Silicon Nanocrystals: Hong Yu, Jie-Qiong Zeng and Zheng-Rong Qiu
Silicon Nanocrystals: Hong Yu, Jie-Qiong Zeng and Zheng-Rong Qiu
Silicon Nanocrystals: Hong Yu, Jie-Qiong Zeng and Zheng-Rong Qiu
Silicon Nanocrystals
Hong Yu, Jie-qiong Zeng and Zheng-rong Qiu
Southeast University
China
1. Introduction
Silicon has many advantages over other semiconductor materials: low cost, nontoxicity,
practically unlimited availability, and decades of experience in purification, growth and
device fabrication. It is used for almost all modern electronic devices. However, the indirect
energy gap in bulk crystalline Si makes it unable to emit light efficiently and thus unsuitable
for optoelectronic applications. For example, lasers, photodetectors are not constructed
from silicon. On the other hand, although silicon is widely used for solar cell fabrication, the
efficiency can not exceed the Schockley and Queisser limit in single band gap device,
because of its inability to absorb photons with energy less than the band gap and
thermalisation of photon energy exceeding the band gap. One approach for tackling this
disadvantage is to use tandem cells, which can implement the increasing of the number of
band gaps (Conibeer et al., 2006; Cho et al., 2008). Moreover, the band gap in silicon is too
small to interact effectively with the visible spectrum. If the gap could be adjusted, silicon
would be used for either electronic or optical application. In 1990, it was firstly observed
experimentally by Canham (Canham, 1990), that photoluminescence (PL) occurs in the
visible range at room temperature in porous silicon (PS). Since then the silicon clusters or
silicon quantum dots (Si-QDs) or silicon nanocrystals (Si-NCs) have attracted much of
research interest, and many of theoretical models, computations, and experimental results
on band structures, PL and other electronic properties have been reported during the last
decades (Öğüt et al., 1997; Fang & Ruden, 1997; Wolkin et al., 1999; Wilcoxon et al., 1999;
Soni et al., 1999; Vasiliev et al., 2001; Garoufalis & Zdetsis, 2001; Carrier et al., 2002; Nishida,
2004; Biteen et al., 2004; Tanner et al., 2006). The results from these reports show that in low-
dimension silicon structures, such as silicon nanocrystals or silicon quantum dots, electronic
and optical properties can be quite different from those of silicon bulk counterpart, for
instance, free-standing Si-NCs show strong luminescence, the color of which depends on the
size of the Si-NCs, and the gap and energy increase when their size is reduced. Therefore,
the energy gap can be tuned as a function of the size of quantum dots.
We are especially interested in the theoretical study on the band gap and the optical
spectrum with respect to the size of the Si-NCs or Si-QDs and surface terminations and
reconstructions.
The effective mass approximation (EMA) is used by Chu-Wei Jiang and M. Green to
calculate the conduction band structure of a three-dimensional silicon quantum dot
superlattice with the dots embedded in a matrix of silicon dioxide, silicon nitride, or silicon
carbide( Jiang & Green, 2006), and later the EMA is only of partial use in determining the
absolute confined energy levels for small Si-NCs, because it has been found a decreasingly
www.intechopen.com
122 Crystalline Silicon – Properties and Uses
accurate prediction of the confined energy level by the EMA as the Si-QDs size decreases(
Conibeer et al., 2008).
Time dependent density functional theory (TDDFT) has been performed by Aristides D.
