Experimental Set-Up

M.Sc. Project Report – Semester II

Preparation and characterization of silicon nanoparticle by laser ablation in solution

Submitted By Dibyendu Dey Y0121

Under the guidance of Prof. Ramesh Chandra Budhani

Department of Physics Indian Institute of Technology, Kanpur

Certificate

This is to certify that Dibyendu Dey (Y0121) has worked on the project titled “Preparation and characterization of silicon nanoparticle by laser ablation in solution”. The entire work has been carried out the Condensed Matter Low Dimensional Material Laboratory.

Dr. R. C. Budhani Department of Physics IIT Kanpur

2 Contents

1 Abstract 4 2 Introduction 4 2.1 Nanoparticle and Nanotechnology ………………………………………...4 2.2 Quantum Confinement Effect in Silicon Nanoparticle…………………….4 3 Experiment 6 3.1 Previous Set-up……………………………………………………………..6 3.2 Modification in Previous Set-up…………………………………………....6 3.3 New Set-up………………………………………………………………….6 4 Experimental Apparatus 8 4.1 Nd-YAG Laser ……………………………………………………………..9 4.2 Surfactant …………………………………………………………………..9 4.3 PL apparatus ………………………………………………………………..9 4.4 Spectrometer ………………………………………………………………10 5 Results 11 5.1 Absorption Spectrum ……………………………………………………..11 5.2 PL measurement…………………………………………………………..12 5.2.1 PL measurement of nanoparticle sample in solution…………....12 5.2.2 PL of the dried sample…………………………………………..13 5.3 TEM measurement………………………………………………………...14 5.3.1 TEM of sample in SDS solution………………………………...14 5.3.2 TEM of sample in water………………………………………...15 6 Discussions 16 7 Conclusions 17 8 References 17

3 1.0 Abstract Silicon nanoparticles are prepared by laser ablation of boron-doped p-type crystalline silicon wafer in both water and SDS (sodium dodecyl chloride) solution. I also present the Photo-Luminescence, absorption and particle size property of silicon nanoparticle. The PL measurement shows a wide spectrum and the tunneling electron microscopy indicates a particle size of nano-scale order.

2.0 Introduction

Nanotechnology - a broad term, covering the use of techniques to manipulate matter at an atomic scale. The drastic change in properties in particle having nanosize is Quantum Mechanical in nature and Quantum Mechanics is a magic! Thus Nanotechnology is a real revolution because it is changing in a fundamental way how we build things. For example, if we take a hard material, clay or a ceramic, and powder it down to the nanoscale, and mix it with a polymer, we may land up getting something which has a combination of hardness and toughness never seen in the world before. Thus this technology is growing interests in leaps and bounds. Other features that contribute to nanotechnology's promise are the expectation of cheap, low polluting mass manufacturing and the possibility of making things, on the scale of biological building blocks, which could imitate living systems. So far, most applications involve enhancements of preexisting materials, but new developments are in the works. Thus nanoscale material fabrication receives intensive scientific interests as a dynamically growing field of material science.

2.1 Nanoparticle and Nanotechnology

The term “Nanoparticle” is generally used in the area of material science to represent particles of size less than 100 nanometers, lying intermediate between the molecular state and bulk state. Nanoscale material fabrication receives intensive scientific interests in the growing field of material sciences. This is due to the wide applicability of nano-materials that exhibit unique properties, usually differing from the corresponding bulk counterpart. Along with the wide range of applicability, nanoparticle can also be synthesized by a gamut of different processes like laser-assisted chemical liquid-phase decomposition [1], thermal decomposition [2], chemical reaction [3], microemulsion method [4], laser ablation [5-8] etc. Each method has its own importance in terms of control of particle size, cost efficiency and experimental complicacy. Laser ablation is one of the most cost efficient and easiest technique for the preparation nanoparticle.

