- Email: [email protected]
Structural characterization of silicon thin film superlattice grown at low temperature
Accepted Manuscript Structural characterization of silicon thin film superlattice grown at low temperature
Debjit Kar, Debajyoti Das PII:
S0749-6036(17)30773-5
DOI:
10.1016/j.spmi.2017.06.053
Reference:
YSPMI 5102
To appear in:
Superlattices and Microstructures
Received Date:
29 March 2017
Revised Date:
19 June 2017
Accepted Date:
20 June 2017
Please cite this article as: Debjit Kar, Debajyoti Das, Structural characterization of silicon thin film superlattice grown at low temperature, Superlattices and Microstructures (2017), doi: 10.1016/j. spmi.2017.06.053
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Structural characterization of silicon thin film superlattice grown at low temperature Debjit Kar and Debajyoti Das* Nano-Science Group, Energy Research Unit, Indian Association for the Cultivation of Science, Jadavpur, Kolkata – 700 032, INDIA
Abstract: The growth of nc-Si:H from the SiH4 plasma inherently proceeds through a defective and amorphous incubation layer, originating due to the lattice mismatch between the substrate and Si material. In the current investigation, a-Si:H/nc-Si:H thin film super-lattice structures have been fabricated at very low temperature (~180 oC), using single step conventional plasma processing. In this intriguing technique the a-Si:H barrier sub-layer itself operates as an effective incubation layer which facilitates growing the nc-Si:H active sub-layer possessing adequate crystallinity inside a very low thickness (tnc) of even ≥3 nm, and the sudden discontinuity in growth of the active layer confines the size of Si-ncs within the dimension of this active sub-layer thickness. The superlattice structure of the ultimate network has been evidenced from small angle XRD (SAX) measurements, whereas the nature of crystallinity has been studied by ellipsometry, Raman spectroscopy and electron microscopy. This simple yet compelling technique seems to have enormous technological implications taking the advantage of the quantum confinement effect in low dimensional silicon nano-crystallites within very thin superlattice structures. Keywords:
incubation layer, thin film superlattice, low temperature growth, small angle XRD (SAX)
*Author to whom all correspondence should be addressed, E-mail (D. Das): [email protected]; Fax: +91(33)24732805 1
ACCEPTED MANUSCRIPT
1. Introduction: It is highly expected that in a near future, nanocrystalline silicon (nc-Si) thin film based solar cells having high conversion efficiency and low fabrication cost will substitute bulk crystalline silicon based devices for most applications [1]. With that aim numerous attempts have been made by the researchers towards evolution of new technologies and materials. In that direction tandem structure solar cell is the one that can boost the efficiency progressively with the utilization of the extended solar spectrum by stacking more cells on top of the other, combining multiple absorber layers of different optical band gaps [2-3]. The individual band gap of each absorber layer needs to maintain a descending order from the top cell to the bottom one, in order to absorb the photons in the most efficient way [4-5]. In order to achieve compatible transport of charge carriers within different layers in the cell structure, the present day silicon research has got enormous impetus in the development and utilization of various low-dimensional structures of silicon e.g., nanocrystals, quantum dots and superlattice structures in thin films [4,6-9]. By virtue of quantum confinement effects arising within its low dimensional structures with magnitude in the order of corresponding Bohr radius, silicon nanocrystals or quantum dots exhibit numerous novel opto-electronic properties; it can widen the optical band gap without deteriorating the electrical conductivity [10-11]. In general, the silicon nano-crystals or quantum dots are synthesized in a two-step process: in the first step silicon dielectric thin films are deposited followed by annealing at elevated temperature (~1100 oC) for solid-phase crystallization of Si. Conversely, in nano-scaled superlattice structures, silicon layer and Si-dielectric layer (Si3N4/SiO2/SiC) are alternately deposited in stack and subsequently treated with a post deposition high temperature (~1100 oC) annealing [12-15]. The unavoidable high temperature annealing step hinders its successful utilization in the fabrication of thin film solar cells by severely affecting the other pre-deposited layers of the device, e.g., initiating diffusion of dopants across the doped layers, and limits the use of low cost substrates, as well. Moreover, the high temperature annealing step belongs to an uncertain process in regulating the size, distribution and number density of the Si nano-crystals. Furthermore, in order to eliminate the dielectric barrier on both sides of the charge carriers in Si nanocrystals, the dielectric layers of the superlattice structure need to be thin enough and the Sincs in the active layer are required in high density so that the wave functions involving the 2
ACCEPTED MANUSCRIPT
adjacent Si-ncs mutually overlap. Consequently, numerous stringent issues are involved in the commonly employed superlattice growth technology to utilize in solar cell fabrication. In this context, one-step fabrication of superlattice films with the alternate a-Si:H/nc-Si:H sub-layers possessing plenty of tiny Si-ncs grown at device compatible temperature, could be significantly effective in mobilizing the transport of charge carriers across the relatively low dielectric barrier of the a-Si:H sub-layer and, thereby, facilitating device fabrication [16]. In present manuscript a comprehensive structural characterization of such a-Si:H/nc-Si:H thin film superlattice structures developed at ~180 °C by conventional rf-PECVD has been pursued, in view of its future utilization in different layers of the all silicon solar cells with third generation technology at low cost.
2. Experimental Details: Thin films of superlattice structures having alternative layers of a-Si:H and nc-Si:H were fabricated in a capacitively coupled PECVD reactor, using (SiH4 + H2)-plasma. Before proceeding towards the formation of superlattice thin films, the deposition conditions for individual layers have been optimized independently. For growing the a-Si:H layers, the SiH4 and H2 flow rates, the pressure of reacting gases in the plasma and rf-power operated to the parallel plate electrodes were kept at 95 sccm, 5 sccm, 3 Torr and 15 W, respectively; while for the nc-Si:H layers, those parameters were set at 99.5 sccm, 0.5 sccm, 4 Torr and 40 W, respectively. Prior to deposition, the plasma chamber has been evacuated to ~ 1×10-6 Torr and then heated up to 350 oC in the presence of H2-gas at relatively higher gas pressure (50 Torr). These steps were followed in order to eliminate any residual oxygen from the reactor that can contaminate the periodic layers of the superlattice structure. Finally, a constant temperature of the Corning® Eagle2000TM glass and silicon wafer substrates was maintained at 180 oC throughout the entire deposition process. A set of superlattice films was grown with identical aSi:H sub-layers of thickness 4.8 nm and by varying the thickness of nc-Si:H sub-layer from 11 to 1.5 nm, by changing only the corresponding sub-layer deposition time. A gross thickness of around 200 nm was maintained for each superlattice sample with similar a-Si:H sub-layer at its both ends, by accumulating different numbers of bi-layer cycles.
3
ACCEPTED MANUSCRIPT
Small angle X-ray diffraction was obtained by Bruker (D8 Advance) system with source radiation (λ=1.5418 °A) beam incident at glancing angle (GISAX). The Raman spectroscopy was performed by 514 nm wavelength excitation from Ar+ laser (Spectra Physics) operated at ~2 mWcm-2 power density and J-Y Horiba Triple Raman spectrometer. Olympus open stage microscope was used to focus the beam spot inside the sample and the signal was detected in a back-scattering geometry by a TE-cooled CCD camera. Transmission electron microscope data was obtained by a transmission electron microscope (JEOL JSM 2010) working at 200 kV. The samples for TEM studies were prepared on Cu microscope grids previously coated by amorphous carbon. Infra-red spectroscopic studies were carried out on samples deposited on single crystal Si substrates, by Perkin Elmer Spectrum100 FTIR spectrometer.
3. Results and Discussions: 3.1. Small-Angle XRD: Small angle X-ray diffraction (SAX) is an essential non-destructive tool to characterize the periodic structures of the superlattice thin films. Figure 1 shows the SAX pattern of different a-Si:H/nc-Si:H superlattice samples. Diffraction patterns have different 1st and 2nd order (m) peaks, which are the clear evidence of the presence of multi-layers structure in the samples. The first order diffraction peaks are observable for the superlattice films with tnc=1.5, 2 and 3 nm whereas for tnc>3 nm, only the second order diffraction peaks are noticeable due to the instrumental limitations on angle scan within 2-span of 0.5° to 4.0°. It is noteworthy that the peak positions of each first and second order diffraction peaks gradually shift towards lower diffraction angle (2θ) as the periodic thickness of the superlattice films increases by virtue of increasing the active layer thickness, tnc (thickness of barrier layer remaining constant). From the Braggs’ law of diffraction, the periodic thickness of each superlattice film has been calculated from the 1st or 2nd order diffraction peaks. The increasing thickness of nc-Si:H active sub-layer of the periodic structure of different superlattice films are also separately estimated from the sample thickness measured by surface profilometer. Figure 2 demonstrates an excellent linear relation between the periodic thickness of each superlattice films estimated from SAX data and the periodic thicknesses of corresponding nc-Si:H active layer thickness tnc obtained from 4
ACCEPTED MANUSCRIPT
Figure 1: The small angle XRD (SAX) patterns of different superlattice thin films showing 1st and 2nd order diffraction peaks due to the presence of multilayer structure.
profilometer data. Interestingly, the extrapolation of the linear plot to tnc= 0, shown by the blue section in the plot, leads to the superlattice periodic thickness ~4.8 nm which is exactly the thickness of a-Si:H barrier layer in each cycle, thus confirming the accuracy of both experimental process. It may be noted that the peak-widths of the SAX diffraction peaks are much larger (~0.5o) than the resolution (0.02o) of the diffractometer. This implies a blurred edge between the sub-layers which might arise due to the periodic presence of a very thin amorphous incubation section in the nc-Si:H active sub-layer [17-20] which is different from the a-Si:H barrier sub-layer because of their different growth parameters.
5
Periodic thickness of superlattice (nm) estimated from SAX data
ACCEPTED MANUSCRIPT
16
Thickness of a-Si sub-layer =4.8 nm
14 12 10 8 6 4
0
2
4 6 8 10 nc-Si:H sub-layer thickness (nm) estimated from profiler data
12
Figure 2: The as-estimated periodic thickness from the SAX plotted as a function of thickness of the nc-Si:H sub-layer.