Zdetsis and C. S. Garoufalis over the last ten years (Garoufalis & Zdetsis, 2001; Zdetsis &
Garoufalis, 2005). In their calculations the Si dangling bonds on the surface of the Si-NCs are
passivated by hydrogen and oxygen. In the DFT method, they have used the hybrid
nonlocal exchange correlation functional of Becke and Lee, Yang and Parr, which includes
partially exact Hartree–Fock exchange (B3LYP). Their results are in excellent agreement
with accurate recent and earlier experimental data. It is found that the diameter of the
smallest oxygen-free nanocrystal that could emit PL in the visible region of the spectrum is
around 22 Å, whereas the largest diameter falls in the range of 84– 85 Å. The high level and
the resulting high accuracy of their calculations have led to the resolution of existing
experimental and theoretical discrepancies. Their results also clarify unambiguously and
confirm earlier predictions about the role of oxygen on the gap size. More recently, they
report accurate high level calculations of the optical gap and absorption spectrum of ultra
small Si-NCs of 1nm, with hydrogen and oxygen passivation (with and without surface
reconstruction) (Garoufalis & Zdetsis, 2009). They show that some of the details of the
absorption and emission properties of the 1 nm Si nanoparticles can be efficiently described
in the framework of TDDFT/B3LYP, by considering the effect of surface reconstruction and
the geometry relaxation of the excited state. Additionally, they have examined the effect of
oxygen contamination on the optical properties of 1nm nanoparticles and its possible
contribution to their experimentally observed absorption and emission properties.
By performing the same method, TDDFT, the optical absorption of small Si-NCs embedded
in silicon dioxide is studied systematically by Koponen (Koponen et al., 2009). They have
found that the oxide-embedded Si-NCs exhibit absorption spectra that differ significantly
from the spectra of the hydrogen-passivated Si-NCs. In particular, the minimum absorption
energy is found to decrease when the Si-NCs are exposed to dioxide coating. The absorption
energy of the oxide-embedded Si-NCs remains approximately constant for core sizes down
to 17 atoms, whereas the absorption energy of the hydrogen-passivated Si-NCs increases
with decreasing crystal size. They suggest a different mechanism for producing the lowest-
energy excitations in these two cases.
Wang, et al, generate and optimize geometries and electronic structures of hydrogenated
silicon nanoclusters, which include the Td and Ih symmetries by using the semi-empirical
AM1 and PM3 methods, the density functional theory DFT/ B3LYP method with the 6-
31G(d) and LANL2DZ basis sets from the Gaussian 03 package, and the local density
functional approximation (LDA), which is implemented in the SIESTA package(Wang et al.,
2008). The calculated energy gap is found to be decreasing while the diameter of silicon
nanocluster increases. By comparing different calculated results, they conclude that the
calculated energy gap by B3LYP/6-31G(d)//LDA/SIESTA is close to that from experiment.
For investigation of the optical properties of Si-NCs as a function of surface passivation,
they carry out a B3LYP/6-31G(d)//LDA/SIESTA calculation of the Si35 and Si47 core
clusters with full alkyl-, OH-, NH2-, CH2NH2-, OCH3-, SH-, C3H6SH-, and CN- passivations.
In conclusion, the alkyl passivant affects the calculated optical gaps weakly, and the
electron-withdrawing passivants generate a red-shift in the energy gap of silicon
nanoclusters. A size-dependent effect is also observed for these passivated Si nanoclusters.
The optical absorption spectra of SinHm nanoclusters up to 250 atoms are computed using a
linear response theory within the time-dependent local density approximation (TDLDA)
www.intechopen.com
Silicon Nanocrystals 123
(Vasiliev et al., 2001). The TDLDA formalism allows the electronic screening and correlation
effects, which determine exciton binding energies, to be naturally incorporated within an ab
initio framework. They find the calculated excitation energies and optical absorption gaps to
be in good agreement with experiment in the limit of both small and large clusters. The
TDLDA absorption spectra exhibit substantial blueshifts with respect to the spectra obtained
within the time-independent local density approximation.
(a) (b)
Fig. 1. Cross-sectional TEM (a) HRTEM (b) images of Si QDs in oxide. (a) Shows the layered
structure and (b) shows individual nanocrystals in which crystal planes can be seen
Scientists also have studied about the nanostructures of Si QDs in other dielectrics such as
silicon nitride and silicon carbide, see Fig. 2a and b (Conibeer et al., 2006, 2008). Results from
HRTEM images are very promising, showing crystalline nanocrystals in the nitride matrix
and carbide matrix.
