 12. Li, R.; Lee, J.; Yang, B.; Horspool, D.; Aindow, M.; Papad- imitrakopoulos, F. “Amine-Assisted Facetted Etching of CdSe Nanocrystals.” J. Am. Chem. Soc. 2005, Vol 127, No 8, 2524- 2532.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 93 94 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Synthesis of a Polyene EO Chromophore: Using a Diels-Alder Reaction to Form a Side-Chain Structure

Denis Nothern Jingdong Luo, Alex K.-Y. Jen Cornell University University of Washington

INTRODUCTION dependant on its glass transition temperature (Tg). As the material approaches its T it will typically experience thermal relaxation Nonlinear optical (NLO) polymers have drawn significant at- g in which the polymer backbone becomes pliable enough for the tention over the past decade for their growing potential to revolu- chromophores that had previously been aligned noncentrosym- tionize the field of electro-optics (EO). These polymers combine metrically to become free and return to the preferential centro- comparatively simple and low cost processing with the advantag- symmetric state. At high temperatures it is also common that the es of high signal bandwidth and low operating voltages. Recent chromophores and the host-polymer experience phase separation work has focused on improving the stability of the host-polymers due to unequal thermal characteristics, this is evidenced by high via a cross-linking scheme based on a Diels-Alder reaction.[1] In currents during poling. As discussed earlier, forming side-chains this research, a similar concept has been applied to the bonding will help counter thermal relaxation and will also help maintain a between the nonlinear chromophores and the host-polymer, to homogeneous system, but at the same time will make poling more form a side-chain structure utilizing a Diels-Alder reaction. difficult. It is because of this trade-off between poling efficiency The larger motivation behind this research is maximizing and thermal stability that this research explores a solution that two key material characteristics: the poling efficiency and the maximizes them both. If one embeds the chromophores as guests thermal stability. The poling efficiency, while not the only factor in the host-polymer but also functionalizes the chromophores so associated with achieving a high electro-optic coefficient, is an that as the temperature is increased upon poling they undergo a important material property that indicates the polymer’s ability reaction with the host-polymer and attach as side-chains aligned to form a noncentrosymmetric system (a condition imposed by with the applied field, the poling efficiency and thermal stability the second order nature of the electro-optic effect). To achieve a can be balanced. high poling efficiency the chromophores in the polymer must be free to rearrange their alignment when it is desired while avoid- ing chromophore-chromophore interactions that lead to close packing aggregation in the lower energy symmetric configura- tion. With chromophores freely embedded in the host-polymer (a guest-host system) they are able to easily align with the poling Scheme 1. Maleimide undergoes cycloaddition with field, but will also be relatively free to return to the lowest energy anthracene in a Diels-Alder reaction. state once the poling field is removed. Alternatively, the chro- mophores can be embedded into the host-polymer as side-chains The ability to attach the chromophores as side-chains in situ that limit their ability to rearrange, thus avoiding aggregation but during poling is dependant on a Diels-Alder (DA) reaction be- limiting the ability to align with the applied field. This leads to the tween the chromophores and the host-polymer. DA reactions have examination of the thermal stability of the system. been extensively explored and are based on the cyclo-addition of The thermal stability of the material is crucial because of the a diene with a dienophile to form a six-membered ring (scheme demands on such a device. In use it will be exposed to a wide 1)[2]. The use of this type of reaction is promising in that it can be range of temperatures—sources both intrinsic to the operation thermally controlled and produces no additional products beyond of the device and sources extrinsic to the device present in the the desired cyclohexene (this gives it the name “Click Chemis- operating environment. For example, a device is not useful if it try”). If additional products are present in the polymer system cannot be used for extended periods of time due to overheating as they could have adverse and fundamentally unpredictable effects a result of its normal use, nor is it useful if it has only very limited on the operation of the device. Since the system must be heat- implementation resulting from temperature constraints placed on ed close to its Tg in order to achieve maximum alignment dur- the system around it. The polymer’s thermal stability is primarily ing poling, a thermally controlled DA reaction with a very high

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 95 SYNTHESIS OF A POLYENE EO CHROMOPHORE: USING A DIELS-ALDER REACTION TO FORM A SIDE-CHAIN STRUCTURE

Scheme 2. Synthesis of polyene chromophore. temperature required to reverse and an activation energy similar EXPERIMENTAL to the poling temperature is ideal so that complete linking of the Synthesis of the polyene chromophore was carried out with a chromophores to the polymer occurs once they have already been fairly standard procedure (scheme 2). mostly aligned with the field. A DA reaction also exhibits essen- The maleimide functional group to be used as the dienophile tial characteristics when taking processability into account because was synthesized as an acid chloride (scheme 3). Literature indi- the chromophore can be bound to the polymer cleanly and in a con- cated that this reaction would only last a matter of hours however trolled manner without the need for solvents, catalysts or any other it was found to take significantly longer.[4] The first batch was external chemical processes—this is particularly important for more only reacted for 3 hours and had a great deal of impurity, which efficient chromophores that are sensitive to their surroundings. greatly affected the functionalization results. This concept has been shown to work with nonlinear chromo- phores already, but difficulty arises when it is applied to highly ef- ficient nonlinear chromophores. More efficient chromophores, by nature, are highly polar and consequently they are quite sensitive to their chemical environment. This makes it exceedingly difficult to functionalize them without decomposing them completely. With this in mind, the objective was to synthesize several polyene-type chromophores with a dienophile group at R1 so that it may combine with a diene-functionalized host-polymer to form a side-chain system via a DA reaction (figure 1). Other structures Scheme 3. Production of acid chloride. are present at R2 and R3 to benefit from the advantages of further limiting the chromophore-chromophore interactions. Once syn- The finished chromophore 8 was deprotected and the acid thesized, the materials properties were analyzed. chloride 10 was added (scheme 4). The yield of this reaction was negligible due to the poor purity of the acid chloride. Despite extremely low yield, chromatography indicated that the reaction did occur as hoped for. To confirm the success of this reaction scheme, a sample of the AJC-139 chromophore was reacted for 24 hours with purer acid chloride (scheme 5). This reaction was a success with 57% yield. A chromophore R =dienophile group 1 similar to the original was synthesized from an intermediate 16 R2=phenyl group and a slightly different acceptor group (scheme 6). Figure 1. A polyene-type chromophore.

96 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 NOTHERN

Figure 2. Polymer Host.

At this point synthesis was considered complete and chro- mophore 17 was embedded in an anthracene functionalized poly- methyl-methacrylate (PMMA) host polymer with a 25 wt.% loading density (figure 2). It was spin-cast to form a .75um film Scheme 4. Functionalization of polyene chromophore. on a glass substrate half coated with conducting indium-tin-ox- ide (ITO) (120nm) particularly designed to be optically transpar- ent at telecommunications wavelengths (1.3um). Three samples Synthesis of a Polyene EO Chromophore were prepared by sputteringDenis Nothern, Jingdonga gold Luo, electrode Alex K.-Y. Jen onto the top of the polymerAt thisfilm point such synthesis that wasthey considered were half complete over and the ITO and half over chromophore 17 was embedded in an anthracene functionalized thepoly-methyl-methacrylate glass (figure 3). (PMMA This) allowed host polymer a withvoltage a 25 to be.88 applied.12 across wt.% loading density (figure 2). It was spin-cast to form a O OCH O O the.75um electrode film on a glassand substratethe ITO half layer coated withand conducting pole the polymer.3 While the indium-tin-oxide (ITO) (120nm) particularly designed to be polingoptically voltage transparent was at telecommunications being applied, wavelengths current passing through the (1.3um). Three samples were prepared by sputtering a gold polymerelectrode onto was the carefullytop of the polymer monitored film such thatfor they indications were Fig. 2.of Polymer damage Host. to the film.half overDuring the ITO heating, and half over the the current glass (figure was 3). recorded This overtime. Once allowed a voltage to be applied across the electrode and the ITO layer and pole the polymer. poled,While the the poling r voltage value was was being measured applied, current via passing the through reflection the polymer technique was carefully[3] . monitored for indications33 of damage to the film. During heating, the current was recorded [3] Theovertime. results Once canpoled, bethe rseen33 value in was the measured following via the reflection section. technique . The results can be seen in the following section.

Figure 3. Poling setup.

Scheme 5. Functionalization of AJC-139 chromophore RESULTS

Fig. 3. Poling setup. Figure 4. Current Vs. Time. Results: All the samples were exposedFigure to 3 the. Poling poling voltagesetup. and then heated at a rate of 5 C/sec starting at 50 C. The current through the polymer layer varies over time as the sample is heated (figure 4). This gives an indication of when

the T Allg is beingthe samples approached were as the currentexposed reaches to athe local poling voltage and then maxima (or a minima in the case of polymer breakdown and heatedcurrent is at able a to rate jump ofdirectly 5° C/secfrom the goldstarting to the ITO). at 50 °C. The current through the polymerWith a 25 layerwt.% loading varies density, over poling time with as 100V/um the sample is heated (figure at 111 C the r33 value was found to be as high as 44pm/V 4).(Table This 1). Thisgives is nearly an doubleindication the value forof thewhen less efficient the T is being approached chromophores previously implemented in such a side-chain g Fig. 4. Current Vs. Time. asscheme. the current The high poling reaches temperatur a locale is an maximaindication that (or the a minima in the case of side-chains formed and that the thermal stability of the system polymerwill be good. breakdown Additionally, theand very current low current is at able such highto jump temperatures directly (4.5 uA from at 160 theC) indicates that there is little phase separation between the polymer and the chromophores, which goldis in agreement to the ITO).with expectations for a side-chain structure. With a 25 wt.% loading density, poling with 100V/um at Sample Field (V/um) Temp ('C) Current (uA) r33 (pm/V) #1 100.00 160 4.5 42 111°C the r33 value was found to be as high as 44pm/V (Table 1). This is nearly double the value for the less efficient chromophores previously implemented in such a side-chain scheme. The 6high of 7 poling temperature is an indication that the side-chains formed and that the thermal stability of the system will be good. Addi- tionally, the very low current at such high temperatures (4.5 uA Scheme 6. Functionalization of 2nd polyene cromophore. at 160°C) indicates that there is little phase separation between the polymer and the chromophores, which is in agreement with expectations for a side-chain structure.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 97 Synthesis of a Polyene EO Chromophore Denis Nothern, Jingdong Luo, Alex K.-Y. Jen

At this point synthesis was considered complete and chromophore 17 was embedded in an anthracene functionalized poly-methyl-methacrylate (PMMA) host polymer with a 25 .88 .12 wt.% loading density (figure 2). It was spin-cast to form a O OCH3 O O .75um film on a glass substrate half coated with conducting indium-tin-oxide (ITO) (120nm) particularly designed to be optically transparent at telecommunications wavelengths (1.3um). Three samples were prepared by sputtering a gold electrode onto the top of the polymer film such that they were Fig. 2. Polymer Host. half over the ITO and half over the glass (figure 3). This allowed a voltage to be applied across the electrode and the ITO layer and pole the polymer. While the poling voltage was being applied, current passing through the polymer was carefully monitored for indications of damage to the film. During heating, the current was recorded [3] overtime. Once poled, the r33 value was measured via the reflection technique . The results can be seen in the following section.

Fig. 3. Poling setup. SYNTHESIS OF A POLYENE EO CHROMOPHORE: USING A DIELS-ALDER REACTION TO FORM A SIDE-CHAIN STRUCTURE Results: All the samples were exposed to the poling voltage and then heated at a rate of 5 C/sec starting at 50 C. The current through the polymer layer varies over time as the sample is heated (figure 4). This gives an indication of when the Tg is being approached as the current reaches a local maxima (or a minima in the case of polymer breakdown and current is able to jump directly from the gold to the ITO). With a 25 wt.% loading density, poling with 100V/um at 111 C the r33 value was found to be as high as 44pm/V (Table 1). This is nearly double the value for the less efficient chromophores previously implemented in such a side-chain Fig. 4. Current Vs. Time. scheme. The high poling temperature is an indication that the Table 4. Current vs. Time side-chains formed and that the thermal stability of the system will be good. Additionally, the very low current at such high temperatures (4.5 uA at 160 C) Sample Field (V/um) Temp (‘C) Current (uA) r (pm/V) indicates that there is little phase separation between the polymer and the chromophores, which 33 is in agreement with expectations for a side-chain structure.#1 100.00 160 4.5 42 #2 100.00 110 58 35

#3 100.00 111 45 44 Sample Field (V/um) Temp ('C) Current (uA) r33 (pm/V) #1 100.00 160 4.5 42 Table 1. r33 values.

CONCLUSION It has been shown that a highly efficient6 ofpolyene 7 chromo- phore can be functionalized with a dienophile. A feat that was previously unachievable. Furthermore, the resulting compound can be incorporated into a side-chain polymer via Diels-Alder “Click Chemistry” and efficient poling has been achieved in this

polymer system resulting in a reasonable r33 value. Improvements in long-term stability compared to the guest- host structure are yet to be shown for such a system, but previous investigations of Diels-Alder based side-chain structures with other chromophores is promising.

ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

REFERENCES [1] J. Luo, M. Haller, H. Li, T.-D. Kim, A. K.-Y. Jen, Adv. Mater. 2003, 15, No. 19. [2] Francesco Fringuelli and Aldo Taticchi, The Diels-Alder Re- action: Selected Practical Methods. John Wiley & Sons: 2002. [3] P. Gunter (Ed.), Nonlinear Optical Effects and Materials. Springer-Verlag: New York, 2000. (p. 197) [4] Hoyt and Benicewicz. Poly. Sci. Part A: Poly. Chem. 1990, v. 28, p. 3403.

98 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Spectroscopic Investigations of Chromophores in Dyed Salt Crystals

Stacy A. Oliphant Kristin L. Wustholz, Bart Kahr and Philip J. Reid Edmonds Community College Department of Chemistry University of Washington

INTRODUCTION alignment and photophysics of single chromphores in dye-doped crystal hosts using confocal microscopy. Furthermore, by com- The Science and Technology Center (STC) is committed to paring the orientations of dyes to their energetics, the environ- the development of next-generation electro-optic (EO) devices, mental heterogeneity among individual dye molecules within the typically based on dye-doped polymers. Polymer-based switches crystal can be studied. have demonstrated greater device efficiencies when compared to traditional inorganics. Since the EO effect relies on the nonlinear material response, the dye-doped polymers must be noncentro- symmetric. This is typically achieved via the application of a poling field. Ultimately, EO activity is given by:

|Nµβ| where the EO response is proportional to the number of dye mol- ecules (N), the molecular dipole moment (μ), hyperpolarizability (β), and the ordering parameter (cos3 θ) that represents the extent of alignment afforded by poling. Understanding the factors that influence chromophore orientation has taken a backseat to the de- velopment of enhanced molecular properties (i.e. β). Yet, recent work suggests that poling is only partially effective in achieving Figure 1. Dye-Doped KAP Crystals. Violamine R (1) and DCM molecular alignment1 – making it quite clear that the poling pro- (2) ad3sorb to different growth sectors of the KAP crystal host. cedure and ordering parameter are not well understood. To experimentally test the efficacy of poling, and hence EXPERIMENTAL bulk EO device efficiency, it makes sense to begin with a sys- Dyed crystals were grown by slow evaporation from aque- tem in which the chromophores are intrinsically aligned by the ous solution, in a temperature controlled air chamber (30°C) with host. When grown in the presence of many organic dyes, aque- dye concentrations of 10-4 M to 10-8 M. The latter concentration ous solutions of potassium acid phthalate (KAP) and sodium po- was used to ensure single-molecule resolution. tassium tartrate tetrahydrate (commonly known as Rochelle salt) Single-molecule experiments were performed using a con- frequently deposit “dyed crystals” wherein chromophores are ori- focal microscope with the following parameters: an inverted ented and overgrown by the host lattice. In these crystals, dyes microscope (Nikon, TE2000). Excitation from a 532 nm solid- selectively adsorb to particular sub-volumes within the lattice state laser (NovaLux) was filtered (Chroma) and the excitation and bulk spectroscopic studies suggest that they are intrinsically polarization was manipulated using a half-waveplate. The laser aligned during the growth process. For example, Figure 1 shows was focused to a diffraction-limited spot with an objective (100x, that 4-(Dicyanomethylene)-2-methyl-6-(p-dimethylaminosty- Nikon, 1.3 NA). Cleaved crystal samples (~1 mm) were mounted ryl)-4H-pyran (DCM) adsorbs to the {11-1} growth sector, while on a closed loop x-y piezo scanning stage (Queensgate). Emis- Violamine R (VR) adds to the {010} growth sector of the host sion was spectrally filtered by a dichroic mirror and emission crystal, KAP. filters (Chroma) and spatially filtered with a confocal pinhole Here, a natural comparison presents itself in the study of the (ThorLabs: 50 μm). The fluorescence intensity was imaged onto NLO properties of chromophore-polymer composites and dye- a single-photon counting APD (PerkinElmer). Single-molecule doped single crystals. We use single-molecule spectroscopy to fluorescence spectra were measured by adding a monochromator measure the orientations and fluorescence spectra of embedded (Acton) attached to a liquid N2 CCD camera (Princeton Instru- dye molecules. The focus of this work was to interrogate the ments) to the confocal instrument.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 99 SPECTROSCOPIC INVESTIGATIONS OF CHROMOPHORES IN DYED SALT CRYSTALS

less intense on the hillock, and exhibited stronger fluorescence at RESULTS 647nm than 604 nm. Further work will be done to explore the The heavily-dyed crystals were grown and studied by absor- properties of dyes on/off hillocks. Since we were unable to mea- sure single-molecule spectra with this detector, a liquid N cooled bance and fluorescence. By using these techniques, ensemble- 2 average properties of the dyes were measured. Figure 2 shows CCD camera and monochromator was coupled to the confocal polarized absorption spectra of a heavily dyed VR/KAP crystal, microscope. The new setup was characterized using fluorescent demonstrating that chromophores are oriented ~42° from [100]. beads (Molecular Probes) and heavily-dyed VR/KAP crystals. Bulk studies allow one to measure average orientation, but the The procedure for obtained a single-molecule fluorescence spec- extent of alignment (the orientational distribution) is concealed trum was as follows: the crystal was first scanned using the un- 2 at high loading densities. To measure the extent of alignment modified confocal set-up to obtain the 10 x 10 µm image with and single-dye energetics, we chose to study dyed crystals on a 100 ms integration time at a power of 5 µW. Then, the micro- molecular level. By using single-molecule excitation dichroism, scope stage was used to focus the excitation on a single molecule, the orientation of the dyes included in crystals can be measured. and the power was intensified to 10 µW and the integration time Figure 3 shows the fluorescence dichroism of a single-molecule was increased to 10 minutes. Figure 4 shows a fluorescence spec- VR dyed KAP crystal grown from a concentration of 10-8 M dye. trum of a single dye molecule in VR/KAP. With single-molecule -1 resolution demonstrated, future work will be done to optimize The dichroism was measured using the equation θ = tan [√Ia/Ic] and the average orientation for 11 molecules is 39.6° from [100] data collection and correlate the fluorescence spectra with orien- which agrees with the bulk absorption measurement of ~42° from tation. [100]. Our results showed agreement between single-molecule Another interesting system that was studied this summer is and bulk measurements; and to prior single-molecule measure- dyed Rochelle salt crystals. Rochelle salt (RS) is a ferroelectric ments on this system.2 Yet, to study the emission properties of crystal that undergoes a phase transition at certain temperatures single-dye molecules in this system we needed to modify the con- called Curie points. Our goal was to grow dyed RS crystals and focal instrument accordingly. to use the embedded dyes to report on the ferroelectric phase tran- sition. Unfortunately, most of the time was spent determining the optimal growth conditions for the RS crystals, and we were unable to study to begin single-molecule spectroscopic investiga- tions. Yet, preliminary studies on sulforhodamine B (SB) and Chicago Sky Blue (CSB) dyed RS crystals were performed (Fig- ure 5). We successfully reproduced prior work on CSB/RS crys- tals and demonstrated that the fluorescent dye SB incorporates in the {010} growth sector of RS, absorbs at 505 nm, and emits at 583 nm. Future work will be done to learn more about the ferro- electric properties of the crystal by monitoring SB dye orientation during a phase transition at the single-molecule level.

Figure 2. Polarized Absorption Spectra of 10-5 M VR/KAP Crystal. Dyes were measured to be oriented ~42° from

[100] at an absorption maximum (λmax) of ~545nm.

A monochromator/ Hamamatsu CCD camera assembly was added to the confocal microscope and used to measure fluores- cence spectra of dyes in growth hillocks. Preliminary work on lightly-dyed (10-5 M dye) VR/KAP crystals using a Hamamatsu CCD camera suggested environmental heterogeneity between Figure 3. Single-Molecule Orientations in VR/KAP. The crystal was dyes located on or off growth hillocks (Figure x). The fluores- scanned on both the “a” and “c” directions to obtain a 10 x 10 μm2 cence intensities at 604 nm and 647 nm were monitored as a area, over 100 ms integration time, at 100 nm step size at 5 μW power. Fluorescence Intensity (Ia): 506 – 29 / 100 ms, Fluorescence Intensity (Ic): function of location on/off the hillock. Interestingly, dyes were -1 233 – 29 / 100 ms, θavg = 39.6° for 11 molecules using θ = tan [√Ia/Ic].

100 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 OLIPHANT

SUMMARY This summer, dyed crystals were grown and characterized using conventional and single-molecule spectroscopy. The aver- age of single-molecule orientations in VR/KAP crystals agreed with the ensemble-average measurements, yet orientations out- side the ensemble-average were observed. These results are in agreement with prior work on this system. 2 Preliminary work on lightly-dyed (10-5 M dye) VR/KAP crystals suggested envi- ronmental heterogeneity between dyes located on or off growth hillocks. By modifying the confocal microscope to include a

monochromator and liquid N2 cooled CCD camera, single-mol- ecule fluorescence spectra were obtained. In demonstrating the single-molecule resolution of the monochromator/CCD camera setup, we have shown that this instrument can be used to further investigate the environmental heterogeneity among dyes within crystals and eventually, poled polymers. Only limited studies, particularly bulk characterization of dyed Rochelle salt, were completed. Future work will be done on these crystals to study the ferroelectric properties of the host using single fluorescent reporters to monitor a phase transition.

Figure 4. Intensity Map and Corresponding Single-Molecule Fluorescence ACKNOWLEDGEMENTS Spectra in VR/KAP. The single-molecule fluorescence spectrum was obtained using 10 μW incident power and 10 min integration time. The authors thank the National Science Foundation (NSF) and the Center on Materials and Devices for Information Tech- nology Research (CMDITR) for support of this work. S.O. is supported by the CMDITR REU summer program.

REFERENCES 1. Robinson, B.H.; Dalton, L.R. J. Phys. Chem. A. 2000, 2000, 4785. 2. Wustholz, K.L; Kahr, B.; Reid, P.J. J. Phys. Chem. In press Sept. 2005

Figure 5. Rochelle Salt Dyed Crystals; Chicago Sky Blue (left) & Sulforhodamine B illuminated by a short wave ultra-violet lamp (254 nm) (right).

Figure 6. Fluorescence on Growth Hillocks in 10-5 M Dyed VR/KAP. A 5x5μm2 area (box in blue) was scanned, using a 500 nm step size and 18 sec. integration time. The intensity maps show environmental heterogeneity within the crystal. The third intensity map (I647 nm/I604 nm) shows change in relative intensity of these peaks on the hillock versus off the hillock.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 101 102 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Characterization of the Molecular Parameters Determining Charge-Transport in a Series of Substituted Oligoacenes

Robert Snoeberger Seth Marder University of Washington Georgia Institute of Technology

Organic semiconductors are currently receiving much re- on chalcogen-substituted oligoacenes, where the number of fused search interest due to applications in devices such as field-effect benzene rings in the center of the molecule is increased. The mol- transistors (FET), photovoltaic (PV) cells, and organic light- ecules used in the study, system (a) and system (b), are displayed emitting diodes (OLED). These organic devices may be produced in figure 1. more cheaply and efficiently than inorganic counterparts, how- Marcus theory may be used to characterize the charge-trans- ever, more efficient and stable materials need to be designed for fer properties of organic materials which will help the develop- organic devices to become feasible for production. ment of future systems. The mobility of a material may be de- Currently, pentacene is commonly used as an organic semi- scribed with equation 1,4,7,8 at high temperature and assuming the conductor, in crystalline FET applications, due to the large hole transfer process is described by a series of uncorrelated hops. mobility1. It has been seen however, that the mobility of penta- cene based FETs decrease with exposure to air2,3. A remedy is (1) the replacement of the terminal fused benzene rings with thio- phene rings making anthradithiophene (ADT); ADT provides Where e is the electronic charge, a is the distance between 4,5 greater air stability with an acceptable drop in device mobility , molecules, kET is the hopping rate, kB is the Boltzmann constant, with a measured mobility of 0.02 cm2/Vs at room temperature. and T is the temperature in Kelvin. It is shown that the mobility is A further extension to this idea, explored by Takimiya et al6, proportional to the hopping rate, which is expressed by equation was the replacement of the sulfur atoms in the thiophene rings 2,4,7,8 with larger chalcogen atoms. The intent is to increase the electron coupling between adjacent molecules. Takimiya reported the syn- (2) thesis of diphenyl-benzodithiophene (DPh-BDS) along with the selenium and tellurium analogs. Device mobilites were measured Here, λ is the reorganization energy, t is the transfer integral, for hole-transport for the molecules, DPh-BDS gave a mobility and h is the Planck constant. The reorganization energy is com- of 1.6x10-2 cm2/Vs at room temperature compared to DPh-BDT posed of intermolecular (outer) and intramolecular (inner) con- which gave a mobility of 4.6x10-3 cm2/Vs at room temperature6. tributions. The outer reorganization energy arises from the polar- ization and relaxation of the surrounding environment while the inner reorganization energy corresponds to the geometry relax- ations occurring during the charge-transfer reaction. The charge- transfer reaction may be modeled as a simultaneous oxidation Anti (a) Syn and reduction reactions. The oxidation and reduction reactions

include associated vibrational relaxations, whose energy corre- sponds to the relaxation energy. The inner reorganization energy is composed of the sum of the two relaxation energies, one from the oxidation reaction and the other from the reduction reaction. Anti (b) Syn According to equation 2, the activation energy for the hopping rate is equal to one fourth of the reorganization energy. This rela- The goal of our summer work has been to characterize these tion makes the reorganization energy an important parameter to systems using charge-transport theory. The theoretical effort was characterize for charge-transport systems, and is the dominant fo- applied to the oligoacenes without phenyl substituents so that the cus of the current work. The reorganization energy for the systems effect of the chalcogen atom on the oligoacene may be isolated. of interest were calculated using the Gaussian9 98 suite of programs A second type of systems which was characterized was based with Density Functional Theory (DFT)10 using Becke three-param-