3.2. Raman Spectroscopic studies: The crystalline nature of the superlattice structure has been investigated using Raman spectroscopy. The Raman spectrum of each superlattice thin film has been shown in Figure 3. On lowering in the thickness of nc-Si:H sub-layer (tnc), a consistent red-shift of peak frequency (~ 520 cm-1) occurs while the peak width broadens asymmetrically. For tnc=2 nm, the Raman peak is very broad and centered at ~485 cm-1 and mostly presents an amorphous dominated network. For tnc>2 nm, the observed sharp peak at ~ 520 cm-1 corresponds to the TO mode vibration of nc-Si implying the presence of high amount of crystallinity in the superlattice thin films [7]. A close observation on the peak position of nc-Si component, shown at the inset of Fig. 3 reveals that there is a significant peak shift of ~7 cm-1 as the thickness of the nc-Si:H layer is decreased from 10 to 3 nm which signifies the gradual miniaturization of the size of the silicon 6
ACCEPTED MANUSCRIPT
Figure 3: The Raman spectra of different superlattice thin films showing consistent modification of the peak characteristics with the change in tnc. The inset shows the close-look of the dominant peak position where it has been shifted towards lower wave number at lower tnc.
nano-crystallites. In addition, the observed broadening of the ~485 cm-1 component of the Raman signal identifies gradual enhancement of the amorphous component in the superlattice thin films on reducing tnc. The quantitative estimation of the degree of crystallinity and the average size of the Si-nc has been done by the deconvolution of each Raman spectrum into four Gaussian and one Lorentzian satellite components, the typical features of which are shown in the Figure 4. In particular, there are five satellite components at ~310 cm-1, ~420 cm-1, 480 cm-1, 500 cm-1 and 519 cm-1 – the first three components correspond to the Raman signals from LA, LO and TO modes of a-Si, the fourth one corresponds to the thermodynamically stable ultra nanocrystallites of silicon or the defects due to the grain-boundary regions and the sharp Lorentzian component is due to the TO mode of the nano-crystalline silicon (nc-Si) [21-24].
7
ACCEPTED MANUSCRIPT
Figure 4: The representative deconvolution in the Raman spectra of the superlattice films with tnc= 6 nm. The insert shows the reduced size of the Si-ncs with the lowering of the tnc.
The nano-crystalline component of the Raman spectra of the superlattice thin films has been found to shift towards lower wave number with gradual thinning of the nc-Si:H sub-layers which is a consequence of gradual decrease in size of the Si-ncs and the phonon confinement effect arising thereof. In first order Raman scattering of bulk c-Si, the excited phonon is constrained at the centre of the Brillouin zone following the momentum conservation law. However, due to the spatial confinement in the tiny Si-ncs, the phonons which are dispersed over the Brillouin zone starts contributing to the Raman signal due to the presence of structural defects at the boundaries of nano-crystallites. Considering that, first order Raman spectra has been simulated taking into account the phonon confinement effect in the reduced size of Si-ncs having Gaussian size-distribution. A sine wave has been considered as the confinement function that disappears just outside the nc-Si boundary and the phonon dispersion curve has been considered following Paillard [25]. The specific simulation technique has been described in our previous work [26]. The as estimated average size of the Si-nc has been plotted in the inset of Figure 4 and the size of Si-nc has been found to be reduced from 9.1 nm to 1.9 nm with lowering of tnc from 10 to 2 nm. Interestingly, the average grain sizes are in good agreement with the thickness (tnc) of nc-Si:H sub-layer. 8
ACCEPTED MANUSCRIPT
Figure 5: Relative change of the integrated intensities of the unc- and nc- components of the Raman spectra.
Figure 5 represents the relative variation in the ratio of integrated area between the ultrananocrystalline (unc-) to nanocrystalline (nc-) silicon satellite peaks in superlattice structures occurring due to the decrease in tnc. Presence of unc-Si is much lower in the more crystalline superlattice thin films. It is noted that at tnc<2 nm, the unc-Si components dominates over nc-Si. The presence of high volume of ultra-nanocrystallites is two-fold beneficial in material point of view as it can widen the optical band gap because of the strong quantum confinement effect and can also provide better charge carrier conduction path by the processes like variable range hopping or Poole-Frenkel mechanism [27,28]. However, the defects at the amorphous silicon part can hinder the charge conduction process. In view of that, the comparative estimation of those defects have been performed by the ratio of integrated intensities of LA and TO modes of a-Si (ILA/ITO) and bond angle deviation (Δθ) of a-Si. The bond angle deviation Δθ (in degree) in the amorphous silicon network has been calculated from the FWHM (Γ in cm-1) of the TO mode of a-Si using the following relation: Γ=15 + 6Δθ. The ILA/ITO and Δθ have been plotted in Figure 6 as function of the thickness of nc-Si:H layer. The magnitude of ILA/ITO is much lower for superlattice thin films at tnc<6 nm which signifies the lower defects and stress in the amorphous 9
ACCEPTED MANUSCRIPT
network. The magnitude of Δθ is found to be in an increasing order at lower tnc due to more amorphous component in the overall superlattice structures and still the magnitude of Δθ at tnc=2 nm is quite low implying a overall less defective amorphous network [21]. Interestingly, the sizecontrol of Si-ncs by adjusting only the thickness of nc-Si:H layer is emerging to be significantly effective without introducing much defects in the superlattice thin films.
Figure 6: The variation in the ILA/ITO and Δθ of a-Si as estimated from the Raman spectroscopic studies.
3.3. Spectroscopic Ellipsometry studies: The spectroscopic ellipsometry measurement is a sensitive yet non-destructive technique that determines the complex reflection ratio of the light reflected back from the film surface, from which the complex dielectric constant can be acquired. Complex dielectric data derived from this measurement can provide qualitative and also quantitative dielectric properties, nature of the chemical bonds present as well as the structural and the morphological properties of the bulk layer, surface layer and the interfaces of the materials. The fundamental parameters of the spectroscopic ellipsometry are Ψ and Δ, which are the amplitude and phase part of complex reflectance ratio (ρ) which can be expressed as ρ = tan(Ψ)eiΔ [29]. The imaginary part of the 10
ACCEPTED MANUSCRIPT
complex dielectric function, <ε2>, as derived from the Ψ and Δ has been plotted in the Figure 7(a) for various superlattice films having different thickness of nc-Si:H sub-layers. A typical spectrum of <ε2> comprises of two distinct energy regions. The comparatively lower energy part of the spectrum (not shown here) consists of the interference pattern due to the superposition of reflected beams from the air-thin films and thin film-substrate interfaces. This interference pattern carries the information about the thickness of the films and the nature of the substrate. However, at higher energy where the penetration depth of the photon is smaller than the film thickness, the nature of the <ε2> spectra is controlled by the compositional properties of the thin film along with the surface roughness [30].
Figure 7: (a) The spectra of the imaginary part of the pseudo-dielectric function for different superlattice thin films having different tnc. (b) The reduction of the intensity difference of the two critical points due to the gradual decrease of tnc.
The superlattice thin film having tnc=11 nm represents two discrete shoulders, around ~3.4 and ~4.2 eV which have been recognized as the critical points (CPs) in the band structure of bulk crystalline silicon, E1 and E2, respectively [31]. The presence of those two shoulders is the consequence of the high degree of crystallinity in the superlattice thin films. These shoulders can be well-observable for tnc≥3 nm, however at tnc<3 nm a Gaussian-like peak appears at ~3.51 eV with higher intensity, possibly implying formation of amorphous dominated network. Typically, the peak corresponding to a-Si is positioned at 3.2 eV. Such shifting of the peak towards higher 11
ACCEPTED MANUSCRIPT
energy might be due strong quantum confinement effect from the nano-crystals in the thin film. It is notably observed that the intensity of <ε2> decreased gradually with the reduction in tnc until 5 nm, however beyond that the intensity sharply enhanced and attained high magnitude for the superlattice films having mostly bulk a-Si:H. The intensity of <ε2> for nano-crystalline silicon thin films is largely governed by the voids and the size of the silicon nano-crystallites. However, those two parameters have a trade-off relation between them – while the higher void can reduce the intensity, the larger size silicon nano-crystallites (hence higher degree of crystallinity in the films) can enhance the intensity [32]. A qualitative estimation of the degree of crystallinity can be performed by a comparative study in the difference in the intensities of the two shoulders which is reduced systematically at lower tnc. Such reduction in intensity is closely related to the decreasing overall crystalline volume fraction and the size-reduction of silicon nano-crystallites because at the smaller diameter of the Si-ncs the oscillator strength changes between the lower and higher energy states. For the present set of superlattice thin films, the size of Si-ncs are reduced with lower crystalline volume fraction that consistently lead to the significant reduction in the intensity of <ε2>, however, for tnc<5 nm the films become more amorphous dominated, having less fraction of voids which lead to the sharp rise in the intensity [26,30].
3.4. TEM studies: Development of the Si-ncs embedded in the nano-crystalline silicon sub-layer of the superlattice structures has been pursued by TEM studies. In Figure 8(a) and 8(b), the plain-view micrographs of the superlattice structures with tnc= 8 and 5 nm demonstrate circular dark spots arbitrarily dispersed over a relatively brighter matrix. The dark and bright segments arise due to the density difference between the crystalline and amorphous regions, respectively [33]. It is important to observe that the plain-view TEM images show the co-existence of amorphous and crystalline silicon which indicates that the nc-Si:H sub-layer does not only contain the crystalline part, it also includes the amorphous silicon as well and the amount of crystallinity does depend on the thickness of that layer. Accordingly, basic differences can be observed in the superlattice thin films where larger dark spots (hence larger size Si-nc) are present in the thin films with tnc=8 nm compared to the thin films with tnc=5 nm. The degree of crystallinity can be qualitatively estimated by the small area electron 12
ACCEPTED MANUSCRIPT
Figure 8: (a-b) The plain-view TEM images of the superlattice thin films for tnc= 8 and 5 nm. The insets represent the histogram of the size of Si-ncs showing a Gaussian type distribution. (a-i) & (b-i) The small area electron diffraction patterns illustrating brighter diffraction rings for the tnc= 8 nm superlattice thin films. (a-ii) & (b-ii) The high resolution TEM images for the thin films showing silicon nanocrystallites having <111> lattice planes with sharp boundary with the amorphous silicon.
diffraction (SAED) patterns as shown in Figure 8 (a-i) and 8(b-i). The SAED patterns of tnc= 8 nm represent three sharp intense bright rings which correspond to the <111>, <220> and <311> lattice planes of crystalline cubic silicon [7]. Similar rings are also visible for tnc= 5 nm, however having diffused brightness which is due to the lower crystallinity in the later superlattice 13
ACCEPTED MANUSCRIPT
structures. The major difference particularly observed is a gross decrease in size of silicon crystallites with the reduced thickness of nc-Si:H sub-layers.
