Nanocrystalline thin film silicon solar cells: A deeper

1 downloads 0 Views 1MB Size Report
cells was determined in short-circuit condition in the wavelength range of 360 to 1100 nm ..... i1 (50 nm) + i2 (1950 nm). 93. 5.98. 63.2 .... 3093–3095. [18] J. Deng, C.R. Wronski, Carrier recombination and differential diode quality factors in.

Thin Solid Films 591 (2015) 25–31

Contents lists available at ScienceDirect

Thin Solid Films journal homepage: www.elsevier.com/locate/tsf

Nanocrystalline thin film silicon solar cells: A deeper look into p/i interface formation Andriy Lyubchyk a, Sergej Alexandrovich Filonovich a, Tiago Mateus a, Manuel João Mendes a, António Vicente a, Joaquim Pratas Leitão b, Bruno Poças Falcão b, Elvira Fortunato a, Hugo Águas a,⁎, Rodrigo Martins a,⁎ a b

CENIMAT/I3N, Departamento de Ciencia dos Materiais, Faculdade de Ciencias e Tecnologia, Universidade Nova de Lisboa, CEMOP-UNINOVA, 2829-516 Caparica, Portugal Departamento de Fisica/I3N, Universidade de Aveiro, 3810-193 Aveiro, Portugal

a r t i c l e

i n f o

Article history: Received 7 May 2015 Received in revised form 22 July 2015 Accepted 11 August 2015 Available online 13 August 2015 Keywords: Thin film solar cells Nanocrystalline silicon p/i interface Ion bombardment Buffer layer

a b s t r a c t The p/i interface plays a major role in the conversion efficiency of nanocrystalline silicon (nc-Si:H) solar cells. Under plasma-enhanced chemical vapor deposition (PECVD) of the intrinsic (i) nc-Si:H layer, ion bombardment can severely affect the underlying p-doped layer and degrade the solar cell performance. The core of the present work is to investigate the effect of light and heavy ion bombardment on the structural modifications of the p-layer during the p/i interface formation. The properties of the nc-Si:H materials deposited under distinct conditions are analyzed and correlated to the deposition rate and the resulting cell efficiency. To recreate the ion bombardment during the initial stages of the i-layer deposition on the p-layer, hydrogen plasma treatment was performed for 30 s (light ion bombardment), after which a flux of silane was introduced into the deposition chamber in order to initiate the heavy ion bombardment and growth of an ultra-thin (5 nm) i-layer. The structural changes of the p-type nc-Si:H layers were observed by spectroscopic ellipsometry. The obtained results confirm that detrimental structural modifications (e.g. partial amorphization of the sub-surface region and bulk) occur in the p-layer, caused by the ion bombardment. To minimize this effect, a protective buffer layer is investigated able to improve the performance of the solar cells fabricated under increased growth rate conditions. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The efficiency of thin film silicon solar cells, both single- and multijunction, has reached a plateau in previous decades [1] that needs to be surpassed in order for PV electricity to become a cost-competitive energy source. While the quality of the intrinsic and doped silicon thin films can hardly be improved further, new light trapping techniques [2–6], anti-reflective coatings [7] and better control over the cell interfaces [8] can still offer potential room for solar cell efficiency improvement. The standard method of hydrogenated nanocrystalline silicon (nc-Si:H) deposition is by radio-frequency plasma enhanced chemical vapor deposition (RF-PECVD) with a high hydrogen dilution of silane [9]. To decrease the processing time and costs of nc-Si:H solar cells, it is desired to increase as much as possible the deposition rate of the intrinsic layer used as absorber in tandem or single junction nc-Si:H solar cells. The most typical method to achieve this is by increasing the RF power and simultaneously use very high frequency (VHF) plasma

⁎ Corresponding authors. E-mail addresses: [email protected] (A. Lyubchyk), sfi[email protected] (S.A. Filonovich), [email protected] (T. Mateus), [email protected] (M.J. Mendes), [email protected] (A. Vicente), [email protected] (J.P. Leitão), [email protected] (B.P. Falcão), [email protected] (E. Fortunato), [email protected] (H. Águas), [email protected] (R. Martins).

http://dx.doi.org/10.1016/j.tsf.2015.08.016 0040-6090/© 2015 Elsevier B.V. All rights reserved.