2.2 Quantum Confinement effect in Silicon Nanoparticle

4 Crystalline silicon being ubiquitous in microelectronic industry is an indirect bandgap material and therefore incapable of light emission. But for nanostructure porous silicon, amorphous silicon shows photoluminescence properties. It has created significant interests in forming nanoparticle of semiconductor materials. It also promises a revolution in miniaturization of microelectronic industry. Small colloidal semiconductor particle have properties that deviate substantially from those bulk materials. For semiconductor materials when electrons get excited from the valence band and moves to the conduction band forming a hole, an electron-hole pair is formed which is termed as excitons. The electron and hole of the excitons are generally delocalized over a length (order of Bohr radius) much larger that lattice constant, for example the exciton Bohr radius for silicon is 5nm. As the size of the semiconductor cluster approaches a length comparable to this exciton diameter, its electronic properties start to change substantially and it is known as quantum size effect [9-11]. This can be seen from the blue shift (increase) of optical band gap as well as enhancement in absorption cross section. For a small cluster of nanoaprticles, having a few thousands of silicon atoms, we can model it to be a 3-d potential well where the energy gap between the discrete levels increase with decreasing diameter of the cluster. Thus the energy gap decreases as the diameter of nanoparticle gets increased and eventually it becomes continuous as in bulk materials. This explains the blue shift of band gap as the particle size decreases. Photoluminescence occurs when electron in conduction band combines with the holes in the valence band emitting a photon and it happens due to relaxation of electron with emission of photon. The formation of electron-hole pair is caused by the excitation from external source of light. Crystalline bulk silicon has an inefficient photo-luminescence because of the indirect band. For indirect band gap materials, it is required to have a phonon density with certain minimum momentum for the transition to take care of the conservation of crystal momentum. As the particle size decreases we can expect to have an enhancement in photoluminescence and it has been explained by many models eg quasi- direct transition [12], Quantum confinement effect [13] etc. As we know, the band gap increases with the decrease of size, and it may put a question over the prediction of enhancement of photo-luminescence with decreasing particle. This can be explained from the Quantum confinement effect in nanocluster of the silicon nanoparticle. In case of nanoclusters the electrons are confined into a smaller size, so the energy levels become discrete and the minimum energy for the system becomes higher than any bigger system. This particular situation can be termed as Quantum Confinement (QC). For metals the effects of QC isn’t visible because there the electron- hole interaction (excitons) is screened off and the electrons behave as merely free particle. According to QC-effect luminescence is attributed to radiative recombination of band to band electron-hole pair, confined within such nanoparticle [14]. For nanoclusters the quantum confinement factors plays a major role in increasing the radiative recombination rate of the excitons (electron-hole pair). From the expression of the relaxation time [14] we can see that as the energy between electron-hole pair increases the relaxation time for the radiative recombination decreases. In other words it enhances the probability of radiative recombination. For lower band gap energy the “radiative recombination” become quite low and thus for crystalline bulk silicon radiative recombination is absent. Thus we can say that the nanocluster acts as materials between a direct band gap semiconductor and an indirect band gap semiconductor.

5 Theoretically, thus it be predicted that the photo-luminescence of silicon nanoclusters is much brighter than a bulk materials with the same power of excitation. But as the band gap shifts towards higher energy, the excitation frequency increases towards more UV-range.

3.0 Experiment 3.1 Previous Set-up

We have formed the experimental set-up for the laser ablation. It is quite different from our previous set-up [fig: 1], where the laser is mounted horizontally.

Solution

Target

Nd-YAG Laser

Fig 1: Schematic Diagram of the Previous Experimental Set-up

3.2 Modification in Experimental Set-up We have brought quite significant changes from the previous set-up. This makes the laser ablation process more efficient and controlled.

 As in the new set-up [fig:2], the laser is mounted vertically, so the laser light doesn’t have to pass through the glass surface of the beaker as previously and it reduces the loss due to absorption .

 The laser light hits the target and the created nanoparticles sputters out of the surface of the metal plate. If the target is kept horizontal, as in our set-up [Fig: 1], it is always possible that the particle formed from laser ablation can get deposited on the target again. So we have put a magnetic stirrer which continuously stirs the solution so that the deposition can be avoided.