From the introduction stated above, we can see that it is reasonable to suppose that the
interior of the dot has the structure of crystalline silicon while the surface of the dot is
passivated with specific atoms at ideal conditions.
www.intechopen.com
124 Crystalline Silicon – Properties and Uses
(a) (b)
Fig. 2. HRTEM images of Si-QDs in (a) silicon nitride and (b) silicon carbide
www.intechopen.com
Silicon Nanocrystals 125
Fig. 4. Ball and stick representations of four possible oxygen passivation configurations, the
yellow balls represent Si atoms, red balls represent O atoms, and the white balls, H atoms:
(a) initial hydrogen-passivated silicon cluster, (b) double-bonded oxygen passivation
configuration (c) backbonded (d) bridge-bonded (e) inserted oxygen configuration
Actually there may exist other oxygen-contamination on the surface of Si-QDs when
oxidized. Zdetsis and Garoufalis (Zdetsis & Garoufalis, 2005) proposed a structure that can
maintain the Td symmetry of the nanocrystals, considering only double-bonded oxygen
configuration, see Fig. 5.
www.intechopen.com
126 Crystalline Silicon – Properties and Uses
Fig. 5. Structure of oxidized nanocrystals with Td symmetry: The blue spheres are Si atoms,
the red spheres are O atoms, and the white spheres are H atoms
Fig. 6. Structures of Si-QDs embedded in an ionic environment, SiO2, Si3N4, SiC matrix and a
co-valent environment (left to right; top to bottom): SiH4, Si5(CH3)12, Si10(NH2)16, Si14(OH)20,
Si18F24, Si26H32, Si35(CH3)36; Si53(NH2)48, Si84(OH)64, Si165F100
König et al (König et al., 2009) have used F, OH, NH2, CH3 and H groups as the passivating
agent for the Si clusters surface to simulate Si-QDs embedded in an ionic environment, SiO2,
Si3N4, SiC matrix and a co-valent environment, respectively, see Fig. 6. Apparently, these
structures are simply built by replacing all the hydrogen atoms on the surface of hydrogen-
passivated Si clusters with F, OH, NH2 and CH3 groups, respectively.
However, there are many more complex structures when Si-QDs embedded in a SiO2
matrix. For instance, Koponen et al (Koponen et al., 2009) put forward some structures in
which Si clusters are embedded in one or two neutral SiO2 shells and the outermost layer of
the cluster is hydrogen passivated in order to get rid of dangling bonds and to better mimic
www.intechopen.com
Silicon Nanocrystals 127
the effect of bulk silicon oxide. One of these structures is presented in Fig. 7(a). In addition,
Guerra et al (Guerra et al., 2009) obtained a crystalline embedded structure from a β-
cristobalite (BC) matrix by removing all the oxygen atoms included in a cut-off sphere,
whose radius determines the size of the Si-QDs. The final optimized structure of the Si32 in
β-cristobalite matrix is illustrated in Fig. 7(b).
Fig. 7. (a) A stick and ball representation of Si17@Si24O36H60 (Si clusters embedded in one
neutral SiO2 shell), (b) The final optimized structure of the Si32 in β-cristobalite matrix
From the introduction stated above, we can clearly see that different structures should be
used for simulating Si-QDs in different environment. Besides the structures presented
above, there are many other structures used to simulate Si-QDs, we’ll not elaborate here. It
is crucial to choose an appropriate structure for a certain problem, otherwise the results may
be inaccuracy and make no sense.
3. Method of calculation
Local-density approximations (LDA) are a class of approximations to the exchange
correlation energy functional in DFT that depends solely upon the value of the electronic
density at each point in space (and not, for example, derivatives of the density or the Kohn-
Sham orbits). It was used to study on the band gap and PL of Si-QDs in early researches
because of its simplicity and low computation cost. However, it is well known that time-
independent LDA calculations typically underestimate the experimental photo-absorption
gaps of Si-QDs (Vasiliev et al., 2001). TDLDA is then developed by Vasiliev as a natural
extension of the ground state density-functional formalism and LDA, designed to include
the proper representation of excited states.