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 103 CHARACTERIZATION OF THE MOLECULAR PARAMETERS DETERMINING CHARGE-TRANSPORT IN A SERIES OF SUBSTITUTED OLIGOACENES eter Lee, Yang, Parr (B3LYP)12 functional. The results of the reor- The results of the reorganization energy calculations for sys- ganization energy calculations for system (a) are shown in chart 1 tem (b) are shown in chart 2 below, calculated using the 6-31G** below. The LanL2dz basis set13,14 was used during the calculations, basis set. unless otherwise specified, because the 6-31G** basis set15 is not available for tellurium, but would be preferred. Therefore, chart 1 contains values calculated with the 6-31G** basis set for sulfur and selenium as a comparison with the LanL2dz basis set.

Chart 2. The Reorganization Energy for System (b) Calculated with B3LYP/6-31G** For Hole-transport

The results from chart 2 show that the isomer dependence

Chart 1. Reorganization Energy of System (a) becomes negligible at long molecular lengths. The reorganization Calculated with B3LYP/LanL2dz energy for the syn isomer stays approximately constant along the For Hole-transport series which is due to the overall change in geometry being the same. This may be attributed to more of the cation charge being The results on chart 1 show as a general trend, for system (a), localized on the chalcogen atom at small molecular lengths but that the reorganization energy, for hole-transport, decreases mov- becomes less of a factor at longer molecular lengths. ing along the system from the sulfur analog to the tellurium analog. The transfer integral, which is the dominant constituent of the This is attributed to a smaller geometry relaxation for the tellu- exponential prefactor in equation 2, may be approximated as half rium analog than that of the sulfur analog because the bond-length the electronic coupling between the donor and acceptor. How- change of the carbon-carbon bonds becomes smaller moving to- ever, a direct calculation of the transfer integral for a particular ward the tellurium analog (0.10A to 0.068A) showing that larger system is dependent on bulk order, which is difficult to deter- portion of the positive charge in the cation state is being located mine. A general trend may be extracted by changing the dimer on the chalcogen atom. A population analysis confirms that more orientation during the electron splitting calculations, which is of the cation charge is located on the larger chalcogen atoms (0.18 presented in chart 3 below, where the dimer distance is increased. electron charge on S in BDT-syn compared to 0.28 electron charge The calculations were performed on Ampac16 using the AM117 on Te in BDTe-syn) but that the geometry change associated with semi-empirical method. the larger population is minimal (0.090A in BDT-syn compared to 0.084A in BDTe-syn). Chart 1 also displays an isomer dependence in the system, where the syn isomer has a lower reorganization energy. This trend is also caused by the change in cationic character on the chalcogen atom in the syn isomer compared to the anti isomer. The bond- length change of the chalcogen-carbon bonds was larger in the syn isomer than that of the anti isomer (0.09A for the syn isomer and 0.08A for the anti isomer of BDT) while the bond length change of the carbon-carbon bonds was much larger in the anti isomer than the syn isomer (0.10A for the syn isomer and 0.15A for the anti isomer of BDT). A population analysis also confirms the correla- tion between reorganization energy and amount of charge of the Chart 3. Transfer Integral for System (a) chalcogen atoms (0.18 electron charge in BDT-syn compared to Calculated with AM1 0.17 electron charge in BDT-anti). For Hole-transport

104 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 SNOEBERGER

The roll-over in the plots in chart 3 occur at approximate- ACKNOWLEDGEMENTS ly twice the vanderwaals radius of the chalcogen atom, which Research support is gratefully acknowledged from the Na- is 3.7A for the sulfur analog, 4.0A for the selenium analog, and tional Science Foundation Center on Materials and Devices for 4.4A for the tellurium analog. The results of chart 3 show that Information Technology Research (CMDITR), DMR-0120967. the tellurium analog has a larger transfer integral at larger dimer distances than that of the selenium and sulfur analogs, which is consistent with the idea that a larger atomic radii will result in a larger transfer integral. However, this is not seen between the sulfur and selenium analogs which suggests that the interaction between the carbon orbitals dominant the electron splitting. The current work shows that including large chalcogen at- oms in hole-transport materials results in desirably low reorga- nization energies and that the smaller molecules show an isomer dependence, which is more pronounced with the smaller chalco- gen atoms. These systems are interesting for charge-transport and synthesis is recommended.

REFERENCES 1. O.D. Jurchescu, J. Baas, T.T.M. Palstra, Appl. Phys. Lett. 84. 3061-3063. (2004) 2. J.H. Schon, Appl. Phys. Lett. 79. 4163-4162. (2001). 3. M. Yamada, I. Ikemote, H. Kuroda, Bull. Chem. Soc Jpn.61. 1057. (1988). 4. O. Kwon, et. al. J. Chem. Phys. 120, 8186-8194. (2004). 5. J.G. Laquindanum, H.E. Katz, A.J. Lovinger, J. Am. Chem. Soc. 120. 664-672. (1998). 6. K.Takimiya, Y. Kunugi, Y, Konda, N. Niihara, T. Otsubo. J. Am. Chem. Soc. 126. 5084-5085. (2004). 7. V. Coropceanu, J.M. Andre, M. Malagoli, J.L. Bredas. Theor Chem Acc. 110. 59-69. (2003) 8. M. Malagoli, V. Coropceanu, D.A. da Silva Filho, J.L. Bredas. J. Chem. Phys. 120, 7490-7496. (2004). 9. M.J. Frisch, et al. Gaussian98, Revision A. 11, Gaussian, In- corporated: Wallingford, CT, 1998. 10. P. Hohenberg and W. Kohn, Physical Review 136, B864- B871 (1964). 11. R.G. Parr and W. Yang, Density-functional theory of atoms and molecules, Oxford Univ. Press: Oxford, (1989). 12. A.D. Becke, J. Chem. Phys. 98 5648 (1993). 13. T. H. Dunning, Jr. and P. J. Hay, in Modern Theoretical Chem- istry, Ed. H. F. Schaefer, III, Plenum: New York, 1-28, (1976). 14. P.J. Hay and W.R. Wadt, J. Chem. Phys. 82, 270 (1985). 15. R. Ditchfield, W.J. Hehre and J.A. Pople, J. Chem. Phys. 54, 724 (1971). 16. AMPAC 8, © 1992-2004 Semichem, Inc. PO Box 1649, Shawnee, KS 66222. 17. M.J.S. Dewar, E.G. Zoebisch, E.F. Healy, J.J.P. Stewart. J. Am. Chem. Soc. 107. 3902-3909. (1985).

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 105 106 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Optimization of Semiconductor Nanoparticle Synthesis and Integration into Sol-Gel Monoliths

Cindy Taylor Scott Saavedra and Neal Armstrong University of Arizona Saavedra Lab, Department of Chemistry The University of Arizona

INTRODUCTION EXPERIMENTAL PROCEDURES Energy sources today are becoming limited and therefore ex- Two aspects of the proposed device, that of the synthesis and pensive. Gasoline prices are skyrocketing and our nonrenewable optimization of the SC-NP’s and their inclusion in the sol-gel ma- sources of energy such as fossil fuels are running low. Research trix, were the focus of recently conducted research. has been done and continues on for a solution to this problem. One possible answer, hydrogen, has incredible possibilities. Nanoparticle Synthesis Hydrogen fuel cells have been built that successfully convert The procedure for synthesizing nanoparticles developed by hydrogen gas into usable energy forms such as electricity. Hydro- Jiang1 was used to create both cadmium sulfide (CdS) and cad- gen atoms are electrochemically separated from protic solutions, mium selenide (CdSe) SC-NP’s. such as water, and combined to form hydrogen gas. The hydro- Cadmium myristate (CdMA, 1.134g, 2mmol) was placed in gen gas can then be put to work running a car or even heating a a 250 mL round bottom flask with toluene (50mL) and n- house. trioctylphosphine oxide (TOPO, 4.99g, 12.9mmol) or oleic acid These cells are not only attractive for what they do, but how (OA, 5mL) or a combination of the two. This solution was stirred they do it. They are clean, efficient, and renewable. With only vigorously at 100˚C until all solid had dissolved. Following dis- water and heat as byproducts, hydrogen fuel cells are ecological solution, an aqueous solution of thiourea (0.19g, 2.5mmol) in de- as well as economical. The water can also be reseparated into ionized water (50mL) was added. The reaction was performed Figure 2: Reaction scheme for TOPO capped hydrogen and oxygen, contributing to the self-sustaining process Experimental Procedures under reflux at Two 100˚C aspects for six hours. of the proposed device, CdS semiconductor nanoparticles. (Ref. 1) that makes fuel cells so efficient and appealing. Thethat formation of the ofsynthesis the nanoparticles and optimization occurred at the of water/tolthe - The research performed at the Center for Materials and De- SC-NP’s and their inclusion in the sol-gel Once the reaction was complete the uene interface (Figure 2). Here the decomposing thiourea com- mixture was separated using a separatory vices for Information Technology Research (CMDITR) is aimed bined withmatrix, the cadmium were the to focus form CdSof recently molecules conducted that would then research. funnel and the aqueous phase discarded. at creating a hydrogen generating device. The arrangement pro- move into the organic phase. The excess TOPO or OA in this The remaining organic layer was then duces hydrogen through photocatalytic reduction of water/protic layer “capped” or attached to the surface of the CdS molecule washed with ethanol and centrifuged three Nanoparticle Synthesis times. The resulting solid was dispersed in solution using ligand-protected semiconductor nanoparticles creating a protective The procedure ligand layer. for As the synthesizing nanoparticles were 1 dichloromethane and allowed to dry (SC-NP) embedded in a sol-gel matrix and attached to conduct- stirred, nanoparticlesthey returned to developedthe interface by where Jiang more was CdS used molecules overnight yielding yellow to orange ing polymer chains (Figure 1). attachedto to createthe core, both increasing cadmium the diameter sulfide of (CdS) the nanoparticle. and crystalline CdS nanoparticles. It was later cadmium selenide (CdSe) SC-NP’s. noted that carrying out these washings resulted in a substantial loss of product and diminished the desired optical properties of the SC-NP’s. Cadmium selenide nanoparticles were synthesized using the same procedure as the cadmium sulfide NP’s with slight differences. The molar quantities of the reagents were scaled down by a factor of five and selenourea (0.016g, 0.12mmol) replaced thiourea in the aqueous phase. In replacing the sulfur atom with selenium it became necessary to degas the aqueous phase with argon prior to the reaction. This Figure 1. Proposed matrix for photocatalytic reduction of protons prevented oxidation of the precursors. Only oleic acid was used as the capping using SC-NP’s and conducting polymers in a sol-gel matrix. Figure Cadmium 2. Reaction myristate scheme for TOPO (CdMA, capped CdS 1.134g, Optimization of Semiconductor Nanoparticle semiconductor nanoparticles. (Ref. 1) ligand during CdSe formation. Also, the 2mmol) was placed in a 250 mL round formation of CdSe was considerably Synthesis and Integration into Sol-Gel bottom flask with toluene (50mL) quicker than that of CdS due to the faster CMDITRand Review n-trioctylphosphine of Undergraduate Research Vol. oxide 2 No. 1 Summer (TOPO, 2005 107 decomposition rate of selenourea, resulting Monoliths 4.99g, 12.9mmol) or oleic acid (OA, 5mL) in a reaction time of only three minutes at or a combination of the two. This solution 100C. Because of this, the synthesis was was stirred vigorously at 100C until all also performed several times at 80C, Cindy Taylor, University of Arizona solid had dissolved. Following dissolution, resulting in a required time of 16-20 an aqueous solution of thiourea (0.19g, Scott Saavedra, Neal Armstrong minutes to reach the desired nanoparticle 2.5mmol) in deionized water (50mL) was diameter. Due to the product loss that Saavedra Lab, Department of Chemistry, The University of Arizona added. The reaction was performed under occurred during the ethanol self-sustaining process that makes fuel cells reflux at 100C for six hours. washing/centrifuging process of the CdS Introduction The formation of the nanoparticles nanoparticles, only an extraction was Energy sources today are so efficient and appealing. occurred at the water/toluene interface performed on the CdSe reaction mixture becoming limited and therefore expensive. The research performed at the Center (Figure 2). Here the decomposing thiourea and the organic layer containing the for Materials and Devices for Information combined with the cadmium to form CdS nanoparticles was transferred directly to a Gasoline prices are skyrocketing and our molecules that would then move into the Technology Research (CMDITR) is aimed storage vial. nonrenewable sources of energy such as organic phase. The excess TOPO or OA in The SC-NP’s were characterized by fossil fuels are running low. Research has at creating a hydrogen generating device. this layer “capped” or attached to the UV-Vis absorbance and fluorescence The arrangement produces hydrogen surface of the CdS molecule creating a spectra (Figure 3). SC-NP’s typically have been done and continues on for a solution protective ligand layer. As the nanoparticles through photocatalytic reduction of a high molar absorptivity, which was to this problem. One possible answer, were stirred, they returned to the interface observed in the analyses taken of the water/protic solution using ligand-protected where more CdS molecules attached to the samples. Due to the incredibly small size hydrogen, has incredible possibilities. core, increasing the diameter of the Hydrogen fuel cells have been built semiconductor nanoparticles (SC-NP) of the nanoparticles (5-7 nm for CdS & 2-3 embedded in a sol-gel matrix and attached nanoparticle. nm for CdSe), the absorbance spectra were that successfully convert hydrogen gas into blue shifted relative to the bulk material. usable energy forms such as electricity. to conducting polymer chains (Figure 1). Hydrogen atoms are electrochemically 2 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 separated from protic solutions, such as water, and combined to form hydrogen gas. The hydrogen gas can then be put to work running a car or even heating a house. These cells are not only attractive for what they do, but how they do it. They are clean, efficient, and renewable. With only water and heat as byproducts, hydrogen fuel cells are ecological as well as economical. Figure 1: Proposed matrix for photocatalytic The water can also be reseparated into reduction of protons using SC-NP’s and hydrogen and oxygen, contributing to the conducting polymers in a sol-gel matrix.

CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 1 OPTIMIZATION OF SEMICONDUCTOR NANOPARTICLE SYNTHESIS AND INTEGRATION INTO SOL-GEL MONOLITHS Onceanticipated, the reaction proving was complete the the nanoparticlesmixture was separated had usingthe a separatory desired funnel and optical the aqueous propertyphase discarded. of The photoluminescenceremaining organic layer was while then washed the second with ethanol was and centrifugedmuch smaller three times. and The red-shifted. resulting solid was This dispersed was in dichloromethanedue to unpassivated and allowed tosites dry overnighton the yieldingsurface yel -of low tothe orange nanoparticles crystalline CdS nanoparticles. called “surface It was later traps noted ” that carrying out these washings resulted in a substantial loss of productthat trappedand diminished photons the desired and optical released properties them of the at SC-NP’s.higher wavelengths. This second low- Cadmiumenergy peakselenide increased nanoparticles in were size synthesized as the size using of the samethe procedure CdS nanoparticles as the cadmium increased.sulfide NP’s with Surface slight differences.traps were The molar not quantities a problem of the in reagents the spectra were scaled for downthe by a cadmiumfactor of five selenideand selenourea though, (0.016g, indicating0.12mmol) replacedthat thiourea these in the NP’s aqueous had phase. moreIn replacing thoroughly the sulfur atompassivated with selenium surfaces.it became necessary The to intensity degas the aqueous of the phaseCdSe with argon particles prior to wasthe reaction. also much This prevented greater oxida than- tion thatof the ofprecursors. CdS at Only the oleicsame acid concentration was used as the capping giving ligandthem during greater CdSe formation. efficiency. Also, the formation of CdSe was considerably quicker than that of CdS due to the faster decompositionSol-Gel rate Formationof selenourea, resulting in a reaction time of only three Sol-gels minutes at are 100˚C. formed Because by of the this, hydrolysisthe synthe- sis wasof also an performed alkoxide several followed times at by80˚C, condensation resulting in a required(Figure time of 4).16-20 Deionizedminutes to reach water the desired (0.214 nanopar mL,- ticle 12mmol)diameter. Due and to the product hydrochloric loss that occurred acid during (0.56 the ethanol mol) washing/centrifuging were added process toof the CdS tetramethyl nanopar- ticles,orthosilicate only an extraction (TMOS, was performed 1 mL, on the6.7mmol) CdSe reaction in a mixturesmall and the vial organic and layer stirred containing for the 15nanoparticles minutes was to transferredhydrolyze directly the to a methylstorage vial. terminated ends of the The SC-NP’s were characterized by UV-Vis absorbance and alkoxide,fluorescence spectra yielding (Figure the 3). precursorSC-NP’s typically solution. have a highOnce molar completelyabsorptivity, which hydrolyzed was observed (evidenced in the analyses by takenthe of the evolution samples. Due of to the the liquidsincredibly tosmall one size phase) of the nanoparticlesthe precursor (5-7 nm for was CdS & added 2-3 nm for to CdSe), a disposable the absor- banceacrylate spectra were cuvette blue shifted containing relative tophosphate the bulk material. buffer The(20mM, fluorescence pH spectra7, 2 mL) for the and cadmium deionized sulfide nanopar water- ticlesin contained a 1:1 two ratio peaks. with The the first volumehigh-energy of peak solvent was anticipated,containing proving nanoparticles. the nanoparticles had the desired optical property of photoluminescence while the second was much smaller and red-shifted. This was due to unpassivated sites on the surface of the nanoparticles called “surface traps” that trapped photons and released them at higher wavelengths. This Figuresecond 4:low- Schematic energy peak representation increased in size of ashydrolysis the size of the CdSand nanoparticles condensation increased. occurring Surface during traps were sol-gel not a problem in the spectramonolith for the cadmium formation. selenide though, indi- cating that these NP’s had more thoroughly passivated surfac- es. The intensity The first of the variable CdSe particles optimized was also wasmuch thegreater pH than ofthat of the CdS atbuffer the same concentration solution giving used. them greater The Figure 3. Absorbance and Fluorescence Spectra of CdS-1 and CdSe-16 (24 Figure 3: Absorbance and Fluorescence Spectra efficiency. minutes), the nanoparticle samples used in sol-gel inclusion experiments. polymerization (solidifying) rate of the sol- of CdS-1 and CdSe-16 (24 minutes), the gel increased in direct proportion with pH. nanoparticle samples used in sol-gel inclusion This was initially a problem as the sol-gel experiments. 108 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 would polymerize quickly, entrapping air bubbles and weakening the matrix. It was determined that a pH of 6.5 was ideal The fluorescence spectra for the because the sol-gel solidified in a cadmium sulfide nanoparticles contained two peaks. The first high-energy peak was

CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 3 reasonable amount of time while allowing a majority of the air bubbles to escape. The second factor to be modified was reasonablethe solvent. amount The of desired time while characteristics allowing a majoritywere a solventof the air the bubbles nanoparticles to escape. would be soluble in, water miscible, and one that The second factor to be modified wasTAYLOR thewould solvent. not quench The the desired luminescence characteristics of the Sol-Gelwerenanoparticles. aFormation solvent the nanoparticles This luminescence would be quenching effect was determined to be a CdSe-16 in Dioxane soluble Sol-gels are in, formed water by themiscible, hydrolysis and of an alkoxideone that fol- problem with the first solvents used, 1.20 lowedwould by condensation not quench (Figure the 4). luminescence Deionized water (0.214 of themL, 1.00 methanol and ethanol. These solvents, as L 12mmol)nanoparticles. and hydrochloric acid (0.56 This μmol) were luminescence added to tetra- P 0.80 well as tetrahydrofuran (THF) and d methyl orthosilicate (TMOS, 1 mL, 6.7mmol) in a small vial and e CdSe-16 in Dioxane z

quenching effect was determined to be a i l 0.60

dimethylformamide (DMF), acted as a stirred for 15 minutes to hydrolyze the methyl terminated ends 1.20 m

problem with the first solvents used, r 0.40 of theelectron alkoxide, donors,yielding the filling precursor in solution. the exciton Once complete hole - o N 1.00 methanol and ethanol. These solvents, as L left by excited electrons in the SC-NP. P ly hydrolyzed (evidenced by the evolution of the liquids to one 0.20 0.80 well as tetrahydrofuran (THF) and d

This resulted in a lack of emission by the e 0.00 phase) the precursor was added to a disposable acrylate cuvette z i

l 0 7 400.600 450 500 550 600 65 00 dimethylformamidenanoparticles and consequently (DMF), no acted peaks in as a

containing phosphate buffer (20mM, pH 7, 2 mL) and deion- m Wavelength (nm r 0.40 electronthe spectrum. donors, filling in the exciton hole o ized water in a 1:1 ratio with the volume of solvent containing N

left by A excited peak was electrons finally in found the SC-NP. in the 0.20 nanoparticles.Thisfluorescence resulted spectrain a lack of of CdSe-16 emission when by p-the Figure 5. Fluorescence spectrum of CdSe-16.24 in p-dioxane solvent. 0.00 nanoparticlesdioxane was usedand consequentlyas the solvent no(Figure peaks 5). in 400 450 500 550 600 650 700 theThe spectrum. peak indicated that this solvent did not Two additional factors alteredWavele n fromgth ( thenm printed procedure have the same problem as its predecessors. were temperature and the time of nanoparticle addition. Sol-Gels A peak was finally found in the are very sensitive to temperature and humidity. As time passed, fluorescencep-Dioxane also spectra has intermediate of CdSe-16 polarity when so p- the SC-NP was able to be dispersed while the humidity in the air increased due to seasonal changes (mon- dioxane was used as the solvent (Figure 5). soons), affecting the formation of the sol-gels. To counteract this Thethe peaksolution indicated remained that water this misciblesolvent didgiving not this solvent all the necessary characteristics. effect, the temperature at which the precursor was hydrolyzed have the same problem as its predecessors. was lowered to 10°C. Also, to minimize the effects of precursor p-Dioxane also has intermediate polarity so addition on the nanoparticles, the SC-NP’s were introduced at the the SC-NP was able to be dispersed while beginning of the hydrolysis step rather than in the buffer solution. the solution remained water miscible giving Figure 5: Fluorescence spectrum of CdSe-16.24 By including these small changes in the experimental procedure, this solvent all the necessary characteristics. transparent sol-gels with known inclusion of CdSe nanoparticles Figure 4. Schematicin p-dioxane representation solvent. of hydrolysis and condensation occurring during sol-gel monolith formation. were created (Figure 6). Two additional factors altered from the printed The first procedurevariable optimized were was temperature the pH of the andbuffer the solu - tionFiguretime used. of The 5: nanoparticle polymerizationFluorescence (solidifying)addition. spectrum rate ofSol-Gels ofCdSe-16.24 the sol-gel are in - creasedvery in sensitivedirect proportionin p-dioxaneto temperature with pH. solvent. This was and initially humidity. a problem as theAs sol-gel time would passed, polymerize the quickly, humidity entrapping in theair bubbles air and increased weakening Two additional the matrix. due It to factorswas determined seasonal altered that changesfroma pH of the 6.5 wasprinted(monsoons), ideal because procedure the affectingsol-gel were solidified the temperature in formation a reasonable and ofamount the the of timetimesol-gels. while of allowing nanoparticle To a majority counteract addition.of the air bubbles this Sol-Gels effect, to escape. theare verytemperature The sensitivesecond factor atto to whichbetemperature modified the was precursor theand solvent. humidity. The was de - siredAshydrolyzed characteristics time passed, was were lowered a thesolvent humidity theto nanoparticles10 C. in Also, thewould to air be solubleincreasedminimize in, water the miscible, due effects and to oneof seasonal thatprecursor would not addition changes quench the luminescence(monsoons),on the of nanoparticles, the nanoparticles.affecting the This the luminescence formation SC-NP’s quenching of were the effectsol-gels.introduced was determined To at to be counteract thea problem beginning with this the first effect, solvents of the used, the methanoltemperaturehydrolysis and ethanol. step at These which rather solvents, the than as well precursor in as thetetrahydrofuran buffer was (THF) and dimethylformamide (DMF), acted as electron donors, hydrolyzedsolution. By was including lowered these to 10 smallC. changes Also, to fillingin in the the exciton experimental hole left by procedure, excited electrons transparent in the SC-NP. Thisminimize resulted in thea lack effectsof emission of by precursorthe nanoparticles addition and con- onsol-gels the nanoparticles, with known theinclusion SC-NP’s of CdSe were sequentlynanoparticles no peaks in werethe spectrum. created (Figure 6). Figure 6: a) Fluorescence spectrum of a CdSe introduced A peak was finally at found the in the beginningfluorescence spectra of of CdSe- the hydrolysis step rather than in the buffer loaded sol-gel monolith, b) A picture of a sol-gel 16 when p-dioxane was used as the solvent (Figure 5). The peak containing nanoparticles. indicatedsolution. that this By solvent including did not havethese the smallsame problem changes as its predecessors.in the experimental p-Dioxane also hasprocedure, intermediate transparent polarity so the SC-NPsol-gels was able with to be dispersed known while inclusion the solution ofremained CdSe wa- Figure 6. a) Fluorescence spectrum of a CdSe loaded sol-gel ternanoparticles miscible4 CMDITR giving Review this were solvent of Undergraduatecreated all the necessary (Figure Research characteristics. 6). Vol. 1 No. 1 Summer Figure2004monolith, 6:b) A a) picture Fluorescence of a sol-gel containing spectrum nanoparticles. of a CdSe loaded sol-gel monolith, b) A picture of a sol-gel CMDITR Review of Undergraduatecontaining Research nanoparticles. Vol. 2 No. 1 Summer 2005 109

4 CMDITR Review of Undergraduate Research Vol. 1 No. 1 Summer 2004 OPTIMIZATION OF SEMICONDUCTOR NANOPARTICLE SYNTHESIS AND INTEGRATION INTO SOL-GEL MONOLITHS

CONCLUSIONS The research conducted thus far supports the idea that the goal of a hydrogen generating device is a possibility. The cad- mium selenide semiconducting nanoparticles created had the de- sired optical properties theoretically necessary for the proposed arrangement in quantitative amounts for excellent efficiency. Also by modifying the buffer medium pH, reaction temperature, and solvent used in the accepted sol-gel monolith procedure, the successful formation of a nanoparticle loaded sol-gel matrix pushes the dream of the device closer to becoming a reality.