 We have used a Teflon holder to hold the sample. The sample holder is formed in such a shape that the sample doesn’t get toppled while ablation.

6  We also have used a bi-convex lens of focal length 15cm to focus the laser light on the target of metal peace this is hold on a Teflon holder. In the previous set-up also we used this lens for focusing.

3.3 New Set-Up With all those modification mentioned in sec [3.2], the new set-up has been shown schematically in fig [2]. Fig 2: Schematic Diagram of the Experimental Set-up

Nd-YAG Laser

Lens assembly

Target drive

Teflon target holder

Target

Stirring capsule

7 Magnetic Stirrer

4.0 Experimental Apparatus 4.1 Nd-YAG Laser:

We have used an Nd-YAG pulse laser of maximum power of 10mJ/pulse. The lasing medium is one of the most important criterions to characterize lasers. Nd-YAG laser is a type of solid-state laser where neodymium is doped to the yttrium-aluminum-garnet (YAG) host crystal to make the lasing medium. This type of laser is important because it can produce high power. The Nd-YAG laser is Q-switched (gives brief high energy pulses instead of continuous pulse) because of saturable absorber, which only allows the laser light to move between the mirrors for a very small period of time. There are different dyes used as a saturable absorber, which becomes transparent for a considerable exposure to light and becomes opaque again.

1=1064nm;2 = 532nm; 3 =256nm

 / 1 2

Pumping KTM  3

Fig 3: Schematic diagrams for the Nd-YAG laser’s working principle

One of the most important properties of this Nd-YAG laser is its second harmonic generation (SHG). The basic principle of second harmonic generation is based on a non-linear crystal property. When a monochromatic light impinges on a surface with lack of symmetry, light of frequency twice that of the incident light can get generated.

8 Fig 4: The energy level diagram of Nd: YAG [17]

4.2 Surfactant

We have used a sodium dodecyl sulfate (SDS) as the synthetic surfactant. Surfactant is needed to control the size of the nanoparticle and for its chemical and magnetic stability. When the laser hits the metal surface an embryonic metal particle is formed and it is followed by a slow particle growth. The SDS coating involves in termination of this growth and thus with increasing SDS concentration we can expect to have a smaller particle size

Si-nanoparticle

Co SDS

Fig 5: Schematic diagram for the Polymer coating of the nanoparticle

4.3 PL apparatus

I have taken the PL measurement with an apparatus of model number Flurolog- 3.22. It works mainly on three steps.

. The excitation spectrometer filters a particular frequency out of the light source. It uses a double grating for this purpose, which give high sensitivity.

9 . The particular frequency of light excites the sample inside the sample compartment and the output is again passed through an emission spectrometer. It uses the similar principle as the excitation spectrometer to filter light.

. The signal is detected by a detector called PMT (Photo multiplier tube), which is controlled by a computer interface.

. By stepping either or both the spectrometer through a wavelength region and recording the variation of intensity a spectrum is found.

Excitation Emission P Source Spectrometer Sample Spectrometer MT

Controller PC

Fig 6: Schematic diagram for Flurolog-3.22

4.4 Spectrometer (Lambda 40 UV/Visible Spectrometer) This instrument is used to measure the absorption spectrum of the nanoparticle. A grating (1052 lines/mm) is used to split the source light (UV/Visible) into a particular wavelength and falls on the sample of nanoparticle solution to measure the absorption coefficient. Using the absorption coefficient found from a SDS solution only eliminates the background caused by the SDS solution.

Reference Detector Light Source

Detector Grating Beam-Splitter Sample

10 Fig 7: Schematic diagram of the spectrometer’s working principle

5 Results

I have taken PL measurement of the nanoparticle solution along with TEM measurement and absorption spectrum.