The computation software we used is CASTEP module of Material Studio. CASTEP is an ab
initio quantum mechanical program employing DFT to simulate the properties of solids,
interfaces, and surfaces for a wide range of materials classes such as ceramics,
semiconductors, and metals. First principle calculations allow researchers to investigate the
nature and origin of the electronic, optical, and structural properties of a system without the
need for any experimental input. It allows to choosing local, gradient-corrected, and non-
local functionals for approximating exchange and correlation effect, and non-local functional
include screened exchange, HF, B3LYP and PBE0. The screened exchange LDA (sX-LDA) is
used to calculate the band structures of silicon quantum dots by considering the
www.intechopen.com
128 Crystalline Silicon – Properties and Uses
computational cost and accuracy. The structures with hydrogen and oxygen passivations
are partially shown in Fig. 3 and Fig.4, respectively.
www.intechopen.com
Silicon Nanocrystals 129
exposed to air and their PL energies were measured again. We can clearly see from Fig. 9
that after exposure to air, a redshift of the PL is observed, which can be as large as 1 eV for
blue luminescent samples that contain QDs smaller than 2 nm. Besides, authors pointed out
that a PL redshift was also observed as soon as the oxygen-free PS samples were transferred
from Ar to a pure oxygen atmosphere while no redshift at all was detected when the
samples were kept in pure hydrogen atmosphere or in vacuum. Moreover, they found that
there is an upper limit of the emission energy (2.1 eV) which is independent of size. So the
conclusion is that the surface passivation plays an important role, especially the chemistry of
oxygen at the surface can cause an evident change on the PL of Si-QDs and an upper limit of
energy gap.
Fig. 9. Room temperature PL spectra from PS samples with different porosities kept under
Ar atmosphere (a) and after exposure to air (b)
On the whole, despite some inevitable differences, the experimental results show that for
ultra pure Si-QDs as their sizes decrease, there is a considerable blue-shift in the peak PL
energy which is due to the increase of band gap. However, if the Si-QD is not pure enough,
much more complex results will be obtained due to the influence of impurity contamination
and complicate interface situation.
www.intechopen.com
130 Crystalline Silicon – Properties and Uses
Fig. 10. Energy gaps of silicon QDs as a function of the QD diameter with H-termination
from our work (■) & (●), and from the literatures [(▲) & (◄) from Vasiliev et al; (▼) from
Conibeer et al]
www.intechopen.com
Silicon Nanocrystals 131
Fig. 11. Calculated energy gaps as a function of the diameter (D) of the Si -QDs studied.
The dashed line shows the oxidation-induced peak energy (at ∼2.1 eV) in PL spectra
observed in PS
Obviously, the two conclusions stated above contradict each other. For this reason, we have
performed calculations using the CASTEP computation program for all four kinds of
oxygen configurations. The results we obtained based on LDA method are shown in Fig. 12,
from which we can safely come to a conclusion that it is the formation of a Si=O double
bond which is responsible for the PL redshifts in PS and other three types of oxygen
configurations only cause a few decrease in energy gap.
The discrepancy between the above results may be due to the difference of the oxygen
geometries, initial silicon clusters and the models based on different theories. So it is clear
that different structures or models may obtain quite different results.
www.intechopen.com
132 Crystalline Silicon – Properties and Uses
Fig. 12. Calculated energy gaps as a function of the diameter (D) of the Si-QDs
Fig. 13. Band gaps as a function of the diameter of Si-QDs with different interface
terminations. For clarity, the graph is split into two sub-graphs with different energy scale.