processes [10], which usually lead to poorly performing p–i–n solar cells. During the initial stages of the intrinsic nc-Si:H deposition, ion bombardment can easily damage a 20–30 nm thick doped layer, decreasing its crystalline volume fraction and conductivity [11,12]. This leads to solar cells with a degraded performance due to the large increase of defects at the p/i interface region, which is reflected by the increase of the series resistance, the recombination rate and, consequently, the decrease of the fill factor (FF) and efficiency. For the purpose of improving the p/i interface of thin film Si p–i–n solar cells, some methods have been developed including the deposition of a buffer layer between the p-type and intrinsic layers [13]. It was reported that the insertion of a wide bandgap semiconductor layer, such as a-SiC:H, contributed to improve the blue response [14] and efficiency of a-Si:H solar cells [15,16]. Other groups used protocrystalline silicon (pc-Si:H) [17,18] or boron-doped microcrystalline silicon (p-μc-Si:H) [19] as buffer layers instead of a-SiC:H. Bugnon et al. explored SiOx as buffer layer in microcrystalline (μc-Si:H) and amorphous (a-Si:H) solar cells [20], finding that a SiOx buffer can limit boron cross-contamination and promote the nucleation of the i-layer; thus acting as buffer and seed layer simultaneously. Concerning nc-Si:H solar cells, it was shown that wide bandgap buffer layers can enhance the VOC and FF [21] due to a reduction of accumulated charges at the p/i interface, caused by the strengthened electric field at the interface [22,23]. These works employed an a-Si:H buffer layer;

26

A. Lyubchyk et al. / Thin Solid Films 591 (2015) 25–31

Fig. 1. Structure of the fabricated p–i–n nc-Si:H solar cells. a) Schematic representation of the solar cell structure after dry etching; b) SEM image showing a FIB cross section of one of the investigated solar cells. Colors are given as a guide for the eye.

however, nc-Si:H material as buffer has recently attracted a lot of attention since it can provide better optical transmission and electrical conductivity [24,25]. In this paper, we report an investigation of the effects of light and heavy ion bombardment on the p-layer, which occurs during the initial growth stages of the intrinsic layer deposition processed under VHF PECVD conditions. Based on this study, the beneficial effect of implementing an intrinsic nc-Si:H buffer layer at the p/i interface is then demonstrated for nc-Si:H p–i–n solar cells fabricated under increased growth rate conditions of the intrinsic layer. 2. Experimental details Single junction nc-Si:H solar cells with p–i–n configuration were deposited on 3-mm-thick 10 × 10 cm2 glass substrates, covered with 700 nm of SnO2:F (Fig. 1), using a spider-configuration Elettrorava PECVD system. The 30-nm-thick p-type and 50-nm-thick n-type layers were deposited in separated RF-PECVD chambers, at a substrate temperature of 160 °C. All deposition parameters were previously optimized [10,26,27]. The intrinsic absorber and/or buffer layers were directly deposited on the p-layer at 160 °C in another VHF-PECVD chamber. To study the effect of the growth rate of the intrinsic nc-Si:H on the solar cells performance, various deposition conditions were used, labeled i1, i2 and i3 in Table 1. All solar cells have a total i-layer thickness of about 2 μm, including the thickness of the buffer layer. The hydrogen dilution, DH, defined as DH = (FH2 / (FH2 + FSiH4)) × 100%, was varied from 96.10% to 97.01%. To promote the crystallinity of the i-layer, hydrogen plasma treatment (HPT) is applied to the as-grown p-layer while kept under vacuum. The flow of hydrogen (FH2) and the deposition pressure were kept constant during the deposition and HPT. Back contacts of aluminium zinc oxide (AZO 75 nm) and Al (200 nm) were deposited by PVD methods using a mask with holes of 5 mm diameter, which define cell areas of 19.6 mm2. After deposition of the top contacts, reactive ion etching (Trion Phantom III RIE ATCH system) with SF6 plasma was applied to remove the remaining silicon around the top contacts, to avoid lateral side effects [28], resulting in the devices shown in Fig. 1. Finally, the cells were annealed at 160 °C under vacuum for 120 min, to reduce the dangling bonds impact caused by inhomogeneous microstructure [29]. Table 1 The deposition parameters of the nc-Si:H intrinsic layers by PECVD are the applied power (P), power density (Pd), frequency (f), inter-electrode distance (dE), substrate temperature (Tsub), working pressure (pw), gas fluxes (FSiH4 andFH2) and hydrogen dilution (DH). Deposition conditions