5.1 Absorption Spectrum I took absorption spectrum of the silicon nanoparticle in miliq water solution in FEAT laboratory. They tasted the silicon nanoparticle solution with eliminating the background using another beaker of only water. The working principle has been shown in detail in section [4.4]. The error in the wavelength is 0.3nm, which gives the x-axis error bar, and the error in absorbance (i.e. the y-axis error bar) is 0.015 in maximum.

0.85 A 0.80

0.455 200.0 300 400 500 600 700 800 900 1000.0 nm Fig 8: absorption spectrum of Si-nanoparticle solution in water It clearly shows that there is no absorption peak in the visible region, but there seems to be an absorption in the ultraviolet region. I fitted a Gaussian peak function [fig: 9]

11 and the fitting correlation is found to be of the order of 98.6%. This clearly indicates a peak to be in the shorter wavelength region. Due to the effect of Quantum Confinement the band gap for this silicon nanoparticle solution increases substantially and it can be clearly seen from the absorption in the UV region. This QC effect becomes more and sharper as the particle size decreases. But of course for very small particle size (size much lesser than exciton diameter), this QC-effect again fades away, which hasn’t till been properly understood [14].

Analysis of the absorption spectrum 1.4

1.2 Data: TMP_B Model: Gauss

] Chi^2/DoF = 0.00024 u

. R^2 = 0.98622

a 1.0 [

y0 0.47159 ±0.00065

e xc -78.17596 ±44.06816

c w 264.74013 ±18.01273

n A 2166.90474 ±1024.78823

a 0.8 b r o s

b 0.6 A

200 400 600 800 1000 Wavelength [nm]

Fig 9: Plot fitted with a Gaussian distribution 5.2 PL Measurement

I first attempted to take PL measurement on the nanoparticles, both in water and SDS solution. The solution was divided into five parts where each part is oxidized for few seconds using an Ozoniser and then the PL measurement was taken.

5.2.1 PL measurement of Si nanoparticle solution in water The nanoparticle solution in water was first scanned over a continuous wavelength region to see the absorption spectrum and it was found that the excitation wavelength is 440nm. So the sample was excited with 440nm wavelength and a PL signal was detected. Initially the PL signal from the solution seemed to be from the Si- nanoparticles. But when a PL measurement was taken on sample of miliq water without any nanoparticle, the same signal was found.

12 PL measurement of Si nanoparticle in milliq water 10000 PL measurement of Si nanoparticle in milliq water with small slit exposure Excitation wavelength:440nm 700 8000

600 Peak=511.6nm 6000 Excitation wavelength=440nm

500 t t u

p u 4000 t p u t u 400 O O 2000 300

450 500 550 600 650 700 750 450 500 550 600 650 700 750 wavelength(nm) wavelength(nm)

Fig 10: PL measurement of nanoparticle miliq water solution Water has a photoluminescence with same excitation wavelength of 440nm and as the concentration of the nanoparticle solution isn’t very high, the signal form Silicon couldn’t supersede the water signal to get a PL of the nanoparticle solution. Thus it naturally evoked the necessity to bring some process of drying the sample out of water. We also thought of using some other solution which may be implemented in the future work of this project.

5.2.1 PL of dried Si nanoparticles sample

The low concentration of the sample hindered from getting a PL signal from silicon nanoparticles as the signal had to supersede the PL signal from the water. So I had to dry out the silicon nanoparticles out of the miliq water. I put the solution in a Centrifuge to separate out bigger particle from the solution. The Centrifuge was operated at speed of 12000rpm for duration of 35minute. It helps separating major part of the nanoparticles from the solution. The extra water was poured off and the container is then ultrasonically dispersed to mix the nanoparticles in rest of the water. Thus the solution becomes quite dense nanoparticle solution. This small amount of solution is now kept over an aluminum foil and dried using a hot-blower. We have used aluminum foil because it doesn’t give any photoluminescence. The dried sample was first excited with a continuous wavelength region and it showed photoluminescence with an excitation in the UV region (310nm).