The points for the Si10X16 clusters (dQD = 7.3 Å) are shown in both graphs
There are many more complicate calculations simulating Si-QDs in specific environments,
we’ll not list them one by one here. From the results above, we know that many factors will
influence the properties of Si-QDs, other than the size, the interface termination also have a
significant impact, so we should make an comprehensive considerations to study the
properties of Si-QDs.
5. Conclusion
The band gap of Si-NCs or Si-QDs is size-dependent, and the energy of PL can be tuned
from the near infrared to the ultraviolet by changing the size of nanocrystals. The surface
passivation also plays an important role in determining the band gap. The third generation
photovoltaic (PV) solar cell using tandem cells, which is based on Si quantum dots
nanostructures, is proposed by Martin Green group (Conibeer et al., 2006, 2008). A number
www.intechopen.com
Silicon Nanocrystals 133
of theoretical models, computations results on PL have been reported, and much of research
effort on the properties of Si-QDs has been performing currently. Based on CASTEP
quantum mechanical program with choice of sX-LDA/CA-PZ, the band gap of Si-QDs with
H-passivation and O-passivation is calculated. The results show that band gap increases
when the size of quantum dots decreases for both H-passivation and O-passivation
structures. The computation method sX-LDA/CA-PZ is comparable with TDLDA in the
computational accuracy. Further research on the band gap and PL of Si-QDs by choosing
other non-local functional as the exchange correlation is in progress.
6. Acknowlegment
This work is partially supported by National High-Tech Research and Development
Program of China under Grant No 2007AA04Z301.
7. Reference
Biteen, J. S.; Lewis, N. S., and et al. Size-dependen oxygen-related electronic states in silicon
nanocrystals. Applied Physics Letters, Vol. 84 (2004), pp. 5389-5391
Canham, L. T.. Silicon quantum wire array fabrication by electrochemical and chemical
dissolution of wafers. Applied Physics Letters , Vol.57 (1990), pp.1046-1048
Carrier, P. ; Lewis, L. J. and et al. Optical properties of structurally relaxed Si/SiO2
superlattices : the role of bonding at interfaces. Physical Review B, Vol. 65, (2002),
pp.165339-1—165339-11
Cho, E. C., Park, S. W., and et al. Silicon quantum dot/crystalline silicon solar cells.
Nanotechnology, Vol.19, (2008), pp.1-5
Conibeer, G. ; Green, M. and et al. Silicon nanostructures for third generation photovoltaic
solar cells. Thin Solid Films, Vol. 511-512 (2006), pp.654-662
Conibeer, G. ; Green, M. and et al. Silicon quantum dot nanostructures for tandem
photovoltaic cells. Thin Solid Films, Vol. 516(2008), pp.6748-6756
Fang, T. N. & Ruden, P. P.. Electronic structure model for n- and p- type silicon quantum
dots. Superlattices and Microstructures, Vol. 22 (1997), pp.590-596
Garoufalis, C. S. & Zdetsis, A. D.. High level Ab Initio calculations of the optical gap of small
silicon quantum dots. Physical Review Letters, Vol. 87 (2001), pp.276402-1—276402-4
Garoufalis, C. S. & Zdetsis, A. D.. Optical properties of ultra small nanoparticals:potential
role of surface reconstruction and oxygen contamination. Journal of Math. Chem.,
Vol. 46 (2009), pp.952-961
Guerra, R.; Degoli, E. and et al. Size, oxidation, and strain in small Si/SiO2 nanocrystals.
Physical Review B, Vol. 80 (2009), pp.155332-1—155332-5
Jiang, C. W. & Green, M. A.. Silicon quantum dot superlattices: modeling of energy bands,
densities of states, and mobilities for silicon tandem solar cell applications. Journal
of Applied Physics, Vol. 99(2006), pp.114902-1—114902-7
König, D.; Rudd, J. and et al. Impact of interface on the effective band gap of Si quantum
dots. Solar Energy Materials & Solar Cells, Vol. 93 (2009), pp.753- 758
Koponen, L.; Tunturivuori, L. and et al. Effect of the surrounding oxide on the
photoabsorption spectra of Si nanocrystals. Physical Review B, Vol. 79 (2009),
pp.235332-1—235332-6
www.intechopen.com
134 Crystalline Silicon – Properties and Uses
Luppi, M. & Ossicini, S.. Ab initio study on oxidized silicon clusters and silicon nanocrystals
embedded in SiO2: Beyond the quantum confinement effect. Physical Review B, Vol.
