Engineering Solid Electrolytes for Lithium-Ion Battery Applications

Home , Eucryptite, Lithium imide, Lithium perchlorate

UCLA UCLA Electronic Theses and Dissertations

Title Engineering Solid Electrolytes for Lithium-Ion Battery Applications

Permalink https://escholarship.org/uc/item/5st377v1

Author Seegmiller, Trevor David

Publication Date 2015

Peer reviewed|Thesis/dissertation

eScholarship.org Powered by the California Digital Library University of California

UNIVERSITY OF CALIFORNIA

Los Angeles

Engineering Solid Electrolytes for Lithium-Ion Battery Applications

A thesis submitted in partial satisfaction of the requirements

for the degree of Master of Science in Chemical Engineering

Trevor David Seegmiller

2015

ABSTRACT OF THE THESIS

Engineering Solid Electrolytes for Lithium-Ion Battery Applications

Trevor David Seegmiller

Master of Science in Chemical Engineering

University of California, Los Angeles, 2015

Professor Jane P. Chang, Chair

An investigation of lithium aluminum silicate (LASO) synthesized by atomic layer

deposition (ALD) was investigated on various substrates and electrodes. Individual constituent

and compound oxide depositions were characterized using in-situ FTIR to determine surface

ligands after individual metal precursor and water pulses. Individual deposition rate of LiOH

and Al2O3 are 1.2 Å/cycle and 1.4 Å/cycle respectively, while SiO2 depositions by ALD could

not be achieved at low temperatures. LASO depositions have been demonstrated to be

25 Å/global cycle consisting of ten Al2O3 ALD cycles, six LiOH ALD cycles, and four SiO2

ALD cycles. Ionic conductivity values calculated for ALD lithium aluminum silicate (LASO)

were in the range of 2.2×10-9 to 4.7×10-8 S/cm.

The composites of ALD LASO and iCVD polymer solid electrolyte thin film on indium titanium oxide were also investigated and the corresponding ionic conduction was 1.8×10-8 S/cm at 110oC. In-situ electrochemical TEM showed lithium intercalation and deintercalation in

silicon-germanium alloys coated with a 33nm LASO thin film. Galvanostatic charge-discharge

cycling of LASO-coated carbon anodes showed 99.6% Coulombic efficiency, higher than 96.2% ii obtained with uncoated anodes. A solid state full cell comprised of LiCoO2 cathode, LASO solid electrolyte, and aluminum anode demonstrated ion transport between 0.9 V and 2.2 V cycling.

iii

The thesis of Trevor David Seegmiller is approved:

Yi Tang

Bruce Dunn

Jane P. Chang, Committee Chair

University of California, Los Angeles

2015

TABLE OF CONTENTS

CHAPTER 1 INTRODUCTION AND BACKGROUND ...... 1 1.1 Motivation ...... 2 1.2 History and Challenges of Lithium Ion Cells ...... 4 1.3 3D Battery Architecture ...... 11 1.4 Electrolytes for Lithium Ion Battery ...... 19 1.5 Thin Film Deposition Techniques for Solid Electrolytes ...... 32 1.6 Summary ...... 36 CHAPTER 2 METHOD OF APPROACH ...... 37 CHAPTER 3 EXPERIMENTAL METHOD ...... 38 3.1 LASO Atomic Layer Deposition Chamber and In-situ FTIR Chamber...... 38 3.2 iCVD Process ...... 42 3.3 Substrates and Nanowires as Electrodes ...... 45 3.4 Thin Film Characterization Techniques ...... 46 3.5 Electrochemical Characterization Techniques ...... 59 CHAPTER 4 RESULTS ...... 68 4.1 In-situ FTIR of LASO and Its Constituent Oxides ...... 68 4.2 ALD of LASO ...... 77 4.3 In-situ TEM Electrochemical Characterization of LASO on Si/Ge Nanowires ...... 80 4.4 Hybrid ALD-iCVD Solid Electrolyte Films ...... 87 4.5 LASO in Half-Cell Applications ...... 92 4.6 LASO in Full-Cell 2D Applications ...... 94 CHAPTER 5 SUMMARY...... 96 APPENDIX ...... 98 BIBLIOGRAPHY ...... 139

TABLE OF FIGURES

Figure 1.1: Comparison of volumetric and gravimetric energy density of different battery chemistries. Lithium-ion and lithium-polymer ion batteries have high energy and power density compared to lead-acid and nickel-metal hydride chemistries (Tarascon 2001)...... 4

Figure 1.2: Schematic of discharging lithium-ion cell, where lithium metal intercalated into the anode material and travels through the electrolyte to be intercalated into the cathode material...... 7

Figure 1.3: General electrode configuration (left) and 3D array electrode configuration (right).. 12

Figure 1.4: Thin film Li-ion solid state batteries provide favorable properties in energy density and capacity (Kim 2015) ...... 15

Figure 1.5: Schematic of vapor-liquid-solid mechanism for silicon nanowire growth using gold nanodots as a catalyst and silane precursor for silicon metal...... 16

Figure 1.6: Monomer chemical structures for PV3D3 (left) and PV4D4 (right) in solid polymer electrolyte applications (Reeja-Jayan 2015)...... 27

Figure 1.7: Phase equilibrium diagram for Li2O-Al2O3-SiO2 system. Modified from (Roy 1949)...... 30

Figure 1.8: Schematic representation of the atomic layer deposition (ALD) process using self- limiting surface chemistry. A metal-organic precursor flows into the chamber and reacts on the surface-terminated hydroxide group. After which, an oxidant reacts with the surface metal-organic species to form a metal oxide layer (George 2010)...... 34

Figure 3.1: (a) Schematic of the hot-wall atomic layer deposition chamber for synthesizing LASO thin films. (b) Schematic of the in-situ FTIR chamber for observing surface reactions in the ALD process...... 39

Figure 3.2: Schematic of a iCVD chamber where precursor flows into the chamber and decomposes on a wire array above the substrate to begin the polymerization process on the surface (Chen 2015)...... 43

Figure 3.3: Schematic of spectroscopic ellipsometry measurement principles showing the direction changes in polarized light after striking a sample (Ohkoshi 2005)...... 46

Figure 3.4: An example of an optical model to measure thickness of thin film oxides using spectroscopic ellipsometry (Li 2011)...... 47 vi

Figure 3.5: Experimental spectroscopic ellipsometry data obtained from 30 cycles of ALD Al2O3 deposited on Si (110) substrate. Sample thickness was 43 Å...... 49

Figure 3.6: In-situ difference FTIR spectra obtained during the first LTB and H2O exposures during ALD deposition cycles. The absorption regions of interest include the surface hydroxyl region (O-H stretching at 3800-3550 cm-1), surface lithium butoxide species -1 -1 (*LiOC(CH3)3)(C-H region at 3000-2800 cm and C-O stretching in 1215-970 cm )...... 52

Figure 3.7: Survey Spectra of LASO deposited by ALD with the highlighted binding energies for oxygen (~530 eV), carbon (~285 eV), silicon (~100 eV), aluminum (73 eV), and lithium (~55 eV) by dashed lines at the appropriate peaks. This survey scan was completed on a 63 nm LASO sample deposited on Si (100)...... 54

Figure 3.8: A scanning electron microscope (SEM) image of a single carbon pillar electrode coated with a 23nm LiAlSiO coating imaged under an excitation voltage of 10 kV...... 55

Figure 3.9: A high resolution transmission electron microscope (HRTEM) image of a Ge0.4Si0.6 alloy nanowire of diameter ~80 nm with a conformal LASO coating of 33 nm. The operating voltage of the electron microscope was 200 kV...... 57

Figure 3.10: Selected area diffraction (SAD) pattern observed from a Ge0.4Si0.6 alloy nanowire.57

Figure 3.11: Schematic of an in-situ TEM sample setup showing the working electrode (Ge/Si alloy) and counter electrode (Li metal) (Liu 2011)...... 58

Figure 3.12: Example impedance model where the ratio of C2 in the circuit is changed from a ratio or 0.1 (---) to 1 (---)...... 61

Figure 3.13: Model of a simplified Randle’s circuit used to analyze ion conduction of the thin films. Model consists of of a resistor in series with a resistor and capacitor (or Warburg diffusion element) in parallel to model resistance of ion conductivity in the circuit, with the capacitor modeling the capacitance double layers at the interface...... 62

Figure 3.14: A sample setup for electrochemical impedance measurements where the deposited films are between a conducting indium tin oxide (ITO) substrate and electron-beam deposited platinum electrodes (Pt)...... 63

Figure 3.15: Nyquist impedance spectra obtained from 21 nm ALD LASO sample deposited on an indium tin oxide (ITO) substrate (●). Using the simplified Randle's circuit, a model (––– ) is superimposed on the data...... 64

vii

Figure 3.16: Cyclic voltammogram of lithium cobalt oxide (LiCoO) coated with 4 nm LASO as the working electrode in a three electrode system from 0 V to 1.2 V potential back to 0 V. The reference electrode and counter electrode were lithium metal in a 1 M LiClO4 solution with propylene carbonate as the solvent...... 66

Figure 3.17: Galvanostatic titration of 3D carbon pillar array with an ALD LASO coating of 26 nm in 1 M LiClO4 in 1:1 ethylene carbonate : dimethylcarbonate solution with lithium as the counter and reference electrodes...... 67

Figure 4.1: In-situ difference FTIR spectra for Al2O3 film deposited by alternating exposures of o TMA and water oxidant on ZrO2 nanoparticles at 225 C...... 69

Figure 4.2: In-situ difference FTIR spectra for LiOH film deposited by alternating exposures of o LTB and water oxidant on ZrO2 nanoparticles at 225 C...... 71

Figure 4.3: In-situ difference FTIR spectra for SiO2 film deposited by alternating exposures of o TEOS precursor and H2O on ZrO2 nanoparticles at 225 C...... 73

Figure 4.4: In-situ difference FTIR spectra for lithium aluminat film deposited by alternating o exposures of LTB, H2O, TMA, and H2O precursors on ZrO2 nanoparticles at 225 C...... 73

Figure 4.5: In-situ difference FTIR spectra for LASO deposited in TMA, H2O, LTB, H2O, TTBS, o and H2O repeating cycles on ZrO2 nanoparticles at 225 C. Shown here are global cycles 1 and 2...... 75

Figure 4.6: In-situ difference FTIR spectra for LASO deposited in TMA, H2O, LTB, H2O, TTBS, and H2O repeating cycles on ZrO2 nanoparticles at 225 oC. Shown here are global cycles 3 and 4...... 75

Figure 4.8: Growth rate of ALD films from ellipsometry by alternating pulses of TTBS/H2O, o TMA/H2O, and TTBS/H2O/TMA/H2O at 225 C on Si(100) substrate. Growth rate of 2.2 Å/cycle was observed for ALD of aluminum silicate by TTBS/H2O/TMA/H2O. The deposition rate of LASO grown by 10 aluminum cycles, 6 lithium cycles, 4 silicon cycles, and 6 lithium cycles constituting one global cycle is 21 Å /global cycle...... 79

Figure 4.9: XPS spectra of ALD films grown by TTBS/H2O, TMA/H2O, and o TTBS/H2O/TMA/H2O on Ge(100) substrate at 225 C. The presence of Si 2p photoemission peak at ~102 eV for ALD film deposited by TTBS/H2O/TMA/H2O repeated cycles show SiO2 deposition...... 80 viii

Figure 4.10: HRTEM of Ge0.4Si0.6 alloy nanowire (a) before lithiation with LiOH/Li counter electrode at the tip (bottom of frame) and (b) after driving lithium ions into the alloy. Diffraction patterns were also taken (c) before and (d) after lithiation...... 82

Figure 4.11: Cyclic voltammetry of lithiation and delithation of LASO/Ge0.4Si0.6 heterostructure. The cycle started at 0 V, proceeded to -8V at 50mV/s, then up to 4.5V, then returning to 0V...... 83

Figure 4.12: Current response of a LASO/Ge0.6Si0.4 heterostructure when a constant -7.5 V was placed on the system for 3 minutes. A constant current of 1.2 pA is observed until ~130 seconds when a sudden current of -2.68 uA flows into the anode alloy, and then returns to ~1.2 pA...... 84

Figure 4.13: HRTEM images of Ge0.6Si0.4 alloy anode coated with 33 nm LASO (a) before and (b) after lithiation with LiOH/Li counter electrode on the tip (bottom). Diffraction patterns were also taken (c) before and (d) after lithiation...... 86

Figure 4.14: STEM image of Ge0.6Si0.4 alloy after lithiation. The striations in the 33 nm LASO coating suggest the LASO film is preserved after lithiation of the alloy anode...... 87

Figure 4.15: AFM scan of (a) 20 nm iCVD PV4D4 polymer thin film (RMS: 5.1 nm) and (b) 28 nm ALD LASO thin film (RMS: 1.4 nm)...... 88

Figure 4.16: AFM scan of (a) 25nm iCVD PV4D4 polymer electrolyte (RMS: 5.1 nm) and (b) 25nm iCVD PV4D4 coated with 28nm ALD LASO (RMS: 5.5nm )...... 89

Figure 4.17: AFM scan of (a) 50 nm iCVD PV4D4 polymer electorlyte (RMS: 6.8nm) and (b) PV4D4 coated with 28 nm ALD LASO (RMS: 5.7 nm)...... 89

Figure 4.18: Inset graph shows impedance spectroscopy measurements of a 25 nm iCVD PV4D4 polymer electrolyte sample soaked in 1M LiClO4 1:1 EC:DMC solution using hanging mercury drop probes. Two measurements were taken at different locations on the sample. After coating the 25 nm iCVD PV4D4 polymer with 28 nm ALD LASO, the ion conductivity of the film increased substantially...... 90

Figure 4.19: Impedance spectroscopy measurement of 90 nm ALD LASO deposition (32% lithium content) deposited on quartz-ITO substrate. A 20 nm iCVD PV4D4 deposition followed and the ALD-iCVD hybrid electrolyte heterostructures was soaked for 3 days in 1 M LiClO4 1:1 EC:DMC solution. The ionic conductivity was calculated to be 4.78 × 10-11 S/cm...... 91

Figure 4.20: Galvanostatic charge-discharge cycling of 22 nm LASO deposited on 2D carbon slurry in 1M LiClO4 in 50:50 EC/DMC solution with lithium foil as counter and reference electrodes...... 92

Figure 4.21: Galvanostatic charge-discharge cycling of 22 nm LASO-coated 3D carbon pillars in 1M LiClO4 in 1:1 EC/DMC solution with lithium foil as reference and counter electrodes.93

Figure 4.22: Schematic of a full cell solid state battery consisting of a solid state consisting of an aluminum collector back, lithium cobalt oxide cathode, ALD LASO as the solid electrolyte, and a sputtered aluminum anode...... 94

Figure 4.23: Galvanic charge, discharge, and charge curve for a 2D all solid state lithium ion battery consisting of LiCoO2 as the cathode, 5 nm ALD LASO as the solid electrolyte, and aluminum as the anode and current collector...... 95

LIST OF TABLES

Table 1.1. Selected primary and secondary battery chemistries (Palacin 2009)...... 2

Table 1.2. Examples of electrochemical reactions and the standard potential (voltage) associated with each reaction. The reaction of hydrogen ions to hydrogen gas is defined to be zero...... 9

Table 1.3. Theoretical maximum capacity of select anode materials and capacities for cathode materials for Li-ion battery applications (Aifantis 2010; Xu 2013; Nitta 2014)...... 10

Table 1.5. Table of selected organic polymer solid electrolytes...... 24

Table 1.6. Common organic solid electrolyte polymers for lithium ion batteries (Grünebaum 2014)...... 25

Table 1.4. Selected inorganic solid electrolyte ceramics and their ionic conductivities and activation energies (Knauth 2009)...... 28

Table 3.1. Properties of LASO precursor chemicals ...... 39

Table 3.2. Cycle sequences implemented for atomic layer deposition of LASO using TMA/H2O, LTB/H2O and TEOS/H2O (or TTBS/H2O) chemistry...... 41

Table 3.3. Precursor chemical properties for polymer iCVD deposition...... 44

Table 3.4. Optical models and constants for Al2O3, SiO2, and LiOH thin films...... 48

Table 3.5. Infrared vibrational wavenumbers for lithium tert-butoxide (LTB) using in-situ FTIR spectroscopy (Cavanagh 2010)...... 51

Table 3.6. Infrared vibrational wavenumbers for tetraethyl orthosilicate (TEOS) using in-situ FTIR spectroscopy (Ferguson 2004)...... 51

Table 3.7. Infrared vibrational wavenumbers for trimethylaluminum (TMA) using in-situ FTIR spectroscopy (Goldstein 2008)...... 51

Table 3.8. Atomic sensitivity factor values for photoemission peak values in LASO thin film samples (Briggs 1990)...... 54

Table 3.9. List of passive circuit elements used in equivalent circuit modeling in impedance spectroscopy with fundamental relationships and impedance response equations...... 61

xii

ACKNOWLEDGEMENTS

I would like to thank Professor Jane P. Chang for the opportunity to learn and grow under

her tutelage and guidance. By setting high standards, I have learned how to push and achieve

much more than I ever thought I could. With her busy schedule she still gives the time necessary to see the success of her students.

Professor Bruce Dunn has taken time to explain difficult concepts, and given insight to the problems I faced. I could not have moved forward without his help. I am grateful for his words of encouragement, and inspiring role model as a researcher.

I would also like to thank Professor Yi Tang for his help and guidance, as well as a listening ear during the later stages of my degree. I appreciate his patience and I am grateful for his attendance on my committee.

Also, recognition needs to be given to the Office of Naval Research (ONR) for funding of my graduate studies. Thanks also go to Dr. Kathrine Jungjohann and Dr. Jinkyoung Yoo at the

Center for Integrated Nanotechnology (CINT). Thanks need to be given to Professor Karen

Gleason and post doctoral researchers Dr. B. Reeja-Jayen and Dr. Nan Chen for their collaboration in solid electrolytes.

I also want to thank members of Professor Chang’s research group, and most notably Dr.

Jea Cho, to whom I owe gratitude for his guidance and mentorship in research and collaboration.

Special thanks and attention are also needed for my patient wife Heather, and our wonderful daughter Hannah. Through their support, I have been able to accomplish difficult goals. I look forward to continuing our grand adventure together.

xiii

CHAPTER 1 INTRODUCTION AND BACKGROUND

Due to the high power specifications, portable and miniaturized electronics are powered

using lithium-ion (Li-ion) batteries. Further research has introduced anode and cathode materials

that increase the power density by a factor of ten from 1991-2001. However, the pace of

advancement in battery technology capacity is much slower than the advancements made in

microelectronics. Part of the problem comes from traditional 2D electrode construction. As a

Li-ion 2D battery is scaled to fit in miniature power applications, the loss of the energy density is

proportional to the volume of the anode and cathode materials.

New 3D battery architectures can reduce the energy density loss that results from reduced

dimensions. Application of 3D electrodes, however, is limited due to the lack of electrical

insulation between the anode and cathode in liquid and polymer-based electrolyte materials.

Solid electrolyte materials that also can be constructed conformally on the 3D architecture is

currently needed to further the development of small-scale lithium-ion batteries.

Current deposition techniques of metal-oxide based lithium ion conductors include physical vapor deposition (PVD) and chemical vapor deposition (CVD). These methods have produced thin film 2D layers that indicate promising applications to develop thin film Li-ion batteries. The line-of-sight limitation of these deposition techniques means a conformal coating on 3D architectures is extremely difficult. A self-terminating surface reaction deposition technique, called atomic layer deposition (ALD) could synthesize solid electrolytes for 3D Li-ion

battery structures, with the potential of fine composition and thickness control.

1.1 Motivation

Potential energy can be stored through the use of electrochemical cells. Electrochemical

cells are used in a variety of applications, from portable electronic devices to transportation

power systems. Examples of electrochemical cells include capacitors, batteries and fuel cells.

Batteries generate electrical energy from potential chemical energy within a cell, and are the

most common cell in use today due to high energy density and cost effectiveness.

Batteries come in two categories: primary and secondary. Single-use batteries are

primary batteries, while secondary batteries are rechargeable. Commercial battery technology, often called battery chemistries, include lead-acid, nickel-cadmium, and nickel metal hydride.

Table 1.1 displays various characteristics of common battery technology, which are based on cell

chemistry. Many different battery chemistries have advantages and disadvantages that need to

be incorporated into the application. Practical specific energy of the cell illustrates the amount of

energy that a battery or cell has available per unit weight, and usually a larger specific energy is

viewed as positive for most applications.

Table 1.1. Selected primary and secondary battery chemistries (Palacin 2009).

Battery Type Voltage Theoretical Practical Practical Chemistry (V) specific energy specific energy energy density (W h kg-1) (W h kg-1) (W h dm-3) Zn/MnO2 Primary 1.5 358 145 400 (alkaline) Li/I2 Primary 2.8 560 245 900 Pb/acid Secondary 2.1 252 35 70 Ni/Cd Secondary 1.3 244 35 100 Ni/MH Secondary 1.3 240 75 240 Li-ion Secondary 4.1 410 150 400

Figure 1.1 shows the most common secondary battery cell energy density by weight

versus energy density by volume. As chemistries developed, more energy can be placed in a

smaller size and lighter weight than previous chemistries, leading to more common use and

applications of electronic devices. Battery chemistries have been slow to develop, but

commercialization of more powerful chemistries has opened the way to portable electronic

devices such as laptop computers, tablets, and smartphones.

Lead acid batteries were first developed at the end of the 19th century. Although the

chemistry was relatively simple to construct, the specific energy of the cell, or the total energy to

weight ratio, was very low. Over the course of the next few decades, other chemistries and

Rechargeable batteries that have a large energy density to weight ratio are introduced into

consumer electronics. Nickel-cadmium batteries (Ni/Cd) were first invented in 1898, and in the

second half of the 20th century, they were the rechargeable battery of choice for portable power

tools and flashlights. These rechargeable cells had a higher cell energy capacity (higher total

energy) than other chemistries and could be charged and discharged for multiple cycles without large diminishing capacity. However, these cells tend to self-discharge up to 20% per month, tends to output a decreased voltage if not recharged to a full state of charge (“memory effect”), and is more expensive than other battery chemistries (Pop 2005). Table 1.1 shows selected primary and secondary battery chemistries with each cell potential as well as theoretical and practical energy density. However, in the 1990s, the lithium ion cell became the electrochemical cell of choice for personal electronic devices.

1.2 History and Challenges of Lithium Ion Cells

Lithium ion cells are used because of their high voltage, light weight, and higher energy density than other battery technologies, such as nickel-cadmium (Ni-Cd) or nickel metal hydride

(Armand 2008). Lithium ion batteries (LIB) have enabled portable electronic devices to become smaller and more powerful. Electronic technology development has outpaced lithium battery advancements for the last several years, as Moore’s Law and Dennard scaling have had direct impact on electronic performance. Efforts have been made to increase the energy and power density of lithium ion batteries. These efforts would enable more technology development ranging from independent microelectromechanical systems (MEMS) to transportation power systems for electric or hybrid vehicles. LIBs are the most promising battery chemistry currently developed for high-power applications such as transportation as well as portable electronic 4

devices due to the high energy density. The total global demand for lithium ion batteries is

currently valued at $14 billion (Alias 2015).

The development, uses, and abilities of lithium-ion batteries have changed rapidly over a short amount of time. Lithium ion cells were originally proposed in the early 1970s, and primary

cells were developed a few years later. These original applications were for wrist watches and

calculators due to the high energy content and high potential of the cells. Lithium rechargeable

batteries, or secondary cells, were demonstrated afterward, but a poor recharge cyclability, or the

ability to hold a charge, limited applications (Whittingham 1976). The discovery of compounds

that could electrochemically react reversibly with lithium ions were developed over the next few

years. However, information was not widely available. Most data on these new compounds

were only in the form of conference proceedings and not published in circulated journals

(Tarascon and Armand 2001).

The main disadvantage of rechargeable lithium ion batteries (LIB) at this time was poor

cycle performance. Other significant disadvantages had to do with safety and stability, as the

rechargeable batteries were susceptible to thermal runaway and fires. These problems were

attributed to lithium metal crystal formation on the lithium metal electrode called dendrites.

Despite the significant disadvantages, the high cell potential, high energy density, low self-

discharge rate, and wide operation temperature made LIB the focus of research for

electrochemical energy storage. It was not until 1991 when Sony Corporation used materials

which reacted with a high degree of reversibility that allowed lithium ion rechargeable batteries

to be commercialized (Nishi 2001).

After the introduction of lithium ion batteries, materials research has become the main

focus in lithium ion battery improvement. Different electrode material chemistries and

manufacturing techniques have augmented and enhanced lithium ion battery performance. The total storage capacity of lithium ion batteries doubled nine years after its first introduction (Nishi

2000) and continues to make rapid progress.

1.2.1 Lithium-ion Battery Cell Construction

Lithium ion batteries work by using an anode and cathode separated by an electrolyte.

The anode of the battery is the electron source. Current commercial batteries usually have

graphite anodes. The cathode is the positive terminal in the battery, where the lithium ions

migrate when the electrochemical cell is discharging. The separating medium, the electrolyte, is

ionically conductive but electrically insulating to prevent short-circuits.

When a lithium ion battery (LIB) secondary cell is charged, lithium ions unintercalate

from the cathode, travel across the electrolyte, and intercalate into the anode. The

electrochemical reaction in the anode changes the lithium ion to lithium metal. Storing lithium

metal in the anode gives a charged battery its potential energy. As the cell discharges, electrons

from the lithium metal in the anode flow through a load to the cathode. The lithium ion then

travels from the anode to the cathode through the electrolyte. This discharging process is

represented in Figure 1.2.

Figure 1.2: Schematic of a discharging lithium-ion cell, where lithium metal intercalated into the anode material becomes an ion, deintercalates into the electrolyte and travels through the electrolyte to be intercalated into the cathode material.

The primary anode material of LIB primary cells was lithium metal. However, due to its poor cyclability, safety concerns due to dendrite formation, and advancements in materials chemistry, lithium metal is no longer used as an anode material. The discovery of intercalating materials such as graphite introduced an economical and safe anode for LIB. There are many other available anodes for lithium ion batteries, and the materials dictate the characteristics of the battery. The energy density of LIB is mostly limited by the anode materials in the cell (Nishi

2001). As a result, industry pours heavy investment into materials research to find suitable high- capacity anode materials for LIB applications. These materials range from transition metal oxides to metalloids like silicon and germanium (Liang 2013).

The LIB cathodes are usually lithium transition metal oxides, and customarily contain high concentrations of lithium for high-energy applications (Li 2014). The material of choice for most commercial LIB applications is lithium cobalt oxide due to its high capacity. This has disadvantages such as poor thermal stability and high environmental impact. 7

1.2.2 Lithium-ion Cell Electrochemical Reactions

Electrochemical reactions occur in a battery as a sequence of two half-reactions to produce electrical energy. These reactions are thermodynamically favorable to produce electrical energy. For current commercial batteries which contain lithium cobalt oxide (LiCoO2) and carbon (C), the overall electrochemical reaction for a discharging battery can be described as

(Kumar 2010):

+ + 6 ( 1-1)

1−푥 2 푥 6 2 The two퐿푖 half 퐶표reactions푂 퐿푖 associated퐶 → 퐿푖퐶표 with푂 this퐶 equation occur at the cathode and anode respectively. At the anode, lithium metal is intercalated, or imbedded, in the anode favorably separating from its electron to produce a lithium ion.

+ + 6 (1-2) + − 푥 6 The lithium퐿푖 퐶 ion→ 푥퐿푖travels 푥through푒 퐶the electrolyte towards the cathode, while the electron flows from the anode through a load to the cathode. At the cathode, lithium ions are intercalated into a lithium cobalt oxide structure:

+ + (1-3) + − 1−푥 2 2 The process퐿푖 is퐶표 reversed푂 푥퐿푖 to charge푥푒 the→ 퐿푖퐶표lithium푂 battery system; an external power source is connected to the battery instead of a load to drive electrons and lithium ions toward the anode.

0 The theoretical standard cell voltage (E cell) can be calculated by using the difference

0 between the standard electrode potential at the cathode (E cathode) and the standard electrode

0 potential at the anode (E anode):

= (1-4) 0 0 0 퐸푐푎푡ℎ표푑푒 − 퐸푎푛표푑푒 퐸푐푒푙푙

Table 1.2. Examples of electrochemical reactions and the standard potential (voltage) associated with each reaction. The reaction of hydrogen ions to hydrogen gas is defined to be zero.

Reaction E0 (V) + -3.10 + +−2 -0.76 2퐿푖 2++푒2 →− 퐿푖 0.00 푍푛 + + 푒− → 푍푛 +0.80 2 퐻+ + 4−푒 →+퐻2 + 2 +1.47 퐴푔 푒 →+ 퐴푔 − 2+ 푃푏푂2 퐻 푒 → 푃푏 퐻2푂 A true estimate of the open circuit cell voltage can be obtained by modifying the standard

electrode potential by using the Nernst equation, which adds the reacting component in its

nonstandard state:

= (1-5) 0 푎푝푟표푑푢푐푡푠 퐸 퐸 − 푅푇푙푛 � 푟푒푎푐푡푎푛푡푠� where a is the activity of the products푎 or reactants respectively. The open-circuit voltage of a

battery is the maximum voltage in its charged state and giving zero current.