P, W

Pd, mW/cm2

f, MHz

dE, mm

Tsub, °C

pw, Torr

FSiH4, sccm

FH2, sccm

DH, %

i1 i2 i3

20 40 58

139 278 403

75 75 75

13 8 8

160 160 160

1.0 2.1 2.0

10.0 15.5 15.4

246 500 500

96.10 96.99 97.01

To evaluate the influence of the ion bombardment on the p-layer, nanocrystalline p-type films were deposited on 1-mm thick glass substrates as described in Ref. [30]. Then, in order to simulate the initial stages of growth of the intrinsic nc-Si:H on top of the p-type layer, HPT was performed at different conditions (f = 75 MHz; 1.0 b pw b 2.13 Torr; 139 b Pd b 433 mW/cm2) for 30 s (light ion bombardment), followed by the introduction of a silane flux (according to the i1, i2 or i3 conditions of Table 1) into the deposition chamber to initiate the film growth (heavy ion bombardment). The deposition time was adjusted to grow only an ultra-thin (5 nm thick) i-layer. The structural characterization of the p and i-type films was performed using Raman spectroscopy and spectroscopic ellipsometry (SE). Raman scattering experiments were performed at room temperature, in backscattering configuration, using a LabRam Horiba HR800UV spectrometer and a 532 nm laser. SE was measured using a Horiba Jobin Yvon UVISEL ellipsometer, with a fixed 70° incidence angle in the 1.5–6.5 eV range, to determine the crystalline volume fraction, XC, of the films and its variation with thickness. The electrical properties of the films were studied via temperature dependent dark conductivity, from which the room-temperature conductivity, σd, and activation energy, Ea, were calculated. The linearity of the I (V) dependence was confirmed before each conductivity measurement. Low voltages (0.1–1 V) were used to reduce high-field effects, such as field-enhanced hopping transport. Chromium contacts (180 nm thick, 4 mm long and 1 mm apart) were deposited before the Si films by thermal evaporation. In order to determine the photosensitivity, σph/σd, of the intrinsic layers, the photoconductivity, σph, was measured under illumination (~100 mW/cm2 at 300 K). The dark conductivity measurements were performed with a Deep Level Transient Spectroscopy (DLTS) system from Polaron equipped with a cryostat ranging from 200 K to 400 K. The solar cells were characterized by current–voltage measurements at room temperature under AM1.5 illumination conditions (SPIRE SPISun Simulator 240A). The external quantum efficiency (EQE) of the cells was determined in short-circuit condition in the wavelength range of 360 to 1100 nm using a home-made set up [3]. SEM observations were carried out using a Carl Zeiss AURIGA CrossBeam (FIB-SEM) workstation.

Table 2 Structural (crystalline volume fraction, XC(RS)) and electrical (photosensitivity, σph/σd, and activation energy, Ea) properties of the intrinsic layers deposited with the conditions of Table 1, resulting in different growth rates (rd). Deposition conditions

Rd, Ǻ/s

XC(RS), %

σph/σd

Ea, meV

i1 i2 i3

2.4 3.5 4.5

51 ± 2.5 57 ± 2.8 60 ± 2.9

8 × 103 7 × 103 2 × 102

571 554 528

A. Lyubchyk et al. / Thin Solid Films 591 (2015) 25–31

27

crystalline volume fraction, XC(RS), was determined using the procedure reported by Yue et al. [31]: X C ðRSÞ ¼

Fig. 2. SE reference data of the imaginary part of the dielectric function (εi) for p-doped amorphous (a-Si), hydrogenated amorphous (a-SiH) (used for intrinsic layers modeling) and nanocrystalline (nc-Si) silicon.

3. Results and discussions Firstly, the structural and electrical properties of intrinsic nc-Si:H films, deposited with the conditions presented in Table 1, were studied. The corresponding properties are presented in Table 2. Raman spectroscopy was employed to estimate the crystalline volume fraction of the i-layers (Fig. 3a). The power of the incident beam was fixed at 1.7 mW to avoid recrystallization of the films. The

AC þ AGB  100%; AC þ AGB þ γAA

ð1Þ

where AC, AA and AGB are the integrated intensities of the nanocrystalline, amorphous and grain-boundaries peaks (dotted lines in Fig. 3a), respectively, obtained by deconvolution of the Raman spectra. γ is the ratio of the cross sections for the amorphous and crystalline phases, which for mixed phase silicon is taken to be γ = 0.8 [32,33]. The estimation of the γ coefficient is rather complex and, so far, no consensus exists in the literature on its value. Nevertheless, although the uncertainty in γ may lead to errors in the absolute crystalline volume fractions, it does not affect the comparison of crystal volume fractions in samples grown under similar deposition conditions. A 5% uncertainty was considered in the fitting procedure to account for possible distortions that may affect lineshapes and positions of the deconvoluted peaks (Fig. 3a), related to the amorphous and crystalline material contribution. As a result, the calculated values of the crystalline volume fraction shown in Table 2 are affected by an uncertainty determined by applying standard error propagation formulas to Eq. (1). The growth rate increases from 2.4 for i1 to 4.5 Å/s for i3. Despite this increase of the growth rate, the films maintain approximately the same good structural and electrical properties, typical of intrinsic nc-Si:H films. For a more detailed structural characterization, SE measurements were conducted. Modeling was performed by fitting the data considering material references [34,35] (Fig. 2), using a three-layer model to determine the crystalline fraction, XC, of the films and its variation with the

Fig. 3. (a) Raman spectra of the intrinsic films deposited with the conditions of Table 1. XC(RS) is the crystalline volume fraction calculated with (1). The dash-dotted peaks correspond to the three deconvoluted material contributions: amorphous (green) 480 cm−1, grain boundary (red) 490 cm−1 and nanocrystalline (blue) grain components 520 cm−1 (example given for the intrinsic layer deposited under condition i2); (b) SE data of the imaginary part of the pseudo dielectric function bεiN for the three films; (c) multilayer modeling of the SE data; XC, XA, XV represent crystalline, amorphous and void volume fractions, respectively. δ are the uncertainties of the volume fraction distribution for each individual layers. The values of the thickness of the individual layers are indicated on their sidewalls.