At tnc= 8 nm, silicon nano-
crystallites have larger size of ~7.6 nm as shown by the subsequent histogram in the inset of Figure 8(a). The high-resolution TEM images (Figure 9(b-ii)) show large size Si-ncs having <111> lattice planes of Si with sharp boundary from the amorphous regions. Lowering of the tnc to 5 nm has drastically reduced the size of Si-ncs to ~5.2 nm as found in the histogram, shown in the inset of Figure 8(b). The high resolution TEM image also has demonstrated the presence of tiny nano-crystallites. It can be significantly noted that the typical size of Si-ncs are closely comparable to the nc-Si:H sub-layer thickness. Therefore, the effective size-control of the Si-ncs and subsequent tuning of properties appears possible only by adjusting the thickness of nc-Si:H sub-layer in superlattice thin films. 3.5. Infrared absorption studies:
Figure 9: The infrared absorption spectra in two different regions, plotted separately at (a) 500–800 cm-1 and (b) 1850–2250 cm-1.
Investigations on the nature of chemical bonds between silicon and hydrogen in the superlattice thin films have been carried out by the infrared absorption studies with samples deposited on the Si-wafer substrates. The typical infrared absorption coefficient spectra of films for different thickness of nc-Si:H sub-layers (tnc) are represented in Figure 9(a) and 9(b) where two distinct wave number regions are plotted separately at 500–800 cm-1 and 1850–2250 cm-1. In the range of wagging mode vibration of SiH (Figure 9(a)), there is dominant peak at ~640 cm-1 [34]. 14
ACCEPTED MANUSCRIPT
However, an elongated tail can also be observed at higher wave number which is due to the SiH rocking mode vibration. It is well-observed that the intensity of the absorption coefficient has been gradually enhanced with the lowering of the thickness of nc-Si:H sub-layers, which has qualitatively signified the overall increase in the hydrogen content in the thin films, as a consequence of increasing amorphous component in the structure. Similar trend is noticeable in the stretching mode vibration range as well (Figure 9(b)). In this range, the individual peaks comprise of two satellite peaks at ~2000 and 2100 cm-1, which correspond to the stretching mode vibrations of silicon mono-hydride (SiH) and silicon di-hydride (SiH2), respectively [34, 35]. The relative increase in the intensity of SiH stretching component compared to the SiH2 counterpart implies the transformation from nano-crystalline to amorphous dominated network structure in the overall superlattice thin films on reducing tnc. In order to get the quantitative estimation of structural properties, the wagging mode absorption coefficient spectra has been deconvoluted to two satellite components: SiH wagging and SiH rocking whereas the stretching mode absorption coefficient spectra has been split into SiH and SiH2 stretching components.
Figure 10:(a) The increase in the bonded hydrogen content as a function of nc-Si:H layer thickness, (b) The comparative change in integrated intensities of the SiH and SiH2 stretching mode vibrations.
The hydrogen content has been calculated from the integrated intensity of the wagging mode vibration following equation:
C H A N Si d
100at.%
where, α is the absorption coefficient corresponding to the wave number ω. Aω= 1.6×1019 cm-2 is the Si-H wagging mode oscillation strength and NSi=5×1022 cm-3 is the atomic density of c-Si 15
ACCEPTED MANUSCRIPT
[36]. The as-estimated hydrogen content has been found to enhance from 1 at.% to 14 at.% as the tnc is reduced from 10 to 2 nm, as seen in Figure 10(a). Adding to that the stretching mode analysis, shown in Figure 10(b) has revealed that fractional percentage of SiH has been increases to ~90% at tnc= 2 nm, implying a highly amorphous dominated superlattice thin films, while at tnc= 11 nm similar high fraction of SiH2 component identifies mostly nc-Si dominated superlattice structure [7]. It is important to mention that in the superlattice structures, only thickness of nc-Si:H sub-layers has been varied keeping all other parameters unchanged. Consequently, the sharp changes in the bonding nature in the superlattice thin films can only be attributed to the structural change occurring in nc-Si:H sub-layers.
4. Discussions: In the field of silicon based nanotechnology, major challenge lies in attaining tailor-made optoelectronics properties. One way to accomplish such properties may emerge through proper utilization of the quantum confinement effect arising out of the reduced size of silicon nanocrystallites towards few nanometers. In doing so, the control over average size of the Si-ncs and its size-distribution appears as the foremost endeavor during the fabrication process. The most common approach, in this regard, is the formation of Si-ncs in various dielectric matrix of silicon e.g. a-SiO, a-SiN or a-SiC by plasma enhanced CVD (inductively and capacitively coupled), layer-by-layer deposition mostly using sputtering or by hotwire CVD [4, 9, 37-39]. For Si-ncs in dielectric matrix, the size is predominantly controlled by the inclusion of O/N/C in the pristine hydrogenated crystalline silicon network and the plasma enhanced CVD is one of the best options for its relatively low temperature growth [6-8]. The increasing O/N/C content could provide the wider optical band gap by means of both quantum confinement effect in the reduced Si-ncs and the effect of dielectric matrix itself. However, by virtue of the mutual trade-off relation existing between the optical and electrical properties in silicon dielectrics materials, the electrical property of the films deteriorates at higher optical band gap which is unfavorable in photovoltaics [40]. Alternatively, the growth of Si-ncs in layer-by-layer structure, in general, involves a very high temperature post-deposition annealing in which crystallization occurs in the silicon-rich dielectric layers between two highly amorphous dielectric matrices. This process is again not compatible to the silicon solar cell fabrication technology because of its high temperature involvement. 16
ACCEPTED MANUSCRIPT
In the present investigation, a-Si:H/nc-Si:H super-lattice structures have been fabricated by a low temperature yet relatively simpler and most popular capacitively coupled radio frequency plasma aided CVD technique. The formation of superlattice thin films with layer-bylayer structure has been evidenced from the SAX measurements, whereas the presence of crystallinity is confirmed by the Raman spectroscopy and spectroscopic ellipsometry. However, more interesting to reveal is that adequate crystallization occurs even at very thin layer (≥3 nm) of nc-Si:H. In general, the growth of nc-Si:H from the silane plasma consists of an initial typically defective and amorphous incubation layer on the substrate, inherently appearing due to the lattice mismatch linking the substrate material and nano-crystalline silicon, and that impede crystallization within very thin layer of few nanometer [41]. In the plasma CVD process, the thickness of incubation layer can be reduced using highly hydrogenated silane plasma where the high atomic-H density over the growing film surface promotes the crystallization by means of hydrogen etching. Interestingly, for the samples under present investigation, this hindrance for nc-Si:H growth are further surpassed by the presence of a primary a-Si:H layer on the substrate that operates as the effective incubation layer and that facilitates growing the nc-Si:H sub-layer possessing considerably high crystallinity inside a very low thickness of ≥3 nm. From the Raman studies, it has been identified that the size of Si-ncs is comparable to thickness of nc-Si:H sublayers attained through the termination of growth of Si-ncs within two neighboring a-Si:H layers on each side. This method of crystallization by extending a-Si:H barrier layer itself as the effective incubation layer and then sudden termination of growth of nc-Si:H layer being repeated for number of cycles to produce the superlattice structure. The variation in nc-Si:H layer thickness, tnc, consequences the overall structural changes in the superlattice thin films. From TEM studies, the superlattice films are found to possess crystallinity embraced within amorphous matrix. On systematic lowering in tnc, the a-Si component dominates over the crystalline part along with simultaneous enhancement of silicon mono-hydride (SiH) bond density over silicon di-hydride (SiH2) which indicates the reduced microstructure factor and hence reduced defect density in the overall superlattice structure that embraces even a high density of tiny silicon nano-crystallites. Such tiny Si-ncs in QD configuration within superlattice structures are highly promising for further improvement of solar cells, as reported in the recent literature [42-44].
17
ACCEPTED MANUSCRIPT
5. Conclusions: Superlattice thin films having layer-by-layer structures of a-Si:H and nc-Si:H are prepared with varying the thickness of nc-Si:H sub-layer by capacitively coupled plasma enhance CVD at low substrate temperature of 180 oC from (SiH4 + H2)-plasma. The presence of superlattice structures of the thin films has been confirmed by small angle XRD studies. The Raman spectroscopy, spectroscopic ellipsometry and TEM studies have corroborated the presence of silicon nanocrystallites. The average sizes of Si-ncs in the superlattice structures have been around the dimension of the individual nc-S:H sub-layer thickness. In general, diminishing thickness of ncSi:H layer consequences gradual increase in amorphous component, however, even at a very low tnc ≥ 3 nm, tiny silicon nano-crystallites are present which are beneficial for both widening of the band gap and providing suitable charge conduction path. Raman studies as well as the infra-red spectroscopy have revealed the presence of lower defects in the superlattice structures with low thickness of the nc-Si:H active layer. The effective growth of Si-ncs is facilitated by the initial aSi:H layer which serves as the actual incubation layer to the growing nc-Si:H sub-layer. This intriguing technique of forming Si-ncs in thin-layer superlattice structures at very low temperature using single step conventional plasma processing could open up a convenient avenue for the fabrication of devices e.g., all-silicon third generation solar cells.
Acknowledgements: The work has been done under nano-silicon projects funded by the Department of Science and Technology (Nano-Mission Program) and the Council of Scientific and Industrial Research, Government of India.
18
ACCEPTED MANUSCRIPT
References: [1]
M.A. Green, Silicon solar cells: The ultimate photovoltaic solution?, Prog. in Photovoltaics 2 (1994) 87-94.
[2]
T.I. Kamins and K.L. Chiang, Properties of Plasma‐Enhanced CVD Silicon Films, J. Electrochem. Soc. 129 (1982) 2326-2331.
[3]
C. Qijin, X. Shuyan and O. Kostya, Structural evolution of nanocrystalline silicon thin films synthesized in high-density, low-temperature reactive plasmas, Nanotechnology 20 20 (2009) 215606.
[4]
X.J. Hao, A.P. Podhorodecki, Y.S. Shen, G. Zatryb, J. Misiewicz and M.A. Green, Effects of Si-rich oxide layer stoichiometry on the structural and optical properties of SiQD/SiO2 multilayer films. Nanotechnology 20 (2009) 485703.