Current is the flow of electrons (or charge) per unit time and is measured in amperes, or charge per second. Electrochemical reactions in the cell dictate the electron flow in the system.

Low currents in the cell are attributed to activation losses, or internal resistances to reactions are

very high. On the other hand, the maximum current characteristic of a battery is usually

determined by mass transfer limitations of ions. This transfer can occur at any position in the

cell, but the limiting factor in most battery applications is the electrolyte.

Energy density is the amount of energy that can be extracted per unit volume of the working cells, while specific energy density is the energy derived per unit weight of the cell (or the active electrode material). Usually energy density is limited by the amount of ions that are available in the system (Sarakonsri 2010). 9

Power density is how much current per unit time can be extracted from the cell. This

amount is usually normalized per unit weight or Watts per kg. Large power density batteries are

usually required in automobile and stationary power storage systems to supplement electrical

infrastructure.

The capacity of the battery is the total amount of electricity that can be extracted from the

electrochemical reactions in the cells, often described in mAh/g. This measure is material

specific, and is an intrinsic property. The theoretical maximum capacities of various anode

materials are shown in Table 1.3.

Table 1.3. Theoretical maximum capacity of select anode materials and capacities for cathode materials for Li-ion battery applications (Aifantis 2010; Xu 2013; Nitta 2014).

Negative Electrode Capacity Positive Electrode Capacity (Anode) (mAh/g) (Cathode) (mAh/g) Silicon 3579 Oxygen ~1500 Germanium 1384 Sulfur ~1277 Aluminum 993 Li2(Fe or Mn)SiO4 ~250 Tin 990 LiCoO2 ~150 Antimony 660 LiNi(1-x-y)CoxMnyO2 ~100 Bismuth 385 385 Carbon 300 300

There are also several terms that explain the life of the battery. Shelf-life describes the amount of time a battery can be inactively stored before its capacity falls to 80%. Similarly,

cycle life is a similar measure. Cycle life describes the number of battery charges and discharges

before the battery capacity fades to 80% of its initial reversible value. The service life of a

battery describes the time a battery can be used at various loads and operating temperatures.

This time is normalized for ampere per mass or liter.

There are several obstacles that are facing lithium ion batteries. Even with advancements

in LIB technology, the pace of lithium ion battery advancement has not kept up with portable

electronic devices (Tarascon and Armand 2001). This leads to a so-called performance gap

where devices demand more power and energy than current battery technologies can handle

(Armand and Tarascon 2008).

The energy stored in lithium ion batteries has increased over 300% in the last twenty

years. However, most of this increase has been due to cell engineering, or by controlling the size

and morphology of active materials. Increasing efficiency in this way has almost reached its

limits (Goodenough 2013).

There is also a focus of LIB for integrated systems at the micro scale. Thus, the

development of microbatteries for small integrated sensors and wireless devices is becoming

more interesting (Mukaibo 2010). Having integrated sensors and their power sources the size of

1 mm2 for autonomous nodes is a reasonable and exciting direction for microbatteries (Long

2004).

Because total energy capacity is related to amount of total lithium transport, merely shrinking the footprint of modern batteries only reduces the amount of energy they hold. In current technology, reduced surface area of 2D battery reduces energy as well. Miniaturization does not meet power and energy requirements for integrated systems. Thus, novel chemistries and engineering are needed to improve performance of Li-ion batteries at the microscale.

1.3 3D Battery Architecture

Most batteries are planar, meaning that a flat cathode and flat anode are separated with the electrolyte. However, 3D configurations were proposed to increase the capacity per area footprint, usually associated with units µA h cm-2, as battery scales are decreased (Roberts 2011). 11

This 3D battery architecture overcomes the challenges of decreased power density without

excessively sacrificing energy density, or in other words, optimizes both energy and power

densities of LIB. There are many configurations of 3D architectures. Some examples include

array of cylindrical cathodes and anodes, array of trenches, and a rod array of cylindrical anodes

coated with an electrolyte with the remaining free volume filled with cathode material (Long,

Dunn, Rolison and White 2004).

A schematic of a 3D architecture is shown in Figure 1.3. Limited areal footprint is one challenge of shrinking batteries. In a traditional 2D battery, shrinking the area of the planar electrodes not only decreases the surface area of the battery, but also the amount of electrode material. This results in reduced energy and power density in the battery. Changing to a 3D structure, for example an array of pillars, the surface area and volume of the battery electrodes increases by increasing the length of the pillars.

Figure 1.3: General electrode configuration (left) and 3D array electrode configuration (right) with the same areal footprint (A). Enhanced electrode surface area is a feature of 3D architecture batteries.

Short diffusion length for ions increases power density and significantly lowers ohmic

losses that occur inside the cell. Thin film batteries demonstrated that solid state electrolytes can

be used as long as electrical insulation is maintained at the nano-length scale. Batteries that scale

to 1-10 mm2 and still contain an energy density of 2 J/mm3 would be optimal in these small-scale

micro devices (Dunn 2008).

For the concentric tube 3D electrode configuration shown in Figure 1.3, the surface area

can be calculated using hexagonal close-packed arrangement with a planar footprint of W2:

= + = + (1-6) 2 휋푑 퐴3퐷 푊퐻 휋푑퐿 ∙ 푁푎푛표푑푒 퐴2퐷 2 퐴푅 ∙ 푁푎푛표푑푒 where d is the diameter of the electrode feature, A2D is the surface area of an equivalent planar

electrode, and AR is the aspect ratio defined as L/d.

These calculations show that 3D architecture provides greater surface area per planar footprint when increasing L, which increases power density in the cell. Another way to describe the features in 3D batteries is aspect ratio. This describes the height of a feature (cylinder or trench) in relation to its width. By convention, increasing the aspect ratio of a device feature is usually associated with increasing height, keeping the width of the feature constant. Increasing the aspect ratio of electrodes in a 3D cell also increases its volume, which is proportional to its capacity. However, the length of the electrodes cannot be increased without limit, as the increase in electrical resistance in the electrode material would offset the advantages of increased capacity

(Long, Dunn, Rolison and White 2004). Therefore, it is necessary to engineer 3D electrodes to

optimize energy capacity and power output while balancing ohmic losses.

Most 3D architecture batteries require the manufacture of a conformal electrolyte to

completely cover the electrodes. Liquid electrolytes are not used in these microbatteries because

electrode/electrolyte stability does not permit their use at these length scales (Quartarone 2011). 13

Solid state batteries address some of the challenges in miniaturizing Li-ion batteries.

Comparisons of thin film lithium ion battery energy density with other lithium ion battery characterizations are outlined in Figure 1.4.

There are numerous advantages for all solid state Li-ion battery chemistries: longer product life due to less wear and tear during operation, shock and vibration resistant, and a larger operating temperature range (often reaching temperatures up to 200oC). Disadvantages include low power and current output at ambient temperature conditions, and large internal resistances.

Stresses at the electrode-electrolyte interface can also contribute to reduced performance.

Figure 1.4: Volumetric energy density comparison of different battery chemistries and selected Li-ion battery structures. Thin film Li-ion solid state batteries provide favorable properties in energy density and capacity (Kim 2015)

There are many proposed approaches on manufacturing 3D LIB. These approaches must

take into account the electron and ion conduction of electrodes (Notten 2007), interfacial reaction

kinetics (i.e. the reactions between the electrolyte and the electrodes), and electrode capacity

(Long, Dunn, Rolison and White 2004). The methods to construct 3D architectures also

contribute to the difficulty of building these types of batteries.

1.3.1 Nanowires and Atomic Layer Deposition in Energy Storage

Atomic layer deposition (ALD) has many applications in energy storage. Because the

deposition technique yields conformal and pinhole free films, ALD has many applications in coating of nanofeatures. One such feature is nanowires, which have a very large aspect ratio.

Usually 1-100 nm in diameter, these rods can reach microns in length, and are assumed to be one-dimensional. These rods are below the characteristic length scale of many different phenomena, including exciton Bohr radius, wavelength of light, and phonon mean free path

(Hochbaum 2010).

There are various 1D morphologies available for nanostructures including nanowires,

nanorods, nanoribbons, and nanotubes. The most common materials for nanostructure

manufacture include Si, Ge, GaN, GaAs, CdS, ZnO, and SnO2. Most structures are formed using

a classical VLS (vapor-liquid-solid) mechanism. An example of this process is the manufacture

of silicon nanowires. Gold nanoparticles are placed on a silicon substrate. Using chemical vapor

deposition (CVD), the substrate is heated to ~600oC (the melting point of gold nanoparticles) and

silane gas (SiH4) is introduced into the chamber. The gold acts as a catalyst and silane gas

decomposes to form hydrogen gas, and a gold-silicon eutectic. As more silane decomposes on

the surface of the gold nanoparticle, the silicon concentration increases until a critical

composition of about 4Au:1Si, and silicon begins to crystallize on one side of the gold

nanoparticle. The silicon usually crystallizes above the silicon substrate, forming a single-crystal nanofeature. This feature grows in length with more silane decomposition at the surface of the gold particle, and the diameter is fixed with the size of the gold particle.

Figure 1.5: Schematic of vapor-liquid-solid mechanism for silicon nanowire growth using gold nanodots as a catalyst and silane precursor for silicon metal nanowire synthesis.

Other unique properties of nanowires include a large surface-to-volume ratio that makes distinct structural and chemical behavior. This increased surface area allows greater chemical reactivity and faster kinetics, which are an important component of lithium ion batteries. As mentioned before, the one-dimensional nanowires allow the direct conduction of quantum

particles, such as electrons. This allows nanowires to be an excellent platform to study transport

phenomena in confined and very small dimensions. Lastly, another advantage of these

nanowires is that the lengths are sufficient for top-down fabrication process that have already

been proven to be large-scale, such as photolithography (Hochbaum and Yang 2010).

Nanowires have many applications in the energy industry. One application is in solar

cells. This is because nanostructured silicon shows enhanced light trapping, efficiency, high

carrier collection efficiency, and a potential for low costs (Lu 2010). CVD-grown silicon

nanowires tested in solar cells gave an efficiency of 10%, a short-circuit current density of

40 mA/cm2, and open circuit voltage of 0.44 V (Yoo 2013). This is comparable to the highest

efficiency (10.8%) of ion implanted Si radial p-n junctions ordered in arrays using photolithography (Lu and Lal 2010).

1.3.2 Nanowires in Lithium Ion Batteries

There are advantages in using nanowires in lithium ion batteries. The first advantage is

that nanowires have direct current pathways, in which electrical transport can be realized

compared to particle electrodes (Chan 2008). Ion diffusion length is shorter in such nanofeatures.

This may lead to an increase in rate performance, as the characteristic time is dependent on the

diffusion length and diffusion (Dasgupta 2014). Nanowires can accommodate the volume

expansion, help manage mechanical degradation, and enhancing long-life cycling (Szczech 2011;

Kennedy 2014). Uniform arrays of nanowires can be grown directly on metal or carbon surfaces and applications include additive-free and binder-free battery manufacture (DiLeo 2011).

Nanowires can also be building blocks to construct complex architectures, combining the

advantages of 1D characteristics of the nanowire with the proposed 3D structure of full battery

cells. (Cui 2009). Lastly, nanowires can be isolated and studied using in-situ electrochemical

probing (i.e. cylic voltammetry). Single nanowire electrodes can be built and tested in-situ to

investigate high-resolution structural and electrical evolution during operation (Huang 2010).

Nanowires made of next generation Li-ion battery show promise to be incorporated into

3D architecture batteries. The difficulty of next generation Li-ion battery anodes consist of the

volumetric expansion upon lithium intercalation, around 300% in silicon (Qu 2011). Bare

silicon nanowires, which have a theoretical capacity of 4800 mAh/g, pulverize into small

particles due to stresses of volumetric expansion upon lithium insertion, causing electrical conductivity in the electrode to cease. This causes reduced capacity of the battery after several cycles. One way of increasing the cyclability of nanowires is to have a porous structure. For porous silicon example, after 250 cycles, the capacity remains stable at 2000, 1600, and 1100 mAh/g for charging rates of 2, 4, and 18 A/g (Ge 2012).

Researchers have started conducting extensive studies of atomic layer deposition in

lithium ion batteries in recent years (Liu 2015). For example, using hollow silicon nanotubes as the anode, ALD coatings of TiO2, TiN, or Al2O3 on the surfaces improves cycling performance

drastically. Lithium-active TiO2 improves capacity retention from 1287 mAh/g (uncoated base)

to 1700 mAh/g after 200 cycles at 0.2 C. Steady state coulombic efficiency and capacity

retention of 50% was achieved when cycling times were shortened from 0.2 C to 5 C (Lotfabad

2014). Also, the authors stated that transmission electron microscope (TEM) images provided

insight into cycling-induced failure mechanisms that are linked to microstructure and location of

ALD layers (Lotfabad, Kalisvaart, Kohandehghan, Cui, Kupsta, Farbod and Mitlin 2014).

These studies are limited in the thickness of the ALD coating due to the reduced ionic conductivity of lithium ions in the ALD films. Ionic conductivity is a measure of movement of the ions in a medium and is usually measured in Siemens per centimeter (S/cm). Thus, a conformal coating of a better lithium ion conductor could bring more applications to lithium ion batteries.

The first lithium-ion nanowire battery used a titanium/platinum-coated silicon nanowire as the anode, lithium phosphorus oxinitride (LiPON) as the solid electrolyte, and lithium cobalt oxide as the cathode (Ruzmetov 2012). The LiPON solid electrolyte was deposited using physical vapor deposition (PVD), which limits the application in 3D structures.

1.4 Electrolytes for Lithium Ion Battery

The electrolyte of a battery needs to be conductive to ions, yet insulating electronically.

Most LIB electrolytes in commercial use today are liquid organic solvents with dissolved lithium salts, such as lithium perchlorate. This allows lithium ions to flow between the anode and the cathode. Organic solvents are widely used in current commercial batteries.

Most electrolytes are a combination of ethylene carbonate, dimethyl carbonate, and ethyl methyl carbonate. Propylene carbonate is another organic solvent used in electrolyte chemistries.

The three most common salts are lithium perchlorate, lithium tetrafluoroborate, and lithium hexafluorophosphate (Park 2010). These liquid electrolytes are used because their lithium ion conductivity ranging from 0.1-1 S/cm is relatively high (Kumar and Sarakonsri 2010).

There are disadvantages for using these liquid electrolytes in lithium systems. One

problem is the safety concern of having organic solvents in commercial applications. Another

common problem is irreversible capacity loss due to side reactions that form products that

consume available lithium ions. This occurs in the formation of the solid electrolyte interface

(SEI) layer on the anode upon first charging (Palacin 2009).

1.4.1 Formation of Solid-Electrolyte Interphase (SEI)

Solid-electrolyte interphase (SEI) forms on the negative electrode in Li-ion batteries,

occurring as a result of electrolyte decomposition, usually during the first cycle. Other performance characteristics of the battery are highly dependent on the quality of this SEI layer

(battery performance, rate capability, cyclability, safety, etc.). Understanding the SEI layer in greater detail could provide insight and ability to tune and improve battery performance (Verma

2010).

Batteries are manufactured in a discharged state as lithiated carbons are not stable in air.

Electrolyte solutions are thermodynamically unstable at low potentials vs Li/Li+ with the first

charge of the cell the electrolyte solution reduces or degrades on the graphite surface and forms the SEI. Although the reduction occurs at the same potential, all surfaces are affected, including the current collector. However, due to low surface area, the reaction on graphite is predominant.

The reduction process results in a number of decomposition products that are deposited on the surface. This layer is kinetically stable to the electrolyte and guards against further reduction of the liquid electrolyte (Jung 2013).

Although the onset potential is not fixed, the general consensus of literature values range from 1-2 V and may require several cycles to completely form. The SEI layer is very complex

and is composed of lithium salt degradation products, organic solvent constituents, and inorganic

impurities of the anode. The thickness may vary from a few to hundreds of angstroms and SEI

layer measurement is usually characterized by impedance spectroscopy. In liquid cells,

complete SEI formation before Li-ion intercalation begins in the anode allows best Li-ion battery

performance (Jung, Lu, Cavanagh, Ban, Kim, Lee, George, Harris and Dillon 2013).

There are many factors affecting the SEI including the type of carbon, specific surface

area, crystallographic structure and particle morphology. Edges and surface imperfections also

highly affect decomposition products due to the high current density on these sites.

Pretreatment of the carbon anode was conducted to deposit by ALD Al2O3 to control the

SEI layer growth. The addition of this layer enhanced the cyclability of standard LCO and

graphite lithium-ion batteries. After 200 room temperature cycles, the cell measured a loss of

only 20% of initial capacity. At the same time, the bare electrodes showed ~80% capacity loss from the same number of cycles (Jung, Lu, Cavanagh, Ban, Kim, Lee, George, Harris and Dillon

2013).

Features of an ideal SEI include minimum electron and maximum Li+ conductivity. The

formation kinetics of the SEI layer should be fast before Li+ intercalation. Uniform morphology and composition on the anode should also be highly desirable (Aurbach 2003). With many complications the SEI layer demonstrates in modern Li-ion batteries, it is desirable for all solid- state batteries to be built to eliminate the need to study the complexity of the SEI layer.

Fabricating working cell are limited because conformal solid state electrolytes are difficult to manufacture (Baggetto 2010). There are several considerations when constructing a solid-state electrolyte. A high ionic conductivity (on order of 10-3 S/cm or better) would be

desirable for better ion transport and higher power density. Lithium ion batteries have high

voltages, so the chemical stability with these high voltage lithium cathodes, along with a high

electrochemical decomposition voltage of 6 V vs elemental lithium or lithium alloy anodes is

needed.

As mentioned earlier, the electrolyte of choice should demonstrate chemical stability

against reaction with the anode or cathode during manufacture or operation to prevent

undesirable side reaction products at the electrode-electrolyte interface. Other manufacturing

considerations such as products that are economical, as well as environmental friendliness

(Thangadurai 2014). In order cover these factors, usually solid electrolytes need to be as thin as

possible for short ion-diffusion lengths. Solid electrolytes also need to be conformal over the

entire nanofeature without pin holes or other major flaws that might create shorts. Due to surface energy effects, thin film properties are much different than their bulk properties and need to be taken into account.

There are two main categories of solid electrolytes for batteries: inorganic ceramics and organic polymers. Inorganic ceramics are generally metal oxides with high dielectric constants and already find applications in proton exchange membranes for solid-oxide fuel cells. Organic polymers are mechanically robust due to their low elastic moduli, but generally have low tolerance from thermal variations in the cell (Fergus 2010).

1.4.2 Application of Solid Electrolytes

In order for lithium ion batteries to be feasible, lithium ions need to move from one electrode to the other through the electrolyte. Lithium ion movement through a medium (either

solid or liquid) can be expressed as either a diffusion coefficient (D) or ion mobility (μ). Lithium

ion mobility is defined as the velocity attained by an ion through a medium under an electric

field.

Lithium ion transport in solid electrolytes, either polymer or inorganic ceramic, can be

described as short-term “hops” in the solid matrix. Ion movement in solid electrolytes can be

defined as (Tuller 2007):

(1-6) = = (1 ) 2 푚 표 −훥퐺 표 훥푆 −퐸 퐷 퐷 푒푥푝 � 푏 � 훾 − 푐 푍푎 푣 푒푥푝 � � 푒푥푝 � 푏 � where γ is the geometry correlation,푘 푇 (1-c)Z is a factor that푘 defines the푘 푇 number of neighboring unoccupied sites, a is the jump distance, υo is that attempt frequency and Em is the migration

energy.

Theoretical and experimental results suggest that ion diffusion through the solid

electrolyte interphase follows a two step mechanism. The first step involving the diffusion of the

ion through the bulk material, while the second step involves the capacitive double-layer at the

phase boundary (Shi 2012).

1.4.3 Polymer Solid Electrolytes

Polymer electrolytes are promising candidates for rechargeable batteries. While dry solid

polymer electrolyte films have poor ionic conductivity that are often lower than 10-5 S/cm, soaking polymers in liquid electrolytes raise the ionic conductivity to 10-4 S/cm and some values

of polymer electrolytes even have ionic conductivities as high as 10-3 S/cm (Agrawal 2008), as

shown in Table 1.4.

Table 1.4. Selected organic polymer solid electrolytes.

Polymer Ion Conductivity Reference (S/cm) PVDF-HFP/PSx-PEO3 4.2 × 10-2 (Seidel 2015) -3 PMMA-LiBF4 2.2 × 10 (Osman 2012) PAN/PMMA 2.3 × 10-3 (Rao 2012) + -4 PMMA-SiO2(Li ) 8.6 × 10 (Lee 2012) P(STFSILi) 1.3 × 10-5 (Bouchet 2013) PV4D4 7.5 × 10-8 (Chen 2015)

There are numerous advantages of polymer electrolyte material in addition to a higher

ionic conductivity. The polymer electrolyte has the opportunity to preferentially transfer all

+ single-ions (the Li cation) in the system, or the ionic transference number close to unity (tion~1)

(Agrawal and Pandey 2008). Current literature suggests that the cationic transference number is

approximately 0.5, which has been improved over the last few years.

Polymer films also have high chemical and electrochemical stabilities. These properties

are important in lithium ion batteries as the cell voltage is on the order of 4V. Some polymer

films have good electrochemical stability domain extending as high as 5V (Agrawal and Pandey

2008), as shown in Table 1.5. Other properties that allow polymer films to be used as

electrolytes include mechanical stability and elasticity, which allows compatibility with a variety

of electrode materials.

Most polymer electrolytes have additional lithium salts soaked into the structure. Most

common examples are lithium bis(fluorosulfonyl)imide (LiFSI) and lithium bis(oxalato)borate

(LiBOB). These larger delocalized anions play the dominant role for salt-in-polymer electrolytes.

This is compared to other lithium salts, in which the ionic conductivity of the polymer electrolyte

has to be increased by higher operation temperatures. Some salts are very unstable at these high

o temperatures, such as LiPF6, which is unstable above 60 C (Grünebaum 2014).

Table 1.5. Common organic solid electrolyte polymers for lithium ion batteries (Grünebaum 2014).

Poly(ethylene oxide) Poly(propylene oxide) Poly(ethylenimine)

Poly(organo siloxane) Poly(acrynitrile) Cyclotetrasiloxane

Poly(organophosphazene) Poly(vinylidene fluoride) Cyclotriphosphazene

One important specification of polymeric electrolytes is the glass transition temperature.

Higher glass transition temperatures are usually associated with longer polymer chains, but also

change based on the O:Li content of the polymer electrolyte (Grünebaum 2014).

Most solid polymer electrolyte studies are based on poly(ethylene oxide) (PEO) for its

high solubility for lithium salts (Christie 2005). Ion conductivity is usually in the range of 10-8 -

10-4 S/cm at room temperature, and the value of 10-3 S/cm is usually attributed to be required for practical lithium-ion battery applications (Diddens 2010). PEO is a semi-crystalline polymer and forms various complexes with lithium salts. Strong coupling effects between the Li+ ion and

the oxygen atoms in the PEO chains may be the source of increased ionic conductivity (Wright

2002; Diddens, Heuer and Borodin 2010) and mechanical and safety properties (Hallinan 2013). 25

Ion conductivity depends on physicochemical properties of the polymer matrix including

compositions, processing method, states of ions in the amorphous phase, percent crystal region

character, and interphase of polymer. Throughout the past decade, researchers have explored

dynamics of Li+ in polymer to improve electrolyte properties (Agapov 2011; Wang 2015).

Most results suggest the amorphous phase with activated chain segments (above the glass transition temperature) can aid ion transportation. Increasing the amorphous phase of PEO electrolytes can be one of the most effective ways to increase ionic conductivity. Thus, the

addition of plasticizers to the polymers would be beneficial to increase ionic conductivity

through interphase pathways (Nan 2003).

Copolymers with conductive PEO blocks are promising candidates as well. Interestingly,

copolymer electrolytes have self-assembled microstructures, which result in properties with good

balance of ion conductivity and mechanical performance. Microstructures can include

hexagonally packed cylinders, and have been explored as solid electrolytes (Young 2012).

Polysiloxane organic electrolytes have ionic conductivities that range from 10-7 to 10-5 S/cm

(Grünebaum 2014). Composite lithium ion conducting polymers have been explored in lithium-

air batteries in attempts to reduce dendrite formation at the lithium electrode (Wang 2013). The

combination of lithium ion conducting polymers were found to have limited negative impact of

lithium ion conductivity but seemed to enhance the mechanical expansion of the material without

failure, giving a swelling ratio of 794% (Wu 2013).

Although there are numerous polymer solid electrolytes to choose, possible candidates

include poly(1,3,5-trivinyl-1,3,5-trimethylcyclotrisiloxane) (PV3D3) and poly(1,3,5,7-tetravinyl-

1,3,5,7-tetramethylcyclotetrasiloxane) (PV4D4). These polymers have high electrical resistance

on order of 1015 Ω · cm, which is good for solid electrolyte applications. Other properties of

these polymers include a dielectric constant of 2.5 and a refractive index of 1.465

(O'Shaughnessy 2007).

Figure 1.6: Monomer chemical precursor structures for PV3D3 (left) and PV4D4 (right) in solid polymer electrolyte applications (Reeja-Jayan 2015).

To be used in Li-ion battery applications, PV3D3 and PV4D4 need lithium ions inserted into the polymer film. This is done by soaking the polymers in a liquid electrolyte for an extended time. In previous studies of these polymers in nanoscale electrolytes, samples of polymer thin films ranging from 10 nm to 40 nm were soaked in 1 M LiClO4 propylene carbonate solutions for three days. Non-soaked samples demonstrate dielectric character, while ionic conductivity after soaking were up to 10-7 S/cm at room temperature and activation energy of 0.025 eV (Reeja-Jayan 2015). It should be noted that the smaller ion conductivity of these polymer electrolytes can be offset by shorter diffusion lengths with thinner films at room temperature for a variety of applications (e.g. medical implants) (Chen, Reeja-Jayan, Lau, Moni,

Liu, Dunn and Gleason 2015).

1.4.4 Inorganic Solid Electrolytes

Inorganic solid electrolytes (ISE) are usually ceramics that are strong, have high electrical resistance, and can operate at elevated temperatures. ISE are promising candidates of 27

solid electrolytes because they do not suffer from leakage, volatilization, or flammability

(Hayashi 2012). Compared to organic electrolytes, ISE are shock and vibration resistant.

Lithium single ion conductors can have a lithium transference number of unity, which means that no concentration gradient exists inside the cell while it is operating. This is beneficial in lowering the over-potential in a full cell (Quartarone and Mustarelli 2011).

Most ISE are metal oxides, which already have numerous applications in electronic devices, such as optical coatings, barrier layers, and high dielectric constant insulators. These applications are due to intrinsic chemical and thermal stability. Some metal oxides have unique properties such as superconductivity, negative thermal expansion, and ionic conductivity. Metal oxides demonstrate ion conductivity through movement of cations through the crystal lattice

(Table 1.6). The most common ions for movement include proton-exchange membranes in fuel cells.

Table 1.6. Selected inorganic solid electrolyte ceramics and their ionic conductivities and activation energies (Knauth 2009).

Short Name Type Composition RT ionic conductivity Activation (S/cm) Energy (eV) -3 LLTO Crystalline Li3xLa(2/3)-xTiO3 10 0.3-0.4 -3 NASICON Crystalline Li1.3Al0.3Ti1.7(PO4)3 3 × 10 0.3-0.5 -6 LISICON Crystalline Li14ZnGe4O16 10 0.4-0.6 -4 Thio-LISICON Crystalline Li3.4Si0.4P0.6S4 6.4 × 10 0.5-0.6 -5 Garnet Crystalline Li6La2BaTa2O12 4 × 10 0.4-0.6 -4 Li ion Composite LiI-Al2O3 2.6 × 10 0.4-0.5 conductor- mesoporous oxide -3 Sulfide glass Amorphous GeS2+Li2S+LiI 10 0.4-0.5 +Ga2S3 and La2S3 -6 LiPON Amorphous Li2.88PO3.73N0.14 3.3 × 10 0.45-0.55

More advantages of solid inorganic electrolytes include increased safety and reliability,

and most inorganic solid electrolytes are single ion conductors. However, there are a few

challenges that remain for high-performance electrolytes. One is to create favorable solid-solid

interfaces between electrode and electrolyte (Ohta 2012), and to obtain high ionic conductivity at

room temperature (Cao 2014).

The fast lithium ion conductor lithium lanthanum titanate (LLTO) has attracted much

attention because of the high conductivity that can be achieved, which is on order of 10-3 S/cm

(Bohnke 2008). The high ionic conductivity is attributed to multiple equivalent sites in the

perovskite structure. While this compound does demonstrate high ionic conductivity, the

material reacts easily with lithium metal which increases the electric conductivity (Inaguma

1993).

Another inorganic lithium ion conductor under consideration is lithium phosphorous

oxynitride (LiPON). This solid electrolyte has an ion conductivity on order of 1.7 × 10-6 S/cm at

room temperature and can be deposited in thin films using sputtering (Suzuki 2011). It is

possible that the three-coordinated nitrogen atoms are responsible for the ionic conductivity.

Although both LLTO and LiPON demonstrate ionic conductivity that is favorable, synthesizing

techniques for these films are difficult to be incorporated with other Li-ion battery materials that

can be scaled in 3D architectures. Other deposition techniques are being explored, such as

plasma-assisted direct vapor deposition. Nevertheless, a decrease in LiPON ionic conductivity

was observed using this method (10-7 to 10-10 S/cm) (Kim 2009).