13 PL of dried-Si nanoparticle sample PL of dried-Si nanoparticle Sample

5500 Slit width:0.1mm 200000 Integration time:0.1ms 5000 Excitation Wavelength:310nm 180000

4500 160000

4000 t t 140000 u u p p t t

u 3500 u 120000 O O 3000 100000 Slitwidth:0.3nm 2500 80000 Excitation Wavelength:310nm 2000 60000 1500 350 400 450 500 550 350 400 450 500 550 Wavelength [nm] Wavelength[nm]

Fig 11: PL of the dried sample of silicon nanoparticle

I have also taken a PL of a simple aluminum foil for the shake of completeness of the experiment. It doesn’t show any photoluminescence expectedly.

300000 :Al Foil Slit Width:0.3nm :Sample 250000 Excitation Wavelength:310nm

200000 t u p

t 150000

100000

0 350 400 450 500 550 Wavelength [nm]

Fig 12: PL of the dried sample with an aluminum background 5.3 TEM measurement I have taken Transmission Electron Microscopy (TEM) of the nanoparticle sample at ACES. TEM works in the principle that it shines a beam of electron through the Cu-grid. The beam is collimated to a monochromatic wave, using lenses. When it hits the grid, some parts get transmitted and some parts get hindered from the particles. Whatever part is transmitted is projected on a Phosphor screen and thus giving an enlarged image of the object. 5.3.1 TEM image of Si-nanoparticle in SDS solution

14 I have already mentioned that SDS (sodium dodeycl sulfate) has been used to form the polymer coating over nanoparticle [Sec: 4.2]. It helps in separating one nanoparticle from other. But in the image that isn’t clearly visible. This is due to fault in making the film over copper grid. Several layers have been formed while our intension was to make single layer of nanoparticle film. The image was taken in a resolution of 50K; the particle size can’t be detected due to fault in making the film.

Fig 13: TEM image of Si- nanoparticle in SDS 5.3.2 TEM image of Si-nanoparticle in water As I encountered problem in making film using SDS solution, I used nanoparticle in water for making film in second attempt. This time the film quality has improved substantially as can be seen from [Fig:14]. The particle size has been detected to be between 40-50nm. It was observed that spherical nanoparticles have been struck on the microgrid.

Fig 14: TEM image of Si-nanoparticle in water

15 Fig 15: TEM image of Si-nanoparticle in water

6 Discussions Silicon is an indirect band gap semiconductor. Bulk crystalline silicon is an inherently poor light-emitter. This low-quantum-efficiency of photoluminescence in bulk crystalline silicon is attributed to the unavailability of phonons needed for an indirect band gap semiconductor for a transition. However recently, efficient photoluminescence from bulk crystalline silicon has been reported [15, 16]. Nanocrystalline silicon can be a very good alternative to get photoluminescence in visible region for such indirect band gap silicon also. For nanoaparticle of size in the nanometer range, we can consider a 3-d potential well where the particle has been confined, termed as quantum confinement effect as mentioned in [Sec: 2.2]. Thus the band gap increases with the size reduction. Thus the energy of light required for excitation becomes higher and thus get the absorption peak in UV region as been seen in [Sec: 5.1]. Photoluminescence occurs when electron in the conduction band relaxes to the valence band emitting a photon. But for indirect band gap semiconductor the radiative recombination is negligible for the indirect band gap nature. As the particle size decreases the band structure changes substantially. Thus we can expect to see some different nature of photoluminescence in this nano-structure. Due to the enhancement of the band gap the photoluminescence shifts to the visible region and we see PL in room visibly at room temperature. These phenomena can be explained by quantum confinement effect. The PL is caused by radiative recombination whose probability is indirectly proportional to the relaxation time of the electrons in the conduction band. The relaxation time for radiative recombination rate decreases substantially as the band gap increases. It has been mathematically shown in ref [14]. Thus we can see the photoluminescence in the visible