[5]
I.A. Yunaz, K. Hashizume, S. Miyajima, A. Yamada and M. Konagai, Fabrication of amorphous silicon carbide films using VHF-PECVD for triple-junction thin-film solar cell applications. Sol. Energ. Mat. Sol. Cells 93 (2009) 1056–1061.
[6]
A. Samanta and D. Das, Effect of RF power on the formation and size evolution of nCSi quantum dots in an amorphous SiOx matrix. J. Mat. Chem. 21 (2011) 7452–7458.
[7]
D. Kar and D. Das, Conducting wide band gap nc-Si/a-SiC:H films for window layers in nc-Si solar cells. J. Mat. Chem. A. 1 (2013) 14744–14753.
[8]
B. Sain and D. Das, Tunable photoluminescence from nc-Si/a-SiNx:H quantum dot thin films prepared by ICP-CVD. Phys. Chem. Chem. Phys. 15 (2013) 3881–3888.
[9]
Q. Cheng, S. Xu, S. Huang and K. Ostrikov, Effective Control of Nanostructured Phases in Rapid, Room-Temperature Synthesis of Nanocrystalline Si in High-Density Plasmas. Crys. Growth Design. 9 (2009) 2863–2867.
[10]
M.B. Sahana, C. Sudakar, A. Dixit, J.S. Thakur, R. Naik and V.M. Naik, Quantum confinement effects and band gap engineering of SnO2 nanocrystals in a MgO matrix, Acta Mater. 60 (2012) 1072–1078.
19
ACCEPTED MANUSCRIPT
[11]
T. Rajagopalan, X. Wang, B. Lahlouh, C. Ramkumar, P. Dutta and S. Gangopadhyay, Low temperature deposition of nanocrystalline silicon carbide films by plasma enhanced chemical vapor deposition and their structural and optical characterization. J. Appl. Phys. 94 (2003) 5252–5260.
[12]
S. Park, E. Cho, D. Song, G. Conibeer and M.A. Green, n-Type silicon quantum dots and p-type crystalline silicon heteroface solar cells. Sol. Energ. Mat. Sol. Cells 93 (2009) 684–690.
[13]
G. Scardera, T. Puzzer, I. Perez-Wurfl and G. Conibeer, The effects of annealing temperature on the photoluminescence from silicon nitride multilayer structures. J. Cryst. Growth 310 (2008) 3680–3684.
[14]
D. Song, E.C. Cho, G. Conibeer, Y. Huang, C. Flynn and M.A. Green, Structural characterization
of
annealed Si1−xCx/SiC multilayers
targeting
formation
of
Si
nanocrystals in a SiC matrix. J. Appl. Phys. 103 (2008) 083544. [15]
B.G. Lee, D. Hiller, J.W. Luo, O.E. Semonin, M.C. Beard, M. Zacharias and P. Stradins, Strained Interface Defects in Silicon Nanocrystals. Adv. Funct. Mater. 22 (2012) 3223– 3232.
[16]
Y. Rosenwaks, M.C. Hanna, D.H. Levi, D.M. Szmyd, R.K. Ahrenkiel and A.J. Nozik, Hot-carrier cooling in GaAs: Quantum wells versus bulk. Phys. Rev. B. 48 (1993) 14675– 14678.
[17]
G. Chen, Y. Guo, J. Yao, Z. Song and F. Zhang, Interface properties of a-Si:H/SiCx:H superlattice. Thin Solid Films 258 (1995) 132–136.
[18]
B. Abeles and T. Tiedje, Amorphous Semiconductor Superlattices. Phys. Rev. Lett. 51 (1983) 2003–2006.
[19]
G. Han, P. Du, J. Shou and D. Zhao, Effect of hydrogen radical exposure on the interfacial and bulk structure of a-Si:H-SiNx:H multilayer films. Thin Solid Films 334 (1998) 6–10.
[20]
K. Akatsuka, Y. Ebina, M. Muramatsu, T. Sato, H. Hester, D. Kumaresan, R.H. Schmehl, T. Sasaki and M. Haga, Photoelectrochemical Properties of Alternating Multilayer Films Composed of Titania Nanosheets and Zn Porphyrin. Langmuir 23 (2007) 6730–6736. 20
ACCEPTED MANUSCRIPT
[21]
Z. Li and W. Li, Y. Jiang, H. Cai, Y. Gong and J. He, Raman characterization of the structural evolution in amorphous and partially nanocrystalline hydrogenated silicon thin films prepared by PECVD, J. Raman Spectrosc. 42 (2011) 415–421.
[22]
S. Guha, G. Hendershot, D. Peebles, P. Steiner, F. Kozlowski and W. Lang, Photoluminescence and Raman studies of porous silicon in polymethyl methacrylate, Appl. Phys. Lett. 64 (1994) 613-615.
[23]
Y. He, Y. Wei, G. Zheng, M. Yu, M. Liu, An exploratory study of the conduction mechanism of hydrogenated nanocrystalline silicon films, J. Appl. Phys. 82 (1997) 34083413.
[24]
D. Raha and D. Das, Controlling the growth of nanocrystalline silicon by tuning negative substrate bias, Sol. Energ. Mater. Sol. Cells 95 (2011) 3181-3188.
[25]
V. Paillard, P. Puech, M.A. Laguna, R. Carles, B. Kohn and F. Huisken, Improved onephonon confinement model for an accurate size determination of silicon nanocrystals. J. Appl. Phys. 86 (1999) 1921–1924.
[26]
D. Das and D. Kar, Spectroscopic and microscopic studies of self-assembled nc-Si/a-SiC thin films grown by low pressure high density spontaneous plasma processing, Phys. Chem. Chem. Phys. 16 (2014) 25421-25431.
[27]
D. Das and D. Kar, Investigation of the vertical electrical transport in a-Si: H/nc-Si: H superlattice thin films, Phys. Chem. Chem. Phys. 17 (2015) 17063-17068.
[28]
D. Kar and D. Das, Opto-electronic properties of P-doped nc-Si-QD/a-SiC: H thin films as foundation layer for all-Si solar cells in superstrate configuration, J. Appl. Phys. 120 (2016) 025102.
[29]
R.B. Bjorklund, S. Zangooie and H. Arwin, Adsorption of surfactants in porous silicon films. Langmuir 13 (1997) 1440–1445.
[30]
D. Das, Micro-Raman and ultraviolet ellipsometry studies on μc-Si:H films prepared by H2 dilution to the Ar-assisted SiH4 plasma in radio frequency glow discharge. J. Appl. Phys. 93 (2003) 2528–2535.
21
ACCEPTED MANUSCRIPT
[31]
I. Stenger, B. Gallas, L. Siozade, C.C. Kao, S. Chenot, S. Fisson, G. Vuye and J. Rivory, Evolution of the optical properties of Si nanoparticles embedded in SiO2 as function of annealing conditions. J. Appl. Phys. 10 (2008) 114303.
[32]
A. Fontcuberta i Morral, P. Roca i Cabarrocas, and C. Clerc, Structure and hydrogen content of polymorphous silicon thin films studied by spectroscopic ellipsometry and nuclear measurements. Phys. Rev. B 69 (2004) 125307.
[33]
D. Das and D. Kar, Self-assembled nc-Si-QD/a-SiC thin films from planar ICP-CVD plasma without H2-dilution: a combination of wide optical gap, high conductivity and preferred < 220> crystallographic orientation, uniquely appropriate for nc-Si solar cells, RSC Adv. 6 (2016) 3860-3869.
[34]
G. Lucovsky, J. Yang, S.S. Chao, J.E. Tyler and W. Czubatyj, Oxygen-bonding environments in glow-discharge-deposited amorphous silicon-hydrogen alloy films, Phys. Rev. B 28 (1983) 3225-3233.
[35]
K. Bhattacharya
and D. Das, Effect of deposition temperature on the growth of
nanocrystalline silicon network from helium diluted silane plasma, J. Phys. D: Appl. Phys. 41 (2008) 155420. [36]
A. Samanta and D. Das, Studies on the structural properties of SiO:H films prepared from (SiH4+CO2+He) plasma in RF-PECVD, Sol. Energ. Mater. Sol. Cells 93 (2009) 588-596.
[37]
B. Rezgui, A. Sibai, T. Nychyporuk, M. Lemiti, G. Bremond, D. Maestre and O. Palais, Appl. Phys. Lett. 96 (2010) 183105.
[38]
M. Zacharias, J. Heitmann, R. Scholz, U. Kahler, M. Schmidt and J. Blasing, Appl. Phys. Lett. 80 (2002) 661.
[39]
P. Shen, N. Uesawa, S. Inasawa and Y. Yamaguchi, J. Mater. Chem. 20 (2010) 1669.
[40]
D. Das and B. Sain, Electrical transport phenomena prevailing in undoped nc-Si/a-SiNx:H thin films prepared by ICP-CVD, J. Appl. Phys. 114 (2013) 073708.
[41]
M Jana, D Das, S.T Kshirsagar and A.K Barua, Control of Crystallization at Low Thickness in µc-Si: H Films Using Layer-by-Layer Growth Scheme, Jpn. J. Appl. Phys. 38 (1999) L1087. 22
ACCEPTED MANUSCRIPT
[42]
S. Janz, P. Löper and M. Schnabel; Silicon nanocrystals produced by solid phase crystallisation of superlattices for photovoltaic applications, Mat. Sci. and Engg. B 178 (2013) 542–550.
[43]
J. Ma, J. Ni, J. Zhang, Q. Liu, G. Hou, X. Chen, X. Zhang and Y. Zhao; In situ grown size-controlled silicon nanocrystals: A p type nanocrystalline-Si:H/a-SiCx:H superlattice (p-nc-Si:H/a-SiCx:H) approach, Sol. Energy Mater. & Sol. Cells 123 (2014) 228–232.
[44]
S. Yamada, Y. Kurokawa, S. Miyajima1 and M. Konagai; Silicon quantum dot superlattice solar cell structure including silicon nanocrystals in a photogeneration layer, Nanoscale Research Letters 9 (2014) 246.