Among the solid lithium ion electrolytes, lithium aluminum silicate (LASO) is a promising candidate for solid lithium ion conductors (Nagel 1982). With an ionic conductivity

of 10-5 S/cm, it is comparable to other solid ceramic electrolytes.

Figure 1.7: Phase equilibrium diagram of the Li2O-Al2O3-SiO2 system (Roy 1949).

Amorphous LASO is a moderate Li+ ion conductor (Pechenik 1988). This means that there is little grain boundary resistance, an important characteristic of a solid electrolyte in a lithium ion cell. Grain boundary resistance reduces the efficiency of the cell, and a good candidate for solid state LIB should have low lithium ion resistance in a polycrystalline or amorphous film. Amorphous films are generally easier to manufacture than crystalline structures.

Crystalline, or β-eucryptite LASO, is a quartz structure and has a one-dimensional channel

(called a quartz channel) along the crystal c-axial direction. Because Li+ ions alternately occupy the site in the quartz channel, there is strong anisotropic ionic conduction in this direction. There is also evidence to suggest that there is high ionic conduction at boundaries between β-eucryptite and other materials (Shin-ichi 2004).

LASO thin films have been produced using pulsed laser deposition. Activation energy of

LASO is calculated to be 0.79 eV for a film thickness of 38 nm in the range of 500-700K

(Jochum 2010). However, pulsed laser deposition is a thin-film growth technique that is not

scalable for mass manufacturing.

1.4.5 Composite Lithium Ion Conductors

Efforts to further improve the capacity and efficiency of Li-ion batteries have led to the

exploration of composite organic-inorganic solid electrolytes to combine the material properties

of both inorganic solid electrolytes and organic solid electrolytes. These properties are

particularly important for integration with the next-generation Li-ion battery materials, which

typically have a large volumetric expansion upon lithium intercalation (e.g. ~300% for silicon)

(Ge, Rong, Fang and Zhou 2012). In order to mitigate the stresses that occur upon lithium

intercalation of these materials, robust solid electrolyte materials need to be explored. A

combination of organic and inorganic solid electrolytes have previously been applied as hybrid

heterojunctions in solar cells (Heo 2013). The advantages of these hybrid structures are the low-

cost solution based on construction methods for producing high-efficiency solar cells. The cells

constructed of mesoporous TiO2 coated in poly-triarylamine yields a power conversion

efficiency of 12.0% (Heo 2013).

Lithium ion conductive Li1.5Al0.5Ge1.5(PO4)3 coated with polypropylene was used as an

inorganic-organic composite separator for Li-ion battery applications. The lithium ion

conducting glass has a high bulk ion conductivity (10-4 S/cm), and the application of the glass

increased the eletrochemical window to be applied into 5V Li-ion battery cells (Shi 2015). The ion conductivity of the polypropylene separator was measured as 3.02 × 10-4 S/cm, and combined impedance of the composite coating increases due to resistance for lithium ion transport (2.49×10-5 S/cm). Although the combined impedance of the composite structure

decreases, the combination of the material properties such as expansion, higher thermal stability,

and higher wettability could offset the decrease in lithium ion transport.

1.5 Thin Film Deposition Techniques for Solid Electrolytes

Chemical vapor deposition (CVD) is a type of deposition technique in which chemical

precursors are introduced into a reaction vessel and chemical reactions produce the film on the

substrate. The films can be grown by one of two mechanisms. One mechanism is the chemical

precursors react in the vapor phase, then the products condense on the substrate and grow into

the film. The other mechanism is chemical reactions by the precursors on the surface that grow

the film directly on the substrate.

1.5.1 Atomic Layer Deposition

Atomic layer deposition (ALD) is a thin film manufacturing technique that allows precise

control over film characteristics such as thickness and composition. The unique growth

mechanism of saturated surface reactions allows growth of thin films to be conformal and

pinhole free. This deposition method has been used in the semiconductor industry for the last

several years to precisely grow thin-film high-k dielectric films at a very small scale (George

2010).

For ALD, the deposition of aluminum oxide (Al2O3) is well characterized and serves as a standard to how atomic layer deposition progresses. The substrate temperature needs to be within a certain range called an "ALD window" where optimum deposition takes place. Too low

of a substrate temperature could causes condensation and excess deposition of precursors, while

too high of a temperature could decomposes the precursor, leading to uncontrolled film growth.

In the deposition of aluminum oxide, an aluminum precursor gas, usually trimethyl aluminum, is

pulsed into a vacuum chamber towards the substrate; the layer is to be grown epitaxially. This aluminum precursor reacts with the surface hydroxyl groups to form O-Al-(CH3)2 on the surface

of the substrate, while methane is the product. Once all the hydroxyl groups on the surface have

reacted, no more reactions take place. The precursor is purged by an inert gas to rid the chamber

of excess aluminum precursor gas. The next stage is for an oxidant to be introduced to the

chamber. Water is the usual oxidant of choice. Water reacts with the -CH3 groups on the surface

of the substrate to form hydroxyl groups, with methane as the product gas. At the end of the

oxidant pulse, a purge of the oxidant gas is required to prepare for the next pulse of aluminum

precursor. Now the aluminum metal is sandwiched between two oxygen atoms, giving one

sequence of a precursor and oxidant, or cycle. One cycle deposits a layer of alumina, and the

cycles are repeated as needed to build the thin film. Growth of aluminum oxide by ALD is

usually 1.2-1.4 Ǻ per cycle (George 2010).

Figure 1.8: Schematic representation of the atomic layer deposition (ALD) process using self-limiting surface chemistry. A metal-organic precursor flows into the chamber and reacts on the surface-terminated hydroxide group. After which, an oxidant reacts with the surface metal-organic species to form a metal oxide layer (George 2010).

Atomic layer deposition of metal oxides such as Al2O3 directly on lithium cobalt oxide

cathodes shows an unexpected increase in cycling stability and capacity retention (Ji Woo 2014).

However, because aluminum oxide is not a good lithium ion conductor, the deposited layer is on

the order of one nanometer or less (Jung, Lu, Cavanagh, Ban, Kim, Lee, George, Harris and

Dillon 2013). This shows that atomic layer deposition is already being used to improve the

quality of lithium ion batteries.

Over the last several years, more insight and knowledge of ALD thin films have been

brought to lithium ion battery production. An increase of research on atomic layer deposition techniques for lithium ion batteries means that the field is growing and can aid in the implementation of better lithium ion battery technology (Liu and Sun 2015).

1.5.2 Initiated Chemical Vapor Deposition

Initiated Chemical Vapor Deposition (iCVD) is a chemical vapor deposition method

suitable for polymer thin film deposition. In iCVD, a tantalum filament array is resistively heated over a substrate. The heated temperature is typically in the range of 200-400 oC

depending on the initiator. Both monomer and initiator precursors flow into the vacuum

chamber. The filament array decomposes the initiator into radicals, and begins the free-radical

polymerization process of the monomer at the substrate surface, which is also typically heated,

but usually less than 50 oC for optimal adsorption and polymerization (Lau 2006).

The mechanism of iCVD polymerization is believed to occur through three major steps.

First, the thermal decomposition of a vaporized initiator forms primary radicals. Next, the

radicals diffuse and adsorb on the surface from the vapor phase. At the same time, monomers

also diffuse ad adsorb onto the surface. Lastly, monomers polymerize on the surface with typical

free-radical polymerization. The process of free-radical polymerization includes initiation,

propagation and termination events to form a continuous coating (Lau and Gleason 2006).

Unlike bulk polymerization, iCVD has the advantage of separating the initiator

decomposition temperature (through the filament array above the substrate), and the

polymerization temperature (via substrate temperature). This additional degree of freedom can

allow more control over the properties of the polymer coating (Lau and Gleason 2006).

Polymer siloxanes have been deposited in thin films for many years (Kwan 1997).

Although this deposition method does produce uniform polymer thin films, these films are not

inherently solid electrolytes for Li-ion battery applications. Before the polymer film can be

applied to Li-ion battery applications, these polymer films need to be soaked in lithium salts.

This process, called lithiation, usually involves soaking the polymer thin films in a liquid

electrolyte, such as 1 M LiClO4 in propylene carbonate to introduce lithium ions into the polymer film (Reeja-Jayan 2015).

1.6 Summary

Advancing of Li-ion battery technology for incorporation into smaller footprints needs to incorporate advanced materials and 3D architectures. In order for this to be accomplished, an all solid state battery overcoming the solid-electrolyte interface that hinders liquid electrolyte battery systems is needed. Lithium aluminum silicate (LASO) can be deposited conformally on

3D structures using atomic layer deposition. The polymer electrolyte PV4D4 soaked in liquid electrolyte salts have the advantage of mechanical robustness when incorporated into next generation Li-ion battery materials that tend to increase in volume upon lithium intercalation.

The goal of this work is to investigate the feasibility of combining the mechanical and electrical properties of both LASO and PV4D4 to produce a hybrid solid electrolyte suitable for incorporation into all solid state Li-ion cells.

CHAPTER 2 METHOD OF APPROACH

The goal of this study is to demonstrate the viability of using lithium aluminum silicate

based solid electrolytes in Li-ion battery materials. First, the growth rates of individual

component metal oxides grown by atomic layer deposition were explored. Component metal

oxide thin films were characterized for film thickness growth rates, and metal cation composition.

Incorporating the component metal oxide ALD cycles into LASO thin films were characterized

for growth rate, composition, and ion conductivity.

A hybrid ceramic-polymer solid electrolyte was constructed to combine the flexibility of

polymer solid electrolytes and the electrical resistance of the metal-oxide ceramic. The ceramic

was ALD LASO, and polymer solid electrolyte was iCVD PV4D4. Polymer solid electrolyte

was first deposited on indium tin oxide (ITO) substrate, soaked in lithium salt liquid electrolytes,

then ALD LASO was deposited. Ionic conductivities of the LASO film was compared with the

PV4D4-LASO hybrid heterostructure. Alternatively, PV4D4-LASO hybrid heterostructures

were also constructed with ALD LASO deposited on ITO substrates, followed by iCVD PV4D4

deposition, then the soaking in lithium salt. Ionic conductivity was also measured and compared

with the hybrid solid electrolyte iCVD PV4D4 as the first layer in the heterostructure.

Lastly, incorporation of LASO solid electrolyte thin films into current-generation and

next-generation lithium-ion battery applications was carried out. Applications include both 2D and 3D architectures. In-situ TEM electrochemical characterizations were conducted on LASO-

coated next-generation anode materials. Lithium-ion half- and full-cells were constructed and

tested using cyclic voltammetry and galvanostatic cycling to characterize LASO as a solid-state electrolyte in lithium-ion battery applications. 37

CHAPTER 3 EXPERIMENTAL METHOD

The atomic layer deposition chamber for synthesizing lithium aluminum silicate (LASO) thin films was completed in a custom built vacuum chamber. Schematics, precursors and experimental parameters are outlined, as well as more details in the Appendix. An in-situ

Fourier transform infrared spectroscopy (FTIR) ALD chamber was constructed to help identify surface reaction chemistry through the ALD process. Metrology techniques such as spectroscopic ellipsometry, x-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), and transmission electron microscopy (TEM) are used to verify film thickness and composition. Electrochemical properties of the ALD and iCVD synthesized film were obtained through electrochemical impedance spectroscopy, cyclic voltammetry (CV), and galvanic cycling (GC).

3.1 LASO Atomic Layer Deposition Chamber and In-situ FTIR Chamber

A low vacuum hot-wall reactor was used to deposit lithium aluminum silicate (LASO) thin films.

The operating pressure ranges from 30-70mTorr. Figure 3.1 is a schematic of the chamber. A

2.75" outer-diameter 6-way cross served as the chamber, with one port connected to the pumping line, three ports connected to precursors, one port connected to nitrogen purge gas inflow, and one connected to a loading door. Heating wires and type K thermocouples were connected on the outer wall of the chamber for heating, with seven channel temperature controller from

Omega Engineering. The walls of the chamber were heated to 150oC to limit precursor condensation on the chamber walls.

Lithium tert-butoxide (LTB) and trimethyl aluminum (TMA) were used as the lithium and aluminum precursors, respectively. Tris-(tertbutoxy) silanol (TTBS) and tetraethyl orthosilicate (TEOS) were implemented as silicon precursors. Deionized water was used as the oxidant to the organometallic precursors. More chemical property information is found in Table

3.1.

Table 3.1. Properties of LASO precursors.

Lithium tert- Trimethyl Tris(tert-buxoxy) Tetraethyl butoxide (LTB) aluminum silanol (TTBS) Orthosilicate (TMA) (TEOS) Stream Company Sigma-Aldrich Sigma-Aldrich Sigma-Aldrich Chemicals Purity 98% 97% 99.999% 99.999% CAS Number 1907-33-1 75-24-1 18166-43-3 78-10-4 Phase Solid Liquid Solid Liquid

Formula LiOC4H9 Al(CH3)3 ((CH3)3CO)3SiOH Si(OCH2CH3)4

Structure

160 oC Melting Point 15 oC 63-65 oC - (decomp) o Boiling Point - 126 oC 205-210 oC 205-210 C 39

To deliver sufficient precursor into the chamber, solid precursor housings were heated.

The LTB housing was heated to 150 oC, with gas delivery lines heated to 160oC. The TTBS

housing was heated to 45 oC, and the gas delivery lines were heated to 75 oC. Liquid precursor

housings (TMA and TEOS) were not heated due to the high vapor pressure at room temperature

(>1 Torr for TMA), and gas lines were heated to 60 oC. The aluminum precursor was controlled

by a mass flow controller to 0.15 sccm. The deionized water oxidant was enclosed in a 100 mL

glass cylinder and vaporized flow was throttled using a needle valve.

For a complex solid oxide to be deposited using ALD, multiple precursor cycles were

used. A precursor cycle is defined as a sequential order of precursor pulse into the chamber:

ultra-high purity nitrogen gas flowed to purge the chamber from excess precursor species, an oxidant pulse into the chamber, and lastly another inert purge gas to drive excess oxidants from the reactor.

For the deposition of aluminum oxide in the chamber, a five second pulse of TMA was

followed by a 45 seconds of pump down time where the roughing pump vacates the chamber.

The nitrogen purge gas flowed into the chamber at 50 sccm for 20 seconds, after which the

chamber was pumped down for 40 seconds. Deionized water vapor was then pulsed into the

chamber for ten seconds, followed by 50 seconds of pump down time. Nitrogen purge gas

flowed through the system for 20 seconds, and a 40 second pump down time completed the cycle. An outline of the deposition sequences is provided in Table 3.2.

Deposition of LASO was approached as a solid solution of each constituent metal oxide:

LiOH, Al2O3, and SiO2. Each individual cycle is combined to produce one global cycle. One

global cycle consists of a number of Al2O3 cycles, followed by LiOH cycles, then SiO2 cycles

and lastly by LiOH cycles. For example, the LASO deposition notated by global cycle of

(10Al:3Li:4Si:3Li), designates 10 aluminum oxide cycles followed by three lithium oxide cycles,

then four silicon oxide cycles and finally three lithium oxide cycles. Each global cycle is then

repeated to give added thickness at constant composition. Deposition of LASO was performed on silicon substrates for thickness analysis, and on germanium substrates to allow composition analysis of the deposited films. Aluminum oxide was usually deposited first due to its high reactivity to surface hydroxyl groups.

Table 3.2. Cycle sequences implemented for atomic layer deposition of LASO using TMA/H2O, LTB/H2O and TEOS/H2O (or TTBS/H2O) chemistry.

Chemistry TMA/H2O LTB/H2O TEOS/H2O or TTBS/H2O Precursor Exposure (sec) 5 10 10 Precursor Pump-down (sec) 45 50 50 Purge (sec) 20 20 20 Purge Pump-down (sec) 40 40 40 Oxidant Exposure (sec) 10 10 10 Oxidant Pump-down 50 50 50 Purge (sec) 20 20 20 Purge Pump-down (sec) 40 40 40

A hot-wall ALD chamber was built inside a Nexus 670 FTIR spectrometer. This allowed

spectra to be obtained during the atomic layer deposition process to evaluate surface chemistry

and chemistry changes that occured with each pulse of precursor chemistries and oxidant

reaction. The custom made in-situ FTIR chamber consisted of a 2.75" OD six-way cube welded

to two 2.75" OD gate valves. The gate valves isolated the IR transparent KBr windows from the

chamber to prevent ALD reaction on the surface. The two 38 mm diameter, 6 mm thick KBr

windows were held in place using two 2.75" OD window flanges. The volume between the gate

valves and the KBr windows were constantly pumped using a Leybold mechanical rotary vane

pump achieving a base pressure of 30 mTorr.

A high surface area was needed to obtain sufficient signal of the surface species in ALD

in-situ FTIR. Substrates were prepared using zirconium oxide (ZrO2) nanoparticles with an average particle size of 20 nm, hydraulically pressed for 10 seconds onto corrosion resistant grade stainless steel mesh. The hydraulic press was operated at 7,000 lb/in2 pressure, and the line

density of the stainless steel mesh 100 lines per inch. Excess nanoparticles were removed from

the surface of the grid with a razor blade. The ZrO2 nanoparticles yielded a specific surface area

greater than 25 m2/g. This preparation allowed sufficient IR absorption in transmission mode for

surface species analysis during metal oxide ALD. The mesh was mounted to an 18 mm OD, 12 mm ID stainless steel disk spot-welded to two nichrome wires for resistive heating. A type-K

thermocouple was welded to the sample holder and the temperature is monitored using an

Omega 7 Channel Temperature PID controller at a deposition temperature of 225oC.

The in-situ FTIR chamber was set in a Thermo Nicolet Nexus 670 FTIR spectrometer

using a germanium on cesium iodide beam splitter. IR wavenumbers from 4000 cm-1 to 400 cm-1

were detected by a deuterated triglycine sulphate (DTGS) pyroelectric IR detector. The spectrometer is purged with dry and CO2-free air from a purge gas generator.

3.2 iCVD Process

Initiated chemical vapor deposition (iCVD) of polymer films has been studied for several

years (Kwan and Gleason 1997). For this study, poly(1,3,5,7-tetravinyl-1,3,5,7-

tetramethylcyclotetrasiloxane) (PV4D4) was deposited using a custom built chemical vapor

deposition (CVD) chamber 10 inches in diameter and 1.25 inches in height. The top of the

chamber had a glass window that facilitates real-time monitoring of the deposition process. The

chamber had an array of Chromaloy O filaments consisting of 14 filaments spaced 1.5 cm apart 42

and suspended 2.5 cm over the substrate of 100-mm-diameter Si (100) wafer (Lau 2000). The array was resistively heated using a DC power supply and substrate temperature was controlled by a backside-recirculating water heater/chiller. The substrate temperature was monitored using embedded thermocouple sensors (Lau 2006). Pressure of the reactor was monitored using a

Baratron capacitance manometer in conjunction with a down-stream throttle valve. Vacuum was achieved using a combination of a Leybold roots blower and BOC Edwards dry pump (Lau and

Gleason 2006).

Figure 3.2: Schematic of a iCVD chamber where precursor flows into the chamber and decomposes on a wire array above the substrate to begin the polymerization process on the surface (Chen 2015).

The 1,3,5,7-tetravinyl-1,3,5,7-tetramethylcyclotetrasiloxane (V4D4) monomer precursor housing was heated to 110 oC. Both V4D4 monomer and tertbutyl peroxide (TBP) initiator were delivered into the reactor at 1.0 and 0.5 sccm, respectively using regulated needle valves on one side of the reactor. The substrate temperature was maintained at 55 oC, and the pressure of the reactor during deposition was maintained at 350 mTorr. The filament array was held at a constant 300 oC to ensure decomposition of initiator molecules (Trujillo 2010). A JDS Uniphase

He-Ne laser interferometer system monitored the deposition rate in real time (Lau and Gleason

2006). 43

Samples of PV4D4 were deposited on quartz substrates with 1 mm indium tin oxide (Q-

ITO). The ITO served as a back electrode for electrochemical characterization of the PV4D4 films (Chen, Reeja-Jayan, Lau, Moni, Liu, Dunn and Gleason 2015). Before deposition, one section of the Q-ITO substrate was masked with Kapton tape to allow access to the back electrode.

Samples were then soaked for three days in a 1 M LiClO4 in propylene carbonate solution. The soaking process occurred in an argon-filled glovebox with O2 and H2O concentrations below 0.1 ppm. After the soaking films, the PV4D4 samples were rinsed in fresh propylene carbonate and annealed in the glovebox for 1 hour at 110 oC to remove solvent (Reeja-

Jayan 2015).

Table 3.3. Precursor chemical properties for polymer iCVD deposition.

1,3,5,7-tetravinyl-1,3,5,7- di-tert-butyl peroxide tetramethylcyclotetrasiloxane (TBP) (V4D4) Company Gelest, Inc. Sigma-Aldrich Product code SIT7900.0 168521 Purity >95% 98% CAS Number 2554-06-5 110-05-4 Phase Liquid Liquid

Formula C12H24O4Si4 (CH3)3COOC(CH3)3

Structure

Melting Point -43 oC - Boiling Point 110 oC (10 mmHg) 109 oC

3.3 Substrates and Nanowires as Electrodes

Several substrates were used to deposit ALD LASO for characterization. One such substrate was silicon (001) with a 20 nm native oxide layer. Silicon pieces 2 cm × 2 cm were cut from a three inch wafer, which had a resistivity rating of 0.1-0.001 Ω/cm from p-doping. In order to verify silicon deposition in the LASO film, 1 cm × 1 cm pieces from a three inch germanium (111) EPI-rated wafer with a resistivity of >40 Ω/cm, 0.4 mm thick were used.

Native surface germanium oxide was factored into thickness and composition calculations.

Current-generation Li-ion battery materials were also used as electrode and substrate surfaces. Porous carbon consisted of 83% mesocarbon microbeads, 7% carboxymethyl cellulose,

8% graphite, and 2% Ketjen black. Multiple synthesis methods were used to create 2D substrates and 3D pillar arrays. The 2D substrates consisted of an aluminum current collector with the carbon slurry deposited on top using the doctor blade technique. 3D pillar arrays were constructed using a custom designed mold, where the carbon slurry was injected and vacuumed concurrently (Daniel 2015). Lithium cobalt oxide cathodes for 2D half-cell and full-cell

lithium-ion battery testing was drop-cast into a Kapton-tape mask to isolate the active material

and limit effects from attaching leads to the micro-scale thin film electrochemical cell.

Germanium/silicon (GeSi) nanowires were obtained from the Center for Integrated

Nanotechnology (CINT) and Sandia National Laboratories (Dayeh 2011). These next-generation

Li-ion battery anodes were manufactured using vapor-liquid-solid chemical vapor deposition to manufacture nanowires approximately 100 nm in diameter and 2 μm in length. These GeSi nanowires were used as anodes, electrodes and substrates to deposit ALD LASO for in-situ electrochemical characterizations.

3.4 Thin Film Characterization Techniques

Several methods were used to help characterize the deposited material. Thicknesses were

measured using spectroscopic ellipsometry, and composition of the films was calculated using x-

ray photoelectron spectroscopy. Other characterization methods, such as in-situ Fourier

Transform Infrared Spectroscopy, characterized surface ligands during the atomic layer

deposition process.

3.4.1 Spectroscopic Ellipsometry

Spectroscopic ellipsometry is a nondestructive optical measurement technique that uses

polarized light to investigate the dielectric properties of transparent thin films. Light was

polarized using a rotating polarizer. After reflection off a sample, the change in polarization was

measured and quantified as amplitude ratio (Ψ) and phase difference (Δ) that were parallel (p-)

and perpendicular (s-) to the sample surface.

Figure 3.3: Schematic of spectroscopic ellipsometry measurement principles showing the direction changes in polarized light after striking a sample (Ohkoshi 2005).

The amplitude ratio and phase difference were related to the Fresnel reflection coefficients by:

(3-1) = tan( ) = 푖훥 푅푝 휌 훹 푒 푠 푅 46

Thin film and multilayer structures have multiple interfaces which reflect and refract light. The

Fresnel reflection coefficients result from multiple reflections of incident light in the thin film

and lead to an infinite series. The infinite series can be converged into reflection coefficients for reflected light, designated Rp and Rs.

Figure 3.4: An example of an optical model to measure thickness of thin film oxides using spectroscopic ellipsometry (Li 2011).

Film thickness can be back-calculated from these convergent infinite series and the

optical constants refractive index (n) and extinction coefficient (k) from the following equation:

( ) ( ) (3-2) = and = 푝 푝 ( ) 푠 푠 ( ) 푝 푟01+푟12exp −푖2훽 푠 푟01+푟12exp −푖2훽 푝 푝 푠 푠 푅 1+푟01푟12exp −푖2훽 푅 1+푟01푟12exp −푖2훽 In the previous equation, β is a function of film thickness d1, and is directly proportional to the

complex refractive index of the film by:

(3-3) = 2 cos 푑푖 훽 휋 � � 푛푖 θi For the materials휆 used in this study, spectroscopic ellipsometry data of thin films were

gathered and analyzed by fitting the empirical correlation of Cauchy or Sellmeir dispersion

models. The Cauchy dispersion model is defined as:

(3-4) ( ) = + + 퐵푛 퐶푛 푛 휆 퐴푛 2 where the terms An, Bn, and C휆n are휆 coefficients that describe a material’s index of refraction over a range of wavelengths. The Sellmeier dispersion model is defined as :

1 (3-5) = + 1 −퐴 2 2 퐵 where A is the푛 slope− of 휆 vs and B is the intercept of the plot at λ = ∞. The optical 1 1 2 2 푛 −1 휆 constants for each material are listed in Table 3.4.

Table 3.4. Optical models and constants for Al2O3, SiO2, and LiOH thin films.

Material Model Constants Reference Al2O3 Cauchy A = 1.62 (Langereis 2009) -3 Bn = 2.6 ×10 μm -6 4 Cn = 2.0 ×10 μm

SiO2 Cauchy A = 1.46 (Fujiwara 2007) -3 Bn = 2.38 ×10 μm -6 4 Cn = 9.75 × 10 μm

LiOH Sellmeier A = 1.10 ×10-6 μm2 (D. Shannon 2002) B = 0.96

Measurements were performed using a J. A. Woollam Co. M-88 spectroscopic ellipsometer with wavelengths from 280-760 nm. Empirical models for the constituent metal oxides were used for calculating film thickness using Woollam Wvase32 software and chi- squared and minimized squared error fitting techniques.

A spectroscopic ellipsometry model must be written to extract the thickness of the thin film, as only polarization change is measured. Figure 3.5 shows an ALD-deposited aluminum

oxide film fitted to an aluminum oxide Cauchy dispersion model that resulted in a thickness

calculation of 43 Å.

35 Ψ Al O Film 2 3 140 Ψ Model Fit 30 ∆ Al O Film 2 3 130 ∆ Model Fit 25 120 ) 20 110

degrees ( (degrees) 15

100 ∆ Ψ

10 90

5 80

300 400 500 600 700 800 Wavelength (λ) (nm)

Figure 3.5: Experimental spectroscopic ellipsometry data obtained from 30 cycles of ALD Al2O3 deposited on Si (110) substrate. Sample thickness was 43 Å.

3.4.2 Fourier Transform Infrared Spectroscopy

An FTIR spectrometer operates by irradiating a sample with infrared wavelengths, which

trigger resonant molecular thermal vibration. These vibrations depended on the bond strength (or

energy) in the molecule. When the sample is irradiated, a specific energy photon is absorbed

into a resonate bond, usually wavelengths ranging from 2.5-25 μm. These wavelengths are at the

same energy range as molecular bond vibration modes. The photons that are detected after

transmission through the sample result in an absorption signal, which is associated with certain characterized bonds. Observable modes in FTIR are generally limited to stretching or bending modes that result in dipole changes in the molecule.

An FTIR spectrometer works by shining a polychromatic infrared beam into a beam splitter. One beam reflects off a stationary mirror, while the other is reflected by a moving mirror attached to a motor. The moving mirror changes the length of the reflection. When the two beams converge an interferogram is produced. The beam is transmitted through the sample,

where certain wavelengths are absorbed, and a detector quantifies the wavelengths as a function

of time. Using Fourier transform, the time-space interferogram is reconstructed into a

frequency-space spectrum, and can be easily analyzed for absorbance peaks.

There are several advantages of using FTIR spectroscopy. Because the information from

all wavelengths is collected simultaneously, the signal-to-noise ratio is higher than for a given

scan-time of a single wavelength. The throughput is also higher due to the interferometer rather

than monochromatic light. These advantages allow for more precise measurement in less time.

Molecular bonds can be modeled as a spring system, or a set of linear harmonic

oscillators, which can be simplified using Hooke’s law:

+ (3-6) =

푚1 푚2 푣 푘�푓 � � 푚1푚2 where ν is the vibrational frequency, k is a scaling constant, f is the force constant, and m1 and

m2 are masses of the atom in the molecule. There are two main types of molecular vibrations:

stretching and bending. For example, water has three fundamental vibrations due to its nonlinear molecular structure: symmetrical stretching, asymmetrical stretching, and scissoring. Each of these modes have a specific wavelength for their respective bond energy (Griffiths 2007).

Expanding this idea to the precursors, lithium tert-butoxide, trimethylaluminum and

tetraethyl orthosilicate, there are various wavenumbers that are associated with their respective

bond energies (see Table 3.5). To better aid in identifying surface species during the ALD

process, in-situ FTIR spectroscopy was used. In each half-reaction of the ALD process, the reaction is terminated with a different ligand. Acquiring FTIR spectra following each half cycle represents the changes that occur on the surface during atomic layer deposition.