28

A. Lyubchyk et al. / Thin Solid Films 591 (2015) 25–31

Fig. 4. Structural composition and thickness of the layers, extracted from the multilayer modeling of SE measurements, after light and heavy ion bombardment of the p-type films. XC, XA, XV represent crystalline, amorphous and void volume fractions, respectively. δ are the uncertainties of the volume fraction distribution for each individual layer. The values of the thickness of the individual layers are indicated on their sidewalls.

thickness. This model consisted in dividing the film into three layers: bulk, subsurface and surface layers [30]. In these layers, a variable percentage mixture (i.e. volume fraction) of two material references (nc-Si and a-Si, corresponding to the optical properties in Fig. 2) is used. The damage caused by plasma is accounted for by another parameter in the layers: the voids volume fraction [36]. With this method, the film structure was modeled as a function of its thickness, using three independent variables per layer, making a total of 9 variables. The best results were obtained with a variable Bruggeman Effective Medium Approximation (BEMA) mixture of the references for nc-Si, a-Si and voids. It should be noted, however, that in SE analysis the referenced data used for modeling

may correspond to materials that can slightly differ in structure (grain size, etc.) from our actual films, which can lead to inaccuracies in the quantification of the crystalline fractions. Nevertheless, since all samples were modeled with the same SE conditions, the relative trends discussed here are anyway valid. Moreover, the figures of merit based on the minimum χ2 error obtained with the fitting were in all cases inferior to 0.5, which are indicative of good fits [37]. The fitting results are presented in Fig. 3b,c. Each fitting (Fig. 3c) indicates the presence of an incubation layer at the substrate-film interface, with a typical thickness of 19.9–25.1 ± 1.3 nm, followed by a dense nc-Si:H bulk region. The top layer (7.8–10.3 ± 0.5 nm thick)

Fig. 5. (a) J/V measurements under AM1.5 illumination of the nc-Si:H solar cells (SC) with 2 μm-thick intrinsic layers deposited under the conditions i1 (black), i2 (red) and i3 (green) of Table 1; (b) corresponding external quantum efficiency (EQE) spectra of the solar cells.

A. Lyubchyk et al. / Thin Solid Films 591 (2015) 25–31

29

Table 3 Main characteristics of the p–i–n solar cells with 2 μm-thick intrinsic layers deposited under different conditions. Intrinsic layer

i1(2000 nm) i2( 2000 nm) i1 (50 nm) + i2 (1950 nm) i2 (50 nm) + i1 (1950 nm) i3 (2000 nm) i1 (50 nm) + i3 (1950 nm)

Deposition

η

FF

JSC

JSC (EQE)

VOC

RS

time, min

%

%

mA/cm2

mA/cm2

V

Ω cm2

139 111 93 133 75 76

6.59 2.79 5.98 2.1 2.65 4.66

66.3 44.4 63.2 27.9 47.4 53.7

20.95 15.7 21.6 14.6 12.5 19.4

20.3 15.2 22 14.2 12.5 18.4

0.48 0.4 0.42 0.4 0.45 0.45

3.2 11.5 3.3 17.1 10.1 6.2

with high void fraction represents the surface roughness of the films. The growth of such tri-layer structure is common in nc-Si:H films deposited under high hydrogen dilution conditions, with the presence of a small amorphous phase [38,39]. It is worth noting that the crystalline fraction of the layers in our study increases with the growth rate, but results in a decrease of photosensitivity (σph/σd) as shown in Table 2. The next step was to investigate the p/i interface. According to previous studies [30], the p-layer can be damaged during the initial stages of the intrinsic layer deposition due to light and heavy ion bombardment. Consequently, the higher the growth rate of the intrinsic nc-Si:H material the more severely can the previously deposited p-layer be affected. In order to verify this, the deposited p-layers, without any amorphous incubation layer [30], were exposed for 30 s to HPT (light ion bombardment) and to the growth of a 5-nm i-layer (heavy bombardment) under the different conditions of Table 1. Fig. 4 shows the resulting p-type films structure evolution, calculated from the multilayer modeling of spectra extracted from spectroscopic ellipsometry measurements. It can be seen that the entire structure of the p-type film is significantly modified + by the light (typically H+, H+ 3 ) and heavy (such as Sin N 2Hm [10]) ion bombardment, as the penetration depth of the ions into the nc-Si:H material can reach about 20 nm [9]. This is evidenced by the decrease of the crystalline volume fraction through the entire film thickness, observed for all the deposition conditions (Fig. 4). The XC decreases as the p-type nc-Si:H suffers more severe bombardment, e.g. when higher plasma excitation power is applied (up to 403 mW/cm2). The largest decline of the crystalline volume fraction in the bulk of the p-type film is observed for the i3 and i2 (ΔXC are 14.3 ± 2.1% and 13 ± 2.3%, respectively) deposition conditions, while the smallest decrease of 6.4 ± 2.0% is observed for the i1 conditions. Therefore, from the measured structural properties, the p-layer is less modified under the treatments with i1 deposition conditions, while during the treatments corresponding to the i2 and i3 conditions it clearly suffers ion-induced damage. It should be noted that, in each layer, there is a gradient of damage caused by the plasma ions along the layer depth; so, the values resulting from the SE model concerning the bombardment effect represent only averaged changes along the layers thickness.