23
ACCEPTED MANUSCRIPT
Highlights: # a-Si:H/nc-Si:H thin film superlattice structures grown at ~180 oC by single step plasma process # lower defects in superlattice structures with lower thickness (tnc) of nc-Si:H active layer # barrier layer incubates Si-ncs facilitating growth of active layer with adequate Si-ncs at tnc ≥3 nm # sharp control in active layer growth confines Si-nc size within dimension of tnc # easy yet skillful method with huge technological impact of Si-ncs in thin superlattice structures
Debjit Kar, Debajyoti Das PII:
S0749-6036(17)30773-5
DOI:
10.1016/j.spmi.2017.06.053
Reference:
YSPMI 5102
To appear in:
Superlattices and Microstructures
Received Date:
29 March 2017
Revised Date:
19 June 2017
Accepted Date:
20 June 2017
Please cite this article as: Debjit Kar, Debajyoti Das, Structural characterization of silicon thin film superlattice grown at low temperature, Superlattices and Microstructures (2017), doi: 10.1016/j. spmi.2017.06.053
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Structural characterization of silicon thin film superlattice grown at low temperature Debjit Kar and Debajyoti Das* Nano-Science Group, Energy Research Unit, Indian Association for the Cultivation of Science, Jadavpur, Kolkata – 700 032, INDIA
Abstract: The growth of nc-Si:H from the SiH4 plasma inherently proceeds through a defective and amorphous incubation layer, originating due to the lattice mismatch between the substrate and Si material. In the current investigation, a-Si:H/nc-Si:H thin film super-lattice structures have been fabricated at very low temperature (~180 oC), using single step conventional plasma processing. In this intriguing technique the a-Si:H barrier sub-layer itself operates as an effective incubation layer which facilitates growing the nc-Si:H active sub-layer possessing adequate crystallinity inside a very low thickness (tnc) of even ≥3 nm, and the sudden discontinuity in growth of the active layer confines the size of Si-ncs within the dimension of this active sub-layer thickness. The superlattice structure of the ultimate network has been evidenced from small angle XRD (SAX) measurements, whereas the nature of crystallinity has been studied by ellipsometry, Raman spectroscopy and electron microscopy. This simple yet compelling technique seems to have enormous technological implications taking the advantage of the quantum confinement effect in low dimensional silicon nano-crystallites within very thin superlattice structures. Keywords:
incubation layer, thin film superlattice, low temperature growth, small angle XRD (SAX)
*Author to whom all correspondence should be addressed, E-mail (D. Das): [email protected]; Fax: +91(33)24732805 1
ACCEPTED MANUSCRIPT
1. Introduction: It is highly expected that in a near future, nanocrystalline silicon (nc-Si) thin film based solar cells having high conversion efficiency and low fabrication cost will substitute bulk crystalline silicon based devices for most applications [1]. With that aim numerous attempts have been made by the researchers towards evolution of new technologies and materials. In that direction tandem structure solar cell is the one that can boost the efficiency progressively with the utilization of the extended solar spectrum by stacking more cells on top of the other, combining multiple absorber layers of different optical band gaps [2-3]. The individual band gap of each absorber layer needs to maintain a descending order from the top cell to the bottom one, in order to absorb the photons in the most efficient way [4-5]. In order to achieve compatible transport of charge carriers within different layers in the cell structure, the present day silicon research has got enormous impetus in the development and utilization of various low-dimensional structures of silicon e.g., nanocrystals, quantum dots and superlattice structures in thin films [4,6-9]. By virtue of quantum confinement effects arising within its low dimensional structures with magnitude in the order of corresponding Bohr radius, silicon nanocrystals or quantum dots exhibit numerous novel opto-electronic properties; it can widen the optical band gap without deteriorating the electrical conductivity [10-11]. In general, the silicon nano-crystals or quantum dots are synthesized in a two-step process: in the first step silicon dielectric thin films are deposited followed by annealing at elevated temperature (~1100 oC) for solid-phase crystallization of Si. Conversely, in nano-scaled superlattice structures, silicon layer and Si-dielectric layer (Si3N4/SiO2/SiC) are alternately deposited in stack and subsequently treated with a post deposition high temperature (~1100 oC) annealing [12-15]. The unavoidable high temperature annealing step hinders its successful utilization in the fabrication of thin film solar cells by severely affecting the other pre-deposited layers of the device, e.g., initiating diffusion of dopants across the doped layers, and limits the use of low cost substrates, as well. Moreover, the high temperature annealing step belongs to an uncertain process in regulating the size, distribution and number density of the Si nano-crystals. Furthermore, in order to eliminate the dielectric barrier on both sides of the charge carriers in Si nanocrystals, the dielectric layers of the superlattice structure need to be thin enough and the Sincs in the active layer are required in high density so that the wave functions involving the 2
ACCEPTED MANUSCRIPT
adjacent Si-ncs mutually overlap. Consequently, numerous stringent issues are involved in the commonly employed superlattice growth technology to utilize in solar cell fabrication. In this context, one-step fabrication of superlattice films with the alternate a-Si:H/nc-Si:H sub-layers possessing plenty of tiny Si-ncs grown at device compatible temperature, could be significantly effective in mobilizing the transport of charge carriers across the relatively low dielectric barrier of the a-Si:H sub-layer and, thereby, facilitating device fabrication [16]. In present manuscript a comprehensive structural characterization of such a-Si:H/nc-Si:H thin film superlattice structures developed at ~180 °C by conventional rf-PECVD has been pursued, in view of its future utilization in different layers of the all silicon solar cells with third generation technology at low cost.
2. Experimental Details: Thin films of superlattice structures having alternative layers of a-Si:H and nc-Si:H were fabricated in a capacitively coupled PECVD reactor, using (SiH4 + H2)-plasma. Before proceeding towards the formation of superlattice thin films, the deposition conditions for individual layers have been optimized independently. For growing the a-Si:H layers, the SiH4 and H2 flow rates, the pressure of reacting gases in the plasma and rf-power operated to the parallel plate electrodes were kept at 95 sccm, 5 sccm, 3 Torr and 15 W, respectively; while for the nc-Si:H layers, those parameters were set at 99.5 sccm, 0.5 sccm, 4 Torr and 40 W, respectively. Prior to deposition, the plasma chamber has been evacuated to ~ 1×10-6 Torr and then heated up to 350 oC in the presence of H2-gas at relatively higher gas pressure (50 Torr). These steps were followed in order to eliminate any residual oxygen from the reactor that can contaminate the periodic layers of the superlattice structure. Finally, a constant temperature of the Corning® Eagle2000TM glass and silicon wafer substrates was maintained at 180 oC throughout the entire deposition process. A set of superlattice films was grown with identical aSi:H sub-layers of thickness 4.8 nm and by varying the thickness of nc-Si:H sub-layer from 11 to 1.5 nm, by changing only the corresponding sub-layer deposition time. A gross thickness of around 200 nm was maintained for each superlattice sample with similar a-Si:H sub-layer at its both ends, by accumulating different numbers of bi-layer cycles.
3
ACCEPTED MANUSCRIPT
Small angle X-ray diffraction was obtained by Bruker (D8 Advance) system with source radiation (λ=1.5418 °A) beam incident at glancing angle (GISAX). The Raman spectroscopy was performed by 514 nm wavelength excitation from Ar+ laser (Spectra Physics) operated at ~2 mWcm-2 power density and J-Y Horiba Triple Raman spectrometer. Olympus open stage microscope was used to focus the beam spot inside the sample and the signal was detected in a back-scattering geometry by a TE-cooled CCD camera. Transmission electron microscope data was obtained by a transmission electron microscope (JEOL JSM 2010) working at 200 kV. The samples for TEM studies were prepared on Cu microscope grids previously coated by amorphous carbon. Infra-red spectroscopic studies were carried out on samples deposited on single crystal Si substrates, by Perkin Elmer Spectrum100 FTIR spectrometer.
3. Results and Discussions: 3.1. Small-Angle XRD: Small angle X-ray diffraction (SAX) is an essential non-destructive tool to characterize the periodic structures of the superlattice thin films. Figure 1 shows the SAX pattern of different a-Si:H/nc-Si:H superlattice samples. Diffraction patterns have different 1st and 2nd order (m) peaks, which are the clear evidence of the presence of multi-layers structure in the samples. The first order diffraction peaks are observable for the superlattice films with tnc=1.5, 2 and 3 nm whereas for tnc>3 nm, only the second order diffraction peaks are noticeable due to the instrumental limitations on angle scan within 2-span of 0.5° to 4.0°. It is noteworthy that the peak positions of each first and second order diffraction peaks gradually shift towards lower diffraction angle (2θ) as the periodic thickness of the superlattice films increases by virtue of increasing the active layer thickness, tnc (thickness of barrier layer remaining constant). From the Braggs’ law of diffraction, the periodic thickness of each superlattice film has been calculated from the 1st or 2nd order diffraction peaks. The increasing thickness of nc-Si:H active sub-layer of the periodic structure of different superlattice films are also separately estimated from the sample thickness measured by surface profilometer. Figure 2 demonstrates an excellent linear relation between the periodic thickness of each superlattice films estimated from SAX data and the periodic thicknesses of corresponding nc-Si:H active layer thickness tnc obtained from 4
ACCEPTED MANUSCRIPT
Figure 1: The small angle XRD (SAX) patterns of different superlattice thin films showing 1st and 2nd order diffraction peaks due to the presence of multilayer structure.
profilometer data. Interestingly, the extrapolation of the linear plot to tnc= 0, shown by the blue section in the plot, leads to the superlattice periodic thickness ~4.8 nm which is exactly the thickness of a-Si:H barrier layer in each cycle, thus confirming the accuracy of both experimental process. It may be noted that the peak-widths of the SAX diffraction peaks are much larger (~0.5o) than the resolution (0.02o) of the diffractometer. This implies a blurred edge between the sub-layers which might arise due to the periodic presence of a very thin amorphous incubation section in the nc-Si:H active sub-layer [17-20] which is different from the a-Si:H barrier sub-layer because of their different growth parameters.
5
Periodic thickness of superlattice (nm) estimated from SAX data
ACCEPTED MANUSCRIPT
16
Thickness of a-Si sub-layer =4.8 nm
14 12 10 8 6 4
0
2
4 6 8 10 nc-Si:H sub-layer thickness (nm) estimated from profiler data
12
Figure 2: The as-estimated periodic thickness from the SAX plotted as a function of thickness of the nc-Si:H sub-layer.