Table 3.5. Infrared vibrational wavenumbers for lithium tert-butoxide (LTB) using in-situ FTIR spectroscopy (Cavanagh 2010).

Region Wavenumber (cm-1) O-H stretch 3720-3320 Li-OH hydroxyl stretch 3672 H2O in LiOH·H2O 3570-3474 C-H stretching 3000-2817 Asymetric C-O stretch (from Li2CO3) 1429-1496 C-O stretch (unreacted t-butoxide groups) 1215-1174 Symetric C-O stretch (from Li2CO3) 1088-1000

Table 3.6. Infrared vibrational wavenumbers for tetraethyl orthosilicate (TEOS) using in-situ FTIR spectroscopy (Ferguson 2004).

Region Wavenumber (cm-1) SiO-H stretch 3743 C-H stretch in *Si(OCH3CH3)x on surface 2979-2901 SiO2 longitudinal optical phonon 1250-1075 Asymetric C-O stretch 1484-1298 C-O stretch 1250 Symetric C-O stretch 1170-970

Table 3.7. Infrared vibrational wavenumbers for trimethylaluminum (TMA) using in-situ FTIR spectroscopy (Goldstein 2008).

Region Wavenumber (cm-1) O-H stretch 3730-3640 AlO-H stretch 3770-3650 C-H stretching from *AlCH3 2980-2820 Bulk Al2O3 phonon 1000-900

In-situ FTIR spectra were collected from a dedicated chamber build inside a Nexus 670

FTIR spectrometer from Thermo Nicolet shown in Figure 3.1. Figure 3.6 shows in-situ FTIR spectra obtained during the first LTB and H2O exposures in the ALD of lithium oxide using LTB 51

and water chemistry on a pressed ZrO2 nanoparticle substrate. There are three main areas of

absorptions to observe during the ALD process: the generation or consumption of surface

-1 hydroxyl (O-H stretching at 3800-3500 cm ) and Li-OC(CH3)3 surface species (C-H stretching

-1 at 3000-2800 cm ), as well as the region of C-O stretching particular to Li2CO3 (peaks at ~1200

and ~1000).

H2O (1)

Absorbance (A.U.) Absorbance LTB (1)

4000 3500 3000 2500 2000 1500 1000 Wavenumber (cm-1)

Figure 3.6: In-situ difference FTIR spectra obtained during the first LTB and H2O exposures during ALD deposition cycles. The absorption regions of interest include the surface hydroxyl region (O-H stretching at 3800-3550 cm-1), surface -1 lithium butoxide species (*LiOC(CH3)3)(C-H region at 3000-2800 cm and C-O stretching in 1215-970 cm-1).

3.4.3 X-ray Photoelectron Spectroscopy (XPS)

The major technique used to analyze the composition of these complex metal oxide films

is X-ray photoelectron spectroscopy (XPS). This surface-sensitive quantitative technique can

measure elemental composition at parts per thousand on the top portion of the sample (within

three mean-free path-lengths of the scattered photon). 52

Spectra are obtained by irradiating a material with X-rays while simultaneously

measuring kinetic energy and electron quantity. When a sample is irradiated, core electrons are

ejected based on the photoelectric effect. Ejected electrons have specific kinetic energy based on

the element they originated. All elements except hydrogen and helium can be analyzed using

this technique. Using the conservation of energy, atomic identification in the sample can be

obtained by calculating the binding energy of the core electrons:

= ( + ) (3-7)

푏푖푛푑푖푛푔 푝ℎ표푡표푛 푘푖푛푒푡푖푐 where Ebinding퐸 is the binding퐸 energy− 퐸 of the electron,휙 Ephoton is the energy of the X-ray being used,

Ekinetic is the kinetic energy of the electron as measured by the instrument, and ϕ is the work

function dependent on both the spectrometer and the material.

XPS can also provide qualitative chemical bonding information, as binding energy is

influenced by the molecular environment such as oxidation state and electronegativity of the

surrounding elements. Measurements were carried out in an Axis Ultra DLD by Kratos using an

x-ray source of non-monochromatic Al Kα at 1486.6 eV. Survey spectra were collected using a

pass energy of 160 eV, a step size of 1 eV, and a dwell time of 0.1 µs. Detail scans were

obtained using a pass energy of 20 eV, step size of 0.1 eV and a dwell time of 0.2 µs. The

collected spectra were all referenced to C-C 1s binding energy of 284.8 eV.

For LASO film deposited using ALD, the characteristic Si 2p, Al 2p, and Li 1s

photoemission peaks occur at ~100 eV, ~75 eV, and ~55 eV respectively. Integrating the

intensity of a specific peak is proportional to the flux of photoelectrons. The intensities are

dependent on the x-ray flux of the instrument, mean free path of the sample photons, and

efficiencies of the measurement device. Using empirical methods, surface composition can be

calculated by incorporating the atomic sensitivity factors (ASF) of the respective elements:

/ = (3-8) 푘 / 푘 푘 푛퐼 퐴푆퐹 퐶 푘=1 푘 푘 where I refers to the∑ integrated퐼 퐴푆퐹 intensity of the photoemission peak for element k. The relative sensitivity factors are empirically developed, and a summary of the sensitivity factors used for the Kratos system are listed in Table 3.8 (Wagner 1981).

Table 3.8. Atomic sensitivity factor (ASF) values for photoemission peak values in LASO thin film samples (Briggs 1990).

Z Element Line B.E. (eV) ASF 3 Li 1s 54.8 0.025 6 C 1s 284.0 0.278 8 O 1s 529.0 0.780 13 Al 2p 72.9 0.193 14 Si 2p 99.7 0.328

O 1s

C 1s Li 1s

Al 2p

Counts (CPS) Si 2p

600 500 400 300 200 100 0 Binding Energy (eV)

3.4.5 Scanning Electron Microscopy

Scanning electron microscopy (SEM) uses accelerated electrons from an anode in a field potential on order of 20 kV, condensed using electromagnetic coils in a pole piece to focus the 54

electrons to a diameter of approximately 6 nm. The condensed electron beam strikes the sample,

and generates secondary electrons. These secondary electrons are then detected by a scintillator-

photomultiplier to produce an image corresponding to the topography of the sample.

SEM images were obtained using an FEI Nova 600 Dual Beam-SEM/FIB system under

10 kV tension. Figure 3.8 shows a SEM image of a carbon nanopillar array coated in 23 nm

LASO solid electrolyte.

Figure 3.8: A scanning electron microscope (SEM) image of a single carbon pillar electrode coated with a 23nm LiAlSiO coating imaged under an excitation voltage of 10 kV.

3.4.6 Transmission Electron Microscopy

Transmission electron microscopy allows the imaging of nanomaterials by utilizing electrons and their properties. These microscopes transmit a beam of electrons through a very thin specimen (maximum thickness is usually 100 nm). As the electron beam interacts with the sample, an image is magnified and focused onto an imaging medium fluorescent screen or a sensor such as a CCD camera. Resolutions on the order of Angstroms are achievable due to the small de Broglie wavelength of electrons (Williams 2009). For a 300-kV electron, the electron relativistic wavelength is on order of 1.97 pm according to the following formula:

(3-9) = ( + ) ( ) ℎ푐 2 2 2 2 휆 푒 where T is the tension� 푇 푚(or푐 accelerating− 푚 푐 voltage of the instrument), me is the electron mass (9.109

× 10-31 kg), c is the relativistic speed of light (2.998 × 108 m/s), and h is Plank’s constant (6.626

× 10-34 N·m·s). The resolution of a microscopic image can be approximated using Rayleigh’s formula:

0.61 (3-10) = ( ) 휆 훿 where μ is the refractive휇 푠푖푛 훽 index of the medium, is the wavelength, and β is the semi-angle

collection of the magnifying lens. The best visible휆 light microscopes can give a resolution on

order of 350 nm, while current transmission electron microscopy instruments can on order of less

than 0.9 Å.

A transmission electron microscope produces an electron beam from a field emission gun

(FEG) or thermionic source (LaB6). The beam then passes through condenser lenses to help

focus the beam to the specimen. Some portions of the beam are diffracted from the sample,

while others are transmitted. The transmitted electrons are focused using the objective and

projection lenses to make the image.

TEM is used to confirm conformality and uniformity of ALD LASO films deposited on

nanowires and nanoparticles. High-resolution TEM (HRTEM) and scanning transmission electron microscopy (STEM) are characterization techniques which help resolve detail in the

LASO coating and the substrates. Standard images of the transmitted beam are bright field characterization techniques. Within the same system, diffraction patterns can also be obtained.

Figure 3.10: Selected area diffraction (SAD) pattern observed from a Ge0.4Si0.6 alloy nanowire.

Using a NanoFactory single-tilt sample holder available at the Center for Integrated

Nanotechnology at Sandia National Labs, electrochemical characterization of single nanowires coated in LASO solid electrolyte was completed. The nanowire substrates with LASO coating were attached to one end of the sample holder. A tungsten rod was scratched with lithium metal in a glove box and attached to the piezo-manipulator section of the sample holder. This allowed the lithium metal to be driven towards the sample substrate with sub-nanometer precision. A

LASO-coated nanowire was isolated and physical contact was made. 57

A TF 30 Tecnai TEM was used for imaging the nanowires while lithium was driven into

the material. A Solartron Analytics Modulab equipped with a Potentiostat and Femto Ammeter

were used to measure the current and voltage of the open cell in the microscope. A schematic of

the cell is shown in Figure 3.12. Driving lithium into and out of the nanowires occurred using

potentiostatic (constant voltage) or cyclic voltammetry techniques (see Section 3.5.2).

Figure 3.11: Schematic of an in-situ TEM sample setup showing the working electrode (Ge/Si alloy) and counter electrode (Li metal) (Liu 2011).

3.4.7 Atomic Force Microscopy

In order to study the surface morphology of thin film deposits, atomic force microscopy was used. Atomic force microscopes (AFM) are capable of high resolution scanning probe microscopy on order of nanometers to evaluate the surface morphology of a thin film surface.

AFM works through the attractive or repulsive molecular interactions on a silicon tip a few nanometers in diameter. This tip is on a cantilever a few microns long with an aluminum backing which allows a laser to reflect into a photodiode sensor. The probe tip scans along the surface, and the forces slightly bend the cantilever and the laser deflection is measured to output the topology of the surface. This deflection measurement is incorporated into a proportional- integral-derivative (PID) feedback controller using probe height or probe tip force as the setpoint.

Most microscopes have three modes of operation: contact, non-contact, and tapping. In contact mode, the probe tip is dragged along the surface to generate topological features from the deflected laser beam on the cantilever tip. This mode usually achieves high resolution but degrades the probe tip rapidly under hard surfaces. Non-contact mode maintains a height of

10 - 100 Å above the surface. Occasionally the intermolecular forces of the surface have interference from surface moisture adsorbed on the surface, though probe life is extended. Under tapping mode, the probe is tapped against the surface with high speed, typically 50-500 kHz after optimizing on the resonant frequency of the cantilever, generating a topological image based on the height of the probe tip (from the deflected laser measurement).

Measurements were taken on a VEECO DI3100 Atomic Force Microscope with a scan size of 1 μm2 at a scan rate of 0.5 Hz with a resolution of 2 Å. Probe tips were Bruker aluminum-coated silicon tips with a tip diameter of about 50 nm, operating at a frequency of

~318 kHz. The surface roughness of a sample is calculated by the root mean square equation:

(3-91) ( ) = 2 ∑ 푍푖 − 푍푎푣푒 푅푀푆 � 푁 where Zi is the height of a sampled point and Zave is the height average of N sampled points. The

RMS value is obtained from analyzing 1 μm2 scans normalized for surface curvature using

Bruker Corporation Nanoscope Analysis AFM software.

3.5 Electrochemical Characterization Techniques

In order to verify the thin films as capable in electrochemical cells, specialized characterization methods were implemented. These techniques aid in measuring ionic conductivity as well as verify electrochemical properties in the solid electrolyte or battery 59

prototypes. Cyclic voltammetry aid in characterizing electrochemical reactions, while galvanic

titration or galvanic cycling measures the performance of battery cells after multiple charging or

discharging cycles.

3.5.1 Electrochemical Impedance Spectroscopy

Electrochemical Impedance spectroscopy (EIS) is a powerful electrical characterization

technique that allows properties of materials and their interfaces to be studied. These properties

can include ionic conduction, dielectric constants, semiconductor properties such as band gap,

and mixed electronic-ionic conductors (Barsoukov 2005).

A monochromatic voltage signal is applied to the sample at a set frequency, and the measured response is current. A frequency sweep is applied in which the voltage signal is applied over a variety of frequencies. The current is then calculated into impedance (resistance and phase shift) of the applied alternating voltage to better characterize the electrical properties of the sample. Impedance is represented as a complex quantity Z:

= + (3-102)

where R is the푍 real 푅part 푗푋of impedance (resistance), and the imaginary is reactance (X).

The impedance is characterized by using equivalent circuits to model the electrical

character of the sample. Equivalent circuits consist of three basic passive ideal circuit elements:

resistor (R), inductor (L), and capacitor (C). Fundamental relationships and the derived

impedance responses are listed in Table 3.9:

Table 3.9. List of passive circuit elements used in equivalent circuit modeling in impedance spectroscopy with fundamental relationships and impedance response equations.

Element Symbol Relationship Impedance Response Resistor (R) ( ) = ( ) = = Inductor (L) 푉 푡 푅퐼 푡 푅 훥푉 ( ) = ( ) 푍 = = 푅 훥퐼 Capacitor (C) 휕( ) 퐿 훥푉 1 푉(푡) = 퐿 퐼 푡 푍 = = 푗휔퐿 휕푡( ) 훥퐼 휕푞 푡 훥푉 퐶 푡 푍퐶 휕푉 푡 훥퐼 푗휔퐶 Care in building equivalent circuits is needed for impedance analysis, as more than one model can fit the same experimental data. Thus, the equivalent circuit should also represent a plausible physical model of the sample. Figure 3.12 to 3.13 demonstrate examples of impedance

spectra models and the associated circuit diagram.

160 C2 = 0.1

120

C2 = 1 ) Ω

80 -Z'' ( -Z'' R1=100 40 C1=0.1

0 0 40 80 120 160 Z' (Ω)

Figure 3.12: Example impedance model where the ratio of C2 in the circuit is changed from a ratio or 0.1 (---) to 1 (---).

1.5

R1=1 C=1

R =2 1.0 2 ) Ω

Z'' ( Z''

0.5 R2=1

0.0 1.0 1.5 2.0 2.5 3.0 Z' (Ω)

(ITO) on a quartz substrate (CEC020Q, Präzisions Glas & Optik GmbH, Germany). The ITO serves as the back electrode for the measurements. The top electrodes comprised of a 10 nm adhesion layer of titanium followed by 100 nm of platinum were deposited using electron beam deposition into an array of 660 µm dots. Titanium probes were used to contact the platinum electrodes and the back electrode. A schematic of the sample setup is shown in Figure 3.14.

Measurements on samples with top electrodes were taken using an HP 4284A Precision

LCR Meter capable of measuring 20 Hz to 1 MHz at 13 mVrms amplitude using

Micromanipulator Fine-tip 7A tungsten probe tip with a diameter of 0.005” with a bendable nickel shank.

Measurements using hanging mercury drop electrodes were done by using a Solartron

Analytical 1252A frequency response analyzer connected with a Solartron SI 1287 electrochemical interface. Data was collected using Scribner Associates ZPlot, and ZView impedance analysis software was used to verify equivalent circuit modeling. A Nyquist impedance plot obtained from 21 nm ALD LASO sample on indium tin oxide (ITO) is shown in

Figure 3.15.

2000

1500 )

Ω 1000

-Z'' ( -Z''

500

0 500 1000 1500 2000 2500 3000 3500 Z' (Ω)

Figure 3.15: Nyquist impedance spectra obtained from 21 nm ALD LASO sample deposited on an indium tin oxide (ITO) substrate (● ). Using the simplified Randle's circuit, a model (–––) is superimposed on the data. Ionic conductivity for the LASO films were calculated by the relation in Pouillet’s Law:

(3-113) = 푡 휎 푏푢푙푘 where t is the푅 thickness퐴 of the thin film, Rbulk is the extrapolated resistance (dc) from the fitted

circuit model, and A is the area of the electrodes. Using the model in Figure 3.15, the calculated

impedance of the 21 nm LASO film is 2.1 × 10-8 S/cm.

3.5.2 Cyclic Voltammetry

The complete electrochemical behavior of a system can be obtained at different potentials

and recording the current-time curves to yield a three-dimensional i-t-E surface. The

accumulation of these curves can be tedious and time-intensive. In an effort to reduce time in gaining information, sweeping the potential over time can give more information as long as the

potential steps are closely spaced. Cyclic voltammetry (CV) was one such method to record

current versus time (Bard 2001).

A standard CV used a three-electrode setup, which consisted of a lithium foil reference electrode, a working electrode (the electrode being studied), and a lithium foil counter all

submerged in a 1 M LiClO4 solution with 1:1 ethylene carbonate: dimethyl carbonate as the

solvent. The reference electrode detected the potential on the working electrode, while the counter electrode measured the current flow through the working electrode. In this study a potential sweep from an initial voltage linearly increased at 50 mV/s to 1.2 V and cycled to 0 V.

The potential is applied between the working electrode and the reference electrode, while the current is measured between the working electrode and the counter electrode.

For in-situ TEM electrochemical CV measurements, the potential sweep started at 0 V

and proceeded to -8 V at a scan rate of 50 mV/s. The potential then increased to 4.5 V, and returned to 0 V.

0.6

0.4

0.2 ) µΑ

0.0 Current ( -0.2

-0.4

0.0 0.2 0.4 0.6 0.8 1.0 1.2 Potential (V)

3.5.3 Galvanostatic Cycling

A galvanostat, or amperostat, controls the amount of current through an electrolytic cell to be constant between working and reference electrodes. By measuring the amount of charge that is transferred in an electrochemical system, information regarding electrochemical reactions on the surface of electrodes, and total chemical reactions, can be quantified. The potential in this system is a dependent variable as a function of time. Experiments were run to determine charge transfer upon lithiation intercalation into Li-ion battery materials (such as carbon or lithium cobalt oxide), and efficiency of the cycling material (Bard and Faulkner 2001).

The coulombic efficiency and irreversible capacity were measured using total capacity in galvanostatic titrations:

(3-14) = 푑푖푠 푐 푄 휂 푐ℎ =푄 (3-125)

푖푟푟 푐ℎ 푑푖푠 where is the푄 coulombic푄 − efficiency푄 (%), is the irreversible capacity (mAh/g), is the

푐 푖푟푟 푐ℎ charge 휂capacity, and is the discharge capacity.푄 푄

푑푖푠 A three electrode푄 cell with 1 M LiClO4 dissolved in a 1:1 mixture of ethylene carbonate

and dimethyl carbonate solution was used for galvometric cycling. Li-ion battery materials

incorporating ALD LASO were cycled using lithium foil as the counter and reference electrodes.

0.7

0.6

0.5

0.4

(V) WE

E 0.3

0.2

0.1

50 100 150 200 250 300 Time (h)

Figure 3.17: Galvanostatic titration of 3D carbon pillar array with an ALD LASO coating of 26 nm in 1 M LiClO4 in 1:1 ethylene carbonate:dimethylcarbonate solution with lithium as the counter and reference electrodes.

CHAPTER 4 RESULTS

The first section discusses experiments verify atomic layer deposited constituent oxides

of lithium, aluminum, and silicon. In-situ FTIR was used to help identify surface chemistry as the films were deposited using atomic layer deposition. In-situ TEM electrochemical characterization was used to identify morphological changes that occur to silicon/germanium alloy nanowire anodes upon lithium intercalation using LASO as a solid electrolyte in a full-cell

setup. Further portions of this study contained the construction of a heterostrucutred solid electrolyte comprising of ALD LASO ceramic solid electrolyte and iCVD PV4D4 solid polymer electrolyte. Lastly, integration of the ALD LASO solid electrolyte incorporation into lithium-ion battery anode and cathode materials were investigated. The fabrication of a full 2D battery prototype was conducted and tested using cyclic voltammetry and galvanostatic cycling.

4.1 In-situ FTIR of LASO and Its Constituent Oxides

In-situ Fourier transform infrared spectroscopy (FTIR) was used to view possible surface chemistries contributing to atomic layer deposition of lithium aluminum silicate and its constituent oxides. Aluminum oxide was first studied using trimethylaluinum (TMA) and water

o on ZrO2 nanoparticles at 225 C. Figure 4.1 shows in-situ difference FTIR spectra for the first

five ALD cycles. Background was captured of the heated ZrO2 nanoparticles.

H2O (5)

TMA (5)

H2O (4)

TMA (4)

H2O (3)

TMA (3)

H2O (2)

TMA (2) Absorbance (A.U.)

H2O (1)

TMA (1)

4000 3500 3000 2500 2000 1500 Wavenumber (cm-1)

Figure 4.1: In-situ difference FTIR spectra for Al2O3 film deposited by alternating o exposures of TMA and water oxidant on ZrO2 nanoparticles at 225 C.

Upon the first TMA exposure (1A), IR absorption response showed a decrease of the

surface hydroxyl species (-OH) by a decrease in absorbance in the 3725 cm-1 region. The

-1 increase in the C-H stretching region at 2895 cm also showed an increase in the Al-CH3 surface

species, with more absorption peaks at the 1211 cm-1 region. The absorbance regions correspond

to previous literature reports (Goldstein, McCormick and George 2008). It follows that the

surface hydroxide species reacted with the methyl ligands in TMA to form water, which was

desorbed and purged with the reactant gasses.

When water was exposed to the surface, the IR absorbance was reversed as relative to the

initial TMA exposure. A decrease in the 1211 cm-1 and 2895 cm-1 regions, as well as a

regeneration of the surface hydroxyl groups on the surface (3725 cm-1) resulted. It follows that successful desorption of the methyl ligand (–CH3 species) occurred on the surface. It was

expected that water reacts with the surface methyl species to form methane and surface hydroxyl

groups. The methane was effectively released from the surface and the cycle is ready to start

again.

Over the course of 5 cycles, both the –OH and C-H regions had consistent absorbance

over the course of repeated TMA and H2O cycles. During the TMA exposure, C-H regions

showed positive absorbance, while a decrease in –OH absorbance was observed. When water

was exposed to the surface, the increase in –OH absorbance showed a regeneration of surface -OH species, and a decrease in C-H absorbance regions showed an evolution of the methyl species. Both the TMA and H2O show self-limited chemical reaction character by the

number of available sites for the reactions that occurred.

To study the growth of LiOH, in-situ FTIR studies were then conducted to characterize the alternating exposures of lithium tert-butoxide (LTB) and water. Figure 4.2 shows the difference spectra for the deposition of LiOH on ZrO2 nanoparticles.

H2O (4) LTB (4)

H2O (3)

LTB (3)

H2O (2)

LTB (2)

Absorbance (A.U.) H2O (1)

LTB (1)

4000 3500 3000 2500 2000 1500 Wavenumber (cm-1)

Figure 4.2: In-situ difference FTIR spectra for LiOH film deposited by alternating o exposures of LTB and water oxidant on ZrO2 nanoparticles at 225 C.

When the surface was exposed to LTB, absorption peaks were observed at 1200 cm-1 and

-1 2971 cm corresponding to C-O and C-H stretches of the tert-butoxy ligand (-OC(CH3)3)

(Comstock 2013). A decrease in the absorption peaks in the 3700cm-1 region, indicative of a

decrease in surface hydroxyl groups was observed as well. This showed successful LTB

adsorption to the surface.

After water exposure, an increase in the –OH absorbance is observed, as well as a

decrease in the absorbance of the 2971 cm-1 and 1200 cm-1 regions. This suggests the surface

tert-butoxy ligands that were present after the LTB exposure were afterward no longer present, and a regeneration of the surface hydroxyl species occurred. Continuing the exposure of alternating LTB/H2O cycles resulted in similar patterns for every LTB exposure, and water

exposure. Self-limited surface half reactions are present and the combination of in-situ FTIR and

depositions that demonstrate constant deposition rate suggest near-ideal atomic layer deposition mechanisms can exist for LTB and H2O chemistry.

In the attempt of the atomic layer deposition of silicon oxide film deposited with TEOS and water, the in-situ difference spectra were obtained during 4 cycles on ZrO2 nanoparticles

(Figure 4.3). When TEOS is exposed to the surface, negative absorption was observed at 3739

-1 cm . This indicates that the surface hydroxyl groups that exist on the ZrO2 nanoparticles have reacted with the TEOS precursor on the surface. Positive absorbance peaks at 2979, 2937, and

-1 2900 cm , vibration modes of C-H on the silicon ethyloxy ligand (SiOCH2CH3), also are evidence of adsorption of the silicon precursor. The in-situ FTIR difference spectra after the water exposure shows an increase in the absorption peak at 3739 cm-1, as well as negative absorption peaks in the C-H region.

The magnitude of the absorption peaks at 3739 cm-1, the 2900 cm-1 range, and 1200 cm-1 decreased after every TEOS precursor exposure. After ten exposures, the absorptions peaks in the difference spectra were severely diminished, signifying little change in the ligands on the

o surface. This signifies that TEOS and H2O chemistries do not deposit SiO2 at 225 C.

In-situ FTIR study of lithium aluminum oxide was conducted next. Figure 4.4 shows the deposition of aluminum and lithium oxides using TMA and LTB with water as the oxidant.

Depositions of lithium aluminum oxide consisted of one cycle of TMA and water, followed with a LTB pulse and water. The TMA/H2O/LTB/H2O designate a global cycle which was repeated.

Similar to the individual TMA/H2O and LTB/H2O chemistries, the assigned peaks at

3700 cm-1, 2900 cm-1, and 1200 cm-1 showed similar deposition in regard to the TMA or LTB ligands. The deposition of lithium aluminum oxide appeared to continue consistently from global cycle to global cycle, in a manner consistent with atomic layer deposition.

H2O (4)

TEOS (4)

H2O (3) TEOS (3)

H2O (2)

TEOS (2)

Absorbance (A.U.) H2O (1)

TEOS (1)

4000 3500 3000 2500 2000 1500 Wavenumber (cm-1)

Figure 4.3: In-situ difference FTIR spectra for SiO2 film deposited by alternating o exposures of TEOS precursor and H2O on ZrO2 nanoparticles at 225 C.

H2O (4) LTB (4)

H2O (3)

TMA (3)

H2O (2)

LTB (2)

Absorbance (A.U.) H2O (1)

TMA (1)

4000 3500 3000 2500 2000 1500 Wavenumber (cm-1)

Figure 4.4: In-situ difference FTIR spectra for lithium aluminum oxide film deposited by alternating exposures of LTB, H2O, TMA, and H2O precursors on o ZrO2 nanoparticles at 225 C.

Figure 4.5 shows the in-situ FTIR spectra for the deposition of lithium aluminum silicate using pulses of TMA, H2O, LTB, H2O, TEOS, H2O repeated over 4 global cycles on ZrO2

nanoparticles at 225 oC.

The initial TMA and water pulses started the deposition similar to pure aluminum oxide

deposition at the beginning of this section: after TMA exposure, a decrease in the –OH

absorbance peaks (3700 cm-1) and increase in the C-H peaks (2900 cm-1 and 1200 cm-1) , while

after H2O exposure, -OH is regenerated and a decrease in C-H peaks was observed.

When lithium precursor was next exposed to the substrate, there is an increase in the tert- butoxy ligand peaks (1200 cm-1 and 2971 cm-1) and a decrease in the hydroxyl ligand peaks,

similar to the individual LTB precursor deposition. Water exposure decreases the tert-butoxy

ligand peaks and appeared to regenerate the hydroxyl surface species (3700 cm-1).

When TEOS is exposed to the surface, the absorbance of the –OH peak decreased

(3700 cm-1), while there were three absorbance peaks in the C-H stretch region at 2979, 2933, and 2858 cm-1, each coordinating with a C-H stretch in the tetra-ethoxy ligand (Ferguson, Smith,

Weimer and George 2004). Upon water exposure, a desorption of the tetra-ethoxy ligand was

observed with a decreased in the C-H peaks, and a regeneration of the –OH surface species was

shown by a peak at 3700 cm-1.

H2O (6)

TEOS (6)

H2O (5) LTB (5) H2O (4) TMA (4)

H2O (1)

TEOS (3) H O (2) Absorbance (A.U.) 2 LTB (2)

H2O (1) TMA (1)

4000 3500 3000 2500 2000 1500 Wavenumber (cm-1)

Figure 4.5: In-situ difference FTIR spectra for LASO deposited in TMA, H2O, LTB, H2O, TTBS, and H2O repeating global cycle on ZrO2 nanoparticles at 225 oC. Shown here are global cycles 1 and 2.

H2O (12) TEOS (12)

H2O (11) LTB (11)

H2O (10) TMA (10)

H2O (9)

TEOS (9)

H2O (8) Absorbance (A.U.) LTB (8)

H2O (7)

TMA (7)

4000 3500 3000 2500 2000 1500 Wavenumber (cm-1)

Figure 4.6: In-situ difference FTIR spectra for LASO deposited in TMA, H2O, LTB, H2O, TTBS, and H2O repeating global cycles on ZrO2 nanoparticles at 225 oC. Shown here are global cycles 3 and 4.