The observed modifications of the p-layer after the i2 and i3 depositions should have a detrimental effect when fabricating the full solar cells structure, as the induced defects/amorphization are expected to increase the recombination rate of the photo-generated carriers. To confirm this, solar cells with p–i–n configuration (Fig. 1) were produced with the intrinsic nc-Si:H layers grown under the i1, i2 and i3 deposition conditions. All the other layers in the solar cells were identical. The J/V characteristics of the devices are shown in Fig. 5a. The cell with the intrinsic layer i1 achieved the highest efficiency (η) of 6.6% (Table 3) and the cells with intrinsic layer deposited at higher growth rate (i2 and i3 with the same thickness) yielded significantly lower performance (Fig. 5a). The changes of the open-circuit voltage (VOC decreases up to 17%) and series resistance (RS increases up to 3.6 times), observed for the cells with i2 and i3 relative to i1, indicate that a degradation of the p/i interface region and/or of the p-layer itself has occurred as expected. This is in agreement with a decrease of the built-in potential at the p/i interface as a result of defects that cause higher carrier recombination and, consequently, also reduction of JSC and FF [40,41]. The external quantum efficiency (EQE) measurements shown in Fig. 5b support this observation. There is a decrease in the EQE in the entire spectral range, which indicates that carrier recombination is occurring mainly at the defective p/i interface [42,43]. The short circuit current density calculated from the EQE data [ JSC(EQE)] is in good agreement with values obtained from J/V measurements ( JSC), with an error below 5% (see Table 3) attributed to small spectral mismatches between the light source of the sun simulator and the AM1.5 solar spectrum used to determine JSC(EQE). Moreover, the SEM cross sections presented in Fig. 6 show that there is no visible difference in the microstructure of the cells bulk grown at different deposition conditions. The SEMs of i1, i2 and i3 cells show dense and compact silicon films devoid of visible defects. This supports that the misbehavior observed for the i2 and i3 solar cells should be originated from electronic defects created at the p/i interface and damage of the p-layer. From the set of results achieved we notice that the decrease of solar cells performance is related with the decrease of crystalline

Fig. 6. Comparison of the solar cells microstructure obtained by SEM cross section measurements. i1, i2 and i3 correspond to solar cells with intrinsic layers grown under the distinct deposition conditions of Table 1.

30

A. Lyubchyk et al. / Thin Solid Films 591 (2015) 25–31

Fig. 7. J/V characteristics (a,c) and EQE (b,d) of the solar cells with buffer and absorber layers deposited with distinct combinations of conditions, as shown in Table 3. The red and green lines correspond to the solar cells of Fig. 5 with i2 and i3 intrinsic layers, respectively, repeated here for comparison. The absolute efficiency difference (Δη) between these cells and those with a 50 nm-thick i1 buffer layer are indicated in the J/V plots.

volume fraction in the bulk of the p-film (XC) upon the i-layers deposition. Hence, as the XC decline (ΔXC) increases we observe an almost monotonically decrease of the solar cell efficiency: i1 ΔXC = 6.4 ± 2%, η = 6.6%; i2 ΔXC = 13 ± 2.3%, η = 2.8% and i3 ΔXC = 14.3 ± 2.1%, η = 2.7%. According to the results discussed above, the i1 intrinsic layer deposition conditions that did not modify significantly the p-layer and its p/i interface, are good conditions to be used for a buffer layer that can protect the p-layer from the light and heavy ion bombardment promoted during subsequent depositions performed at higher growth rates, as occurring with the i2 and i3 conditions. In order to verify such hypothesis, additional solar cells were fabricated with an i2 or i3 intrinsic layer, but previously inserting a 50 nm-thick buffer layer from the i1 conditions between the p and i2/3 intrinsic layers. The results of the solar cells performance with intrinsic layers formed by distinct combinations of buffer/absorber deposition conditions are shown in Fig. 7. The inclusion of the i1-buffer before the deposition of the i2 absorber layer leads to a pronounced increase of the current density, fill factor (FF), open circuit voltage (VOC) and, consequently, efficiency (see Table 3) relative to the cell without buffer layer. In the same way, the i1-buffer applied to the solar cell with the absorber layer i3 (Fig. 7c,d) leads to a significant decrease of RS and improved current collection efficiency (Table 3). To confirm that the observed improvements were caused by the softlydeposited buffer layer, a solar cell with the “reverse configuration” of the absorber and buffer layers was produced. This cell consisted of an i1 absorber layer deposited on a 50-nm-thick buffer layer from the hard i2 conditions. The results are presented in Table 3 and Fig. 7a,b, showing that the overall performance of such cell becomes similar to the cell of i2 intrinsic layer without any buffer.