3.2. Raman Spectroscopic studies: The crystalline nature of the superlattice structure has been investigated using Raman spectroscopy. The Raman spectrum of each superlattice thin film has been shown in Figure 3. On lowering in the thickness of nc-Si:H sub-layer (tnc), a consistent red-shift of peak frequency (~ 520 cm-1) occurs while the peak width broadens asymmetrically. For tnc=2 nm, the Raman peak is very broad and centered at ~485 cm-1 and mostly presents an amorphous dominated network. For tnc>2 nm, the observed sharp peak at ~ 520 cm-1 corresponds to the TO mode vibration of nc-Si implying the presence of high amount of crystallinity in the superlattice thin films [7]. A close observation on the peak position of nc-Si component, shown at the inset of Fig. 3 reveals that there is a significant peak shift of ~7 cm-1 as the thickness of the nc-Si:H layer is decreased from 10 to 3 nm which signifies the gradual miniaturization of the size of the silicon 6
ACCEPTED MANUSCRIPT
Figure 3: The Raman spectra of different superlattice thin films showing consistent modification of the peak characteristics with the change in tnc. The inset shows the close-look of the dominant peak position where it has been shifted towards lower wave number at lower tnc.
nano-crystallites. In addition, the observed broadening of the ~485 cm-1 component of the Raman signal identifies gradual enhancement of the amorphous component in the superlattice thin films on reducing tnc. The quantitative estimation of the degree of crystallinity and the average size of the Si-nc has been done by the deconvolution of each Raman spectrum into four Gaussian and one Lorentzian satellite components, the typical features of which are shown in the Figure 4. In particular, there are five satellite components at ~310 cm-1, ~420 cm-1, 480 cm-1, 500 cm-1 and 519 cm-1 – the first three components correspond to the Raman signals from LA, LO and TO modes of a-Si, the fourth one corresponds to the thermodynamically stable ultra nanocrystallites of silicon or the defects due to the grain-boundary regions and the sharp Lorentzian component is due to the TO mode of the nano-crystalline silicon (nc-Si) [21-24].
7
ACCEPTED MANUSCRIPT
Figure 4: The representative deconvolution in the Raman spectra of the superlattice films with tnc= 6 nm. The insert shows the reduced size of the Si-ncs with the lowering of the tnc.
The nano-crystalline component of the Raman spectra of the superlattice thin films has been found to shift towards lower wave number with gradual thinning of the nc-Si:H sub-layers which is a consequence of gradual decrease in size of the Si-ncs and the phonon confinement effect arising thereof. In first order Raman scattering of bulk c-Si, the excited phonon is constrained at the centre of the Brillouin zone following the momentum conservation law. However, due to the spatial confinement in the tiny Si-ncs, the phonons which are dispersed over the Brillouin zone starts contributing to the Raman signal due to the presence of structural defects at the boundaries of nano-crystallites. Considering that, first order Raman spectra has been simulated taking into account the phonon confinement effect in the reduced size of Si-ncs having Gaussian size-distribution. A sine wave has been considered as the confinement function that disappears just outside the nc-Si boundary and the phonon dispersion curve has been considered following Paillard [25]. The specific simulation technique has been described in our previous work [26]. The as estimated average size of the Si-nc has been plotted in the inset of Figure 4 and the size of Si-nc has been found to be reduced from 9.1 nm to 1.9 nm with lowering of tnc from 10 to 2 nm. Interestingly, the average grain sizes are in good agreement with the thickness (tnc) of nc-Si:H sub-layer. 8
ACCEPTED MANUSCRIPT
Figure 5: Relative change of the integrated intensities of the unc- and nc- components of the Raman spectra.
Figure 5 represents the relative variation in the ratio of integrated area between the ultrananocrystalline (unc-) to nanocrystalline (nc-) silicon satellite peaks in superlattice structures occurring due to the decrease in tnc. Presence of unc-Si is much lower in the more crystalline superlattice thin films. It is noted that at tnc<2 nm, the unc-Si components dominates over nc-Si. The presence of high volume of ultra-nanocrystallites is two-fold beneficial in material point of view as it can widen the optical band gap because of the strong quantum confinement effect and can also provide better charge carrier conduction path by the processes like variable range hopping or Poole-Frenkel mechanism [27,28]. However, the defects at the amorphous silicon part can hinder the charge conduction process. In view of that, the comparative estimation of those defects have been performed by the ratio of integrated intensities of LA and TO modes of a-Si (ILA/ITO) and bond angle deviation (Δθ) of a-Si. The bond angle deviation Δθ (in degree) in the amorphous silicon network has been calculated from the FWHM (Γ in cm-1) of the TO mode of a-Si using the following relation: Γ=15 + 6Δθ. The ILA/ITO and Δθ have been plotted in Figure 6 as function of the thickness of nc-Si:H layer. The magnitude of ILA/ITO is much lower for superlattice thin films at tnc<6 nm which signifies the lower defects and stress in the amorphous 9
ACCEPTED MANUSCRIPT
network. The magnitude of Δθ is found to be in an increasing order at lower tnc due to more amorphous component in the overall superlattice structures and still the magnitude of Δθ at tnc=2 nm is quite low implying a overall less defective amorphous network [21]. Interestingly, the sizecontrol of Si-ncs by adjusting only the thickness of nc-Si:H layer is emerging to be significantly effective without introducing much defects in the superlattice thin films.
Figure 6: The variation in the ILA/ITO and Δθ of a-Si as estimated from the Raman spectroscopic studies.
3.3. Spectroscopic Ellipsometry studies: The spectroscopic ellipsometry measurement is a sensitive yet non-destructive technique that determines the complex reflection ratio of the light reflected back from the film surface, from which the complex dielectric constant can be acquired. Complex dielectric data derived from this measurement can provide qualitative and also quantitative dielectric properties, nature of the chemical bonds present as well as the structural and the morphological properties of the bulk layer, surface layer and the interfaces of the materials. The fundamental parameters of the spectroscopic ellipsometry are Ψ and Δ, which are the amplitude and phase part of complex reflectance ratio (ρ) which can be expressed as ρ = tan(Ψ)eiΔ [29]. The imaginary part of the 10
ACCEPTED MANUSCRIPT
complex dielectric function, <ε2>, as derived from the Ψ and Δ has been plotted in the Figure 7(a) for various superlattice films having different thickness of nc-Si:H sub-layers. A typical spectrum of <ε2> comprises of two distinct energy regions. The comparatively lower energy part of the spectrum (not shown here) consists of the interference pattern due to the superposition of reflected beams from the air-thin films and thin film-substrate interfaces. This interference pattern carries the information about the thickness of the films and the nature of the substrate. However, at higher energy where the penetration depth of the photon is smaller than the film thickness, the nature of the <ε2> spectra is controlled by the compositional properties of the thin film along with the surface roughness [30].
Figure 7: (a) The spectra of the imaginary part of the pseudo-dielectric function for different superlattice thin films having different tnc. (b) The reduction of the intensity difference of the two critical points due to the gradual decrease of tnc.
The superlattice thin film having tnc=11 nm represents two discrete shoulders, around ~3.4 and ~4.2 eV which have been recognized as the critical points (CPs) in the band structure of bulk crystalline silicon, E1 and E2, respectively [31]. The presence of those two shoulders is the consequence of the high degree of crystallinity in the superlattice thin films. These shoulders can be well-observable for tnc≥3 nm, however at tnc<3 nm a Gaussian-like peak appears at ~3.51 eV with higher intensity, possibly implying formation of amorphous dominated network. Typically, the peak corresponding to a-Si is positioned at 3.2 eV. Such shifting of the peak towards higher 11
ACCEPTED MANUSCRIPT
energy might be due strong quantum confinement effect from the nano-crystals in the thin film. It is notably observed that the intensity of <ε2> decreased gradually with the reduction in tnc until 5 nm, however beyond that the intensity sharply enhanced and attained high magnitude for the superlattice films having mostly bulk a-Si:H. The intensity of <ε2> for nano-crystalline silicon thin films is largely governed by the voids and the size of the silicon nano-crystallites. However, those two parameters have a trade-off relation between them – while the higher void can reduce the intensity, the larger size silicon nano-crystallites (hence higher degree of crystallinity in the films) can enhance the intensity [32]. A qualitative estimation of the degree of crystallinity can be performed by a comparative study in the difference in the intensities of the two shoulders which is reduced systematically at lower tnc. Such reduction in intensity is closely related to the decreasing overall crystalline volume fraction and the size-reduction of silicon nano-crystallites because at the smaller diameter of the Si-ncs the oscillator strength changes between the lower and higher energy states. For the present set of superlattice thin films, the size of Si-ncs are reduced with lower crystalline volume fraction that consistently lead to the significant reduction in the intensity of <ε2>, however, for tnc<5 nm the films become more amorphous dominated, having less fraction of voids which lead to the sharp rise in the intensity [26,30].
3.4. TEM studies: Development of the Si-ncs embedded in the nano-crystalline silicon sub-layer of the superlattice structures has been pursued by TEM studies. In Figure 8(a) and 8(b), the plain-view micrographs of the superlattice structures with tnc= 8 and 5 nm demonstrate circular dark spots arbitrarily dispersed over a relatively brighter matrix. The dark and bright segments arise due to the density difference between the crystalline and amorphous regions, respectively [33]. It is important to observe that the plain-view TEM images show the co-existence of amorphous and crystalline silicon which indicates that the nc-Si:H sub-layer does not only contain the crystalline part, it also includes the amorphous silicon as well and the amount of crystallinity does depend on the thickness of that layer. Accordingly, basic differences can be observed in the superlattice thin films where larger dark spots (hence larger size Si-nc) are present in the thin films with tnc=8 nm compared to the thin films with tnc=5 nm. The degree of crystallinity can be qualitatively estimated by the small area electron 12
ACCEPTED MANUSCRIPT
Figure 8: (a-b) The plain-view TEM images of the superlattice thin films for tnc= 8 and 5 nm. The insets represent the histogram of the size of Si-ncs showing a Gaussian type distribution. (a-i) & (b-i) The small area electron diffraction patterns illustrating brighter diffraction rings for the tnc= 8 nm superlattice thin films. (a-ii) & (b-ii) The high resolution TEM images for the thin films showing silicon nanocrystallites having <111> lattice planes with sharp boundary with the amorphous silicon.
diffraction (SAED) patterns as shown in Figure 8 (a-i) and 8(b-i). The SAED patterns of tnc= 8 nm represent three sharp intense bright rings which correspond to the <111>, <220> and <311> lattice planes of crystalline cubic silicon [7]. Similar rings are also visible for tnc= 5 nm, however having diffused brightness which is due to the lower crystallinity in the later superlattice 13
ACCEPTED MANUSCRIPT
structures. The major difference particularly observed is a gross decrease in size of silicon crystallites with the reduced thickness of nc-Si:H sub-layers.