When TMA is exposed to the substrate, the –OH surface species decrease, and C-H

increase. However, the peak is not as large as the C-H peak observed in the initial TMA pulse.

After the water pulse, an increase in the absorption peak at 3700 cm-1 is expected. However, in

the C-H region, in addition to the C-H peak of the methyl ligand at 2895 cm-1, there was also

observed desorption peaks at 2979, 2933 and 2858 cm-1. This suggests that not all the tetra-

ethyoxy ligand from the TEOS precursor pulse was completely desorbed after the accompanying

water pulse. The smaller C-H absorption peak upon TMA exposure at 2895 cm-1 compared to

the 2895 cm-1 C-H peak at the initial TMA pulse could also be explained in this manner: less

TMA reacted with the surface –OH species due to surface sites hindered by the existence of the

tetra-ethoxy ligand. The addition of the aluminum precursor seems to aid the reaction with the tetra-ethyoxy ligand on the surface, furthering desorption of the species after water exposure

(subsequent to the TMA exposure).

Continuation of the global cycle with the lithium pulse resulted in a decrease in –OH absorption, and an increase in C-H absorption similar to LTB adsorption onto the surface. Water exposure also created a predictable increase in the absorption peak at 3700 cm-1 and a decrease in

the C-H absorption peak at 2971 cm-1. Upon continuation of the global cycle, the pattern seemed

to continue with extra negative absorption peaks after the aluminum water pulse indicative of tetra-ethoxy ligand desorption. However, the repetition and consistency of the peaks after the first cycle indicate predictable deposition patterns suggesting constant growth rates per global cycle of lithium, aluminum and silicon precursors.

4.2 Atomic Layer Deposition of LASO

The atomic layer deposition of lithium aluminum silicon oxide is modeled as a solid mixture of constituent oxides deposited via ALD. Individual constituents were studied before binary and tertiary oxides. Previous studies on LASO used TEOS as the silicon precursor , while the use of tris(tert-butoxy)silanol (TTBS) was also studied in this work for ALD LASO.

The deposition of aluminum oxide was completed on Si (100) substrates using trimethylaluminum as the metal precursor and water as the oxidant. Using spectroscopic ellipsometry, the aluminum oxide depositions are consistent at 1.4 Å/cycle, as shown in Figure

4.6 This depositions rate is consistent with other given literature values in the range of 1.2-1.3

Å/cycle (George 2010).

The deposition of lithium oxide on Si(100) and substrates was also studied using lithium

tert-butoxide as the precursor. Spectroscopic ellipsometry fitting to lithium oxide thin film

models gives a resulting growth rate of 1.2 Å/cycle, also shown in Figure 4.6. Literature reports lithium oxide ALD growth rates range from 0.86 Å/cycle (Cavanagh, Lee, Yoon and George

2010) to 2.2 Å/cycle (Comstock and Elam 2013). The inconsistent measurement can be due to formation of surface LiCO3 and LiOH when lithium oxide thin films are exposed to atmospheric

conditions, as well as spectroscopic modeling.

The binary oxide of lithium and aluminum were deposited using a sequential cycle of LTB,

H2O, TMA, and H2O. This series designated the global cycle that was then repeated for the

deposition of the binary oxide. As shown in Figure 4.6, the binary oxide growth rate given by ellipsometry was 2.2 Å/cycle. The ellipsometry model was constructed using Bruggeman’s effective medium approximation model of lithium oxide and aluminum oxide individual

component models. X-ray spectroscopy confirmed the presence of both aluminum and lithium cations in the individual constituent oxides, as well as the binary oxide.

2.3 Å/cycle TMA+H O+LTB+H O 400 2 2

300

) 1.4 Å/cycle TMA+H O Å 2

200 Thickness ( 1.2 Å/cycle LTB+H O 100 2

0 0 50 100 150 200 ALD Cycles

Figure 4.7: Growth rate of ALD films from spectroscopic ellipsometry of Al2O3, LiOH and lithium aluminate films. Growth rate of 2.3 Å/cycle was observed for ALD of lithium aluminate by TMA/H2O/LTB/H2O, while rates of 1.4 Å/cycle and 1.2 Å/cycle were measured for Al2O3 (TMA/H2O) and LiOH (LTB/H2O) depositions respectively.

To better observe the atomic layer deposition of silicon oxide ALD, depositions were conducted with Si(100) and Ge(100) substrates. XPS scans were carried out for samples grown on Ge(100) to confirm deposition of silicon. Both TEOS and TTBS precursor chemistries did not produce an observable thickness via spectroscopic ellipsometry using water as the oxidant on silicon substrate (see Figure 4.8). This is confirmed with no silicon 2p peak at 102 eV (as shown in Figure 4.9). This is consistent with other studies using silicon precursors at low-temperature

ALD (Hämäläinen 2011).

However, when silicon precursors were incorporated into binary oxides, such as aluminum silicate, the atomic layer deposited thin film was grown at a consistent 2.2 Å/cycle, 78

and the presence of silicon was observed (see Figure 4.9 for the aluminum silicon deposition).

Further investigation using in-situ FTIR helps illuminate the reasons why deposition of aluminum silicate is successful while the growth of silicon oxide is not.

21 Å/cycle LASO

150 2.2 Å/cycle TTBS+H2O+TMA+H2O ) Å 100

1.4 Å/cycle TMA+H2O

Thickness ( 50

0 Å/cycle TTBS+H2O 0 0 50 100 150 ALD Cycles

Figure 4.8: Growth rate of ALD films from ellipsometry by alternating pulses of o TTBS/H2O, TMA/H2O, and TTBS/H2O/TMA/H2O at 225 C on Si(100) substrate. Growth rate of 2.2 Å/cycle was observed for ALD of aluminum silicate by TTBS/H2O/TMA/H2O. The deposition rate of LASO grown by 10 aluminum cycles, 6 lithium cycles, 4 silicon cycles, and 6 lithium cycles constituting one global cycle is 21 Å /global cycle.

Al 2s Al 2p Si 2p O 2s Ge 2p Ge 3d

TMA/H O/TTBS/H O

2 2 Counts (CPS)

TMA/H2O

TTBS/H2O

140 120 100 80 60 40 20 0 Binding Energy (eV)

Lithium aluminum silicate was deposited on Ge(100) substrates to determine thickness and

composition using XPS. The deposition was found to deposit at a rate of 21.3 Å/cycle, while the composition varied widely.

4.3 In-situ TEM Electrochemical Characterization of LASO on Si/Ge Nanowires

Silicon has been studied as a next generation Li-ion battery material, but due to its high volumetric expiation, (up to 400%), applications have been limited (Liu 2012). Silicon germanium alloys are under current investigation as potential candidates for applications into Li- ion batteries due to the high lithiation capacity of silicon and germanium (3579 mAh/g and 1384 mAh/g respectively) (Bogart 2013). This is due to germanium’s smaller volumetric expansion upon lithiation, as well as higher electrical conductivity.

In-situ transmission electron microscopy (TEM) lithiation study was used to investigate

ALD LASO film as a solid electrolyte for high capacity anodes for lithium ion battery applications. These next-generation anode materials were coated with 33 nm LASO solid electrolyte. Lithium metal was scraped onto a counter electrode and inserted into a pizomanipulator and brought to proximity of the nanowire under investigation. The lithiation of the nanowire was observed and recorded. The nanowire and LASO coating appeared to be sensitive to beam damage at a dosage of 20 nA, so a reduced beam current of ~7 nA was used.

This sacrificed resolution, but mostly preserved the morphology of the heterostructures as the experiments continued.

4.3.1 LASO-coated GeSi Alloy

Figure 4.10 shows high-resolution TEM (HRTEM) micrographs of a 33 nm LASO- coated germanium/silicon (GeSi) alloy nanowire with a concentration of Ge0.4Si0.6. The silicon

substrate is on the bottom of the frame, and the lithium counter electrode is the mass at the top of

the frame. Lithiation of the Ge0.4Si0.6 nanowire was induced by a ramping potential of 50 mV/s from 0 V to -8 V, and then returned to 0V. The lithiated nanowire is shown in Figure 4.10 b.

Figure 4.10: HRTEM of a 33 nm LASO-coated Ge0.4Si0.6 alloy nanowire (a) before lithiation with LiOH/Li counter electrode at the tip (bottom of frame) and (b) after driving lithium ions into the alloy. Diffraction patterns were also taken (c) before and (d) after lithiation.

Examination of the lithiated Ge0.4Si0.6 alloy nanowire showed the ALD LASO coating

remained intact. However, there are some regions along the Ge0.4Si0.6/LASO heterostructure that

showed morphological changes, most likely due to the internal stresses of the lithiated nanowire.

This contrasts with a previous study of 6 nm LASO coating tin oxide (SnO2) high-capacity anode material after lithiation in which the LASO coating fractured and was pulverized upon expansion

of the SnO2 lithiated anode (Zhang 2012).

Selected area diffraction (SAD) was also conducted to study changes occurring before

lithiation and after lithiation (Figure 4.9 c and d). The crystal structure of the silicon/germanium

alloy nanowire was observed prior to the lithiation, while the amorphous structure occurred upon

lithiation. These amorphous regions are similar to other reported crystallographic observations made in the lithiation of silicon and germanium alloy nanowires (Amato 2014).

The cyclovoltammetric response curve of the lithiation and delithiation of the Ge0.4Si0.6 nanowire is shown in Figure 4.11. A slight curve is observed as the potential is ramped from 0V to -8 V. However, once -8V is observed, the current drastically increases. For the experimental run, the current was capped at 50 pA, which corresponded to a current density of 1.2 A/cm2. The total integrated charge into the structure corresponded to 5.99 × 10-9 C, while the charge in the reverse direction (delithiation) amounted to 9.5 × 10-10 C.

-20 Current (pA)

-40

-60

-8 -6 -4 -2 0 2 4 Voltage

Figure 4.11: Cyclic voltammetry of lithiation and delithation of 33 nm LASO- coated/Ge0.4Si0.6 heterostructure. The cycle started at 0 V, proceeded to -8V at 50mV/s, then up to 4.5V, then returning to 0V.

A different GeSi alloy with the composition of Ge0.6Si0.4 was also used in the in-situ

TEM experiments. The LASO-coated Ge0.6Si0.4 heterostructure was given a constant -7.5 V potential between the nanowire and lithium metal counter electrode. The circuit was open for the first 60 seconds to verify the heterostructure before the potential was applied. The current

through the structure for the first minute appeared to be a constant 1.2 pA giving an internal

resistance of 79 TΩ. With very little warning, the current spiked to 2.68 μA before returning to

1.2 pA, as shown in Figure 4.12. The current density spike was equivalent to 2.53 × 105 A cm-2.

Nanowires have previously been found to withstand abnormally high currents up to 1011 A cm-2

(Zhang 2005). This is a combination of factors including electron mean free path much longer than

the diameter of the nanowire, as well as probable excellent heat transfer qualities in a 1D

structure.

There is an energy barrier in crystalline silicon that may contribute to the high over-

potential in lithiating silicon-based anodes. (Ohara 2004; Szczech and Jin 2011) There are reports that the activation energy for lithiating silicon can reach as high as 0.75 eV, which occurred on evaporated silicon thin film electrodes with a Li3PO4 solid electrolyte film with 1 M

LiClO4 in propylene carbonate liquid electrolyte (Baggetto 2009).

0.0

-0.5

-1.0 A) µ

-1.5

Current ( -2.0

-2.5

126 128 130 132 134 136 138 Time (s)

Figure 4.12: Current response of a 33 nm-coated LASO/Ge0.6Si0.4 heterostructure when a constant -7.5 V was placed on the system for 3 minutes. A constant current of 1.2 pA is observed until ~130 seconds when a sudden current of -2.68 uA flows into the anode alloy, and then returns to ~1.2 pA.

Figure 4.13 shows the studied LASO-coated nanowire before lithiation (a) with the substrate at the top of the frame. After lithiation (b), the morphology of the Ge0.6Si0.4 nanowire changes substantially, mirroring a morphology similar to the lithiation of pure germanium nanowires (Liu 2011; Kennedy 2014).

Of note is the crystallographic information using selected area diffraction (Figure 4.12 from (c) and (d)). The crystalline character of the nanowire can be seen through the amorphous

LASO thin film. Upon lithiation, however, the structure of the lithiated Ge0.6Si0.4 nanowire becomes more crystalline in character, but indexing the structure is difficult. This may be a complete lithiation of the nanowire similar to complete lithiation of silicon and germanium to the form of Si4Li15 and Ge4Li15 respectively. Complete analysis of the lithiated structure is beyond

the scope of this study.

Figure 4.13: HRTEM images of Ge0.6Si0.4 alloy nanowire anode coated with 33 nm LASO (a) before and (b) after lithiation with LiOH/Li counter electrode on the tip (bottom of frame). Diffraction patterns were also taken (c) before and (d) after lithiation.

The scanning TEM (STEM) image of the Ge0.6Si0.4 alloy nanowire demonstrated striations in the LASO film that are characteristic of the global cycle: ten Al precursor pulses, three lithium precursor pulses, followed by four silicon precursor pulses, and finishing with three lithium precursor pulses. There were twelve global cycles deposited on the nanowires, which is shown in the layers of the solid electrolyte thin film (see Figure 4.14).

Ge Si Nanowire 0.6 0.4

LASO Thin Film

Figure 4.14: STEM image of a 33nm LASO-coated Ge0.6Si0.4 alloy nanowire after in-situ TEM electrochemical characterization drove lithium ions into the heterostructure. The striations in the 33 nm LASO coating suggest the LASO film is preserved after lithiation of the heterostructure.

4.4 Hybrid ALD-iCVD Solid Electrolyte Films

Alternative solutions to accommodate volumetric expansion of next-generation anodes are needed for more applications of solid state lithium ion battery technology with high capacities. One approach is the synthesis of hybrid organic-inorganic thin film solid electrolytes, where structural weaknesses of the ceramic films can be accommodated by polymer co-films.

Lithium conductive polymer films have flexibility and have already been tested to accommodate next generation anodes such as silicon and germanium with high rates of cyclability and coulombic efficiency (Yao 2012). However, polymer electrolytes by themselves do not have sufficient electrical resistance at the nanoscale to be applied in 3D architectures. However, combining the attributes of both may provide a suitable film for such applications.

The study on the synthesis of a hybrid composite electrolyte was conducted by integrating ALD LASO ceramic films with poly-(tetravinyltetramethylcyclotetrailsoxane) 87

(PV4D4) films (Courtesy of Gleason Group, Massachusetts Institute of Technology). These

films were synthesized by initiated chemical vapor deposition (iCVD). Figure 4.15 shows AFM

topography of a 20 nm iCVD PV4D4 films on indium-titanium oxide (ITO) substrate, compared

to a 28 nm thick ALD LASO film. The as-deposited iCVD PV4D4 film had a root mean squared

(RMS) roughness of 5.1 nm, in contrast with the ALD LASO film, which had a RMS roughness of 1.4 nm on the 2D ITO substrates.

Figure 4.15: AFM scan of (a) 20 nm iCVD PV4D4 polymer thin film (RMS: 5.1 nm) and (b) 28 nm ALD LASO thin film (RMS: 1.4 nm).

The synthesis of the hybrid ALD LASO and iCVD PV4D4 films were conducted first by soaking the iCVD PV4D4 samples in 1M LiClO4 in 1:1 solution of ethylene carbonate and propylene carbonate. These samples were then taken into the vacuum chamber and 28 nm ALD

LASO was deposited on top of the iCVD PV4D4 film. Figure 4.16 shows AFM topography of a

25 nm iCVD PV4D4 film and the hybrid film after deposition of 20 nm of ALD LASO. Figure

4.17 shows the AFM topography of a 50 nm iCVD PV4D4 film before, and after a deposition of

20 nm ALD LASO film.

Figure 4.16: AFM scan of (a) 25nm iCVD PV4D4 polymer electrolyte (RMS: 5.1 nm) and (b) 25nm iCVD PV4D4 coated with 28nm ALD LASO (RMS: 5.5nm ).

Figure 4.17: AFM scan of (a) 50 nm iCVD PV4D4 polymer electorlyte (RMS: 6.8nm) and (b) PV4D4 coated with 28 nm ALD LASO (RMS: 5.7 nm).

Impedance was tested on initial iCVD V4D4 lithiated films using hanging mercury drop impedance spectroscopy to verify the ionic conduction character of the bottom layer of the heterostructure. After the iCVD V4D4 film was characterized, 28 nm LASO films were deposited in the ALD chamber. Both hybrid iCVD PV4D4-ALD LASO films were analyzed for ionic conductivity using impedance spectroscopy using probe measurements. Top electrodes were constructed using electron beam evaporation physical vapor deposition. The top electrodes consisted of 100 nm platinum with 10 nm titanium as an adhesion layer.

3.0

2.5

2.0 )

Ω 1.5

-Z'' (M 1.0

0.5

0.0

0.00 0.05 0.10 0.15 0.20 0.25 0.30 Z' (MΩ)

Figure 4.18: Inset graph shows impedance spectroscopy measurements using hanging mercury drop probes of a 25 nm iCVD PV4D4 polymer electrolyte sample soaked in 1M LiClO4 1:1 EC:DMC solution. Two measurements were taken at different locations on the sample. After coating the 25 nm iCVD PV4D4 polymer with 28 nm ALD LASO, the ion conductivity of the film increased substantially.

The ion conductivity of the iCVD film increased with the addition of the 28 nm ALD

LASO film from 1.5 × 10-5 S/cm to 4.2 × 10-10 S/cm. Subsequent samples constructed in this

way showed similar decrease in ionic conductivity upon the addition of the ALD LASO films

grown on iCVD polymer films. Ionic conductivity of these films was calculated from impedance

spectroscopy measurements obtained from hanging mercury drop probes.

It was supposed that the elevated temperature and the vacuum chamber would evacuate

the solvent from the polymer film, limiting the ion conduction effectiveness of the iCVD PV4D4

film. To understand the effects of the solvent in the implementing of hybrid inorganic-organic

solid state electrolytes, composite samples were constructed using ALD LASO deposited on the

quartz-ITO substrate followed by a deposition of iCVD PV4D4 then soaked in liquid electrolyte

solution. Figure 4.19 shows the impedance spectra of a composite inorganic-organic solid

electrolyte with 90 nm ALD deposited on a quartz-ITO substrate, with 20 nm iCVD PV4D4

deposited on top. The subsequent sample was then soaked in a 1 M LiClO4 electrolyte solution with 1:1 ethyl carbonate: dimethyl carbonate (EC:DMC) as the solvent.

100

) 50 Ω

-Z'' (k 25

0 25 50 75 100 Z' (kΩ)

Figure 4.19: Impedance spectroscopy measurement of an ALD LASO – iCVD PV4D4 heterostructure. The base of the structure consisted of 90 nm ALD LASO (32% lithium content) film deposited on quartz-ITO substrate. A 20 nm iCVD PV4D4 deposition followed, and the ALD-iCVD hybrid electrolyte heterostructure was soaked for 3 days in 1 M LiClO4 1:1 EC:DMC solution. The ionic conductivity was calculated to be 4.78 × 10-11 S/cm.

The initial 90 nm ALD LASO film ion conductivity was calculated based on impedance

spectra obtained using mercury drop probes before the iCVD film was deposited on the sample.

The ionic conductivity of the ALD LASO film was 2.1 × 10-8 S/cm, while the ionic conductivity

of the heterostructures was calculated to be 4.78 × 10-11 S/cm. Deposition order of the ALD

LASO and iCVD V4D4 thin films did not appear to have a significant impact on ionic

conductivity.

4.5 LASO in Half-Cell Applications

In order for LASO solid electrolyte to be applied as possible candidates for solid

electrolytes in all solid state lithium ion batteries, application half-cell and full-cell

configurations need to be demonstrated.

Galvanostatic charge-discharge cycling was performed on a 2D carbon anode material

coated with 22 nm LASO film. The carbon anode is composed of 83% mesocarbon microbeads,

7% carboxymethyl cellulose, 8% graphite, and 2% Ketjen black. The responses obtained from the half-cell is shown in Figure 4.20 where the carbon-LASO composite was submerged in 1 M

LiClO4 in 1:1 ethylene carbonate (EC): dimethyl carbonate (DMC) solution with lithium foil as

the counter and reference electrodes charging and discharging at a C/20 rate.

0.6

0.5

0.4

(V) 0.3 WE E 0.2

0.1

0.0 0 40 80 120 Time (s)

Figure 4.20: Galvanostatic charge-discharge cycling of 22 nm LASO deposited on 2D carbon slurry in 1M LiClO4 in 50:50 EC/DMC solution with lithium foil as counter and reference electrodes. The half-cell carbon-LASO composite showed lithium intercalation and deintercallation

with total lithiation capacity of 301 mAh/g, compared with 351 mAh/g on bare carbon electrodes.

The delithiation capacity was calculated to be 294 mAh/g, with 319 mAh/g being calculated

from bare carbon electrode slurries. Compared with bare carbon electrodes, the coulombic

efficiency of the LASO coated carbon electrode was 97.8 %. This higher value than bare 2D

carbon slurry (90.9 %) demonstrates that LASO-coated carbon could have extended capacity

after many cycles.

As LASO-coated 2D carbon electrodes have demonstrated promise in electrochemical

cells, expansion into 3D arrays of carbon electrodes with the same composition have been

conducted to demonstrate 3D solid state half-cells can be cycles with increased coulombic efficiency. Galvanic cycling of the LASO-coated 3D array is shown in Figure 4.21, with the cycling rate at C/50.

0.7

0.6

0.5

0.4

(V) WE

E 0.3

0.2

0.1

100 200 300 400 500 600 700 Time (h)

Figure 4.21: Galvanostatic charge-discharge cycling of 22 nm LASO-coated 3D carbon pillars in 1M LiClO4 in 1:1 EC/DMC solution with lithium foil as reference and counter electrodes. Cycling of the 3D carbon pillar array coated in 22 nm LASO shows consistent cycling

over 8 charge/discharge cycles. Unexpectedly, the coulombic efficiency of the sample was

calculated to be 75.6%, which does not match 2D LASO-coated carbon experiments. One

possible reaons for the unexpected loss of lithium is through interference of the silver paste in the

construction of the carbon array, as silver is a possible anode material (Taillades 2004). With further refinement, optimization of the LASO-coated 3D carbon arrays to produce 3D microarrays for carbon-based electrodes in lithium-ion batteries.

4.6 LASO in Full-Cell 2D Applications

A full solid state lithium ion battery cell was constructed using 25 μm lithium cobalt oxide

as the cathode, 5 nm ALD LASO as the solid electrolyte, and 1 μm sputtered aluminum as the

anode and current collector. Upon connecting the lead wires to the anode and cathode, resistance

was measured to be 64 MΩ.

Preliminary galvanostatic titration measurements suggest ion transfer upon an induced

current of 1.2 nA, reaching 2.2 V (vs. Li). Although there appears to be significant over potential, ionic conductivity has been demonstrated in a full cell constructed of the LASO solid electrolyte (Daniel, Leland, Kyeong-Sik, Chi On and Bruce 2015).

0.98

0.96

0.94

(V)

WE 0.92 E

0.90

0.88

500 1000 1500 2000 2500 3000 3500 Time (s)

Figure 4.23: Galvanic charge-discharge cycling curve of a 2D all solid state lithium ion cell consisting of LiCoO2 as the cathode, 5 nm ALD LASO as the solid electrolyte, and 1 μm sputtered aluminum layer as the anode and current collector.

CHAPTER 5 SUMMARY

In-situ Fourier transform infrared spectroscopy (FTIR) studies led to insight of atomic layer depositions of aluminum oxide and lithium hydroxide. The in-situ FTIR studies did

highlight limitations of tetraethyl orthosilicate and water chemistry at low temperature were not

continuous, but the deposition of silicon oxide in complex oxides were attainable from aided

hydroxylation by the presence of a more electropositive atom near the ethoxy ligand.

Growth rates of individual component oxides were found to be 1.4 Å/cycle and 1.2

Å/cycle for Al2O3 and LiOH deposited through atomic layer deposition. Silicon oxide thin films

could not be grown by atomic layer deposition; however, binary and tertiary oxides containing

another constituent could deposit silicon at an increased deposition rate: 2.3 Å/cycle for lithium aluminate, and 2.2 Å/cycle for silican aluminate. Lithium aluminum silicate atomic layer deposition consisting of ten aluminum oxide depositions, six lithium hydroxide ALD cycles, four silicon oxide ALD cycles and six lithium hydroxide ALD cycles sequentially in a global cycle yielded 21.3 Å/global cycle. ALD LASO thin films were characterized for composition of metal cations by XPS, and ion conductivity of ALD LASO was found to be on order of 9.8 × 10-8 S/cm.

Construction of a hybrid ceramic-polymer solid electrolytes demonstrated interface

resistances that were high when constructing the hybrid structure with polymer film as the base.

Preliminary results of hybrid films with ceramic ALD LASO as the first layer with PV4D4

polymer soaked in liquid electrolyte solution demonstrates ionic conductivity.

Lastly, LASO solid electrolyte thin films were integrated into current-generation and

next-generation lithium-ion battery applications. In-situ TEM electrochemical characterizations were conducted on LASO-coated germanium-silicon alloy nanowires and demonstrated promising mechanical strength with a 33 nm coating.

LASO was deposited on 2D carbon slurry and 3D carbon pillars and demonstrated cyclability and improved performance in liquid electrolyte. A 2D full was constructed using lithium cobalt oxide as the cathode, LASO as the solid electrolyte, and sputtered aluminum as the anode and current collector. Galvanostatic cycling demonstrated ion migration in the sample, but poor capacity due to limited lithium ion penetration into the aluminum anode of a 2D full cell

Li-ion battery.