4. Conclusions We have studied the influence of various deposition conditions of intrinsic nc-Si:H on the structural properties of the underlying p-type nc-Si:H film, and their effects on the solar cells performance. It was found that, at the initial stages of nc-Si:H growth over the p-type layer, under “hard deposition conditions” (fast growth rate) the p-type films suffer morphological modifications that affect their opto-electronic properties. On the other hand, the “softer deposition conditions” (low growth rate) indicated as i1 do not affect significantly the p-type films. Furthermore, the damage of the p-layer had a strong influence on the p–i–n solar cell's performance. Namely, it was observed the increase of RS (from 4.0 to 14.2 Ω cm2) and reduction of FF (from 63 to 39%). It was established that such detrimental effects, mainly induced by the partial amorphization of the p-type layer, were caused by the ion bombardment. It was observed that a greater decline of the crystalline volume fraction in the bulk of the p-film (XC) leads to a lower performance of the solar cell. Therefore, the XC value of the p-layer is a crucial quantity for the performance of nc-Si:H cells. It was shown the beneficial use of an intrinsic buffer layer, deposited at “softer deposition conditions”, on the solar cells performance. The application of a 50-nm-thick nc-Si:H buffer layer, deposited with i1 conditions, led to a significant decrease of RS (5.3 Ω cm) and, consequently, to the increase of fill factor (up to 63%) and efficiency (by N 50%). Although the efficiencies of the cells with i1-buffer plus i2 or i3 absorber layers are lower (~6 or 4.7%, respectively) than that with only i1 (6.6%), the i-layer deposition time is pronouncedly reduced from ~ 139 min (for the cell with only i1) to 93 or 76 min for the cells with i1-buffer + i2 or i3 layers, respectively. As such, this indicates that the use of softly-

A. Lyubchyk et al. / Thin Solid Films 591 (2015) 25–31

deposited nc-Si:H buffer layers can potentially allow major time savings and cost reductions in the solar cells production. The efficiency values attained in this work are lower than the present record (10.5% [44]) for this type of devices, since a basic deposition recipe was chosen using standard substrates and with no advanced light trapping methods, in order to clearly show relative changes due to bombardment effects and simplify the deposition procedure. The development of further optimized buffer layers or an applied power profiling technique can be key solutions for controlling the nanocrystalline evolution and maximizing the cell performance/time ratio of the PECVD process. Acknowledgments The authors acknowledge for the financial support, under project “Si-QuaDot” PTDC/CTM-ENE/2514/2012 from the Portuguese Fundação para a Ciência e a Tecnologia (FCT-MEC), Strategic Project UID/CTM/ 50025/2013 and RECI/FIS-NAN/0183/2012 (COMPETE: FCOMP-010124-FEDER-027494). S. A. Filonovich acknowledges the support from FCT-MEC through the post-doc scholarship SFRH/BPD/91281/2012. M. J. Mendes acknowledges funding by the EU Marie Curie Action FP7PEOPLE-2013-IEF through the DIELECTRIC PV project (Grant No. 629370). A. Vicente acknowledges the support from FCT and MITPortugal program through the scholarship SFRH/BD/33978/2009. References [1] http://www.nrel.gov/ncpv/. [2] M.J. Mendes, S. Morawiec, F. Simone, F. Priolo, I. Crupia, Colloidal plasmonic back reflectors for light trapping in solar cells, Nanoscale 6 (2014) 4796. [3] S. Morawiec, M.J. Mendes, S.A. Filonovich, T. Mateus, S. Mirabella, H. Águas, I. Ferreira, F. Simone, E. Fortunato, R. Martins, F. Priolo, I. Crupi, Broadband photocurrent enhancement in a-Si:H solar cells with plasmonic back reflectors, Opt. Express 22 (S4) (2014) A1059. [4] R.S.A. Sesuraj, T.L. Temple, D.M. Bagnall, Optical characterization of a spectrally tunable plasmonic reflector for application in thin-film silicon solar cells, Sol. Energy Mater. Sol. Cells 111 (2013) 23–30. [5] S. Geißendorfer, M. Vehse, T. Voss, J.-P. Richters, B. Hanke, K. von Maydell, C. Agert, Integration of n-doped ZnO nanorod structures as novel light-trapping concept in amorphous thin film silicon solar cells, Sol. Energy Mater. Sol. Cells 111 (2013) 153–159. [6] M.J. Mendes, S. Morawiec, T. Mateus, A. Lyubchyk, H. Águas, I. Ferreira, E. Fortunato, R. Martins, F. Priolo, I. Crupi, Broadband light trapping in thin film solar cells with self-organized plasmonic nano-colloids, Nanotechnology 26 (2015) 135202. [7] H.K. Raut, A.S. Nair, S.S. Dinachali, V.A. Ganesh, T.M. Walsh, S. Ramakrishna, Porous SiO2 anti-reflective coatings on large-area substrates by electrospinning and their application to solar modules, Sol. Energy Mater. Sol. Cells 111 (2013) 9–15. [8] R. Martins, L. Raniero, L. Pereira, D. Costa, H. Águas, S. Pereira, L. Silva, A. Gonçalves, I. Ferreira, E. Fortunato, Nanostructured silicon and its application to solar cells, position sensors and thin film transistors, Philos. Mag. 89 (2009) 2699. [9] R.A. Street, Hydrogen chemical potential and structure of a-Si:H, Phys. Rev. B 43 (1991) 2454. [10] H. Águas, V. Silva, E. Fortunato, S. Lebib, P. Roca i Cabarrocas, I. Ferreira, L. Guimaraes, R. Martins, Large area deposition of polymorphous silicon by plasma enhanced chemical vapor deposition at 27.12 MHz and 13.56 MHz, Japan, J. Appl. Phys. 42 (2003) 4935. [11] B. Kalache, A.I. Kosarev, R. Vanderhaghen, P. Roca i Cabarrocas, Ion bombardment effects on microcrystalline silicon growth mechanisms and on the film properties, J. Appl. Phys. 93 (2003) 1262. [12] S. Nunomura, M. Kondo, Positive ion polymerization in hydrogen diluted silane plasmas, Appl. Phys. Lett. 93 (2008) 231502. [13] L. Raniero, S. Zhang, H. Águas, I. Ferreira, R. Igreja, E. Fortunato, R. Martins, Role of buffer layer on the performances of amorphous silicon solar cells with incorporated nanoparticles produced by plasma enhanced chemical vapor deposition at 27.12 MHz, Thin Solid Films 487 (2005) 170. [14] W.Y. Kim, H. Tasaki, M. Konagai, K. Takahashi, Use of a carbon‐alloyed graded‐band‐ gap layer at the p/i interface to improve the photocharacteristics of amorphous silicon alloyed p‐i‐n solar cells prepared by photochemical vapor deposition, J. Appl. Phys. 61 (1987) 3071. [15] B. Rech, C. Beneking, H. Wagner, Improvement in stabilized efficiency of a-Si:H solar cells through optimized p/i-interface layers, Sol. Energy Mater. Sol. Cells 41/42 (1996) 475–483. [16] S. Ogawa, M. Okabe, T. Itoh, N. Yoshida, S. Nonomura, Amorphous Si1− xCx:H films prepared by hot-wire CVD using SiH3CH3 and SiH4 mixture gas and its application to window layer for silicon thin film solar cells, Thin Solid Films 516 (2008) 758–760.