REFERENCES 1 Jiang, S., L. An, D. Pan and B. Jiang. 2004. Controllable Syn- thesis of Highly Luminescent and Monodisperse CdS Nanocrys- tals by a Two-Phase Approach under Mild Conditions. Advanced Materials. 16:982-985

ACKNOWLEDGEMENTS Clayton Shallcross, Graduate Student, Chemistry, University of Arizona Dr. Neal Armstrong, Professor, Chemistry, University of Arizona Dr. Scott Saavedra, Professor, Chemistry, University of Arizona Zhijie Sui, Post Doctorate, Chemistry, University of Arizona Sam Phimphivong, Graduate Student, Chemistry, University of Arizona Muditha Senarathyapa, Graduate Student, Chemistry, University of Arizona

Funding provided by the Center on Materials and Devices for Information Technology Research (CMDITR), an NSF Sci- ence and Technology Center No. DMR 0120967

Cindy Taylor will complete her studies at the University of Arizona in May 2006 and plans on pursuing a doctoral degree in Analytical Chemistry in the fall of the same year.

110 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Characterization of the Photodecomposition of the CF3-FTC Chromophore

Jillian Thayer Jason Benedict, Dawn Cohen and Dr. Bart Kahr Olympic College University of Washington

INTRODUCTION parallel to the direction of polarization should be destroyed. This will result in a sample having differential absorption for mutu- Demands for improved electro-optic devices support ongoing ally orthogonal input polarization, known as linear dichroism or research in photonics. Non-linear optical (NLO) materials have absorption anisotropy. so far been the focus of meeting such demands, as they show sig- Previous experiments conducted in the Kahr lab with polar- nificant optical effects when an electrical field is applied. Newer ized light however, resulted in isotropic decay of the chromo- NLO materials are organic chromophores, light absorbing mol- phore, and no dichroism. This unexpected behavior implied the ecules which consist of an electron donor, pi system conjugation presence of interactions amongst the chromophores. One hypoth- bridge, and an electron acceptor. Chromophores are often studied esis proposed a transfer of energy from molecules excited by the while suspended in a polymer, which is spin coated and hard- polarized light to molecules out of the plane of polarization, un- ened on a glass slide. They have increasingly complex structures able to be excited by the polarized light directly. Polarized light and proportionally increased susceptibility to decomposition. absorption spectrophotometry was the primary tool in assessing For instance, many bonds can be broken in an organic molecule decay of chromophores, while various forms of spectroscopy with a molecular weight of about 850 atomic mass units. Among were used to examine resulting photodecomposition products. the ten characteristics of ideal NLO materials are concerns over strength, stability, and laser damage thresholds1. These areas of METHODS AND MATERIALS interest prove to be a problem for NLO chromophores, and there developed the idea of a non-linearity/stability tradeoff. For chro- Methods of analysis involved using slides spin coated with a mophores, inherently sensitive to light, photodecomposition is of thin film of varying concentrations of dye-doped polymer. Elec- particular concern. trodes attached to conductors on both sides of the film subject a region of the chromophores to electric field poling and partial reorientation parallel to the plane of the applied electric field. The poled region should then be anisotropic relative to the unpoled region. For analysis of photodecomposition products, previously poled films of AJL-8 (chromophore) were used and subjected to various forms of spectroscopy. In order to look for clues as to the structure of photodecomposition products, clean spectra and convenient use with the thin films were important require- ments. Flakes of polymer doped with chromophore were scraped Figure 1.1. A new NLO chromophore, CF3-FTC from poled slides, separated by poled and unpoled regions, and mixed with KBr to form pellets. The pellet was used with an in- Understanding the mechanisms of photodecomposition is frared (IR) spectrometer, and a spectrum was obtained for both imperative if improvement of chromophores and the lengthening the poled and unpoled regions. Next, poled slides were left intact, of electro-optic device lifetime are expected. Photodecomposition and IR spectra were obtained for both poled and unpoled regions occurs when light energy absorbed by a chromophore is sufficient on an attenuated total reflectance spectrometer. Finally, intact to break molecular bonds, or allow high energy reactions with poled slides were viewed under a Raman microscope coupled to their environment to occur. Controlled experiments are required a Raman spectrometer, and Raman spectra were taken again for to research photodecomposition. In this project linearly polarized both the poled and unpoled regions. light was directed at a sample of chromophores doped into a poly- The next group of experiments also used thin films dye doped mer. In the polymer matrix, molecules are randomly dispersed. polymer, specifically containing the chromophore CF3-FTC, de- Only those molecules whose dipole moment is approximately signed and synthesized by the University of Washington’s Dalton

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 111 CHARACTERIZATION OF THE PHOTODECOMPOSITION OF THE CF3-FTC CHROMOPHORE

Lab. No electric poling was applied to these slides, and concen- trations of 5%, 10%, 15%, and 25% chromophore in amorphous polycarbonate (APC) were made to illustrate the behavior of de- cay as chromophore interactions change directly with concentra- tion. A slide was placed at a precise location on the stage of a po- larizing microscope, and at a precise angle, denoted as 0 degrees. The microscope light source was a tungsten bulb, set at a constant intensity setting. Other light sources considered and tested were the differential interference contrast (DIC) microscope and a xe- non arc lamp. A beam of light was focused onto the thin film of CF3-FTC/APC and a spectrum was taken by the charge coupled Figure 1.3. The changes in absorbance of CF3-FTC in 24 device (CCD) absorbance spectrophotometer. The rotating stage hours is shown at approximately 660 nanometers. was then turned 90 degrees counterclockwise, and a second spec- trum was taken. This set of absorbance measurements at 0 and 90 The timed acquisitions of absorbance by the chromophore degrees were repeated to improve statistical error, and the stage did provide some interesting data. It can be concluded that no was left at the angle of 90 degrees. A comparison of the orthogo- anisotropy is induced by spin coating the films. Measurements nal measurements will assay the degree of dichroism in the photo- of absorption at 0 and 90 degrees before photolysis show statisti- lyzed region of the films. For the next 24 hours, this setup was left cally insignificant differences; no dichroism is observed. As time undisturbed, while a timed acquisition program gathered spectra progressed in the experiments the absorbance at 660 nanometers every hour. At the conclusion of 24 hours, four more spectra were (indicative of CF3-FTC) decreased steadily while the absorbance taken at 0, 90, 0 and 90 degrees. This was repeated for samples in the region of the decomposition products (around 440 nm) in- of all concentrations. A few samples were also tested in 3.5 hour creased. Most trials exhibited predictable exponential decay of the experiments using the polarizing microscope’s condensing lens. chromophore, but many trials showed a poor fit to an exponential growth model for the decomposition products. Measurements of absorption at 0 and 90 degrees after photolysis showed no signs of induced anisotropy. As there should be induced dichroism, the fact that there is none at all is curious and raises questions.

Figure 1.2. Experimental Design

RESULTS AND DISCUSSION Knowledge of decomposition product structure will guide other photodecomposition research. The spectroscopic studies used in this research held potential for being convenient to use Figure 1.4. No anisotropy is shown after 24 hours of polarized photolysis. with the poled or unpoled films of dye doped polymer. In actual- ity, practical limitations kept these three forms from providing useful information. All three sets of spectra yielded from these techniques were indecipherable. The infrared spectrometer pro- duced a very weak spectrum due to both moisture in the KBr pellet and heterogeneous dispersion of the polymer film. The ATR spectrometer also gave very weak spectra because of poor contact between the internal reflectance element and the film of chromophore doped polymer. Lastly, the Raman microscope and spectrometer gave very weak spectra, caused by fluorescence of the chromophore. Raman spectrometers are very susceptible to Figure 1.5. All concentrations of chromophore decay exponentially. interference by fluorescence.

112 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 THAYER

CONCLUSIONS Several conclusions may be drawn from this research. Limi- tations on spectroscopic techniques used with poled films are sig- nificant and these methods need revision before further testing. Specifically, a laser of a different wavelength could minimize fluorescence of chromophores when using a Raman microscope. No statistically significant anisotropy is induced by the process of spin coating, before poling has occurred. Additionally, after pho- tolysis using the setup followed in this research, no significant anisotropy is induced. To pursue the hypothesis of energy transfer among the chromophores, the current experiment of photolysis should be altered. Taking measures to cool the experiment or us- ing lasers as a light source would inhibit sample heating that may be allowing chromophore mobility. Nuclear magnetic resonance (NMR) spectra will be taken at stages of photolysis to look for decomposition products. Experiments of photolysis will also be conducted under vacuum to test the possibility of reaction be- tween the chromophores and oxygen. The underlying principles surrounding the photodecompo- sition of organic chromophores overlap with that of dendrimers and dendronized polymers. Research in this area will therefore be continued, to contribute knowledge to these systems as well.

ACKNOWLEDGMENTS l Thank you to Thank you to members of the Kahr research group; special thanks to mentors Jason Benedict, Dawn Cohen, and Dr. Bart Kahr for their guidance and knowledge. l Thank you to the Dalton lab, for providing supplies of CF3- FTC and APC; special thanks to Phil Sullivan for spin coating. l Thank you to the University of Washington Department of Chemistry, the National Science Foundation, and the Materials and Devices for Information Technology Research-Science and Technology Center. l Special thanks to the organizers and supporters of the 2005 Hooked on Photonics Undergraduate Research Program, specifi- cally Kristin Wustholz, Dr. Sara Selfe, and Dr. Phil Reid

REFERENCE 1Ewy, T. R. Design and Synthesis of High Performance Nonlinear Optical Chromophores. University of Washington, Seattle, WA, 1996.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 113 114 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Electroluminescent Properties of Organic Light-Emitting Diodes (OLEDs) with 2,5-Bis(9,9-dimethylfluoren-2-yl)-1,1,3,4-tetraphenylsilole

Evans Thompson Benoit Domercq and Bernard Kippelen Georgia Institute of Technology School of Electrical and Computer Engineering Georgia Institute of Technology

A new organic compound, 2,5-Bis(9,9-dimethylfluoren-2-yl)-1,1,3,4-tetraphenylsilole (DTS), derived from the silole family has been electroluminescently characterized. Func- tional organic light-emitting diode (OLED) devices were fabricated using DTS as electron- transport layer (ETL), hole-transport layer (HTL), emission layer (EML), and single layer (SIL). DTS was found to be a bright green emitter with a peak wavelength of 535 nm and a maximum half life of 4000 s.