APPENDIX

A1.1 Diagram of LASO Atomic Layer Deposition Hot Wall Reactor

University of California, Los Angeles, Department of Chemical Engineering Part Hot Wall Atomic Layer Material Stainless steal Deposition Reactor Filename LASO Chamber.STEP Scale No Units Quantity 1 Designer Jea Cho Date 08/08/2014 Revised by Date Contractor MDC Vacuum Date Comments Tolerance +/- .0001

A1.2 Diagram of in-situ FTIR Atomic Layer Deposition Hot Wall Reactor

University of California, Los Angeles, Department of Chemical Engineering Part In-situ FTIR ALD Reactor Material Stainless steal Filename UCLA-in-situ FTIR Chamber.STEP Scale No Units Quantity 1 Designer Jea Cho Date 08/08/2014 Revised by Date Contractor MDC Vacuum Date Comments Tolerance +/- .0001

A1.3 List of Parts for LASO Chamber Vendor Catalog No. Description Chamber Hardware Duniway DST-531 Thermocouple Sensor Tube, Nickel-plated, Mild Steel, 1/8” NPT Port TCG-531 Thermocouple Type 531 Gauge Controller IFT-NW40-4 Molecular Sieve Trap, In-Line FT-4-MS Molecular Sieve Replacement Material, Zeolite 13- X Leybold-D16A Leybold D16A Mechanical Direct Drive Pump Kurt J. Lesker TFT1KY2C302 TC/Power Feedthrough, 1.33” Laminar UFC-1100A Mass Flow Rate controller Technologies MDC 407002 6-way cross, 2.75” 310029 Angle Valve, 2.75” CF 110000 Flange, 1.33”, Blank 110008 Flange, 2.75”, Blank 409004 Multi-port flange, 5-1.33” to 2.75” 402000 Nipple, 1.33” 402002 Nipple, 2.75” 311074 Pneumatic Angle Valve, NW40 665205 Quick-Access Viewport Door, 2.75” 9392007 TC/Power Feedthrough, 1.33” 404002 Tee, 2.75” 150001 Zero-length Reducer 2.75” to 1.33” 414006 CF to Male VCR, 1.33” 414007 CF to Male VCR, 2.75” 732003 Nipple Reducer, NW25 to NW40 700002 Flange Assembly, NW25 700003 Flange Assembly, NW40 991539-01 Push-on Electrical Connectors 463000 Sealed glass tube (1.33” flange) 731001 Adapter, NW25x.125" FPT Nor-Cal LFT-075-1-025 Single Liquid Feedthrough, ¼” VCR to 1.33” CF Products Swagelok SS-4MG-VCR-MH SS Medium-Flow Metering Valve, Vernier Handle SS-4BG-V51 SS Bellows-Sealed Valve, Spherical Stem Tip SS-DLV51 SS High-Purity High-Pressure Diaphragm-Sealed Valve SS-BNVCR4-P-C 16L SS High-Purity Bellows-Sealed Valve, ¼” 316L-4-VCR-3AS 316 SS VCR ¼” Butt Weld Gland SS-4-VCR-1 316 SS VCR ¼” Female Nut SS-4-VCR-4 316 SS VCR ¼” Male Nut

100

SS-4-HCG ¼” Stainless Steel Female NPT Hex Coupling Chamber Electronics Ari Industries BXX09B88-4T Inconel 600 Sheath Heater (88” Length, 0.09” Inc. Diameter) Omega CN1507TC 7 Channel Temperature Controller Engineering SSR330DC25 Solid State Relay (DC 25 Amp) CO1-K072 INCH Type K Thermocouple, Cement-on TT-K-24-100 Type K Thermocouple, PFA Insulated, 24AWG McMaster 69295K81 Plug, Compact Push-in Connector, 250V AC, 0.093” 69295K61 Receptacle, Compact Push-in Connector 250V AC, 0.093” 69295K23 Pin, Compact Push-in Connector 250V AC, 22-18 AWG, 0.093” 69295K33 Socket, Compact Push-in Connector 250V AC, 22-18 AWG, 0.093” 7587K941 Wire, Stranded Single-Conductor, UL 1007/1569, 20 AWG 300V AC, Black 3641K26 High-Temperature Heat Cable, 96” Length, 400 Watts 2641K24 High-Temperature Heat Cable, 36” Length, 125 Watts 3641K23 High-Temperature Heat Cable, 24” Length, 125 Watts 69145K68 Spade Terminal, Block, Vinyl Insulated, 22-18 AWG, #8 Screw/Stud 3869L34 Type K Thermocouple Connector, Female 3869K35 Type K Thermocouple Connector, Male 6994K24 Variable Voltage Output Transformer, Bench Top 1440(Volts × Amps), 120V AC Input, 60 Hz 7227K84 Butt Splice, Nylon-Insulated Double-Crimp, 22-18 AWG Chamber Automation Hardware Controlled PS2932BP Solenoid: 3 Pin Connector Kit – Series Motion Solutions, Inc. PS2982B53P Solenoid: 120/60 Coil Sol Kit PSTFTN0N10NP 10 State Manifold McMaster 75065K14 Indoor Steel Enclosure with Knockouts (NEMA 1), 10” H x 8” W x 4” D 7343K711 Toggle Switch, Off-On, 15 Amps, Quick-Disconnect 5339T24 LED Panel Light, 13mm, Conical Lens, 120 V AC 101

7060K63 Insulated Barrel Quick-Disconnect Terminal, Double Crimp Female, 22-18 AWG, 0.25”W x 0.032” Thick Tab 7087K15 Panel-Mount Fuse Holders for 3AG/3AB Fuse, 0.25” Straight Angle Terminal 7527K51 300 VAC/VDC Terminal Block, 10 Circuits, 3/8” Center-to-Center, 20 Amps National NI USB-9481 DAQ USB-based High-Voltage Relay Output Instruments Miscellaneous

MDC 9924004 Inline Electrical Connections (Diameter: 0.059") McMaster 9548K23 Stainless Steel Shim Roll, Type 316, 0.004” Thick, 6” Width, 100” Length 9317T67 Corrosion-Resistant Type 304 Stainless Steel Wire Cloth Discs, 200 Mesh, 5/8” Diameter, 0.0021” Wire Diameter 5272K292 Brass Yor-Lok Tube Fittings, Straight Adapter for 1/8” Tube OD x 1/8 NPT Male 5384K51 Moisture-Resistant Polyethylene Vacuum Tubing, 1/16” ID, 1/8” OD, Semi-Clear, White, 50’ Length 5454K61 Miniature Brass Fitting, Straight for 1/16” ID x 10- 32 Male Pipe Size 9162K191 Chrome-Plated Brass Thread Pipe Plug, 1/8 Pipe Size, Square Head 5454K81 Miniature Brass Slotted Head Plug, 10-32 Thread 51875K61 Brass Compression Tube Fitting with Tube Support, Straight Adapter for ¼” Tube OD x 1/8 Male Pipe 5272K101 Front Sleeve for ¼” Tube OD Brass Yor-Lok Tube Fitting 5272K111 Back Sleeve for ¼” Tube OD Brass Yor-Lok Tube Fitting 5272K121 Nut for ¼” Tube OD Brass Yor-Lok Tube Fitting Omega NI80-020-200 Resistance Heating Wire, Ni-Cr Alloy (80/20) AWG Engineering 24 NI80-012-200 Resistance Heating Wire, Ni-Cr Alloy (80/20) AWG 28 FS-260-12-500 Single Hole Fish Spine, 500/pk (AWG 12/0.081”) FS-330-10-500 Single Hole Fish Spine, 500/pk (AWG 10/0.102”) FS-110-20-500 Single Hole Fish Spine, 500/pk (AWG 20/0.032”)

102

A1.4 List of Parts for in-situ FTIR ALD Chamber Vendor Catalog No. Description Chamber Hardware Duniway DST-531 Thermocouple Sensor Tube, Nickel-plated, Mild Steel, 1/8” NPT Port TCG-531 Thermocouple Type 531 Gauge Controller IFT-NW40-4 Molecular Sieve Trap, In-Line FT-4-MS Molecular Sieve Replacement Material, Zeolite 13- X Leybold-D16A Leybold D16A Mechanical Direct Drive Pump McAllister Differentially Pumped IR Window Flange, 2.75 in. Technical DPW 275 OD Services Kurt J. Lesker TFT1KY2C302 TC/Power Feedthrough, 1.33” Laminar UFC-1100A Mass Flow Rate controller Technologies MDC 409004 Multi-port flange, 5-1.33” to 2.75” 408001 6-WayCube, 2.75" 402000 Nipple, 1.33” 9392007 TC/Power Feedthrough, 1.33” 150001 Zero-length Reducer 2.75” to 1.33” 414006 CF to Male VCR, 1.33” 414007 2.75" OD CF to 1/4" MVCR Adapter 300001 2.75'' GateValve, 1.5"HV, Manual 190057 12-Pt, Bolt, 7/8"Lg, Silver Plated 190000 SocketHeadScrew, .50"Lg, Tapped 190094 SocketHeadScrew/PlateNut, .75"Lg 310029 Angle Valve, 1.5" HV Manual, 2 3/4"CF 409004 2-3/4’’Multiport Swagelok SS-4BG-V51 SS Bellows-Sealed Valve, Spherical Stem Tip SS-4MG-VCR-MH SS Medium-Flow Metering Valve, Vernier Handle SS High-Purity High-Pressure Diaphragm-Sealed SS-DLV51 Valve 316L-4-VCR-3AS 316 SS VCR ¼” Butt Weld Gland SS-4-VCR-1 316 SS VCR ¼” Female Nut SS-4-VCR-4 316 SS VCR ¼” Male Nut SS-4-HCG ¼ in. Stainless Steel Female NPT Hex Coupling Chamber Electronics Omega CN1507TC 7 Channel Temperature Controller Engineering SSR330DC25 Solid State Relay (DC 25 Amp) CO1-K072 INCH Type K Thermocouple, Cement-on TT-K-24-100 Type K Thermocouple, PFA Insulated, 24AWG 103

McMaster 69295K81 Plug, Compact Push-in Connector, 250V AC, 0.093” 69295K61 Receptacle, Compact Push-in Connector 250V AC, 0.093” 69295K23 Pin, Compact Push-in Connector 250V AC, 22-18 AWG, 0.093” 69295K33 Socket, Compact Push-in Connector 250V AC, 22-18 AWG, 0.093” 7587K941 Wire, Stranded Single-Conductor, UL 1007/1569, 20 AWG 300V AC, Black 3641K26 High-Temperature Heat Cable, 96” Length, 400 Watts High-Temperature Heat Cable, 24" Length, 100 3641K23 Watts High-Temperature Heat Cable, 36" Length, 125 3641K24 Watts 69145K68 Spade Terminal, Block, Vinyl Insulated, 22-18 AWG, #8 Screw/Stud Butt Splice, Nylon-Insulated Double-Crimp, 22-18 7227K84 AWG 3869L34 Type K Thermocouple Connector, Female 3869K35 Type K Thermocouple Connector, Male Variable Voltage Output Transformer, Bench Top, 6994K24 1440 VA, 120 VAC Input, 60 Hz Miscellaneous Controlled Motion PS2932BP Solenoid: 3 Pin Connector Kit-Series Solutions, Inc. PS2982B53P Solenoid: 120/60 Coil Sol Kit PSTFTN0N10NP 10 State Manifold Fisher AA10354BY Tantalum Wire (Diameter: 2mm) Scientific International Crystal 0002C-144 KBr Polished Disc – 13mm x 2mm Laboratories 0002C-150 KBr Polished Disc – 38 mm x 6mm Laminar UFC-1100A Mass Flow Rate controller Technologies MDC 9924006 Inline Electrical Connections (Diameter: 0.120 inch) 2 x MTS# McAllister 10000974 Technical DPW275 Spare O-ring Kit 1 x MTS# Services 10000977 McMaster 6493A65 Ultra-Gold Hex L-Key, 5/64" Hex, 3-1/4" Length

104

Metric 316 Stainless Steel Shim, 1.0mm Thick, 90214A527 12mm ID, 18mm OD Stainless Steel Socket Head Cap Screw, 2-56 92185A073 Thread, 3/16" Length Corrosion-Resistant 304 Stainless Steel Wire Cloth 9317T67 Disc, 200 x 200 Mesh, 5/8" Diameter, .0021" Wire Diameter Stainless Steel Round Tube, Type 304 Stainless 89495K691 Steel, 0.250" Thick Wall, 1-1/2" OD, 1' Long Indoor Steel Enclosure with Knockouts (NEMA 1), 75065K14 10" Height x 8" Width x 4" Depth Brass Yor-Lok Tube Fitting, Straight Adapter for 5272K292 1/8" Tube OD x 1/8 NPT Male Moisture-Resistant Polyethylene Vacuum Tubing, 5384K51 1/16" ID, 1/8" OD, 1/32" Wall Thickness, Semi- Clear White, 50 ft. Length Miniature Brass Fitting, Straight for 1/16" Tube ID x 5454K61 10-32 Male Pipe Size Chrome-Plated Brass Thread Pipe Fitting, 1/8 Pipe 9162K191 Size, Square Head Solid Plug Miniature Brass Fitting, 10-32 Thread, Slotted Head 5454K81 Plug Brass Compression Tube Fitting with Tube Support, 51875K61 Straight Adapter for 1/4" Tube OD x 1/8 Male Pipe Front Sleeve for 1/4" Tube OD Brass Yor-Lok Tube 5272K101 Fitting Back Sleeve for 1/4" Tube OD Brass Yor-Lok Tube 5272K111 Fitting 5272K121 Nut for 1/4" Tube OD Brass Yor-Lok Tube Fitting 7343K711 Toggle Switch, SPST, 15 Amps, Quick-Disconnect LED Panel Mount Indicating Light, 13 mm, Conical 5339T24 Shaped Lens, 120V AC/DC Insulated Barrel Quick-Disconnect Terminal, 7060K63 Double-Crimp Female, 22-18 AWG, .25" Width x .032" Thickness Tab Panel-Mount Glass/Ceramic Fuse Holder for 7087K15 3AG/3AB Fuse, .25" Straight Angle Terminal 300 VAC/VDC Terminal Block, 10 Circuits, 3/8" 7527K51 Center-to-Center, 20 Amps 2930T63 Super-Corrosion-Resistant Type 316 Stainless Steel

Wire Cloth Discs, 100 Mesh, 1/2" Diameter 90214A527 Type 316 Stainless Steel Round Shim, 1mm Thick,

12mm ID, 18mm OD National NI USB-9481 USB-Based High-Voltage Relay 779453-01 Instruments Output 105

Omega Resistance Heating Wire, Nickel-Chromium Alloy, NI80-020-200 Engineering 80% Nickel/ 20% Chromium, AWG 24 BARE-20-K-12 Bare wire thermocouple, Type K, 20 AWG FS-260-12-500 Single hole fish spine, 500/PK (AWG12/0.081'') FS-330-10-500 Single hole fish spine, 500/PK (AWG 10/0.102'') FS-110-20-500 Single hole fish spine, 500/PK (AWG 20 / 0.032'')

106

A1.5 Custom Parts for FTIR ALD Chamber Set 1: Two Pieces of Purge-Gas Line Fittings

Purge gas line fittings are placed between the vacuum chamber and the FTIR spectrometer. Set 2: Two Pieces of Alignment Base

The alignment base is placed below the vacuum chamber gate valves to align the FTIR beam through the center of the chamber. Set 3: Two Pieces of Hold-down Clamps

Clamps are placed on the outside of the gate valves to hold the chamber onto the spectrometer base.

107

A1.6 LabVIEW and Automation

LabVIEW is a graphical programming language that uses nested “For” and “While” loops and sequences in order to accomplish tasks or gather data. A NI-USB 9162 module (which is referred to as DAQ in the program) is used to translate USB signals from LabVIEW to a switch for automation purposes. LabVIEW and the drivers for the DAQ need to be installed before plugging the device into the computer. The DAQ can have up to 4 channels (or switches) active for automating selected pneumatic valves on the chamber. In LabVIEW programming, the DAQ operates on Boolean (true or false for each channel) within the “For” and “While” loops to activate and deactivate the switch for pneumatic valves on the chamber. Selecting the DAQ wizard in LabVIEW will help set up the module to connect the program. To verify the DAQ in the program represents the connected controller, right-click the DAQ wizard, select “Properties”. A loading bar appears to verify that the drivers were installed correctly. At the top of the display (the user-interface), there is a “Play” button. Press this, and click “OK” to close out of the pop-up box. A loading bar appears to verifie the program is communicating with the device. There are two interface screens in LabVIEW that are primarily used to set up an automation program: the Front Panel, and the Block Diagram. The Block Diagram is used to assemble segments of the program to layout the pattern for a precursor deposition. For every precursor, there should be one “For” loop inside a larger “While” loop. For the Lithium Aluminum Silicate (LASO) chamber, each precursor is followed by a nitrogen purge pulse, as well as an oxidizing pulse, followed by another nitrogen purge. The graphic programming schematic of one aluminum precursor cycle is shown in Figure A1.

Figure A1 - Block Diagram of Aluminum Precursor Deposition. An oxidizer pulse follows to make a full precursor cycle. The second interface screen is the user interface for the program, or “Front Panel” (see Figure A2). This is where inputs regarding step time, duration, and number of repeats are entered for a deposition set. The steps for one cycle of precursor deposition are listed as follows: 1. Precursor exposure time. 2. Precursor pumpdown time. 3. Nitrogen purge time 4. Nitrogen purge pumpdown time 5. Oxidant (water) exposure time.

108

6. Oxidant pumpdown time. 7. Nitrogen purge time. 8. Nitrogen purge pumpdown time. The Nitrogen purge pumpdown time should be no less than 40 seconds due to the size of the chamber and limitations with the vacuum pumping system (Figure A2).

Figure A2 - Front Panel of LabVIEW Program Each cycle for the precursor includes the precursor exposure and the oxidant exposure in atomic layer deposition. The number of repeats for each precursor cycle is indicated at the top of each column. There are four columns of precursors that can be used: aluminum (using trimethyl aluminum, or TMA), silicon (tris(tert-butoxy)silanol or TTBS), or lithium (lithium tert-butoxide or LTB). The columns are arranged in the following order: aluminum, lithium, silicon, lithium. This designates a global cycle. If the precursor cycles are designated 10Al:3Li:6Si:3Li, then the deposition contains 10 aluminum cycles, 3 lithium cycles, 6 silicon cycles, and lastly 3 lithium cycles. If the number of global cycles is 3, then then the pattern of 10Al:3Li:6Si:3Li is to be repeated three times in the total deposition process. The Al capping cycles (the fifth column in the Front Panel view) are activated at the end of the last global cycle. This aluminum capping layer stabilizes the lithium- containing thin film somewhat in ambient atmosphere.

109

A2.1 Electrical Wiring for Temperature Control Heaters

A2.1.1 Introduction

The heaters on the lithium aluminosilicate (LASO) atomic layer deposition (ALD) chamber rely on high-temperature heat cables controlled by an Omega 7 Channel Temperature Controller (7CC). The Controller has a thermocouple connected to the LASO chamber which monitors the temperature of certain portions of the chamber. The Controller uses standard PID controls to ramp and maintain temperature on the LASO ALD chamber.

The output of the controller is intended to apply the necessary power to heating elements. Unfortunately, because the voltage is 5Vdc, the controller circuit is fed to solid state relays. When the temperature controller activates the circuit, the relay activates the 120Vac circuit containing the heating elements on the chamber (Figure A3).

Figure A3 - Schematic of electrical wiring for temperature controlled heaters

A2.1.2 Programming the Temperature Controller Note: Before programming or checking a programmed profile, the Controller must be in “stopped” mode.

1) Enter programming mode of the Temperature Controller by first selecting the channel that needs to be programmed (by pressing CTR SEL on the controller panel).

2) Press and hold the CTR SEL key until the display states “EntEr PASSCOdE”.

3) The front keys are marked with a small digit in the lower right corner. Enter the pass-code combination: 3254.

4) Use the key until “PrOFILE” is displayed. 110

5) Push the PROG key.

6) The unit display one of the channels with each push on the key (“CntrL 1”, “CntrL 2”,“CntrL 3”,“CntrL 4”,“CntrL 5”,“CntrL 6”,“CntrL 7”) Select the desired controller, then push PROG to select it.

a) The display briefly reads “Strt SP” for ‘Start Setpoint’, and then the current value of the starting setpoint. Use and keys to enter a desired value. The key increases and decreases the flashing digit, where the key selects the next digit.

b) Press SETUP to go to next function.

c) The display briefly reads SetPt 1 and then the current value of Setpoint #1. Use and to enter the desired value, then press PROG to go to the following function.

d) The display reads “EntEr t” for ‘Enter time’, and show the current time value. Time entered is the ramping time, or the time that it takes to ramp up to the set point (or soak time, if the previous and current setpoints are the same). The value shown for time is in minutes. Use and to enter the desired value, then press PROG to go to the next function.

Figure A4 - Example Ramp Profile

e) Repeat a-d to program setpoints #2 through #7. Once all 7 segments have been programmed, the display reverts back to the beginning of Programming mode selection. If all the seven segments are not used, the program can be aborted at any segment by pushing the EXT key. The segment in which EXT is pushed is considered the end of the program. During Verify Program mode, the segment and all following segments are labeled as “PrG End”. Note: After entering the time for the last segment, go to the next segment and then hit EXT.

A2.1.3 Verify Mode 1) Verify mode is used to check previously programmed ramp/soak profile, or making small changes to a program. Enter programming mode, and press until “VErIFY” is displayed. 2) Press PROG.

111

3) Select the desired controller to verify by pressing , then press PROG to step through the profile. 4) Use the and

as in steps 6a-e of the programming section to edit a program. 5) At any time, pressing EXT to exit Verify mode. Unlike other programming modes, exiting Verify does not mark the end of a profile program. A2.1.4 Run Mode 1) To run a controller, select the controller by pushing CTR SEL key, followed by a push on RUN/STP. 2) Pushing CTR SEL displays “CntrL 1” for ‘controller 1’. 3) To select another controller, toggle CTR SEL until the desired controller is displayed. 4) To run a controller’s profile, push the RUN STP key until the unit displays “Ctr. rUn”. At this point the desired controller runs its programmed profile. 5) Once running, the controller can be put into an indefinite hold, or stopped. 6) To indefinitely hold a controller at any setpoint, press and hold the CTR SEL key at the desired controller, and simultaneously press SCN/HLD. 7) The controller displays “C# HOLd”. 8) The controller holds this point until the unit is placed back in Run mode. 9) To stop a controller after it has been put in run mode, select the controller that needs to be stopped using the CTR SEL key to select the controller. Press RUN/STP once. 10) To stop all controllers at the same time, push CTR SEL and RUN/STP keys simultaneously.

112

A2.2 Electrical Wiring for Automatic Pneumatic Valve Control

The LASO ALD chamber automated deposition is possible through a LabVIEW program on a computer, and a set of circuits to activate a pneumatic valve on the chamber. A solenoid is used to convert electrical impulses to pressurized air flow. In the LASO ALD chamber, electrical impulses come from an NI-USB-9481 (DAQ) interfaced with the LabVIEW automatic deposition computer program. The pressurized air operate the pneumatic valves on the chamber. A schematic of the solenoid channels are shown in Figure A5. When the circuit is closed (by the manual switch on the solenoid box or the DAQ via LabVIEW), the solenoid valve is opened. Pressurized air then flows to the pneumatic valve by plastic tubing. The pneumatic valve opens when pressurized air is available. There are ten solenoid circuits (channels) in the solenoid box that can activate up to ten valves. Each DAQ can operate only four electrical channels. Two DAQ modules are used to operate the LASO ALD chamber.

NI-USB 9481 … Relay Module Main power

Manual … Switch Manual Switch Green LED Solenoid 120 V ~ 1 9 Red Red LED LED …

Figure A5 - Schematic of Electrical Wiring for Automated Pneumatic Valve Control

113

A3.1 LASO ALD Hot Wall Reactor Chamber Operational Procedures

A3.1.1 Chemicals Used: 1. Trimethyl Aluminum (TMA) 2. Tetraethyl Orthosilicate (TEOS) 3. Lithium tert-butoxide (LTB) 4. Tris(tert-butoxy)silanol (TTBS)

A3.1.2 EMERGENCY SHUTDOWN: 1. Close all the precursor/reactant valves: LTB (Li) , TEOS (Si) , TMA (Al), Water 2. Stop LabVIEW program on the computer 3. Close the valves to the chamber (GV1,2,3) 4. Shut down all the electronics: Transformers, Mass Flow Controller control panel 5. Proceed to emergency exit locations

A3.1.3 OPERATING PROCEDURE:

Figure A6 - LASO Chamber Gas Line Schematic

• GV4, V2 and V12 – Valves controlling TMA flowrate, MFC channel 2 • V5, V8, and V11 –Valves controlling N2 flow rates, MFC channel 4 • V4 – Valve controlling LTB • M1 and V2 – Valves controlling water flowrate

1. Initial Check: Make sure all the valves are closed.

114

2. Venting the reactor: a. Isolate the chamber: Turn off V9 of Solenoid box, and close roughing valve. Untighten door so it can be opened. b. Flow nitrogen by setting N2 MFC Channel 4 to 10-15%. c. Open V10 and V8 on the Solenoid box. Open V5 to vent the chamber. d. Purge chamber with nitrogen for 10 minutes, or until the door can open without resistance. e. After venting, close V5 on Solenoid Box to the chamber.

3. Sample loading: a. Open the door of chamber and load the sample into the chamber. Make sure the sample is placed in the center of the reactor holder. b. Close the door and tighten the knob. Note: After loading, be sure to nitrogen purge before starting experiment. This removes water in the chamber that would interfere with water-sensitive precursors.

4. Base pressure: Slowly open the roughing valve, and allow the chamber pressure return to a base pressure lower than 65mTorr.

5. Nitrogen Purge: (*) a. Set channel 4 of MFC controller to 10-15%. b. Open V10 and V8 on the Solenoid box. c. Open V5 for 2 seconds, then close V5. d. Pressure goes up to 500 mTorr, then wait ~45 seconds until pressure returns to base pressure. e. Once base pressure is reached, repeat the procedure (* b-d) at least 2 more times (for a total of 3 times). f. Close V5, V8 and V10. Set N2 MFC to 0%.

6. Heating: (Note: Heat chamber, gas lines, and precursor housings concurrently) a. Chamber: Heat the chamber wall to 150ºC by activating Ch7 of 7Channel Controller- 1 (Denoted as 7CC-1 from now on). Also, activate Ch6 and Ch7 on 7Channel Controller-2 (denoted as 7CC-2 henceforth).

115

Figure A7 – Seven-Channel Controller 1 b. Gas Line heating: Heat the gas lines by activating the following channels: i. 7CC-1:Ch4: TMA/Water Gas line ii. 7CC-2:Ch2: LTB Gas line iii. 7CC-2:Ch4: TTBS Gas line c. Sample stage heating: Heat the sample stage by turning Variable Transformer-1 (Denoted as VT-1 from now on), followed by activating Ch3 of 7CC-1

Figure A8 - Variable Transformer 1 (Left - Sample Stage) and Variable Transformer 2 on (Right - Doser Heater)

116

d. Precursor heating: Heat the precursors by activating corresponding channels of 7CC-1 listed below: i. Lithium Tert-butoxide (LTB): Ch1 of 7CC-1 ii. Tris(tert-butoxy)silanol (TTBS): Ch3 of 7CC-2 Note: TMA and TEOS is maintained at room temperature e. Doser heating: Heat the doser by turning Variable Transformer-2 (Denoted as VT-2 from now on), followed by activating corresponding channel of interest on 7-Channel Controller-2 (7CC-2) i. Ch1: LTB Doser

Figure A9 – Seven-Channel Controller 2 f. In summary, the following channels need to be activated for LASO deposition: i. 7CC-1: 1, 3, 4, 7 ii. 7CC-2: 1, 2, 3, 4, 6, 7

Setpoint/ramping info of both 7CC-1 and 7CC-2 are listed below: 08/01/2014 - 7 Channel Controller Setup for LASO Chamber Channel Description Start Pt SP1 t1 SP2 t2 SP3 t3 SP4 t4 SP5 Lithium tert- 1 butoxide (LTB) 25 80 20 80 15 160 20 160 2500 Pg End 2 3 Sample Heater 25 75 20 75 15 225 20 225 2500 Pg End Additional Ch. 4 (Gasline) 25 45 20 45 15 60 20 60 2500 Pg End Cyclopentadineyl Manganese 5 (Cp2Mn) 25 50 20 50 15 75 20 75 2500 Pg End 6 Water (H2O) 25 45 20 45 15 60 20 60 2500 Pg End 7 Chamber 25 150 20 150 15 150 20 150 2500 Pg End

117

7CC-2 (As of 08/01/2014) 10/12/2013 - 7 Channel Controller 2 Setup for LASO Chamber Chnl Description Start Pt SP1 t1 SP2 t2 SP3 t3 SP4 t4 SP5 Lithium tert-butoxide Pg 1 (LTB) Doser 25 80 15 80 10 160 20 160 2500 End Lithium tert-butoxide Pg 2 (LTB) Gasline 25 80 15 80 10 160 20 160 2500 End Tris(tert- butoxy)Silanol Pg 3 Gasline 25 50 15 50 10 75 20 75 2500 End Tris(tert- butoxy)Silanol Pg 4 Housing 25 30 15 30 10 40 20 40 2500 End Cyclopentadineyl Manganese (Cp2Mn) Pg 5 Doser 25 50 15 50 10 85 20 85 2500 End Chamber Nipple Pg 6 Heater 25 100 20 100 15 150 20 150 2500 End Chamber Pumpline Pg 7 Heater 25 100 20 100 15 150 20 200 150 End

7. Manual Deposition: For LiAlSiO4, alternate cycles of Al oxide (b cycles for each global cycle), Li oxide (a cycles for each global cycle), and Si oxide (c cycles for each global cycle). Deposit total of # global cycles. (Note: Aluminum deposit first because have good surface adhesion.) The procedures listed below are only for one cycle each! (Note: pump down time can be extended if need be, but cannot extend pulse time.) a. LiOH Deposition i. Slightly open 45 of lithium precursor Li-t-butoxide, LTB, for 10 seconds. Pressure increases by 5-10 mTorr. ii. Close valve V4 for 50 seconds for pump down time. Make sure chamber goes back to base pressure. iii. Before the pump down time of 60 seconds are up, open water valve M1 to setting of 10. iv. After the pump down time of 60 seconds, open valve V2 for 10 seconds. Watch the pressure, should increase to 100-200 mTorr, but make sure it does not go above 200 mTorr. (Note: water does not have MFC so vapor pressure can be high.) v. Close V2, then close M1 for water. vi. Pump down for 60 seconds till base pressure has been reached. (Note: if chamber has not reached the base pressure after pump down time of 60 seconds, then do a Nitrogen Purge Step by pulsing nitrogen in for 10 seconds.) 118

b. Al2O3 Deposition i. Open the valve V1 of Grey Solenoid Box for 5 seconds. ii. Close V1. Pump down for 45 seconds. Make sure chamber goes back to base pressure. iii. Before the pump down time of 45 seconds are up, open water valve M1 to setting of 10. iv. After the pump down time of 45 seconds, open valve V2 for 5 seconds. Watch the pressure, should increase to 100-200 mTorr, but make sure it does not go above 200 mTorr. (Note: water does not have MFC so vapor pressure can be high.) v. Close V2, then close M1 for water. vi. Pump down for 45 seconds till base pressure has been reached. (Note: if chamber has not reached the base pressure after pump down time of 60 seconds, then do a Nitrogen Purge Step by pulsing nitrogen in for 10 seconds.) c. SiO2 Deposition (Note: if nitrogen purge was used in previous deposition, need to reset the MFC channel 4 to 3% before beginning deposition.) i. Open V3 ii. Half way open the valve GV2 for 20 seconds. Pressure increases to 200 mTorr. iii. Close GV2, then close V3. Pump down for 60 seconds. Make sure chamber goes back to base pressure. iv. Before the pump down time of 60 seconds are up, open water valve M1 to setting of 10. v. After the pump down time of 60 seconds, open valve V2 for 15 seconds. Watch the pressure, should increase to 100-200 mTorr, but make sure it does not go above 200 mTorr. (Note: water does not have MFC so vapor pressure can be high.) vi. Close V2, then close M1 for water. vii. Pump down for 60 seconds till base pressure has been reached. (Note: if chamber has not reached the base pressure after pump down time of 60 seconds, then do a Nitrogen Purge Step by pulsing nitrogen in for 10 seconds.)

8. Automated Deposition a. Load LabVIEW program on the computer i. Open file “072913 – LASO and Al Capping Layer.vi” ii. Set number of cycles for the appropriate element. 1. Each cycle includes a precursor pulse, nitrogen purge, water pulse, and another nitrogen purge. 2. Times listed on each pulse step are not normally changed. iii. Set number of global cycles on the top of the worksheet. This loops the precursor pulses. iv. Check the temperatures on the Channel Controllers. v. Check sample in the sample holder. vi. Turn V9 on the Solenoid Box to “off”, and open V8 and V10 on the Solenoid box.