31

[17] J.M. Pearce, R.J. Koval, A.S. Ferlauto, R.W. Collins, C.R. Wronski, J. Yang, S. Guha, Dependence of open-circuit voltage in hydrogenated protocrystalline silicon solar cells on carrier recombination in p/i interface and bulk regions, Appl. Phys. Lett. 77 (2000) 3093–3095. [18] J. Deng, C.R. Wronski, Carrier recombination and differential diode quality factors in the dark forward bias current–voltage characteristics of a‐Si:H solar cells, J. Appl. Phys. 98 (2005) 024509. [19] C.H. Lee, J.W. Jeon, K.S. Lim, Band-gap profiling in amorphous silicon–germanium solar cells, J. Appl. Phys. 87 (2000) 8778. [20] G. Bugnon, G. Parascandolo, S. Hänni, M. Stuckelberger, M. Charrière, M. Despeisse, F. Meillaud, C. Ballif, Silicon oxide buffer layer at the p–i interface in amorphous and microcrystalline silicon solar cells, Sol. Energy Mater. Sol. Cells 120 (2014) 143–150. [21] G. Yue, B. Yan, C. Teplin, J. Yang, S. Guha, Optimization and characterization of i/p buffer layer in hydrogenated nanocrystalline silicon solar cells, J. Non-Cryst. Solids 354 (2008) 2440–2444. [22] J.W. Lim, D.J. Lee, S.J. Yun, Semi-transparent amorphous silicon solar cells using a thin p-Si layer and a buffer layer, ECS Solid State Lett. 2 (2013) Q47–Q49. [23] J. Ramanujam, A. Verma, Photovoltaic properties of a-Si:H films grown by plasma enhanced chemical vapor deposition, a review, Mater. Express 2 (2012) 177. [24] H. Águas, S.K. Ram, A. Araújo, D. Gaspar, A. Vicente, S.A. Filonovich, E. Fortunato, R. Martins, I. Ferreira, Silicon thin film solar cells on commercial tiles, Energy Environ. Sci. 4 (2011) 4620–4632. [25] S. Liu, X. Zeng, W. Peng, H. Xiao, W. Yao, X. Xie, C. Wang, Z. Wang, Improvement of amorphous silicon n–i–p solar cells by incorporating double-layer hydrogenated nanocrystalline silicon structure, J. Non-Cryst. Solids 357 (2011) 121–125. [26] S.A. Filonovich, H. Águas, I. Bernacka-Wojcik, C. Gaspar, M. Vilarigues, L.B. Silva, E. Fortunato, R. Martins, Highly conductive p-type nanocrystalline silicon films deposited by RF-PECVD using silane and trimethylboron mixtures at high pressure, Vacuum 83 (2009) 1253. [27] H. Águas, S.A. Filonovich, I. Bernacka-Wojcik, E. Fortunato, R. Martins, Role of trimethylboron to silane ratio on the properties of p-type nanocrystalline silicon thin film deposited by radio frequency plasma enhanced chemical vapour deposition, J. Nanosci. Nanotechnol. 10 (2010) 2547. [28] R. Martins, E. Fortunato, Dark current‐voltage characteristics of transverse asymmetric hydrogenated amorphous silicon diodes, J. Appl. Phys. 78 (1995) 3481. [29] W. Beyer, D. Lennartz, P. Prunici, H. Stiebig, Annealing effects of microstructure in thin-film silicon solar cell materials measured by effusion of implanted rare gas atoms, MRS Proc. (2011) 1321. [30] S.A. Filonovich, H. Águas, T. Busani, A. Vicente, A. Araújo, D. Gaspar, M. Vilarigues, J. Leitão, E. Fortunato, R. Martins, Hydrogen plasma treatment of very thin p-type nanocrystalline Si films grown by RF-PECVD in the presence of B(CH3)3, Sci. Technol. Adv. Mater. 