Background of OLEDs in high resolution displays OLEDs, due to its 100% photoluminescent efficiency [3]. Siloles

The demand for high resolution large area flat panel displays also have a low electron affinity (EA) comparable to that of Alq3, is large. Liquid Crystal Displays (LCDs), the leading flat panel a widely used electron transport material. DTS, in particular, has display technology, hold that title because of their long life of an Ionization potential comparable to that of TPD, an organic operation, light weight, low operation costs and brightness. How- compound used for hole-transport. Figure 1 shows the chemical ever, LCDs are not without their disadvantages. LCDs require a structure of DTS. backlight, causing low contrast ratios due to an inability to pro- duce a true black image. Additionally, LCDs are expensive, es- pecially for displays exceeding 30 inches, and have low viewing angles. A new display technology using OLEDs, has the potential to supplant LCDs as the leading display technology. OLEDs convert electricity directly into light. By creating their own light these de- vices will require no backlighting, reducing the operating power requirements. They will be very efficient and have low operat- ing costs. The organic materials are inexpensive and the actual fabrication of large displays will be cheaper than large LCDs. In Figure 1. Chemical structure of 2,5-Bis(9,9-dimethylfluoren- addition, flexible displays can be made with OLEDs since they do 2-yl)-1,1,3,4-tetraphenylsilole (DTS). not require a rigid substrate. Currently OLED displays still have a few hurdles to over- Device Structure come before they make their way to the living room. The efficien- The devices consist of multiple layers of organic and inor- cy and lifetime of the organic compounds are still too low, and ganic compounds. The layers are deposited on an Indium Tin fabrication is still too expensive. Researchers have been working Oxide (ITO) coated glass slide using high vacuum physical to find new compounds with high efficiencies and longer life- vapor deposition (PVD). PVD is a deposition technique where times as well as new OLED fabrication techniques. the material is vaporized by heating to condense on a substrate Background of siloles Siloles, or silacyclopentadienes, are forming a uniform thin film. Si-containing five-membered cyclic dienes with conjugate rings The glass slides went through 4 stages of ultra sonic having different structures attached.[1,2] The highest external cleaning in which they were submerged in soap water, deion- quantum efficiency (EQE) for a singlet emitting OLED is be- ized water, acetone and ethanol. The slides were then baked tween 1.1 and 1.5% for a TPD-Alq OLED device. While for a 3 in a vacuum oven to remove any left over ethanol. Silicon ox- silole the maximum EQE is roughly three times that of TPD-Alq 3 ide was then deposited on the slides and again ultrasonically

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 115 ELECTROLUMINESCENT PROPERTIES OF ORGANIC LIGHT-EMITTING DIODES (OLEDS) WITH 2,5-BIS(9,9-DIMETHYLFLUOREN-2-YL)-1,1,3,4-TETRAPHENYLSILOLE cleaned with acetone and ethanol. The silicon oxide slides were then placed in a microwave vacuum chamber where they were oxygen-plasma treated. The silicon oxide served as an insulation layer to limit area of emission of each OLED de- vice. The order of deposition on the silicon oxide slides was organic layers, lithium fluoride (LiF), and aluminum. The or- ganic and inorganic layers and their thicknesses are listed as follows: TPD (40 nm), Alq3 (40 nm), LiF (1 nm), and Al (300 nm). The thickness of DTS varied. The LiF was used to help the injection of electrons by the aluminum cathode. Four com- binations of deposition of the organic layers were made cor- Figure 2. Normalized EL spectra of ETL, HTL, EML, and SIL devices. responding to the four different types of devices: ETL, HTL, EML and single layer (SIL). ETL, HTL EML devices had a DTS thickness of 40 nm. SIL device had a DTS thickness of 120 nm. Tests An Ocean Optics fiber spectrometer was used to measure the electroluminescent (EL) spectra of each device [4]. Figure 2 shows the normalized EL spectra of each type of device. To measure the EL characteristics of the devices each device was placed in an electrical contact box fixated direct- ly in front of a photodetector. The electrical contact box and photodetector where then covered with a black box to block outside light. The applied voltage was varied using a Keithley 2400 sourcemeter. The current through each device and the Figure 3. Averaged luminance and EQE of ETL and light emitted by the device was measured. All tests were au- HTL devices as a function of applied voltage. tomated using Labview software. The luminance and the ex- ternal quantum efficiency (EQE) was then calculated by nor- malizing the measurement of the photodiode to the photopic response of the human eye and the sensitivity response of the photodiode. Figures 3 through 6 show the mean luminescence, EQE, and current density curves for each device structure as a function of applied voltage [5]. The luminescent lifetime of each device was measured at a constant current density of 50 mA/cm2. The voltage needed to keep that current was automatically adjusted by the source meter. Figures 7 and 8 show the lifetime measurements of the HTL, EML, and SIL devices. Figure 4. Averaged current density of ETL and HTL RESULTS devices as a function of applied voltage. All devices emit bright green light. The ETL, HTL, and EML Device Vappl. Luminance (±Er) EQE(±Er) Current Density(±Er) devices emit light at a wavelength of 535 nm, while the SIL de- (V) (cd/m2) (%) (mA/cm2) vice emits light at a wavelength of 540 nm. ETL 8.0 1300(±100) 0.3(±0.2) 115(±5) Table 1 shows the applied voltage, luminance, EQE, and current density and their respective errors at the maximum EQE. HTL 10.8 850(±150) 0.3(±0.1) 100(±50) For a constant current density of 50 mA/cm2 the HTL, EML, EML 9.4 1550(±150) 1.0(±0.1) 45(±5) and SIL had half lives of 840, 1100, and 4000 s, respectively. SIL 13.1 800(±200) 0.3(±0.1) 85(±5) Lifetime measurements of the ETL devices could not be obtained Table 1. Maximum EQE results. due to a significant amount of leakage as can be seen in figure 4.

116 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 THOMPSON

CONCLUSION Ext. Quantum Efficiency (%) The ETL devices do not appear to operate significantly bet- ter than the HTL devices. The ETL devices do have a lower turn on voltage but the current densities are very similar. The nearly horizontal shape of the current density plot shows there is current leakage. DTS appears to be ambipolar meaning it can be used to transport both holes and electrons.

ACKNOWLEDGEMENTS This work was supported by the MDITR, a Science and

Figure 5. Averaged luminance and EQE of EML and Technology Center of the National Science Foundation (NSF) SIL devices as a function of applied voltage. under Agreement Number DMR-0120967.

REFERENCES [] Chen, H. Y.; Lam, W. Y.; Luo, J. D.; Tang, B. Z.; Zhu, D. B.; Wong, A.; Kwok, H. S. App. Phys. Let. 2002, 81, 574-576. [2] Lee, J.; Liu, Q.; Bai, D.; Kang, Y.; T, Y.; W, S. Organometal- lics. 2004, 23, 6205-6208. [3] Murata, H.; Kafafi, Z. H.; Uchida, M. App. Phys. Let. 2002, 80, 189-191. [4] Dini, D. Chem. Mater. 2005, 17, 1933-1944. [5] Forrest, S. R.; Bradley, D. D. C.; Thompson, M. E. Adv. Mater.

Figure 6. Averaged current density of EML and SIL 2003, 13, 1043-1048. devices as a function of applied voltage.

Evans Thompson is currently studying mechanical engineering at Georgia Institute of Technology. Evans intends to obtain a master’s degree in mechanical engineering upon graduating from Georgia Tech.

Figure 7. Luminance of HTL and EML devices as a function of time.

Figure 8. Luminance of SIL device as a function of time.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 117 118 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Determination of Molecular Orientation of Self-Organized Aggregates of New Liquid-Crystal Perylene Dye

Natalie Thompson Neal R. Armstrong, P. Alex Veneman, Georgia Institute of Technology Armstrong Lab, Department of Chemistry The University of Arizona

INTRODUCTION that is high enough that the reaction with oxygen in the elec- tron transport is not significant, as compared with hole transport Organic compounds with π-conjugated systems which often materials, indicating an ability to use organic electronics in air. exhibit high charge mobilities have a growing interest within the The perylene-dianhydride-bisimide was made into thin films us- scientific community because of the possibility of creating organ- ing Langmuir-Blodgett methods and was then characterized by ic films that can act as conducting materials. These compounds atomic force microscopy (AFM), UV-Vis transmission, Fourier are expected to be used to create cheaper electronics through all- Tranform Infrared (FTIR), reflective-absorptive infrared spec- plastic circuits (1). Also, these materials could be used in other troscopy (RAIRS), and X-ray reflectivity (XRR). organic electronic applications, such as organic light emitting diodes, organic photovoltaics, and organic field effect transis- tors. Liquid-crystalline organic materials are of interest due to their ability to self-aggregate into columns. It is thought that the charge mobilities of the aggregates are related to the cofacial overlap of the conjugated system. Brédas, et al. reported that the largest splitting between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) correlates with the cofacial overlap and then influences the hole (electron) transport of the aggregate. (1) Donley, et al. reported that cofacial overlap results in low activation energies for charge migration. However, this migration, or hopping, prefers to move along a certain axis of the aggregate. (2) Flora, et al. reported a much higher charge mobility along the axis that is parallel to Scheme 1 the columnar axis, as opposed to perpendicular to it. (3) It has been reported that discotic mesophase materials, such as phthalo- EXPERIMENTAL cyanines, triphenylenes, and hexabenzocoronenes, organize into columnar aggregates within films which are then stiff enough to Langmuir-Blodgett film preparation transfer via horizontal transfer on a Langmuir-Blodgett trough. The substrates for Langmuir-Blodgett thin films were sili- (4) Often these π-conjugated cores have surrounding side chains con wafers, highly ordered pyrolytic graphite (HOPG), glass, and which are used to help control the self-organization process. It commercial gold. The silicon wafers and glass slides were modi- is also thought that the side chains may help prevent electrical fied via sonication for thirty minutes in a solution of 1,1,1,3,3,3- cross-talk between columnar aggregates due to their insulating hexamethyldisilazane (HMDS) and 1,3-diphenyl-1,1,3,3-tetra- nature. (5) methyldisilazane (DPTMDS) in chloroform (5 : 5 : 90 :: HMDS

: DPTMDS : CHCl3). After sonication, the substrates remained OBJECTIVES / THESIS in solution until LB deposition. The preparation of the HOPG involved cleaving the substrate to obtain a flat clean surface. The This summer’s research focused on the characterization of a gold surface was rendered hydrophobic by sitting in a solution of relatively new perylene-dianhydride-bisimide. Scheme 1 benzyloxyethanethiol in ethanol overnight. The Derivatives of perylene-dianhydride-bisimide are of in- Monolayer and multilayer films of N,N’-1,2,3-tridodecoxy- terest because of their high electron mobility, which would aid phenyl-3,4,9,10-perylenetetracarboxylicdiimide (PTCDI) were in the efficient transport of electrons in a thin film comprised of prepared on a Nima model 611D LB trough using chloroform as this molecule. Also, these derivatives have an electron affinity

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 119 DETERMINATION OF MOLECULAR ORIENTATION OF SELF-ORGANIZED AGGREGATES OF NEW LIQUID-CRYSTAL PERYLENE DYE the spreading solvent. The PTCDI solution (1 mM, 250 μL) was phile with distinct isolated polar (hydrophilic) and apolar (hydro- spread onto the surface of a deionized water subphase at room phobic) regions. This behavior was expected due to the small temperature and pressure. The solvent was given about 45 min- aromatic region in the side chains. Also, the compression of the utes to evaporate completely. Once, evaporated, any aggregates LB thin films showed that PTCDI forms stiff films which can be that formed in the spreading of the compound were removed from deposited via horizontal transfer. The horizontal transfer could the water surface via suction. Compression was performed at a prove to be useful in easing the process for the creation of organic rate of 45 mm/s until a condensed phase monolayer or multilayer electronic devices. was formed. This phase was observed by the presence of one or multiple shoulders exhibited on a pressure-area (π-A) isotherm. The LB thin films were deposited onto the modified surfaces via horizontal deposition in which the surface of the substrate is parallel to the surface of the trough. After each deposition, the sample was dried with compressed N2. Annealed films were held at 120°C for 4 hours and then cooled to room temperature over 6 hours. The samples were then stored in a sealed container in air until analysis.