119

vii. On the computer, press the white arrow on the toolbar to begin the programmed deposition.

Figure A10 - LabVIEW Front Panel View for Automated Deposition 9. Finishing the deposition: a. Close all MFC and valves. b. Turn off all the heating for chamber and precursor containers. c. Wait for the chamber to cool down to at least 50ºC. (This takes about ~2 hours and it takes ~1 hour to cool down to ~80oC.) d. Nitrogen Purge 3 times (See Step 5).

10. Sample unloading: a. Perform Steps 1 and 2. b. Open the chamber to unload the sample. c. If not loading sample in soon, then pump down the chamber (no need to do nitrogen purge if there is no sample).

120

A3.2 Operating Procedure of in-situ Fourier Transform Infrared Spectroscopy (FTIR) Chamber

A3.2.1 Chemicals Used: 1. Trimethyl Aluminum (TMA) 2. Tetraethyl Orthosilicate (TEOS) 3. Lithium tert-butoxide (LTB)

A3.2.2 Emergency Shutdown: 1. Close all the precursor/reactant valves: LTB (Li) , TEOS (Si) , TMA (Al), Water 2. Close the valves to the chamber 3. Shut down all the electronics: Transformers, Mass Flow Controller control panel 4. Proceed to emergency exit locations

A3.2.3 OPERATING PROCEDURE:

Figure A11 - Overview of FTIR ALD Chamber

1. Initial Check: Make sure all the valves are closed a. Roughing Valve (RV), both Gate Valves (GV1 & GV2) b. Precursor Valves: LTB, TEOS, TMA valves

2. Venting the reactor: a. Isolate the chamber by closing the roughing valve (RV1) b. Flow nitrogen into the chamber by opening Nitrogen Valve and turning switches V5,V6,V10 of the solenoid box c. Wait about ~10 min for the chamber to be vented

121

3. Sample loading: a. Un-do the power and thermocouple lines to the feedthrough b. After undoing the bolts, rotate the sample stage 60º clockwise to obtain enough clearance to pull the entire sample stage assembly out c. Once enough clearance has been obtained, pull the entire sample holder assembly straight out d. Using 5/64’’ Allen key, disassemble the sample mount by undoing the bolts e. Load the sample (e.g. KBr or ZrO2 substrate) and hold the sample into the position by tightening the bolts back f. Reassemble the sample stage by following steps a through c in REVERSE (c. → b. → a.)

4. Base pressure: Open the roughing valve, and pump the chamber pressure to the base pressure of ~30mTorr

5. Nitrogen Purge: (*) a. Set channel 4 of MFC controller to 45%. b. Open the manual valve of N2 line c. Open V10, followed by V6, followed by V5 from solenoid d. Pressure goes up to 200 mTorr then wait till it goes down to base pressure so line is pumped down.

6. Heating: (Note: Heat chamber and gas lines concurrently) a. Chamber & Precursor heating: Heat the system and precursors to the set temperatures by activating the channel of interest. Refer to the table below for the assignment of channels and the corresponding set points.

Figure A12 - Seven-Channel Controller 3

b. Sample stage heating: Heat the sample stage by turning Variable Transformer-3, followed by activating Ch7 of 7CC-3

122

c. At the time of heating, turn on the FTIR spectrometer using the black box at the base of the computer.

Figure A13 - Variable Voltage Transformer

Setpoint/ramping info of 7CC-3 is listed below (as of 11/17/2013) 11/17/2013- 7 Channel Controller 3 Setup (In-situ FTIR Chamber) Start Description SP1 t1 SP2 t2 SP3 t3 SP4 t4 SP5 Ch Pt Pg LTB Gasline 25 80 15 80 10 160 20 160 2500 1 End Pg LTB Housing 25 80 15 80 10 140 20 140 2500 2 End Pg TMA Gasline 25 45 15 45 10 60 20 60 2500 3 End Pg TEOS Gasline 25 45 15 45 10 60 20 60 2500 4 End Pg Water Gasline 25 45 15 45 10 60 20 60 2500 5 End Pg Chamber Wall 25 80 15 80 10 130 20 130 2500 6 End Sample Pg 25 125 15 125 10 225 20 225 2500 7 Heater End

123

7. Deposition for in-situ FTIR monitoring of ALD chemistry

The procedures listed below are only for one cycle each! (Note: pump down time can be extended if need be, but cannot extend pulse time.) IMPORTANT: It is important to isolate the KBr windows on both sides by closing the gate valves before pulsing any precursor/oxidant into the chamber. Failure to do so will cause a deposition of precursor/oxidant to the KBr window and affect the measurements. a. LiOH Deposition i. Close the manual gate valves, and the manual roughing valve. ii. Open the manual valve of lithium precursor Li-t-butoxide, LTB, for 60 seconds. iii. Close the LTB manual valve all the way, open the manual roughing pump. Pressure increases up to ~80mTorr. iv. Over the course of nine minutes, pulse nitrogen three times using the solenoid bypass valves (Figure ), pumping down to base pressure in between each pulse. The bypass switches actuate the pneumatic valves shared with the LASO chamber via parallel connection to the LASO solenoid box. v. IF, at the end of 10 minutes total, the pressure of the chamber is constant (~30 mTorr), then open the window gate valves. Start acquiring a scan for the deposition. vi. After the data has been acquired, close the window gate valves, then close the manual roughing valve. vii. Open the water valve for 60 seconds. viii. Close the manual valve, open the roughing valve. Watch the pressure, should increase to 100-200 mTorr, but make sure it does not go above 200 mTorr. (Note: water does not have MFC so vapor pressure can be high. Use the needle valve to adjust flow rate - 4 revolutions from closed is standard). ix. Over the next nine minutes, purge with N2 three times. Once the chamber pressure is constant, then open the gate window valves. x. Pump down for 60 seconds till base pressure has been reached. (Note: if chamber has not reached the base pressure after pump down time of 60 seconds, then do a Nitrogen Purge Step by pulsing nitrogen in for 10 seconds.)

b. Al2O3 Deposition i. Close the manual gate valves completely, and then the manual roughing valve. ii. Pulse TMA by opening TMA Up & TMA Down switch of Solenoid Bypass Box for 60 seconds. The bypass switches actuate the shared pneumatic valves for TMA flow in the LASO chamber.

124

Figure A14 - FTIR Solenoid Bypass Valves iii. Turn both switches off, then open the manual roughing valve. Over the next nine minutes, purge chamber with nitrogen three times. Make sure chamber goes back to base pressure. iv. If the chamber is at steady state, open the window gate valves. Acquire spectra. v. Close the window gate valves, then close the chamber roughing valve. vi. After the pumpdown, open the water valve for 60 seconds. Close the water valve, and open the chamber roughing valve. Watch the pressure, should increase to 100-200 mTorr, but make sure it does not go above 200 mTorr. (Note: water does not have MFC so vapor pressure can be high.) vii. Over the next nine minutes, purge the chamber with nitrogen three times (for ten seconds each nitrogen purge). viii. If the chamber is at steady state pressure, then open the gate window valves. Acquire spectra. c. SiOx Deposition i. Close the manual gate valves completely, and then the manual roughing valve. ii. Pulse TEOS by opening the ball valve for 60 seconds. iii. Close the TEOS precursor, then open the manual roughing valve (the pressure should go up to ~80mTorr). iv. Over the next nine minutes, purge chamber with nitrogen three times. Make sure chamber goes back to base pressure. v. If the chamber is at steady state, open the window gate valves. Acquire spectra. vi. Close the window gate valves, then close the chamber roughing valve. vii. After the pumpdown, open the water valve for 60 seconds. Close the water valve, and open the chamber roughing valve. Watch the pressure, should increase to 100-200 mTorr, but make sure it does not go above 200 mTorr. (Note: water does not have MFC so vapor pressure can be high.)

125

viii. Over the next nine minutes, purge the chamber with nitrogen three times (for ten seconds each nitrogen purge). ix. If the chamber is at steady state pressure, then open the gate window valves. Acquire spectra.

8. Finishing the deposition: a. Close all valves. b. Turn off all the heating for chamber, water, and precursor containers. c. Wait for the chamber to cool down to at least 50ºC or lower. (This takes about ~1 hour roughly)

9. Sample unloading: a. Perform Steps 3a. through 3e. b. Unload the sample. c. Perform Step 3f. then pumpdown the chamber

10. Data Acquisition: a. After the spectrometer has warmed up and reached steady-state (turn on spectrometer as chamber heating), open in the OMNIC software on the FTIR computer. b. The software goes through two windows to communicate with the spectrometer. c. Open "Jay's Experimental Setup." d. Make sure "Bench" on top right of window is checked. e. Obtain background scan (scans usually take ~8 minutes for 100 scans). f. Before running the first precursor pulse, pulse water into the chamber to have hydroxyl surface groups on ZrO2 nanopowder. g. Press "Ctr-S" to scan. Confirm the scan, and make sure the peak value should be 5- 10 for optimal performance. Post the scan to Window 1. h. Save the spectra under the format "[Date][Time] ([Cycle#][Precursor]).spa". The .SPA format includes the experimental parameters of the spectrometer. Save water pulse spectra as "([Cycle#][Precursor]W)". i. To make a difference spectra, select the most recent scan, and then select the previous scan. Select "Analyze" and "Subtract". Save the difference spectra as "Subtraction of [Date][Time] ([1st Spectra Cycle Name]-[2nd Spectra Cycle Name])".

126

A4 Temperature and Temperature Control Related Troubleshooting

Problem: Temperature Controller shows high temperature when housing at room temperature. Solutions: • While wearing rubber gloves and de-energize the area, carefully unplug thermocouple input bar at the back of the 7-Channel Controller. • Wait a few seconds, plug back in very carefully. • Turn on power, and wait for readout.

Problem: When controller is activated, temperature rapidly rises. Solutions: • De-energize the affected channel. Make sure heating wire does not cross over thermocouple. • Check the adhesion between the thermocouple and the chamber • Also make sure that the flat of the thermocouple probe is facing the chamber. • Energize the channel, and activate to verify proper control.

Problem: With controller on, there is no change in temperature on the chamber Solutions: • Note: Turn off power to connections before inspecting or changing electrical components. • Make sure the thermocouple responds to a change in temperature. • If the light on the appropriate channel's relay is not on, the problem could be between the controller and the relay. De-energize and trace connections between the relay and the controller (this can be done with the resistance setting on a multimeter) • If the light on the channel is constantly on, make sure there is power from the relay to the heating elements. • Inspect the fuse to the power components. • Using a multimeter, inspect the heating element. If there is no resistance (close to 0 Ω on the readout), then the heating wire has shorted. Replace wrap, inspect fuses and solid state relay, and check with multimeter before energizing the system.

Problem: Sample stage does not reach desired temperature. Solutions: • Verify thermocouple is connected to the sample holder assembly. • Make sure power to the heating elements are on. • Inspect the fuse in the Variable Voltage Transformer (VVT). • Measure resistance of sample heater assembly. • Very slowly adjust the VVT to a higher voltage, if more power is needed. (Note: Should not exceed 4%)

127

A5 LASO ALD Chamber Doser Troubleshooting Guide

Problem: No deposition Solutions: • Make sure that there is precursor in the housing. • Switch the lithium housing to the Mn doser assembly side. 1. Vent chamber. 2. Turn off heater power, and unwrap heating wires from the precursor housings and the gas lines. 3. Remove the precursor housing assemblies. 4. Attach the lithium housing to the doser facing the back of the chamber. 5. Wrap with heating wires and insulation. • Clean clogged doser assembly. 1. Vent chamber 2. Turn off heater power, and unwrap heating wires from the precursor housings and the gas lines. 3. Remove the precursor housing assemblies, and the doser assembly. 4. Rinse gas-lines with copious amounts of isopropyl alcohol. 5. If a precursor is blocking flow, it must be removed by a combination of physical removal and isopropyl rinses. 6. Allow chemical solvents to dry out before placing doser assembly back onto the chamber. 7. Dispose of waste according to proper procedures. Problem: No doser heating Solutions: • Make sure power is on to the doser heating wires. • Unplug power, and measure resistance of the doser heating assembly with a multimeter. • If there is not sufficient resistance, or there is too much resistance, then the heating wires may need to be replaced. 1. Remove doser assembly using the steps in the previous section. 2. Replace the heating elements and secure them to the doser. 3. Make sure other electrical connections are maintained. 4. Replace the assembly, and pump down chamber to base pressure.

128

A6.1 Atomic Force Microscopy Operational Procedures

1. Open Nanoscope 6.12r2 from the desktop, and open Workspace file. 2. Open the hood to the microscope, and unlock the AFM head by tightening the lock screw. Disconnect the head cables, and pull out the head completely and place it on the table. Do not drop the head. Take the probe carrier off the head. 3. Place a probe carrier onto the carrier holder. Use tweezers to place a probe onto the carrier. Place the carrier back on the AFM head. Place the carrier head back on to the microscope and plug in the cables from the head to the machine. 4. Single click on "Meter" on the computer console. The system takes a few minutes to start. The window below displays on one of the computer screens: 5. Pull out the AFM head one more time without disconnecting the cables and hold it upright, observing the projected laser beam. Be careful not to drop the head. Find the tip of the cantilever by looking at the projected laser spot and turning the alignment screws located at the top of the head. This usually results in a diffraction pattern in the projected laser. 6. After the tip has been found, replace the head back onto the machine. Unscrew the screw lock in order to lock the head in position. 7. Check the meter window on the computer screen. The signal level (center blue bar) should be above zero. Maximize the signal by slowly turning the laser alignment screws. Center the red dot on the meter window crosshairs using the detector alignment screws located on the left side of the head. 8. Click on the "Navigate" on the screen. This displays the Real Time Navigate window. Move the stage out so that it completely clears the AFM head (use the arrow buttons on the Navigate window). Place the substrate on the center of the holder plate. Move the stage to approximately the area desired to be scanned. Verify that there is enough vertical clearance. If not, adjust the “Z Motor” upward. 9. Click on the “Locate Tip” button. If the cantilever is not visible on the screen inside the window, click on “Zoom Out” button, or decrease/increase the illumination. Center the tip on the crosshairs using the optical camera adjustment screws (on the side of the camera). Click “Zoom In”, and make sure the tip is centered again. Click “OK” when done. 10. Low the head slowly using the Z-motor arrows on the Navigate window to move the optics and AFM head down towards the surface of the substrate. When the head is within ~5mm, close the Isolation lid and lock it (to limit acoustic and electromagnetic noise), and turn to the Navigate window on the computer screen. From the drop-down menu on “Focus On”, selected “Tip Reflection.” Using the “Fine” movement selection, slowly move the head toward the surface until the reflection of the tip is in focus. Change the “Focus On” selection to “Substrate” and make sure that the surface is in focus (adjust the Z-motor if needed). Use caution as the probe may break at this step. Do not allow the tip to touch the surface of the substrate. Using the Navigate arrows, adjust the substrate to the desired measurement point. 11. Click on "Scan-Dual" icon and set Microscope Mode to “Tapping”. Click the Tune icon. On the Tune Window, make sure the survey frequencies are larger than the resonance

129

frequency of the probe (usually set between 100 kHz and 500 kHz).. Click on “Auto Tune”. Wait a few moments until the probe tuning is completed. Adjust drive frequency by using “Offset” to an appropriate linear response region. Click “Execute” and “Zero Phase.” Record the drive frequency selected (the peak offset should be between 3% and 10%). Exit the window. 12. Click "Dual-Scan" again. In the scan parameters on the right of the window, set an appropriate “Initial Scan Size” (usually 1 μm), “x-“ and “y-offsets” to 0, and the “Scan Angle” to 0. Select appropriate Feedback Controls for the scans. (If not sure what settings to use, set “Integral Gain” to 0.4, “Proportional Gain” to “0.6”, and “Scan Rate” to 1 Hz.) 13. Click on the Engage icon. It takes a few minutes for the microscope to engage. Once the probe engages, the AFM starts scanning. During scanning, check to see if the “Trace” and “Retrace” lines are tracking each other. If they are tracking, the lines should look the same, but they necessarily overlap each other. If they are not tracking well, adjust the “Scan Rate”, “Gains”, and/or “Setpoint” to improve the tracking. 14. Change the “Capture” filename under the “Real Time” menu, then “Capture Filename.” Make sure the destination folder is the user. The “Capture” button captures the image after the scan is complete in the vertical direction. “Capture Now” secures the image at any point in the scan. 15. After scans are complete, click on the Withdraw icon to withdraw the microscope. Move the AFM head up using the Z-motor on the Navigate window. Open the hood, and move the stage fully away from the AFM head using the Navigate window. Remove the substrate, and move the stage back under the AFM head. Unlock the AFM head by tightening the lock screw. Disconnect the head cables, and pull the head out completely. Place the head on the table, using care not to drop the head. Remove the probe carrier from the head and place on the carrier holder. Remove the probe with tweezers, and slide the head back into the microscope. Loose the lock screw to fix the head to the microscope. Plug cables back in, and close the hood. Close the Nanoscope program window, saving changes to the user’s workspace. Record session in the log book, and sign the AFM out using the LabRunner program.

130

A6.2 Scanning Electron Microscopy (NOVA 600) Operational Procedures

A6.2.1 Introduction

A beam of high energy electrons bombards the sample and through a series of scattering events, numerous types of particles are ejected from the sample in a limited interaction region. A detector collects specific ejected particles as the beam is scanned across the sample to produce an image. If the sample is insulating or improperly grounded, it charges up due to the continuous influx of charged particles from the beam, and do not convey an accurate scanning image. Common charging effects include image drift, over/under-saturation, and the complete reflection of the beam onto the surrounding chamber.

A6.2.2 Guidelines

1) Fill out the logbook when you come into the room and record your logout time upon leaving.

2) The phone in the room can be used by all individuals for matters related to the use and scheduling of the Nova 600 system only.

3) The Nova 600 room is kept in impeccable order. This is not by accident. You are expected to use and maintain the facility in a similar fashion.

4) Never touch the chamber internals, stage mounts, or sample with your bare hands. Always wear gloves.

5) Checklist a) Check if the chamber is still under vacuum. b) Check the vacuum levels i) Hold the mouse pointer over the status icon. ii) FEG IGP < 5E-9 mbar iii) E Column IGP < 5E-7 mbar iv) I Column IGP < 5E-7 mbar v) Chamber < 9E-5 mbar

A6.2.3 Sample Loading and Removal

1) Prepare the microscope. a) Check that E-beam and I-beam are off (Beam On button should be gray): b) Check the navigation page and set the tilt angle to zero if it is not already

2) Vent the chamber. a) Click Vent in the beam control page. b) Venting takes approximately 5 minutes. c) Once a space between the chamber and door become visible, lightly pull on the door 131

handle. Do not pull hard on the door handle. Once the chamber is completely vented, the door easily slides open. 3) Slowly open the chamber door until the door-stop is hit. Load or remove your sample. (WARNING – All screws, nuts, ect. should be tightened to minimum necessary to hold sample rigid) a) Mount the stub holder. i) Center the stage by setting the X and Y coordinates to 0 if access to the central threaded hole is difficult. ii) Screw the threaded post approximately 1 inch into the base plate if it is not already installed. iii) Ensure that the locking nut is sufficiently raised to prevent binding of the nut during movement of the treaded post. iv) Rotate the post until the setscrew is visible and easily accessible with a hex driver. b) Insert the pin mount stub. i) Gently insert the pin into the top hole of the threaded post. ii) If there is resistance, remove the stub and back the setscrew off 1 turn in the counterclockwise direction. If there is still resistance, your pin diameter may be too large. iii) When inserting the pin, you should feel the pin pushing against the D-spring. c) Check the sample height. i) The height of your sample is extremely important. If it is too high, you may damage the microscope's internal components when closing the door. If it is too low, you may not be able to raise your sample to the eucentric working distance and be unable to operate in mode II. ii) Place the height measuring guide (found in the Ziploc bag on top of the chamber) on the base plate with the tab near the highest point on your sample. The top of your sample should lie between the top and bottom of the tab. iii) To adjust the height of the sample, first remove the height measuring guide and the pin stub mount from the chamber. Rotate the threaded post, reinsert your stub, and recheck the height of the sample. Repeat as necessary. iv) Never leave the height measuring guide in the chamber while rotating the threaded post. This is to prevent knocking it into the chamber. v) Return the measuring guide to the Ziploc bag on top of the chamber. d) Secure the pin mount stub. i) Lightly tighten the setscrew. Warning - This is the #1 part on the system that is damaged. Users are financially responsible for this part if any damage occurs. ii) Finger-tighten the locking nut to the base plate.

4) Pump down the chamber. a) Slowly close the chamber door. b) Click Pump, and gently press against the door for about 5 seconds until the vacuum seal is formed.

132

A6.2.4 E-Beam Startup

1) Turn on the E-beam. a) Select the upper-left quadrant. The data bar should turn blue. b) Turn on the high voltage by clicking Beam On (button should be yellow). i) Default parameters: 10 kV, spot 3 or 4 (0.13 nA or 0.54 nA). c) Un-pause the image. d) Zero the beam shift: right click the write area of the gauge in the beam control panel and select “zero”.

2) Find your sample. a) Reduce the magnification to the minimum value. b) Navigate to your sample. i) Important - When the stage is moved, two distinct 'click' noises can be heard from the cabinets behind the microscope; one before the movement and one following the movement. It is critical that you wait until you hear the second 'click' before proceeding with any additional stage movements. ii) Double-click the left mouse button in the imaging window to center a feature. iii) Click and hold down the mouse wheel with the pointer in the imaging window and drag the mouse. iv) Use the arrow keys to move a single frame over (↑↓←→). v) Stage Align Feature: Drag a line in the imaging window on top of the feature which you want to have rotated to either a vertical or horizontal orientation. vi) Navigation Panel → Coordinates Tab: Coordinates can be manually entered into any of the available axis (X, Y, Z, T, R). c) Adjust the contrast, brightness, and focus to optimize the image.

3) Link the stage (Very Important!) a) After the initial pump-down, the link button should have a question mark on it. This indicates that the software does not know what height the sample is at. In the navigation page the Z value should read zero. b) Increase the magnification to >10,000X on a small feature of high contrast (scratches, dirt and defects are a good place). c) Adjust the contrast, brightness, and focus to optimize the image. d) Press Link Z to FWD and check the navigation page. e) The Z value should now read between 11 and 15 mm. If not, your sample was loaded improperly. You do need to vent and reload your sample.

4) Adjust the sample height to 5 mm. a) Change the Z value to 5 mm. The link button changes indicating that re-linking the stage is recommended. 5 mm is optimal for finding eucentric height and using mode II. Never decrease the Z value to less than 4 mm. b) Repeat the above linking procedure by first refocusing the beam and then pressing Link Z to FWD (ignoring the last bullet). The Z value should update to a value near 5.0000 mm.

133

A6.2.5 Optimizing an Image

1) Focusing Tools a) Course and fine focus knobs are available on the console. Course focus works best for mode I and fine focus best for mode II. b) Reduced scanning mode (F7) provides a lower noise image ideal for fine focusing. The reduced scanning area can be freely repositioned and resized. c) Full screen imaging is accessible by pressing F5.

2) Contrast and Brightness a) The videoscope (F3) presents a graph of the center line scan and how each pixel falls between black and white. It is useful for determining if the grayscale range is being clipped. b) Contrast and brightness numeric values can be found at the Contrast bottom of the beam page (among others). c) Auto contrast and auto brightness buttons are available to you, though not recommended due to their poor results.

3) Astigmatism a) Identifiable by perpendicular streaking of the image during focusing. It is most visible at higher magnifications. b) Adjust the focus until the image is midway between the two streaking directions and appears blurry but not streaked. c) Optimize the image using the two stigmation adjustment knobs. d) Repeat the process of focusing and adjusting the X and Y stigmation at higher and higher magnification until streaking is no longer apparent and optimal clarity has been achieved.

4) Mode I and Mode II a) Similar to "low mag" and "high mag" modes on other SEM systems. However, mode I can operate at all magnifications and produce excellent images. b) If mode II is not available, hold the mouse over the Mode button to receive an explanation as to why. There are stage-Z limitations as well as magnification limitations to mode II. c) Course focus works best in mode I and fine focus works best in mode II, though both are operational in either mode.

5) Dwell Time / Resolution / Averaging a) Each quadrant has its own separate settings for these parameters in both full mode and reduced scanning mode. b) These settings affect the live imaging parameters, not the photo or snapshot parameters. c) Often, samples charge due to a large dwell time. Reduce it, increase the resolution, and save the paused image for better results.

134

A6.2.6 Taking an Image

1) Basics a) After taking an image, the system always becomes paused. Un-pause the image to return to live imaging. b) No images are automatically saved. c) Snapshot (F4): short dwell time image. d) Useful for quick previews of your image optimization settings (see previous section) and for “good-enough” images where precise detail isn’t necessary. Also useful for when your sample is charging. e) Saving of snapshots must be done manually by going to File > Save As.

2) Photo (F2): long dwell time image. a) Used for taking high detail images. b) The Save As window pops open automatically following every photo. c) To cancel a photo scan midway, press the pause button twice.

3) Saving an Image. a) All images are to be saved in the folder for you particular job number in the appropriate month’s folder. b) Folder naming convention: [4-digit job number] + space + “-“ + space + [full name] c) File naming convention: [sample name] + “_001”. By placing “_001” at the end of the file name, all further files can be saved with the same name and the numeric value increasing incrementally until a new file name is manually input.

4) Taking your files with you. a) All files are taken via user supplied USB thumb-drive. b) Always leave the folder behind on the computer even if you decided to delete all of your files from the job. c) You are welcome to leave a copy of your images stored on the computer for backup purposes, although file storage cannot be guaranteed beyond 3 months.

135

A6.3 X-Ray Photoemission Spectroscopy System

Emergency Shutdown Procedure

1. Move the transfer loading arm all the way out of the XPS chamber, then close the gate valve. 2. Turn off X-ray source by pressing HV button (light turns off ) and then STANDBY button (light turns on, and OPER button should become unlit). If there is time, wait for water to flow to cool the anode for ~ 1 minute and press COOLING ON button. Then, turn off the main power and chiller power. 3. Turn off energy analyzer by turning off ESCA kV and power switches on boxes from bottom to top (lens, detector, control and display) 4. Turn off software by pressing the Stop Expt/Kill Data switch until experiment stops.

Operating Procedure

1. Before transferring the sample, make sure the x-ray gun is backed out to a setting of about 1.5 mm on the motion manipulator (~1 cm from the Teflon spacer). Move the sample stage towards the gate valve to a setting of 1-2 on the corresponding manipulator). Lower the stage to a setting of ~4 on the corresponding manipulator. Ensure the stage is vertical. Be careful as the stage rotates easily.

2. Check the pressure on the XPS using the ion gauge controller (base is ~1 x 10-9 Torr, but anything below 3 x 10-9 Torr is acceptable. If the pressure is above the 10-9 Torr range check the ion pump controller for current reading – should read ~105 – 106 A). Check the pressure on the TT using its ion gauge controller, pressure should be ~4 x 10-9 Torr (if it is higher check that the gate valve between the cryo and the TT is open, if it is not, then open it using the controller box (trace the cable to be sure)). Once both pressures are in the requisite range open the gate valve between the TT and XPS using the corresponding controller box (trace the cable if unsure).

3. Slide the transfer arm into the tube slowly, checking the viewport occasionally to make sure that you do not collide with the stage or x-ray source. The manipulator settings on the arm should be X-setting = 5.5 (black) and Y-setting = 5 (black). Make sure the stage is aligned with the sample using the lateral manipulator (pictured in Step 1 on previous page). At this point you may have to raise the sample or lower the stage to provide the necessary vertical clearance. Lower the sample onto the stage. You may have to lower the loading arm to a Y-setting of 3-4 in order to remove the fork from the sample – slowly pull out the fork and after each small movement, increase the Y-setting slightly to avoid excessive straining. Retract the fork completely, being cautious to clear the transfer cart, and close the gate valve between the tube and the XPS.

4. Move the sample stage closer to the source and move source in until it is about 0.25 cm away from the Teflon spacer. Select the angle theta, which is the angle between the sample surface and the detector axis, using the rotational control - 240° corresponds to a 90° take off angle and 270° to a 60° take-off angle. 136

5. Move x-ray source closer if desired and adjust the stage position using the various manipulators until you are satisfied that the beam illuminates the sample. Check the pressure in the XPS chamber – it should be at or below 3 x 10-9 Torr. Turn on the analyzer by turning on the switches labeled in the left picture below – from top to bottom: Display, Control, Detector, Lens, then ESCA kV. Turn on the x-ray source by pressing the power switch on the Control Unit. Press “Cooling On” (this should result in the button beginning to blink with a green light) and then turn on the NESLAB HX-75 Chiller (This should result in the green light becoming steady). Select an anode by pressing the corresponding button (selection results in the button being illuminated by a green light). At this point none of the Interlock Buttons should be lit red. If they are seek assistance from a senior lab member or cautiously check all connections from electronics to the unit.

6. Check the chiller setpoint by pushing in the chiller mode switch. This should read 17 degrees. If it has changed adjust the setpoint dial while keeping the mode switch depressed. Check the flow rate by plugging in the BNC/Banana Plug connection to the multimeter. When set to measure DC Voltage the multimeter should give a reading of over 3.4 V. All the electronics to warm up for 10 minutes before running an experiment.