13 (2012) 045004. [31] G. Yue, J.D. Lorentzen, J. Lin, D. Han, Q. Wang, Photoluminescence and Raman studies in thin-film materials: transition from amorphous to microcrystalline silicon, Appl. Phys. Lett. 75 (1999) 492. [32] E. Bustarret, M.A. Hachicha, M. Brunel, Experimental determination of the nanocrystalline volume fraction in silicon thin films from Raman spectroscopy, Appl. Phys. Lett. 52 (1988) 1675. [33] A.T. Voutsas, M.K. Hatalis, J. Boyce, A. Chiang, Raman spectroscopy of amorphous and microcrystalline silicon films deposited by low‐pressure chemical vapor deposition, J. Appl. Phys. 78 (12) (1995) 6999. [34] Reference spectra are supplied by Jobin Yvon software. Origin of the references: E.D. Palik (Ed.), Handbook of optical constants of solids, Elsevier, 1998; D.E. Aspnes, A.A. Studna, Phys. Rev. B 27 (1983) 985. [35] G.E. Jellison Jr., M.F. Chisholm, S.M. Gorbatkin, Optical functions of chemical vapor deposited thin‐film silicon determined by spectroscopic ellipsometry, Appl. Phys. Lett. 62 (1993) 3348. [36] T. Miyazaki, S. Adachi, Spectroscopic ellipsometry study of si surfaces modified by low-energy Ar+ ion irradiation, Jpn. J. Appl. Phys. 32 (1993) 4941. [37] G.E. Jellison Jr., Spectroscopic ellipsometry data analysis: measured versus calculated quantities, Thin Solid Films 313–314 (1998) 33–39. [38] C.Y. Lin, Y.K. Fang, S.F. Chen, C.S. Lin, T.H. Chou, S.B. Hwang, J.S. Hwang, K.I. Lin, Preferential coalescence of nanocrystalline silicon on different film substrates, J. Non-Cryst. Solids 352 (2006) 44. [39] S. Koynov, S. Grebner, P. Radojkovic, E. Hartmann, R. Schwarz, L. Vasilev, R. Krankenhagen, I. Sieber, W. Henrion, M. Schmidt, Amorphous semiconductors— science and technology, J. Non-Cryst. Solids 198–200 (1996) 1012. [40] J. Ma, J. Ni, J. Zhang, Q. Liu, G. Hou, X. Chen, X.D. Zhang, X. Geng, Y. Zhao, The influence of perpendicular transport behavior on the properties of n–i–p type amorphous silicon solar cells, Sol. Energy Mater. Sol. Cells 120 (2014) 635–641. [41] H. Stiebig, F. Siebke, W. Beyer, C. Beneking, B. Rech, H. Wagner, Interfaces in a-Si:H solar cell structures, Sol. Energy Mater. Sol. Cells 48 (1997) 351–363. [42] L. Feitknecht, O. Kluth, Y. Ziegler, X. Niquille, P. Torres, J. Meier, N. Wyrsch, A. Shah, Microcrystalline n–i–p solar cells deposited at 10 Å/s by VHF-GD, Sol. Energy Mater. Sol. Cells 66 (2001) 397–403. [43] F.A. Rubinelli, R.L. Stolk, A. Sturiale, J.K. Rath, R.E.I. Schropp, Sensitivity of the dark spectral response of thin film silicon based tandem solar cells on the defective regions in the intrinsic layers, J. Non-Cryst. Solids 352 (2006) 1876–1879. [44] H. Sai, T. Koida, T. Matsui, I. Yoshida, K. Saito, M. Kondo, Microcrystalline silicon solar cells with 10.5% efficiency realized by improved photon absorption via periodic textures and highly transparent conductive oxide, Appl. Phys. Express 6 (2013) 104101.

Suggest Documents