Spectroscopic Measurements Figure 1. The π-A isotherm exhibits the formation of a trilayer. This was determined by observation of the shoulders in the isotherm. This isotherm UV-Vis spectra were obtained at room temperature and pres- showed that PTCDI does not have linear regions in between the formation sure. The sample was such that the visible columns were vertical of the layers. This is indicative that PTCDI is not a true amphiphile. with respect to the sample holder. The sample was placed either perpendicular or at 45° with respect to the incident beam. A po- Analysis of the AFM images proved that the horizontal trans- larizer was used to obtain TM and TE polarizations. fer was successful and that PTCDI does self-organize. AFM also Transmission IR was obtained with the sample perpendicular displayed the nature of the aggregates. With a sole deposition, to the IR beam. A polarizer was used to obtain TM and TE. For PTCDI forms islands of aggregates on the surface of the silicon RAIRS, the sample was placed such that the columns were run- (Figure 2). This is expected to be a result of the dewetting pro- ning along the length of the sampling area. cess in the transfer. Further studies showed that multiple deposi- X-ray reflectivity (XRR) was performed on a sample that tions results in a more solid and organized thin film (Figure 3). was composed of five bilayers deposited on modified silicon. AFM analysis of annealed films showed that the annealing process raises the sample temperature to a point where the sample behaves Atomic Force Microscopy more like a liquid and is able to reorganize the aggregates. It also The topography of LB thin films on silicon wafers and HOPG showed that the self-organization of the liquid-like sample has a was examined using atomic force microscopy (AFM) performed tendency to aggregate into a linear fashion, the like of which can in tapping mode, ex situ. The minimum force required to acquire extend for at least 2 μm (Figure 4). The presence of linear aggre- optimal resolution was used in order to minimize distortion of gates hints at the idea that PTCDI forms columns which can extend the sample. The driving amplitude applied was determined by significant distances. These columns with cofacial overlap of the the instrument and in between 100 and 400 mV. Once the tip π-conjugated systems can be used in organic electronic devices. was engaged, imaging began. The setpoint value was increased In particular, PTCDI could be used as an electron transfer material until sinusoidal waves were observed indicating no contact with due to its high electron affinity and mobility. the surface. It was then decreased until contact with the surface Analysis of the UV-Vis data showed little evidence of dichro- was optimized, as monitored by the trace-minus-retrace (TMR) ism in the plane that was analyzed (Figure 5). Though no organiza- values. Optimal TMR was considered to be under 0.2 nm. tion of the molecular transition dipole moment is evident, the data does not lead to the conclusion that there is no organization. Other RESULTS AND DISCUSSION possibilities include: that there is organization in other planes; that there are islands that are coherent in themselves, but not with other Analysis of the π-A isotherm showed that PTCDI forms co- islands; or that the organizational structure does not have aligned herent monolayer and multilayer thin films, though the steps were molecular transition dipole moments, for example if PTCDI forms not as distinct as those exhibited by phthalocyanine derivatives a helical column where each molecule is slightly tilted with respect (Figure 1). This behavior exhibited on the π-A isotherm leads to the surrounding molecules (Figure 6). to the conclusion that PTCDI does not behave as a true amphi-

120 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 THOMPSON

Figure 2. AFM image of one deposited bilayer, unannealed. The sample is NJT-I-53A

Figure 3. AFM image of 2 bilayers, unannealed. The sample is NJT-I-52B Figure 5. UV-Vis spectra of 8 deposited bilayers, unannealed.

The transmission IR data and the reflective absorptive IR (RAIRS) data came to no precise conclusions. Though some dichroism is evident between the spectra, there is not enough information to determine the exact vector orientation of PTCDI (Figure 7). Wavenumber (cm-1)

Figure 4. AFM image of 2 bilayers, annealed. The sample is NJT-I-52B

Figure 7. FTIR and RAIRS data. The red is RAIRS. The helical out of plane blue is transmission IR parallel to the columns. The purple is transmission IR perpendicular to the columns. Figure 6. Possible arrangements for PTCDI that would not show dichroism.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 121 DETERMINATION OF MOLECULAR ORIENTATION OF SELF-ORGANIZED AGGREGATES OF NEW LIQUID-CRYSTAL PERYLENE DYE

The X-ray reflectivity (XRR) data came to no conclusions. It Funding provided by the Center on Materials and Devices was decided that the concentration of the sample was too dilute to for Information Technology Research (CMDITR), an NSF Sci- obtain accurate information on the thickness, density, and rough- ence and Technology Center No. DMR 0120967 ness of the thin film (Figure 8).

Natalie Thompson is finishing her senior year at Georgia Tech and plans to graduate in Spring 2006. She would like to thank Dr. Neal Armstrong, Alex Veneman, and Nira Kumaran for their support and friendship.

Angle (2θ)

Figure 8. XRR graph

CONCLUSIONS LB formation of thin films of PTCDI results in a stiff film that can be transferred to a silanized surface using horizontal trans- fer. These thin films form islands with only one deposition, but multiple depositions results in linearity of the thin film. Also, annealing of the sample allows the molecules to reorganize into linear aggregates. Spectroscopic measurements led to no solid conclusions about the orientation of the molecules within the thin film, but hinted that the molecules might be organized outside of the sample plane (parallel to the substrate).

REFERENCES 1 Brédas, J. L., Calbert, J. P. da Silva Filho, D. A., Cornil, J. Pro- ceedings of the National Academy of Sciences. 2002, 99, 5804. 2 Donley, Carrie L., Xia, Wei, Minch, Britt A., Zangmeister, Re- becca A. P., Drager, Anthony S., Nebesny, Ken, O’Brien, David F., Armstrong, Neal R. Langmuir. 2003, 19, 6512. 3 Flora, Ware, H., Mendes, Sergio B., Doherty, III, Walter J., Saa- vedra, S. Scott, Armstrong, Neal R. Langmuir. 2005, 21, 360. 4 Xia, Wei, Minch, Britt A., Carducci, Michael D., Armstrong, Neal R. Langmuir. 2004, 20, 7998. 5 Donley, Carrie L., et al. J. Mater. Res. 2004, 19, 2087 6 Doherty, III, Walter J., Simmonds, Adam G., Mendes, Sergio B., Armstrong, Neal R., Saavedra, S. Scott. In press.

ACKNOWLEDGEMENTS -Dr. Neal R. Armstrong, Professor of Chemistry and Optical Sciences, University of Arizona -Armstrong Research Group

122 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 Hydrogel Materials for Two-Photon Microfabrication

Mayen Udoetuk Dr. Joe Perry and Vincent Chen Norfolk State University Georgia Institute of Technology

OBJECTIVE substrate, the surface of the substrate was modified with an adhe- sion promoter (95% methanol and 5% 3-(trimethoxysilyl)propyl The purpose of the research is to develop a hydrogel material methacrylate). The adhesion promoter solution was spin-coated system for two-photon microfabrication. Problems that were ad- onto microscope slides and cover slips, and baked on a hotplate dressed in this research project are: at 90°C for 2 minutes. 1) For what composition can sufficient crosslinking be Five drops of the hydrophilic resin were placed on surface achieved by ultra-violet (UV) exposure such that the patterned treated microscope slides and cover slips, and were exposed to structure will not wash away upon development? UV light from a mercury UV lamp (254nm) for various times. 2) Can hydrophilic monomers be crosslinked using current Solubility tests were done in order to find a solvent that would two-photon absorbing dyes? wash away any resin left after exposure, while leaving the po- 3) What are the requirements of the material and the struc- lymerized structure intact. All of the following were found to ture to ensure that pores and other desired features within the perform well as development solvents: dichloromethane, tolu- structure are maintained after swelling? ene, ethanol, dioxane, cyclohexane, dimethyl formamide, and 4- Potential practical applications of these materials are, for ex- methyl-2-pentanone. 4-Methyl-2-pentanone was chosen for use ample, microstructures to be used for cell adhesion, which can in the experiments. find applications in tissue engineering, and structures that respond Several sample geometries were tested but some gave an in- to the solvent environment for sensors or microfluidics. Most adequate amount of crosslinking. The initial setup included the research in microfabrication to date has been carried out with ac- resin on top of a microscope slide modified with adhesion pro- rylate and epoxy type polymers.1 However, in applications that moter, with the UV light shining down on it. Due to the thickness involve biological molecules or require the adhesion of cells onto of the resin, the UV light did not penetrate to the resin-substrate polymeric structures, biocompatible materials need to be used. interface in order to yield sufficient crosslinking. Two geometries Hydrogel systems have hydrophilic surfaces and are of particular that were found to provide sufficient exposure and crosslinking at interest due to their ability to swell and de-swell, depending on the interface are shown in diagram 1.1. their environment. Experiments were conducted to determine the swelling prop- erties of the crosslinked materials. The dimensions of the lines RESEARCH METHODS were measured before and after swelling with water at room tem- The research problems were approached by testing resins based perature. The feature dimensions were measured using optical on hydrophilic monomers and two-photon absorbing dyes or UV ini- microscopy. tiators. Photolithography is a process of transferring spatial patterns Bulk materials of the hydrogels were prepared using setup 1 onto a surface of a material by exposure to light with a mask. Steps in order to better visualize the swelling. Fluorescence and trans- usually involved in this process include substrate cleaning, adhesion mission imaging were used to characterize and document the layer formation, photoresist application, soft baking, mask align- fabricated structures. A laser was used for 3D microfabrication ment, exposure, development, and hard-baking. of structures3. A test structure consisting of a vertical stack of The hydrophilic monomers and crosslinker that were used crossing lines was designed and fabricated using two-photon mi- include N-vinylpyrrolidone (VP), hydroxyethylmethacrylate crofabrication. (HEMA), and ethyleneglycol-bis-methacrylate (EGMA). Resins containing various amounts of these compounds were prepared. The initiator used for UV exposure was 1-[4-(methylthio)phenyl]- 2-methyl-2-morpholinopropan-1-one, (UV initiator #7). The amounts of initiator used were 0.1, 0.2, and 0.5 wt%, see 1-[4-(methylthiophenyl]-2-methyl-2-morpholinopropane-1-one (UV #7) was varied in the compositions. 0.1, 0.2, and 0.5 wt% were used. table 1.1). To ensure that the structures remained attached to the

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 123 HYDROGEL MATERIALS FOR TWO-PHOTON MICROFABRICATION

Solution Ethylene glycol dimethacrylate 1-vinyl-2-pyrrolidinone 2-Hydroxyethyl methacrylate

A 1 1 1

B 2 1 1

C 1 2 1

D 1 1 2

Table 1.1. Hydrogel Compositions

RESULTS A. Table 1.2 contains data on swelling ratios in the bulk material It was found that setup 1 allowed the UV light to penetrate after exposure to UV light at various times. The diameter and the the resin-substrate interface in order to yield sufficient crosslink- thickness of the samples were measured before and after swell- ing. Sample A with a 1:1:1 ratio of EGMA, VP, and HEMA, and ing in water overnight. Setup 1 was used with composition C, C with a 1:2:1 ratio of EGMA, VP, and HEMA gave the most with exposure times of 5, 10, 15, 20, and 25 minutes. The results crosslinking in the least amount of time. Composition C cross- indicate that the bulk materials of composition C swell 4-5 times linked in 3 minutes, whereas composition A required 7 minutes, its original size. Most of the swelling was observed to be in the and compositions B and D could not be crosslinked. This was vertical direction. Setup 2, which includes a mask, was used with determined by observing how much of the sample became solid composition C for 3 minutes. Picture 1.1 shows an example of after UV exposure. Only the part of the sample behind the open- structures obtained by photo-patterning of composition C through ing in the mask should crosslink. Samples that did not yield suf- a mask. The width of the features was measured before and after ficient crosslinking remained in the liquid state after UV expo- swelling in water overnight. No change in width was observed sure. Composition C was also used in the UV mask experiments (see Table1.3). due to its ability to yield higher crosslinking than composition

Setup 1

Setup 2

Diagram 1.1. Experimental Setups for UV Experiments

124 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 UDOETUK

Exposure Time Diameter (dry) Thickness (dry) Diameter (wet) Thickness (wet) Volume Ratios Thickness Ratios (minutes) (inches) (inches) (inches) (inches) (wet/dry) (wet/dry)

5 0.350 0.030 0.358 0.136 3.7 3.5

10 0.340 0.026 0.360 0.138 5.0 4.3

15 0.349 0.032 0.365 0.135 3.6 3.2

20 0.344 0.027 0.361 0.135 4.5 4.0

25 0.343 0.028 0.365 0.131 4.3 3.7

Table 1.2. Results for bulk materials

Picture 1.1 Optical images of structure obtained by UV illumination of resin C through a mask

Condition Width (1) Width (2) Width (3) Average Dry 0.34 0.35 0.36 0.35 Wet 0.36 0.34 0.33 0.34 %Change 5.8% -2.9% -8.3%

Table 1.3. Results from UV Mask (widths are in arbitrary units). Widths (1), (2), and (3) refer to the lines on the wet and dry images. They represent the width before and after swelling the samples with water.

Picture 1.2. (left) SEM Image of Stack of Log Microstructure Obtained by two-photon microfabrication in resin C. (right) Dye #41.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 125 HYDROGEL MATERIALS FOR TWO-PHOTON MICROFABRICATION

The SEM image was obtained by using two-photon micro- ACKNOWLEDGEMENTS fabrication. The following parameters were used: laser scan Dept. of Chemistry and Biochemistry at the Georgia Institute of speed: 20 μm/s; 60 mW; λ= 730nm; 1 %wt of dye #41, composi- Technology tion C. A porous 3D grid (stack of logs) was intended. Dye #41 MDITR REU was used because it has large photon cross sections, and has been National Science Foundation observed to polymerize acrylate systems. The fabricated struc- Dr. Joe Perry ture appears highly deformed, which could be due to inadequate Dr. Mariacristina Rumi crosslinking or laser induced deformations. Vincent Chen Wojtek Haske CONCLUSION Kelly Perry Sufficient crosslinking can be achieved by UV exposure. Perry Group It was found that setup 1 was the most effective in the UV ex- Dr. Keith Oden posure experiments. From the different compositions that were Ms. Olanda Bryant tested, composition C (1:2:1 ratio of EGMA:VP:HEMA) gave Beverly Scheerer the most crosslinking per exposure. In later experiments using the UV mask, composition C with a 1wt% of UV #7 was used. No change in width was observed. This may have been due to a large amount of crosslinking within the structure. At high cross- linking, the swelling of a structure in the presence of water may be limited. In the bulk specimen, the thickness of the samples showed a large change relative to the diameter. The small change may result in the lack of the sample’s ability to expand, as a large area adheres strongly to the substrate. This may also explain the For me, understanding the physical forces in life mean understanding lack of change in width for the features produced lithographi- science. Norfolk State University and Georgia Institute of Technology have afforded me many opportunities to do so. cally. It will be useful explore a method that allows measurement of the height of the fabricated features. This work shows that hydrophilic monomers can be crosslinked using UV initiators and current two-photon absorbing dyes (#41). However, other dyes should be investigated for higher efficiency. Future work will involve investigating more efficient two-photon absorbing dyes, find other biocompatible materials, write functional microstruc- tures, and characterize the swelling ratios in more detail.