At tnc= 8 nm, silicon nano-
crystallites have larger size of ~7.6 nm as shown by the subsequent histogram in the inset of Figure 8(a). The high-resolution TEM images (Figure 9(b-ii)) show large size Si-ncs having <111> lattice planes of Si with sharp boundary from the amorphous regions. Lowering of the tnc to 5 nm has drastically reduced the size of Si-ncs to ~5.2 nm as found in the histogram, shown in the inset of Figure 8(b). The high resolution TEM image also has demonstrated the presence of tiny nano-crystallites. It can be significantly noted that the typical size of Si-ncs are closely comparable to the nc-Si:H sub-layer thickness. Therefore, the effective size-control of the Si-ncs and subsequent tuning of properties appears possible only by adjusting the thickness of nc-Si:H sub-layer in superlattice thin films. 3.5. Infrared absorption studies:
Figure 9: The infrared absorption spectra in two different regions, plotted separately at (a) 500–800 cm-1 and (b) 1850–2250 cm-1.
Investigations on the nature of chemical bonds between silicon and hydrogen in the superlattice thin films have been carried out by the infrared absorption studies with samples deposited on the Si-wafer substrates. The typical infrared absorption coefficient spectra of films for different thickness of nc-Si:H sub-layers (tnc) are represented in Figure 9(a) and 9(b) where two distinct wave number regions are plotted separately at 500–800 cm-1 and 1850–2250 cm-1. In the range of wagging mode vibration of SiH (Figure 9(a)), there is dominant peak at ~640 cm-1 [34]. 14
ACCEPTED MANUSCRIPT
However, an elongated tail can also be observed at higher wave number which is due to the SiH rocking mode vibration. It is well-observed that the intensity of the absorption coefficient has been gradually enhanced with the lowering of the thickness of nc-Si:H sub-layers, which has qualitatively signified the overall increase in the hydrogen content in the thin films, as a consequence of increasing amorphous component in the structure. Similar trend is noticeable in the stretching mode vibration range as well (Figure 9(b)). In this range, the individual peaks comprise of two satellite peaks at ~2000 and 2100 cm-1, which correspond to the stretching mode vibrations of silicon mono-hydride (SiH) and silicon di-hydride (SiH2), respectively [34, 35]. The relative increase in the intensity of SiH stretching component compared to the SiH2 counterpart implies the transformation from nano-crystalline to amorphous dominated network structure in the overall superlattice thin films on reducing tnc. In order to get the quantitative estimation of structural properties, the wagging mode absorption coefficient spectra has been deconvoluted to two satellite components: SiH wagging and SiH rocking whereas the stretching mode absorption coefficient spectra has been split into SiH and SiH2 stretching components.
Figure 10:(a) The increase in the bonded hydrogen content as a function of nc-Si:H layer thickness, (b) The comparative change in integrated intensities of the SiH and SiH2 stretching mode vibrations.
The hydrogen content has been calculated from the integrated intensity of the wagging mode vibration following equation:
C H A N Si d
100at.%
where, α is the absorption coefficient corresponding to the wave number ω. Aω= 1.6×1019 cm-2 is the Si-H wagging mode oscillation strength and NSi=5×1022 cm-3 is the atomic density of c-Si 15
ACCEPTED MANUSCRIPT
[36]. The as-estimated hydrogen content has been found to enhance from 1 at.% to 14 at.% as the tnc is reduced from 10 to 2 nm, as seen in Figure 10(a). Adding to that the stretching mode analysis, shown in Figure 10(b) has revealed that fractional percentage of SiH has been increases to ~90% at tnc= 2 nm, implying a highly amorphous dominated superlattice thin films, while at tnc= 11 nm similar high fraction of SiH2 component identifies mostly nc-Si dominated superlattice structure [7]. It is important to mention that in the superlattice structures, only thickness of nc-Si:H sub-layers has been varied keeping all other parameters unchanged. Consequently, the sharp changes in the bonding nature in the superlattice thin films can only be attributed to the structural change occurring in nc-Si:H sub-layers.
4. Discussions: In the field of silicon based nanotechnology, major challenge lies in attaining tailor-made optoelectronics properties. One way to accomplish such properties may emerge through proper utilization of the quantum confinement effect arising out of the reduced size of silicon nanocrystallites towards few nanometers. In doing so, the control over average size of the Si-ncs and its size-distribution appears as the foremost endeavor during the fabrication process. The most common approach, in this regard, is the formation of Si-ncs in various dielectric matrix of silicon e.g. a-SiO, a-SiN or a-SiC by plasma enhanced CVD (inductively and capacitively coupled), layer-by-layer deposition mostly using sputtering or by hotwire CVD [4, 9, 37-39]. For Si-ncs in dielectric matrix, the size is predominantly controlled by the inclusion of O/N/C in the pristine hydrogenated crystalline silicon network and the plasma enhanced CVD is one of the best options for its relatively low temperature growth [6-8]. The increasing O/N/C content could provide the wider optical band gap by means of both quantum confinement effect in the reduced Si-ncs and the effect of dielectric matrix itself. However, by virtue of the mutual trade-off relation existing between the optical and electrical properties in silicon dielectrics materials, the electrical property of the films deteriorates at higher optical band gap which is unfavorable in photovoltaics [40]. Alternatively, the growth of Si-ncs in layer-by-layer structure, in general, involves a very high temperature post-deposition annealing in which crystallization occurs in the silicon-rich dielectric layers between two highly amorphous dielectric matrices. This process is again not compatible to the silicon solar cell fabrication technology because of its high temperature involvement. 16
ACCEPTED MANUSCRIPT
In the present investigation, a-Si:H/nc-Si:H super-lattice structures have been fabricated by a low temperature yet relatively simpler and most popular capacitively coupled radio frequency plasma aided CVD technique. The formation of superlattice thin films with layer-bylayer structure has been evidenced from the SAX measurements, whereas the presence of crystallinity is confirmed by the Raman spectroscopy and spectroscopic ellipsometry. However, more interesting to reveal is that adequate crystallization occurs even at very thin layer (≥3 nm) of nc-Si:H. In general, the growth of nc-Si:H from the silane plasma consists of an initial typically defective and amorphous incubation layer on the substrate, inherently appearing due to the lattice mismatch linking the substrate material and nano-crystalline silicon, and that impede crystallization within very thin layer of few nanometer [41]. In the plasma CVD process, the thickness of incubation layer can be reduced using highly hydrogenated silane plasma where the high atomic-H density over the growing film surface promotes the crystallization by means of hydrogen etching. Interestingly, for the samples under present investigation, this hindrance for nc-Si:H growth are further surpassed by the presence of a primary a-Si:H layer on the substrate that operates as the effective incubation layer and that facilitates growing the nc-Si:H sub-layer possessing considerably high crystallinity inside a very low thickness of ≥3 nm. From the Raman studies, it has been identified that the size of Si-ncs is comparable to thickness of nc-Si:H sublayers attained through the termination of growth of Si-ncs within two neighboring a-Si:H layers on each side. This method of crystallization by extending a-Si:H barrier layer itself as the effective incubation layer and then sudden termination of growth of nc-Si:H layer being repeated for number of cycles to produce the superlattice structure. The variation in nc-Si:H layer thickness, tnc, consequences the overall structural changes in the superlattice thin films. From TEM studies, the superlattice films are found to possess crystallinity embraced within amorphous matrix. On systematic lowering in tnc, the a-Si component dominates over the crystalline part along with simultaneous enhancement of silicon mono-hydride (SiH) bond density over silicon di-hydride (SiH2) which indicates the reduced microstructure factor and hence reduced defect density in the overall superlattice structure that embraces even a high density of tiny silicon nano-crystallites. Such tiny Si-ncs in QD configuration within superlattice structures are highly promising for further improvement of solar cells, as reported in the recent literature [42-44].
17
ACCEPTED MANUSCRIPT
5. Conclusions: Superlattice thin films having layer-by-layer structures of a-Si:H and nc-Si:H are prepared with varying the thickness of nc-Si:H sub-layer by capacitively coupled plasma enhance CVD at low substrate temperature of 180 oC from (SiH4 + H2)-plasma. The presence of superlattice structures of the thin films has been confirmed by small angle XRD studies. The Raman spectroscopy, spectroscopic ellipsometry and TEM studies have corroborated the presence of silicon nanocrystallites. The average sizes of Si-ncs in the superlattice structures have been around the dimension of the individual nc-S:H sub-layer thickness. In general, diminishing thickness of ncSi:H layer consequences gradual increase in amorphous component, however, even at a very low tnc ≥ 3 nm, tiny silicon nano-crystallites are present which are beneficial for both widening of the band gap and providing suitable charge conduction path. Raman studies as well as the infra-red spectroscopy have revealed the presence of lower defects in the superlattice structures with low thickness of the nc-Si:H active layer. The effective growth of Si-ncs is facilitated by the initial aSi:H layer which serves as the actual incubation layer to the growing nc-Si:H sub-layer. This intriguing technique of forming Si-ncs in thin-layer superlattice structures at very low temperature using single step conventional plasma processing could open up a convenient avenue for the fabrication of devices e.g., all-silicon third generation solar cells.
Acknowledgements: The work has been done under nano-silicon projects funded by the Department of Science and Technology (Nano-Mission Program) and the Council of Scientific and Industrial Research, Government of India.
18
ACCEPTED MANUSCRIPT
References: [1]
M.A. Green, Silicon solar cells: The ultimate photovoltaic solution?, Prog. in Photovoltaics 2 (1994) 87-94.
[2]
T.I. Kamins and K.L. Chiang, Properties of Plasma‐Enhanced CVD Silicon Films, J. Electrochem. Soc. 129 (1982) 2326-2331.
[3]
C. Qijin, X. Shuyan and O. Kostya, Structural evolution of nanocrystalline silicon thin films synthesized in high-density, low-temperature reactive plasmas, Nanotechnology 20 20 (2009) 215606.
[4]
X.J. Hao, A.P. Podhorodecki, Y.S. Shen, G. Zatryb, J. Misiewicz and M.A. Green, Effects of Si-rich oxide layer stoichiometry on the structural and optical properties of SiQD/SiO2 multilayer films. Nanotechnology 20 (2009) 485703.