7. Launch the software program SPECTRA by double clicking the icon labeled SPECTRA in the center of the desktop. To save time before the experiment begins you can define your regions (survey, detailed, etc) by clicking the Edit Region Info Button. Also confirm that the software is configured for the anode you selected in Step 5, by clicking the Tools Button, then clicking “Setup Card” in the menu that pops up. In the dialogue box that pops up you can select either Mg (1253.6 eV) or Al (1486.6 eV) in the Excitation field.

8. Once the electronics have warmed up, check that the Spectrometer Control Unit is set to COMP mode and that the Energy Selector corresponds to the Binding Energy (BE) of the anode you selected earlier in Step 5.

9. Press the “IFIL” button on the X-Ray Source Control Unit, which should light up green. Press the “Stand By” button (display should change to 2.91 A) and monitor the pressure in the chamber, which rises to as high as 2 x 10-8 Torr, but should decrease quickly. When it has returned to 10-9 Torr, press the “HV on” button and adjust the dial so that the display reads ~4.00 – 5.00 kV. Monitor the pressure, which should still be in the 10-9 range (Note: These values are for Mg anode, Al may be higher). Press the “Oper” button, the value of IFIL should rise to ~ 3.6 A and the light should blink briefly before becoming steady again. Press the “IE” button (should light up green) and use the left and right dials to gradually increase the IE and HV values to the desired reading (14 mA for IE and 10 kV for HV are standard, 18 mA for IE and 12 kV are maximum for Mg anode, 14 kV can be used for Al anode). In the SPECTRA window press the circular button marked “Go” to the right of the Edit Regions Button.

10. Once the scans are finished, ramp down the HV and IE values by turning down the dials slowly until they are at about 4 kV and 5 mA respectively, then Press “Oper” (the light 137

should go off and the “Standby” button should light up. Press “HV” (the light should go off), then press “Standby” (the light should go off. If you are going to do more scans but the pressure has increased significantly over the course of your run, the system can be left in this state while it pumps down. If the pressure has not increased significantly, you can adjust the region info and then run more scans at your discretion without having to turn off the HV/IE switches.

11. If you do not run anymore scans, leave the “Cooling On” button on to let the anode cool down for ~5-10 minutes. Press “Cooling On”, then press the power switch. Turn off the chiller. Turn off the analyzer electronics by turning off ESCA kV first, then the boxes from bottom to top (in reverse order from Step 5). Remove your sample from the chamber if desired.

138

BIBLIOGRAPHY

Agapov, A. L. and A. P. Sokolov (2011). "Decoupling Ionic Conductivity from Structural Relaxation: A Way to Solid Polymer Electrolytes?" Macromolecules 44(11): 4410-4414.

Agrawal, R. C. and G. P. Pandey (2008). "Solid polymer electrolytes: materials designing and all-solid-state battery applications: an overview." Journal of Physics D: Applied Physics 41(22): 223001.

Aifantis, K. (2010). Next-Generation Anodes for Secondary Li-Ion Batteries. High Energy Density Lithium Batteries, Wiley-VCH Verlag GmbH & Co. KGaA: 129-164.

Alias, N. and A. A. Mohamad (2015). "Advances of aqueous rechargeable lithium-ion battery: a review." Journal of Power Sources 274: 237-251.

Amato, M., M. Palummo, R. Rurali and S. Ossicini (2014). "Silicon-germanium nanowires: chemistry and physics in play, from basic principles to advanced applications." Chemical reviews 114: 1371-1412.

Armand, M. and J. M. Tarascon (2008). "Building better batteries." Nature 451(7179): 652-657.

Aurbach, D. (2003). "Electrode–solution interactions in Li-ion batteries: a short summary and new insights." Journal of Power Sources 119–121(0): 497-503.

Baggetto, L., H. C. M. Knoops, R. A. H. Niessen, W. M. M. Kessels and P. H. L. Notten (2010). "3D negative electrode stacks for integrated all-solid-state lithium-ion microbatteries." Journal of Materials Chemistry 20(18): 3703-3708.

Baggetto, L., R. A. H. Niessen and P. H. L. Notten (2009). "On the activation and charge transfer kinetics of evaporated silicon electrode/electrolyte interfaces." Electrochimica Acta 54(24): 5937-5941.

Bard, A. J. and L. R. Faulkner (2001). Electrochemical Methods: Fundamentals and Applications. New York, Wiley & Sons, Inc.

Barsoukov, E. and J. R. Macdonald (2005). Impedance Spectroscopy: Theory, Experiment, and Applications, John Wiley & Sons, Inc.

Bogart, T. D., A. M. Chockla and B. a. Korgel (2013). "High capacity lithium ion battery anodes of silicon and germanium." Current Opinion in Chemical Engineering 2: 286-293.

Bohnke, O. (2008). "The fast lithium-ion conducting oxides Li3xLa2/3−xTiO3 from fundamentals to application." Solid State Ionics 179(1–6): 9-15.

Bouchet, R., S. Maria, R. Meziane, A. Aboulaich, L. Lienafa, J.-P. Bonnet, T. N. T. Phan, D. Bertin, D. Gigmes, D. Devaux, R. Denoyel and M. Armand (2013). "Single-ion BAB

139

triblock copolymers as highly efficient electrolytes for lithium-metal batteries." Nat Mater 12(5): 452-457.

Briggs, D. and M. P. Seah (1990). Practical Surface Analysis: Auger and X-ray Photoelectron Spectroscopy. Chichester, UK, John Wiley and Sons.

Cao, C., Z.-B. Li, X.-L. Wang, X.-B. Zhao and W.-Q. Han (2014). "Recent advances in inorganic solid electrolytes for lithium batteries." Front. Energy Res. 2(25).

Cavanagh, A. S. S., Y. Lee, B. Yoon and S. M. M. George (2010). "Atomic Layer Deposition of LiOH and Li2CO3 Using Lithium t-Butoxide as the Lithium Source." ECS Transactions 33: 223-229.

Chan, C. K., H. Peng, G. Liu, K. McIlwrath, X. F. Zhang, R. A. Huggins and Y. Cui (2008). "High-performance lithium battery anodes using silicon nanowires." Nat Nano 3(1): 31- 35.

Chen, N., B. Reeja-Jayan, J. Lau, P. Moni, A. Liu, B. Dunn and K. K. Gleason (2015). "Nanoscale, conformal polysiloxane thin film electrolytes for three-dimensional battery architectures." Materials Horizons 2(3): 309-314.

Christie, A. M., S. J. Lilley, E. Staunton, Y. G. Andreev and P. G. Bruce (2005). "Increasing the conductivity of crystalline polymer electrolytes." Nature 433(7021): 50-53.

Comstock, D. J. and J. W. Elam (2013). "Mechanistic Study of Lithium Aluminum Oxide Atomic Layer Deposition." The Journal of Physical Chemistry C 117: 1677-1683.

Cui, L.-F., Y. Yang, C.-M. Hsu and Y. Cui (2009). "Carbon−Silicon Core−Shell Nanowires as High Capacity Electrode for Lithium Ion Batteries." Nano Letters 9(9): 3370-3374.

D. Shannon, R. (2002). "Refractive Index and Dispersion of Fluorides and Oxides." Journal of Physical and Chemical Reference Data 31: 931.

Daniel, M., S. Leland, S. Kyeong-Sik, C. Chi On and D. Bruce (2015). "A high-energy-density quasi-solid-state carbon nanotube electrochemical double-layer capacitor with ionogel electrolyte." Translational Materials Research 2(1): 015001.

Dasgupta, N. P., J. Sun, C. Liu, S. Brittman, S. C. Andrews, J. Lim, H. Gao, R. Yan and P. Yang (2014). "25th Anniversary Article: Semiconductor Nanowires – Synthesis, Characterization, and Applications." Advanced Materials 26(14): 2137-2184.

Dayeh, S. A., A. V. Gin and S. T. Picraux (2011). "Advanced core/multishell germanium/silicon nanowire heterostructures: Morphology and transport." Applied Physics Letters 98(16): 163112.

140

Diddens, D., A. Heuer and O. Borodin (2010). "Understanding the Lithium Transport within a Rouse-Based Model for a PEO/LiTFSI Polymer Electrolyte." Macromolecules 43(4): 2028-2036.

DiLeo, R. A., S. Frisco, M. J. Ganter, R. E. Rogers, R. P. Raffaelle and B. J. Landi (2011). "Hybrid Germanium Nanoparticle–Single-Wall Carbon Nanotube Free-Standing Anodes for Lithium Ion Batteries." The Journal of Physical Chemistry C 115(45): 22609-22614.

Dunn, B., J. W. Long and D. R. Rolison (2008). "Rethinking Multifunction in Three Dimensions for Miniaturizing Electrical Energy Storage." 49-53.

Fergus, J. W. (2010). "Ceramic and polymeric solid electrolytes for lithium-ion batteries." Journal of Power Sources 195(15): 4554-4569.

Ferguson, J. D., E. R. Smith, A. W. Weimer and S. M. George (2004). "ALD of SiO2 at Room Temperature Using TEOS and H2O with NH3 as the Catalyst." Journal of The Electrochemical Society 151(8): G528-G535.

Fujiwara, H. (2007). Introduction to Spectroscopic Ellipsometry. Spectroscopic Ellipsometry John Wiley & Sons, LTd: 1-11.

Ge, M., J. Rong, X. Fang and C. Zhou (2012). "Porous doped silicon nanowires for lithium ion battery anode with long cycle life." Nano letters 12: 2318-2323.

George, S. M. (2010). "Atomic layer deposition: an overview." Chemical reviews 110: 111-131.

Goldstein, D. N., J. A. McCormick and S. M. George (2008). "Al2O3 Atomic Layer Deposition with Trimethylaluminum and Ozone Studied by in Situ Transmission FTIR Spectroscopy and Quadrupole Mass Spectrometry." The Journal of Physical Chemistry C 112(49): 19530-19539.

Goodenough, J. B. and K.-S. Park (2013). "The Li-Ion Rechargeable Battery: A Perspective." Journal of the American Chemical Society 135(4): 1167-1176.

Griffiths, P. R. and J. A. De Haseth (2007). Fourier Transform Infrared Spectrometry, John Wiley & Sons, Inc.

Grünebaum, M., M. M. Hiller, S. Jankowsky, S. Jeschke, B. Pohl, T. Schürmann, P. Vettikuzha, A.-C. Gentschev, R. Stolina, R. Müller and H.-D. Wiemhöfer (2014). "Synthesis and electrochemistry of polymer based electrolytes for lithium batteries." Progress in Solid State Chemistry 42(4): 85-105.

Hallinan, D. T. and N. P. Balsara (2013). "Polymer Electrolytes." Annual Review of Materials Research 43(1): 503-525.

Hämäläinen, J., J. Ihanus, T. Sajavaara, M. Ritala and M. Leskelä (2011). "Atomic Layer Deposition and Characterization of Aluminum Silicate Thin Films for Optical Applications." Journal of The Electrochemical Society 158(2): P15-P21.

141

Hayashi, A., K. Noi, A. Sakuda and M. Tatsumisago (2012). "Superionic glass-ceramic electrolytes for room-temperature rechargeable sodium batteries." Nat Commun 3: 856.

Heo, J. H., S. H. Im, J. H. Noh, T. N. Mandal, C.-S. Lim, J. A. Chang, Y. H. Lee, H.-j. Kim, A. Sarkar, K. NazeeruddinMd, M. Gratzel and S. I. Seok (2013). "Efficient inorganic- organic hybrid heterojunction solar cells containing perovskite compound and polymeric hole conductors." Nat Photon 7(6): 486-491.

Hochbaum, A. I. and P. Yang (2010). "Semiconductor Nanowires for Energy Conversion." Chemical Reviews 110(1): 527-546.

Huang, J. Y., L. Zhong, C. M. Wang, J. P. Sullivan, W. Xu, L. Q. Zhang, S. X. Mao, N. S. Hudak, X. H. Liu, A. Subramanian, H. Fan, L. Qi, A. Kushima and J. Li (2010). "In situ observation of the electrochemical lithiation of a single SnO2 nanowire electrode." Science (New York, N.Y.) 330: 1515-1520.

Inaguma, Y., C. Liquan, M. Itoh, T. Nakamura, T. Uchida, H. Ikuta and M. Wakihara (1993). "High ionic conductivity in lithium lanthanum titanate." Solid State Communications 86(10): 689-693.

Ji Woo, K., J. J. Travis, H. Enyuan, N. Kyung-Wan, K. Seul Cham, K. Chan Soon, W. Jae-Ha, Y. Xiao-Qing, S. M. George, O. Kyu Hwan, C. Sung-Jin and L. Se-Hee (2014). "Unexpected high power performance of atomic layer deposition coated Li Ni1/3Mn1/3Co1/3O2 cathodes." Journal of Power Sources 254: 190-197.

Jochum, T. and I. Reimanis (2010). "Reactions in Eucryptite-Based Lithium Aluminum Silicates." Journal of the American Ceramic Society 1596: 1591-1596.

Jung, Y. S., P. Lu, A. S. Cavanagh, C. Ban, G.-H. Kim, S.-H. Lee, S. M. George, S. J. Harris and A. C. Dillon (2013). "Unexpected Improved Performance of ALD Coated LiCoO2 /Graphite Li-Ion Batteries." Advanced Energy Materials 3: 213-219.

Kennedy, T., E. Mullane, H. Geaney, M. Osiak, C. O'Dwyer and K. M. Ryan (2014). "High- Performance Germanium Nanowire-Based Lithium-Ion Battery Anodes Extending over 1000 Cycles Through in Situ Formation of a Continuous Porous Network." Nano letters 14: 716-723.

Kennedy, T., E. Mullane, H. Geaney, M. Osiak, C. O’Dwyer and K. M. Ryan (2014). "High- Performance Germanium Nanowire-Based Lithium-Ion Battery Anodes Extending over 1000 Cycles Through in Situ Formation of a Continuous Porous Network." Nano Letters 14(2): 716-723.

Kim, J. G., B. Son, S. Mukherjee, N. Schuppert, A. Bates, O. Kwon, M. J. Choi, H. Y. Chung and S. Park (2015). "A review of lithium and non-lithium based solid state batteries." Journal of Power Sources 282(0): 299-322.

Kim, Y. G. and H. N. G. Wadley (2009). "Plasma-assisted deposition of lithium phosphorus oxynitride films: Substrate bias effects." Journal of Power Sources 187(2): 591-598.

142

Knauth, P. (2009). "Inorganic solid Li ion conductors: An overview." Solid State Ionics 180: 911-916.

Kumar, R. and T. Sarakonsri (2010). A Review of Materials and Chemistry for Secondary Batteries. High Energy Density Lithium Batteries, Wiley-CVH Verlog GmbH & Co. KGaA: 53-80.

Kwan, M. C. and K. K. Gleason (1997). "Pyrolytic CVD of poly(organosiloxane) thin films." Chemical Vapor Deposition 3(6): 299-301.

Langereis, E., S. B. S. Heil, H. C. M. Knoops, W. Keuning, M. C. M. v. d. Sanden and W. M. M. Kessels (2009). "In situ spectroscopic ellipsometry as a versatile tool for studying atomic layer deposition." Journal of Physics D: Applied Physics 42(7): 073001.

Lau, K. K. S. (2000). Chemical Vapor Deposition of Fluorocarbon Films for Low Dielectric Constant Thin Film Applications. Chemical Engineering. Cambridge, MA, Massachusetts Institute of Technology. Ph.D.: 230.

Lau, K. K. S. and K. K. Gleason (2006). "Initiated Chemical Vapor Deposition (iCVD) of Poly(alkyl acrylates): A Kinetic Model." Macromolecules 39(10): 3695-3703.

Lau, K. K. S. and K. K. Gleason (2006). "Initiated Chemical Vapor Deposition (iCVD) of Poly(alkyl acrylates): An Experimental Study." Macromolecules 39(10): 3688-3694.

Lee, Y.-S., S. H. Ju, J.-H. Kim, S. S. Hwang, J.-M. Choi, Y.-K. Sun, H. Kim, B. Scrosati and D.- W. Kim (2012). "Composite gel polymer electrolytes containing core-shell structured SiO2 (Li+) particles for lithium-ion polymer batteries." Electrochemistry Communications 17: 18-21.

Li, B. and X. Liu (2011). Infrared Spectroscopic Ellipsometry for Ion-Implanted Silicon Wafers.

Li, L., K. S. Lee and L. Lu (2014). "Li-rich layer-structured cathode materials for high energy Li-ion batteries." Functional Materials Letters 7(4): 11.

Liang, C., M. X. Gao, H. G. Pan, Y. F. Liu and M. Yan (2013). "Lithium alloys and metal oxides as high-capacity anode materials for lithium-ion batteries." Journal of Alloys and Compounds 575: 246-256.

Liu, J. and X. Sun (2015). "Elegant design of electrode and electrode/electrolyte interface in lithium-ion batteries by atomic layer deposition." Nanotechnology 26(2): 024001.

Liu, X. H. and J. Y. Huang (2011). "In situ TEM electrochemistry of anode materials in lithium ion batteries." Energy & Environmental Science 4: 3844.

Liu, X. H., S. Huang, S. T. Picraux, J. Li, T. Zhu and J. Y. Huang (2011). "Reversible Nanopore Formation in Ge Nanowires during Lithiation–Delithiation Cycling: An In Situ Transmission Electron Microscopy Study." Nano Letters 11(9): 3991-3997.

143

Liu, X. H., J. W. Wang, S. Huang, F. F. Fan, X. Huang, Y. Liu, S. Krylyuk, J. Yoo, S. A. Dayeh, A. V. Davydov, S. X. Mao, S. T. Picraux, S. L. Zhang, J. Li, T. Zhu and J. Y. Huang (2012). "In situ atomic-scale imaging of electrochemical lithiation in silicon." Nature Nanotechnology 7: 749-756.

Long, J. W., B. Dunn, D. R. Rolison and H. S. White (2004). "Three-dimensional battery architectures." Chemical reviews 104: 4463-4492.

Lotfabad, E. M., P. Kalisvaart, A. Kohandehghan, K. Cui, M. Kupsta, B. Farbod and D. Mitlin (2014). "Si nanotubes ALD coated with TiO2, TiN or Al2O3 as high performance lithium ion battery anodes." Journal of Materials Chemistry A 2: 2504.

Lu, Y. and A. Lal (2010). "High-Efficiency Ordered Silicon Nano-Conical-Frustum Array Solar Cells by Self-Powered Parallel Electron Lithography." Nano Letters 10(11): 4651-4656.

Mukaibo, H. and T. Momma (2010). Nanotechnologies for Li Batteries. Electrochemical Nanotechnologies. T. Osaka, M. Datta and Y. Shacham-Diamand. New York, Springer: 7-22.

Nagel, W. and H. Bohm (1982). "Ionic-Conductivity Studies on LiAlSiO4-SiO2 Solid Solutions of the High Quartz Type." Solid State Communications 42(9): 625-631.

Nan, C.-W., L. Fan, Y. Lin and Q. Cai (2003). "Enhanced Ionic Conductivity of Polymer Electrolytes Containing Nanocomposite SiO2 Particles." Physical Review Letters 91(26): 266104.

Nishi, Y. (2000). "Lithium ion secondary battery technologies, present and future." Macromolecular Symposia 156(1): 187-194.

Nishi, Y. (2001). "Lithium ion secondary batteries; past 10 years and the future." Journal of Power Sources 100(1-2): 101-106.

Nitta, N. and G. Yushin (2014). "High-Capacity Anode Materials for Lithium- Ion Batteries: Choice of Elements and Structures for Active Particles." Particle & Particle Systems Characterization 31(3): 317-336.

Notten, P. H. L., F. Roozeboom, R. A. H. Niessen and L. Baggetto (2007). "3-D integrated all- solid-state rechargeable batteries." Advanced Materials 19(24): 4564-4567.

O'Shaughnessy, W. S., S. K. Murthy, D. J. Edell and K. K. Gleason (2007). "Stable Biopassive Insulation Synthesized by Initiated Chemical Vapor Deposition of Poly(1,3,5- trivinyltrimethylcyclotrisiloxane)." Biomacromolecules 8(8): 2564-2570.

Ohara, S., J. Suzuki, K. Sekine and T. Takamura (2004). "A thin film silicon anode for Li-ion batteries having a very large specific capacity and long cycle life." Journal of Power Sources 136(2): 303-306.

144

Ohkoshi, S.-i., T. Nuida, T. Matsuda, H. Tokoro and K. Hashimoto (2005). "The dielectric constant in a thermal phase transition magnetic material composed of rubidium manganese hexacyanoferrate observed by spectroscopic ellipsometry." Journal of Materials Chemistry 15(32): 3291-3295.

Ohta, S., T. Kobayashi, J. Seki and T. Asaoka (2012). "Electrochemical performance of an all- solid-state lithium ion battery with garnet-type oxide electrolyte." Journal of Power Sources 202: 332-335.

Osman, Z., M. I. Mohd Ghazali, L. Othman and K. B. Md Isa (2012). "AC ionic conductivity and DC polarization method of lithium ion transport in PMMA–LiBF4 gel polymer electrolytes." Results in Physics 2: 1-4.

Palacin, M. R. (2009). "Recent advances in rechargeable battery materials: a chemist's perspective." Chemical Society reviews 38(9): 2565-2575.

Park, M., X. Zhang, M. Chung, G. B. Less and A. M. Sastry (2010). "A review of conduction phenomena in Li-ion batteries." Journal of Power Sources 195: 7904-7929.

Pechenik, A., D. H. Whitmore, S. Susman and M. A. Ratner (1988). "Transport in Glassy Fast- Ion Conductors - A Study of LiAlSiO4 Glass." Journal of Non-Crystalline Solids 101(1): 54-64.

Pop, V., H. J. Bergveld, P. H. L. Notten and P. P. L. Regtien (2005). "State-of-the-art of battery state-of-charge determination." Measurement Science and Technology 16(12): R93.

Qu, Y. Q., H. L. Zhou and X. F. Duan (2011). "Porous silicon nanowires." Nanoscale 3: 4060- 4068.

Quartarone, E. and P. Mustarelli (2011). "Electrolytes for solid-state lithium rechargeable batteries: recent advances and perspectives." Chemical Society reviews 40(5): 2525-2540.

Rao, M., X. Geng, Y. Liao, S. Hu and W. Li (2012). "Preparation and performance of gel polymer electrolyte based on electrospun polymer membrane and ionic liquid for lithium ion battery." Journal of Membrane Science 399–400: 37-42.

Reeja-Jayan, B., N. Chen, J. Lau, J. A. Kattirtzi, P. Moni, A. Liu, I. G. Miller, R. Kayser, A. P. Willard, B. Dunn and K. K. Gleason (2015). "A Group of Cyclic Siloxane and Silazane Polymer Films as Nanoscale Electrolytes for Microbattery Architectures." Macromolecules.

Roberts, M., P. Johns, J. Owen, D. Brandell, K. Edstrom, G. El Enany, C. Guery, D. Golodnitsky, M. Lacey, C. Lecoeur, H. Mazor, E. Peled, E. Perre, M. M. Shaijumon, P. Simon and P.- L. Taberna (2011). "3D lithium ion batteries—from fundamentals to fabrication." Journal of Materials Chemistry 21: 9876.

Roy, R. and E. F. Osborn (1949). "The System Lithium Metasilicate Spodumene Silica." Journal of the American Chemical Society 71(6): 2086-2095.

145

Ruzmetov, D., V. P. Oleshko, P. M. Haney, H. J. Lezec, K. Karki, K. H. Baloch, A. K. Agrawal, A. V. Davydov, S. Krylyuk, Y. Liu, J. Huang, M. Tanase, J. Cumings and a. A. Talin (2012). "Electrolyte stability determines scaling limits for solid-state 3D Li ion batteries." Nano letters 12: 505-511.

Sarakonsri, T. and R. Kumar (2010). Introduction to Electrochemical Cells. High Energy Density Lithium Batteries. K. Aifantis, Wiley-VCH Verlag GmbH & Co. KGaA: 1-25.

Seidel, S. M., S. Jeschke, P. Vettikuzha and H. D. Wiemhofer (2015). "PVDF-HFP/ether- modified polysiloxane membranes obtained via airbrush spraying as active separators for application in lithium ion batteries." Chemical Communications 51(60): 12048-12051.

Shi, J., Y. Xia, S. Han, L. Fang, M. Pan, X. Xu and Z. Liu (2015). "Lithium ion conductive Li1.5Al0.5Ge1.5(PO4)3 based inorganic–organic composite separator with enhanced thermal stability and excellent electrochemical performances in 5 V lithium ion batteries." Journal of Power Sources 273: 389-395.

Shi, S., P. Lu, Z. Liu, Y. Qi, L. G. Hector, H. Li and S. J. Harris (2012). "Direct Calculation of Li-Ion Transport in the Solid Electrolyte Interphase." Journal of the American Chemical Society 134(37): 15476-15487.

Shin-ichi, F., S. Satoshi, S. Kaduhiro and T. Hitoshi (2004). "Preparation and ionic conductivity of beta-LiAlSiO4 thin film." Solid State Ionics 167(3-4): 325-329.

Suzuki, N., S. Shirai, N. Takahashi, T. Inaba and T. Shiga (2011). "A lithium phosphorous oxynitride (LiPON) film sputtered from unsintered Li3PO4 powder target." Solid State Ionics 191(1): 49-54.

Szczech, J. R. and S. Jin (2011). "Nanostructured silicon for high capacity lithium battery anodes." Energy & Environmental Science 4(1): 56-72.

Taillades, G. and J. Sarradin (2004). "Silver: high performance anode for thin film lithium ion batteries." Journal of Power Sources 125(2): 199-205.

Tarascon, J. M. and M. Armand (2001). "Issues and challenges facing rechargeable lithium batteries." Nature 414(6861): 359-367.

Thangadurai, V., S. Narayanan and D. Pinzaru (2014). "Garnet-type solid-state fast Li ion conductors for Li batteries: critical review." Chemical Society reviews 43(13): 4714-4727.

Trujillo, N. J., Q. Wu and K. K. Gleason (2010). "Ultralow Dielectric Constant Tetravinyltetramethylcyclotetrasiloxane Films Deposited by Initiated Chemical Vapor Deposition (iCVD)." Advanced Functional Materials 20(4): 607-616.

Tuller, H. (2007). Ionic Conduction and Applications. Springer Handbook of Electronic and Photonic Materials. S. Kasap and P. Capper, Springer-Verlag.

146

Verma, P., P. Maire and P. Novák (2010). "A review of the features and analyses of the solid electrolyte interphase in Li-ion batteries." Electrochimica Acta 55(22): 6332-6341.

Wagner, C. D., L. E. Davis, M. V. Zeller, J. A. Taylor, R. H. Raymond and L. H. Gale (1981). "Empirical atomic sensitivity factors for quantitative analysis by electron spectroscopy for chemical analysis." Surface and Interface Analysis 3(5): 211-225.

Wang, H., D. Im, D. J. Lee, M. Matsui, Y. Takeda, O. Yamamoto and N. Imanishi (2013). "A Composite Polymer Electrolyte Protect Layer between Lithium and Water Stable Ceramics for Aqueous Lithium-Air Batteries." Journal of The Electrochemical Society 160(4): A728-A733.

Wang, Y. and W.-H. Zhong (2015). "Development of Electrolytes towards Achieving Safe and High-Performance Energy-Storage Devices: A Review." ChemElectroChem 2(1): 22-36.

Whittingham, M. S. (1976). "Electrical Energy-Storage and Intercalation Chemistry." Science 192(4244): 1126-1127.

Williams, D. B. and C. B. Carter (2009). Transmission Electron Microscopy: A Textbook for Materials Science. New York, Springer.

Wright, P. V. (2002). "Developments in Polymer Electrolytes for Lithium Batteries." Mrs Bulletin 27(08): 597-602.

Wu, H.-Y., D. Saikia, H.-Y. Chao, Y.-C. Pan, J. Fang, L.-D. Tsai, G. T. K. Fey and H.-M. Kao (2013). "Synthesis and characterization of highly conductive plasticized double core organic–inorganic hybrid electrolytes for lithium polymer batteries." Journal of Power Sources 238: 265-273.

Xu, J., S. Dou, H. Liu and L. Dai (2013). "Cathode materials for next generation lithium ion batteries." Nano Energy 2(4): 439-442.

Yao, Y., N. Liu, M. T. McDowell, M. Pasta and Y. Cui (2012). "Improving the cycling stability of silicon nanowire anodes with conducting polymer coatings." Energy & Environmental Science 5: 7927.

Yoo, J., S. A. Dayeh, W. Tang and S. T. Picraux (2013). "Epitaxial growth of radial Si p-i-n junctions for photovoltaic applications." Applied Physics Letters 102(9): 093113.

Young, W.-S. and T. H. Epps (2012). "Ionic Conductivities of Block Copolymer Electrolytes with Various Conducting Pathways: Sample Preparation and Processing Considerations." Macromolecules 45(11): 4689-4697.

Zhang, C. H., J. Bürki and C. A. Stafford (2005). "Stability of metal nanowires at ultrahigh current densities." Physical Review B 71(23): 235404.

147

Zhang, L. Q., X. H. Liu, Y. C. Perng, J. Cho, J. P. Chang, S. X. Mao, Z. Z. Ye and J. Y. Huang (2012). "Direct observation of Sn crystal growth during the lithiation and delithiation processes of SnO2 nanowires." Micron 43: 1127-1133.

148