REFERENCES (1) Zhou, W.; Kuebler, S.M.; Braun, K.L.; Yu, T.; Cammack, J.K.; Ober, C.K.; Perry, J.W.; Marder, S.R. Science, 2002, 296, 1106-1109 (2) Watanabe, T; Akiyama, M.; Totani, K.; Kuebler, S.M.; Stel- lacci, F.; Wenseleers, W.; Braun, K.; Marder, S.R.; Perry, J.W. Adv. Funct. Mater., 2002, 12, 611-614 (3) Cumpston, B; Sundaravel, A.P; Barlow, S; Dyer, D.; Eh- rlich, J.; Erskine, L.L; Heikal, A.; Kuebler, S.; Lee, S.; McCord- Maughon, D.; Qin, J.; Röckel, H.; Rumi, M.; Wu, X.; Marder, S.; Perry, J. Nature, 1999, 398, 51-54

126 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 The Design of a Fluid Delivery System for Micro-Core Optical Fiber

Greg Winchell Ann Mescher Everett Community College Mechanical Engineering, University of Washington

INTRODUCTION TO POLYMER FIBER DRAW The basic approach to making polymer fiber is to first trim and machine the polymer preform (which is an acrylic rod) to The objective of this research has been to develop a method specifications inherent to the fiber making procedure. Then the for drawing a single-hole polymer optical fiber from a pre-drilled preform is attached via cross pin to the draw mechanism and is acrylic preform, with accurate control ( +/- 2 microns) of the fi- slowly fed downward into the furnace as shown in Figure 1. The ber outer diameter and internal channel diameter. Polymer fiber furnace heats the preform (with wall temperatures peaking at ap- offers superior design capabilities compared to currently manu- proximately 184°C) such that polymer fiber can be drawn me- factured glass fiber, along with potential savings in energy and chanically by the Draw Tower spindle and spooled. While the manufacturing cost. The most significant advantages of polymer fiber is drawn a Laser Diameter Gauge measures the outer diam- fiber are: 1) the ability to incorporate unique organic optical ma- eter of the fiber produced by the process. This is the basic process terials, and 2) much greater flexibility in designing and process- for producing polymer fiber. ing polymer materials as opposed to glass into photonic bandgap If we take a polymer preform and drill a hole in the radial structures. The next paragraph will discuss how polymer fiber is center along the axis of the preform (thus forming a “thick-walled made. tube”), and we then heat this preform, the resultant fiber will have the same aspect ratio (i.e. the inner hole diameter divided by the outer diameter of the fiber) as the original drilled preform before heating. Now if we take another polymer preform and drill a hole in the radial center and this time place fluid in the hole, the aspect ratio of the resultant fiber will not be the same as the preform aspect ratio before heating. Why is this happening?

Fluid Core and the Effect Of Fluid Pressure When the preform with a hole (but no fluid inside the hole) is heated, the experimental evidence consistently shows that the aspect ratio of the fiber produced does not differ from that of the unheated preform. This suggests that mere atmospheric pressure inside the drilled hole is not enough to alter the aspect ratio. However when fluid is added to the drilled hole, this is enough to change the aspect ratio of the fiber indicating that atmospheric pressure in combination with fluid pressure will alter the aspect ratio of the resultant fiber compared to the as- pect ratio of the unheated preform. Fluid Pressure is equal to the density of the fluid multiplied by gravitational acceleration multiplied by height of the fluid column (Fluid Pressure=ρgh). The illustration below (Figure 2) shows that the fluid pressure is highest in the region where the polymer radius is rapidly chang- ing. As the fiber is created below the fluid column, the aspect ratio increases and therefore fluid is gradually removed from the preform. This results in a continuous drop in the fluid column Figure 1. The Polymer Fiber Draw Process height which means that the fluid pressure itself will also begin to decline.

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 127 THE DESIGN OF A FLUID DELIVERY SYSTEM FOR MICRO-CORE OPTICAL FIBER

Since there is a decline in the fluid pressure, the pressure exerted against the inner diameter hole decreases. The decrease in pressure will reduce the aspect ratio over time unless the fluid is replaced as needed. This engenders the need for a reliable fluid delivery system to control the aspect ratio as needed.

Figure 3. The Gravity Feed Apparatus

The Gravity Feed Apparatus was essentially a siphon ar- rangement composed of an adjustable lab jack, a plastic flow tube, a 2000ml beaker, a 1ml capacity graduated cylinder, a top- mounted fluid feed ramp, and a graduated cylinder stand. This device demonstrated excellent flow control. Under laboratory conditions the single most important factor affecting the flow rate of the fluid with this device (i.e. the volumetric flow rate) was the difference in the height of the collection container (Z2) minus the height of the reservoir (Z1) otherwise known as ΔZ (ΔZ= Z2-Z1). However to properly predict volumetric flow rates from this de- Figure 2. Fluid Pressure in Preform due to Fluid Column vice requires math models from Fluid Mechanics [3], combined with an understanding of the components of fluid flow. Thus we Fluid Delivery Apparatus need to understand and utilize the following equations: The fluid delivery system considered first was a type of au- tomated syringe pump known as the NE-1000 Programmable Sy- V = [D2 g / (32 ν L)] * (h + ∆Z) equation 1 ringe Pump retailed by the New Era Pump Systems, Inc. The NE- 1000 Programmable Syringe Pump retails for $995 US Dollars and Q = V π D2 / 4 = [π D4 g / (128 ν L)] * (h + ΔZ) has the following features [1]: equation 2 1. The infusion rates are from 0.73 μl/hr (1 cc syringe) to 2100 ml/hr (60 cc syringe). Equation 1 refers to the average velocity (V) in the flow tube. 2. Stand-alone operation or computer-controlled operation are Average velocity is dependent upon the following parameters: available. 3. Infusion and withdrawal of fluid is performed as needed. 1. D is the internal diameter of the flow tube. 4. The operator can program up to 41 pumping phases that change 2. g is the gravitational acceleration. pumping rates, set dispensing volumes, insert pauses, control 3. ν is the kinematic viscosity of the fluid. and respond to external signals. 4. L is the length of the flow tube. 5. Motor stall detection is included. 5. h is the height of the fluid column in the reservoir. 6. Dispensing accuracy of the device is +/- 1% 6. ΔZ is the difference between Z2-Z1 (see Figure 3) 7. Unlimited lifetime technical support is available. 8. It has a two-year warranty. Now that we have noted the important parameters for the The second fluid delivery system that was considered was a average velocity V, we can use equation 2 to find the volumetric gravity feed design that was developed and tested in the Polymer flow rate Q = V π D2 / 4, where π D2 / 4 is the internal cross-sec- Optics and Processing Laboratory at the University of Washington. tional area of the flow tube. Below is the drawing of the Gravity Feed Apparatus (Figure 3):

128 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 WINCHELL

RESULTS The origin of the inner hole has been measured and is listed at “0” meters on the chart with an initial aspect ratio of about 0.1. Flow Rate vs. ∆Z Other than a few aberrations in the aspect ratio between approxi- All of the variables to find the volumetric flow rate for the pur- mately 30 meters and 120 meters the aspect ratio shows a gradual poses of the lab research were kept constant except for ΔZ. There decline, which is very apparent after 120 meters. How do we ex- is a visual representation of data in Figure 4 that illustrates the di- plain the aberrations? First, there is a sharp rise in the aspect rect proportionality between ΔZ and the volumetric flow rate Q. ratio between 0 and approximately 10 meters: this is the effect of Both the math model and the experimental results of volumetric the tip of the drill bit. Second, in the range of 10 to 120 meters flow rate Q vs. ΔZ show linearity. there are several sharp spikes in the aspect ratio. These are most It is important to note that in some cases, there were several likely caused by trapped air bubbles and/or preform drill shavings different values of Q for the same value of ΔZ. This is because the that were not removed before the fluid was added, thus creating experiments were carried out on more than one day when the lab multiple “plugs” which in turn created abnormal spikes in the as- air temperature (and thus the fluid temperature during fluid flow) pect ratio. It is clear that along the length of approximately 280 was different for each experiment. This resulted in different kine- meters of drawn fiber, the aspect ratio generally decreases from a matic viscosities of the fluid from one experiment to another. The value over 0.25 to a value below 0.20. different viscosity then results in a different volumetric flow rate Δ for a given Z, if all other factors are held essentially constant. FUTURE RESEARCH There are several items of future research that are relevant to this investigation. First, it would be useful to develop a real- time optical scanner for determining the inner diameter of fiber as it is produced. Dr. Kaminsky of the University of Washington Chemistry Department will be assisting personnel in the Poly- mer Optics and Processing Laboratory to develop such a system. Second, it will be imperative to develop a method to eliminate trapped air bubbles and/or debris in the fluid as it is infused into the polymer preform during fiber production. Third, the method

Figure 4. Flow Rate vs. ΔZ by which the Gravity Feed Apparatus is connected to the polymer fiber production system will pose challenges that need to be ad- The Aspect Ratio vs. Distance. dressed. Fourth, the means by which chromophores align in the inner channel of the fiber will need to be researched; chromo- When a polymer preform is drilled with a hole as shown in phore alignment is needed for optimum optical activity in poly- Figure 2, and the hole is filled with a finite amount of fluid, the mer optical fiber. These areas of research plus potentially many aspect ratio of the drawn fiber will show an eventual decline with others lie ahead for researchers. increased distance along the length of drawn fiber. This is the result of decreasing fluid pressure exerted against the internal walls of the inner hole due to the reduction of the fluid column height as CONCLUSION the fiber is being drawn. Figure 5 clearly shows the experimental Polymer optical fiber offers potential advantages over tradi- results of this phenomenon: tional glass optical fiber. First, it is easier from an energy trans- fer standpoint to produce than glass fiber (glass fiber requires ≈ 2000°C to produce vs. ≈ 184°C for polymer fiber). Second, it is possible to make 1-2 micron diameter holes in the fiber (or even smaller) and possibly align chromophores within the holes to achieve optimum optical activity for use in amplifiers and such. Third, since polymer is composed of organic molecules, the po- tential polymer combinations available for fiber experimentation are numerous compared to combinations with glass fiber. These three aspects (plus potentially others) illustrate the need for ad- ditional polymer fiber research.

Figure 5. Aspect Ratio vs. Distance

CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005 129 THE DESIGN OF A FLUID DELIVERY SYSTEM FOR MICRO-CORE OPTICAL FIBER

APPENDIX

Distance Internal Diameter Outer Diameter Aspect Ratio Upper bound Lower bound 0 58 584 0.09931507 0.10273973 0.095890411 5 127 503 0.25248509 0.25646123 0.248508946 10 130 500 0.26 0.264 0.256 33 118 483 0.24430642 0.2484472 0.240165631 40 124 476 0.2605042 0.26470588 0.256302521 47 112 479 0.23382046 0.23799582 0.229645094 61 98 472 0.20762712 0.21186441 0.203389831 68 95 484 0.19628099 0.20041322 0.19214876 80 100 488 0.20491803 0.20901639 0.200819672 93 118 518 0.22779923 0.23166023 0.223938224 104.5 116 526 0.22053232 0.2243346 0.216730038 118 115 514 0.22373541 0.22762646 0.219844358 120 116 503 0.2306163 0.23459245 0.226640159 160 108 491 0.21995927 0.22403259 0.215885947 200 106 520 0.20384615 0.20769231 0.2 240 91 480 0.18958333 0.19375 0.185416667 280 100 536 0.18656716 0.19029851 0.182835821

Excel Table of Values (used to create figure 5)

REFERENCES [1] http://www.syringepump.com/NE-1000.htm [2] Heat Transfer with Applications, Kirk D. Hagen, p. 639. [3] Fundamentals Of Fluid Mechanics, Bruce R. Munson, Donald F. Young, Theodore H. Okiishi, p. 830, Appendix B

ACKNOWLEDGEMENTS Research support is gratefully acknowledged from the Na- tional Science Foundation Center on Materials and Devices for Information Technology Research (CMDITR), DMR-0120967.

130 CMDITR Review of Undergraduate Research Vol. 2 No. 1 Summer 2005