16 region of 400-500nm in Sec [5.2.1]. The small volume of the particle also helps in confining the electron to a small volume thus the conduction electrons can’t diffuse much. It hinders non-radiative recombination. Thus the quantum confinement model quite efficiently explains the recombination mechanism which causes photoluminescence. The PL is also reasoned by pseudo-direct band transition in some papers [14]. The transmission electron micrograph shows a non-uniform particle distribution and the spherical particle size is much higher than expected. It is due to the inability to control the particle size during laser ablation. The higher particle size hinders to see the larger PL and absorption peak. As the exciton diameter of silicon is of the order of 5nm so the quantum confinement effect becomes prominent for the particle size distribution of that order. For higher particle size the QC isn’t that prominent. We must use some gas flow to form a coating and by controlling the flow, we may be able to control the particle size. This has already been implemented by Li et al. [18]. In fact the surface of the nanoparticle is very active and thus many of the silicon nanoparticles become oxidized after ablation. It decreases the PL peak further.

7 Conclusions In this project, I have successfully synthesized nanoparticle by laser ablation in solution. In first semester, I synthesized cobalt nanoparticle (magnetic) and this semester, I have synthesized silicon nanoparticle (semiconductor). I have also been involved in putting the set-up for forming nanoparticle by laser ablation. I characterized the silicon nanoparticle and gave possible explanations for the characters of the sample. This project has exposed me to a wide horizon of research

8 References

[1] PhD thesis by Krisztian Kordas at department of Electrical Engineering at the University of Oulu. [2] J Connolly, T G St Pierre, M Rutnakornpituk and J S Riffle, J.Phys.D: Appl. Phys. 37 2475-2482 (2004) [3] S Misra and H C Padhi, J.Appl.Phys. 75, 4576-4580 (1994) [4] K Nagai, K Okamoto, H Kondoh, T Yokoyama, T Ohta, arxiv: 9A/2000G263 [5] M S Sibbald, G Chumanov and T M Cotton, J.Phys.Chem 100, 4672-4678 (1996) [6] Z Pasti, Z E Horvath, G Peto, A Karacs and L Guczi, Applied Surface Science 109/110, 67-73(1997) [7] J Zheng, C Zhang and R M Dickson Phys.Rev.Lett 93,077402 (2004) [8] H Yin, T Yamamoto, Y Wada, S Yanagida Materials Chemistry and Physics 83 66-70 (2004)

17 [9] Vladimir P. Drachev, E. N. Khaliullin, W.Kim, F Alzoubi, S. G. Rautian, V. P. Safonov, R. L. Armstrong, and V. M. Shalaev, Phys.Rev. B. 69,035318 (2004) [10] G. Ledoux , O. Guillois, D. Porterat and C. Reynaud, Phys.Rev.B.62,15942 (2000) [11] H. Takagi, H. Ogawa, Y. Yamazaki, A. Ishizaki, T. Nakagiri, Appl.Phys.Lett.56, 2379 (1990) [12] M. S. Hybesten , Phys.Rev.lett.72, 1514 (1994) [13] G. Ledoux, J. Gong, F. Huisken and C. Reynaud, Appl.Phys.Lett.80,4834 (2002) [14] C. Delerue, G. Allan, and M. lannoo, Phys.Rev.B.48, 11024 (1993) [15] T. Trupke et al. App. Phys. Letts. 82, 2996 (2003) [16] Martin A. Green et al. Nature. 412, 805 (2001) [17] http://www.lasalle.edu/academ/chem/laser_web/ndyag_laser.html [18] Li et al. Phase Transition 77, 1-2, PP. 131-137, Jan-Feb (2004)

Acknowledgement I would like to may my sincere thanks to Prof. R. C. Budhani for offering me such an exciting Project. I would also like to pay my special gratitude to Mr.Rajeev.K.rakshit, Mr. K. Senapati and to all the members of our lab. I would also like to thanks Prasant who has helped me to take Pl measurement at Dr. A. Pradhan’s laboratory and also to FEAT laboratory for their help.