[5]
I.A. Yunaz, K. Hashizume, S. Miyajima, A. Yamada and M. Konagai, Fabrication of amorphous silicon carbide films using VHF-PECVD for triple-junction thin-film solar cell applications. Sol. Energ. Mat. Sol. Cells 93 (2009) 1056–1061.
[6]
A. Samanta and D. Das, Effect of RF power on the formation and size evolution of nCSi quantum dots in an amorphous SiOx matrix. J. Mat. Chem. 21 (2011) 7452–7458.
[7]
D. Kar and D. Das, Conducting wide band gap nc-Si/a-SiC:H films for window layers in nc-Si solar cells. J. Mat. Chem. A. 1 (2013) 14744–14753.
[8]
B. Sain and D. Das, Tunable photoluminescence from nc-Si/a-SiNx:H quantum dot thin films prepared by ICP-CVD. Phys. Chem. Chem. Phys. 15 (2013) 3881–3888.
[9]
Q. Cheng, S. Xu, S. Huang and K. Ostrikov, Effective Control of Nanostructured Phases in Rapid, Room-Temperature Synthesis of Nanocrystalline Si in High-Density Plasmas. Crys. Growth Design. 9 (2009) 2863–2867.
[10]
M.B. Sahana, C. Sudakar, A. Dixit, J.S. Thakur, R. Naik and V.M. Naik, Quantum confinement effects and band gap engineering of SnO2 nanocrystals in a MgO matrix, Acta Mater. 60 (2012) 1072–1078.
19
ACCEPTED MANUSCRIPT
[11]
T. Rajagopalan, X. Wang, B. Lahlouh, C. Ramkumar, P. Dutta and S. Gangopadhyay, Low temperature deposition of nanocrystalline silicon carbide films by plasma enhanced chemical vapor deposition and their structural and optical characterization. J. Appl. Phys. 94 (2003) 5252–5260.
[12]
S. Park, E. Cho, D. Song, G. Conibeer and M.A. Green, n-Type silicon quantum dots and p-type crystalline silicon heteroface solar cells. Sol. Energ. Mat. Sol. Cells 93 (2009) 684–690.
[13]
G. Scardera, T. Puzzer, I. Perez-Wurfl and G. Conibeer, The effects of annealing temperature on the photoluminescence from silicon nitride multilayer structures. J. Cryst. Growth 310 (2008) 3680–3684.
[14]
D. Song, E.C. Cho, G. Conibeer, Y. Huang, C. Flynn and M.A. Green, Structural characterization
of
annealed Si1−xCx/SiC multilayers
targeting
formation
of
Si
nanocrystals in a SiC matrix. J. Appl. Phys. 103 (2008) 083544. [15]
B.G. Lee, D. Hiller, J.W. Luo, O.E. Semonin, M.C. Beard, M. Zacharias and P. Stradins, Strained Interface Defects in Silicon Nanocrystals. Adv. Funct. Mater. 22 (2012) 3223– 3232.
[16]
Y. Rosenwaks, M.C. Hanna, D.H. Levi, D.M. Szmyd, R.K. Ahrenkiel and A.J. Nozik, Hot-carrier cooling in GaAs: Quantum wells versus bulk. Phys. Rev. B. 48 (1993) 14675– 14678.
[17]
G. Chen, Y. Guo, J. Yao, Z. Song and F. Zhang, Interface properties of a-Si:H/SiCx:H superlattice. Thin Solid Films 258 (1995) 132–136.
[18]
B. Abeles and T. Tiedje, Amorphous Semiconductor Superlattices. Phys. Rev. Lett. 51 (1983) 2003–2006.
[19]
G. Han, P. Du, J. Shou and D. Zhao, Effect of hydrogen radical exposure on the interfacial and bulk structure of a-Si:H-SiNx:H multilayer films. Thin Solid Films 334 (1998) 6–10.
[20]
K. Akatsuka, Y. Ebina, M. Muramatsu, T. Sato, H. Hester, D. Kumaresan, R.H. Schmehl, T. Sasaki and M. Haga, Photoelectrochemical Properties of Alternating Multilayer Films Composed of Titania Nanosheets and Zn Porphyrin. Langmuir 23 (2007) 6730–6736. 20
ACCEPTED MANUSCRIPT
[21]
Z. Li and W. Li, Y. Jiang, H. Cai, Y. Gong and J. He, Raman characterization of the structural evolution in amorphous and partially nanocrystalline hydrogenated silicon thin films prepared by PECVD, J. Raman Spectrosc. 42 (2011) 415–421.
[22]
S. Guha, G. Hendershot, D. Peebles, P. Steiner, F. Kozlowski and W. Lang, Photoluminescence and Raman studies of porous silicon in polymethyl methacrylate, Appl. Phys. Lett. 64 (1994) 613-615.
[23]
Y. He, Y. Wei, G. Zheng, M. Yu, M. Liu, An exploratory study of the conduction mechanism of hydrogenated nanocrystalline silicon films, J. Appl. Phys. 82 (1997) 34083413.
[24]
D. Raha and D. Das, Controlling the growth of nanocrystalline silicon by tuning negative substrate bias, Sol. Energ. Mater. Sol. Cells 95 (2011) 3181-3188.
[25]
V. Paillard, P. Puech, M.A. Laguna, R. Carles, B. Kohn and F. Huisken, Improved onephonon confinement model for an accurate size determination of silicon nanocrystals. J. Appl. Phys. 86 (1999) 1921–1924.
[26]
D. Das and D. Kar, Spectroscopic and microscopic studies of self-assembled nc-Si/a-SiC thin films grown by low pressure high density spontaneous plasma processing, Phys. Chem. Chem. Phys. 16 (2014) 25421-25431.
[27]
D. Das and D. Kar, Investigation of the vertical electrical transport in a-Si: H/nc-Si: H superlattice thin films, Phys. Chem. Chem. Phys. 17 (2015) 17063-17068.
[28]
D. Kar and D. Das, Opto-electronic properties of P-doped nc-Si-QD/a-SiC: H thin films as foundation layer for all-Si solar cells in superstrate configuration, J. Appl. Phys. 120 (2016) 025102.
[29]
R.B. Bjorklund, S. Zangooie and H. Arwin, Adsorption of surfactants in porous silicon films. Langmuir 13 (1997) 1440–1445.
[30]
D. Das, Micro-Raman and ultraviolet ellipsometry studies on μc-Si:H films prepared by H2 dilution to the Ar-assisted SiH4 plasma in radio frequency glow discharge. J. Appl. Phys. 93 (2003) 2528–2535.
21
ACCEPTED MANUSCRIPT
[31]
I. Stenger, B. Gallas, L. Siozade, C.C. Kao, S. Chenot, S. Fisson, G. Vuye and J. Rivory, Evolution of the optical properties of Si nanoparticles embedded in SiO2 as function of annealing conditions. J. Appl. Phys. 10 (2008) 114303.
[32]
A. Fontcuberta i Morral, P. Roca i Cabarrocas, and C. Clerc, Structure and hydrogen content of polymorphous silicon thin films studied by spectroscopic ellipsometry and nuclear measurements. Phys. Rev. B 69 (2004) 125307.
[33]
D. Das and D. Kar, Self-assembled nc-Si-QD/a-SiC thin films from planar ICP-CVD plasma without H2-dilution: a combination of wide optical gap, high conductivity and preferred < 220> crystallographic orientation, uniquely appropriate for nc-Si solar cells, RSC Adv. 6 (2016) 3860-3869.
[34]
G. Lucovsky, J. Yang, S.S. Chao, J.E. Tyler and W. Czubatyj, Oxygen-bonding environments in glow-discharge-deposited amorphous silicon-hydrogen alloy films, Phys. Rev. B 28 (1983) 3225-3233.
[35]
K. Bhattacharya
and D. Das, Effect of deposition temperature on the growth of
nanocrystalline silicon network from helium diluted silane plasma, J. Phys. D: Appl. Phys. 41 (2008) 155420. [36]
A. Samanta and D. Das, Studies on the structural properties of SiO:H films prepared from (SiH4+CO2+He) plasma in RF-PECVD, Sol. Energ. Mater. Sol. Cells 93 (2009) 588-596.
[37]
B. Rezgui, A. Sibai, T. Nychyporuk, M. Lemiti, G. Bremond, D. Maestre and O. Palais, Appl. Phys. Lett. 96 (2010) 183105.
[38]
M. Zacharias, J. Heitmann, R. Scholz, U. Kahler, M. Schmidt and J. Blasing, Appl. Phys. Lett. 80 (2002) 661.
[39]
P. Shen, N. Uesawa, S. Inasawa and Y. Yamaguchi, J. Mater. Chem. 20 (2010) 1669.
[40]
D. Das and B. Sain, Electrical transport phenomena prevailing in undoped nc-Si/a-SiNx:H thin films prepared by ICP-CVD, J. Appl. Phys. 114 (2013) 073708.
[41]
M Jana, D Das, S.T Kshirsagar and A.K Barua, Control of Crystallization at Low Thickness in µc-Si: H Films Using Layer-by-Layer Growth Scheme, Jpn. J. Appl. Phys. 38 (1999) L1087. 22
ACCEPTED MANUSCRIPT
[42]
S. Janz, P. Löper and M. Schnabel; Silicon nanocrystals produced by solid phase crystallisation of superlattices for photovoltaic applications, Mat. Sci. and Engg. B 178 (2013) 542–550.
[43]
J. Ma, J. Ni, J. Zhang, Q. Liu, G. Hou, X. Chen, X. Zhang and Y. Zhao; In situ grown size-controlled silicon nanocrystals: A p type nanocrystalline-Si:H/a-SiCx:H superlattice (p-nc-Si:H/a-SiCx:H) approach, Sol. Energy Mater. & Sol. Cells 123 (2014) 228–232.
[44]
S. Yamada, Y. Kurokawa, S. Miyajima1 and M. Konagai; Silicon quantum dot superlattice solar cell structure including silicon nanocrystals in a photogeneration layer, Nanoscale Research Letters 9 (2014) 246.
23
ACCEPTED MANUSCRIPT
Highlights: # a-Si:H/nc-Si:H thin film superlattice structures grown at ~180 oC by single step plasma process # lower defects in superlattice structures with lower thickness (tnc) of nc-Si:H active layer # barrier layer incubates Si-ncs facilitating growth of active layer with adequate Si-ncs at tnc ≥3 nm # sharp control in active layer growth confines Si-nc size within dimension of tnc # easy yet skillful method with huge technological impact of Si-ncs in thin superlattice structures