71 (2005), pp.035340-1—035340-15
Nishida, M.. Electronic state calculations of Si quantum dots: Oxidation effects. Physics
Review B., Vol. 69 (2004), pp.165324-1—165324-5
Nishida, M.. Calculations of the electronic structure of silicon quantum dots: oxidation-
induced redshifts in the energy gap. Semicond. Sci. Technol., Vol. 21 (2006), pp.443–
449
Öğüt, S. & Chelikowsky, J. R.. Quantum confinement and optical gaps in Si nanocrystals.
Physical Review Letters, Vol. 79 (1997), pp.1770-1773
Soni, R. K.; Fonseca, L. F. and et al. Size-dependent optical properties of silicon nanocrystals.
Journal of Luminescence, vol. 83-84 (1999), pp.187-191
Tanner, M. G.; Hasko, D. G. and et al. Investigation of silicon isolated double quantum-dot
energy levels for quantum computation. Microelectronic Engineering, Vol. 83
(2006), pp.1818-1822
Vasiliev, I.; Chelikowsky, J. and et al. Ab initio absorption spectra and optical gaps in
nanocrystalline silicon. Physical Review Letters, Vol. 86 (2001), pp.1813-1816
Wang, B. C.; Chou, Y. M. and et al. Structural and optical properties of passivated silicon
nanoclusters with different shapes: a theoretical investigation. Journal of Phys. Chem.
A, Vol. 112 (2008), pp.6351-6357
Wilcoxon, J.; Samara, G. and et al. Optical and electronic properties of Si nanoclusters
synthesized in inverse micelles. Physical Review B, Vol. 60 (1999), pp.2704-2714
Wolkin, M.; Jorne, J. and et al. Electronic states and luminescence in porous silicon quantum
dots: the role of oxygen. Physical Review Letters, Vol. 82 (1999), pp.197-200
Zdetsis, A. & Garoufalis, C.. Real space ab initio calculations of exitation energies in small
silicon quantum dots, In : Quantum Dots: Fundamentals, Applications, and Frontiers, B.
A. Joyce et al. (2005), pp.317–332
www.intechopen.com
Crystalline Silicon - Properties and Uses
Edited by Prof. Sukumar Basu
ISBN 978-953-307-587-7
Hard cover, 344 pages
Publisher InTech
Published online 27, July, 2011
Published in print edition July, 2011
The exciting world of crystalline silicon is the source of the spectacular advancement of discrete electronic
devices and solar cells. The exploitation of ever changing properties of crystalline silicon with dimensional
transformation may indicate more innovative silicon based technologies in near future. For example, the
discovery of nanocrystalline silicon has largely overcome the obstacles of using silicon as optoelectronic
material. The further research and development is necessary to find out the treasures hidden within this
material. The book presents different forms of silicon material, their preparation and properties. The modern
techniques to study the surface and interface defect states, dislocations, and so on, in different crystalline
forms have been highlighted in this book. This book presents basic and applied aspects of different crystalline
forms of silicon in wide range of information from materials to devices.
How to reference
In order to correctly reference this scholarly work, feel free to copy and paste the following:
Hong Yu, Jie-Qiong Zeng and Zheng-Rong Qiu (2011). Silicon Nanocrystals, Crystalline Silicon - Properties
and Uses, Prof. Sukumar Basu (Ed.), ISBN: 978-953-307-587-7, InTech, Available from:
http://www.intechopen.com/books/crystalline-silicon-properties-and-uses/silicon-nanocrystals