Letter pubs.acs.org/NanoLett
Dopant-Free All-Back-Contact Si Nanohole Solar Cells Using MoOx and LiF Films Han-Don Um, Namwoo Kim, Kangmin Lee, Inchan Hwang, Ji Hoon Seo, and Kwanyong Seo* Department of Energy Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan 689-798, Korea S Supporting Information *
ABSTRACT: We demonstrate novel all-back-contact Si nanohole solar cells via the simple direct deposition of molybdenum oxide (MoOx) and lithium fluoride (LiF) thin films as dopant-free and selective carrier contacts (SCCs). This approach is in contrast to conventionally used high-temperature thermal doping processes, which require multistep patterning processes to produce diffusion masks. Both MoOx and LiF thin films are inserted between the Si absorber and Al electrodes interdigitatedly at the rear cell surfaces, facilitating effective carrier collection at the MoOx/Si interface and suppressed recombination at the Si and LiF/Al electrode interface. With optimized MoOx and LiF film thickness as well as the all-back-contact design, our 1 cm2 Si nanohole solar cells exhibit a power conversion efficiency of up to 15.4%, with an opencircuit voltage of 561 mV and a fill factor of 74.6%. In particular, because of the significant reduction in Auger/surface recombination as well as the excellent Si-nanohole light absorption, our solar cells exhibit an external quantum efficiency of 83.4% for short-wavelength light (∼400 nm), resulting in a dramatic improvement (54.6%) in the short-circuit current density (36.8 mA/cm2) compared to that of a planar cell (23.8 mA/cm2). Hence, our all-back-contact design using MoOx and LiF films formed by a simple deposition process presents a unique opportunity to develop highly efficient and low-cost nanostructured Si solar cells. KEYWORDS: All-back-contact, dopant-free, selective carrier contact, molybdenum oxide, lithium fluoride
S
device front surface is protected from dopant diffusion by the diffusion barrier during the ABC processes, the Si-nanostructure dopant concentration is equivalent to the low concentration (≤1016 cm−3) of the base Si substrate, and Auger recombination is significantly reduced. The surface passivation issue can also be solved by providing excellent chemical and field-effect passivation at the lightly doped nanostructured Si surfaces. In addition, the shading loss from the front contact is also prevented, because both the emitter and back-surface field (BSF) contacts are formed in an interdigitated manner at the ABC solar-cell rear surfaces. Based on this beneficial and unique design, nanostructured Si solar cells with ABC technology have achieved high external quantum efficiency (EQE) of approximately 96% at short wavelengths (∼400 nm).10 However, the conventional ABC solar-cell fabrication process is not costeffective, because implementation of the ABC design adds complexity.12 The emitter and BSF layers should be formed separately via high-temperature doping processes using diffusion masks fabricated via patterning technology. Furthermore, additional alignment processes are required to separately realize two different contacts on the emitter and BSF region.
i nanostructures are potential candidates for application in highly efficient and cost-effective solar cells, as ultralow reflection of a broad range of wavelengths without the need for additional antireflection coatings is achieved by gradually reducing the impedance mismatch between air and silicon.1−8 Despite their superior light absorption, the power conversion efficiencies (PCEs) of nanostructured Si solar cells remain relatively low compared to those of conventional crystalline Si solar cells. One possible source of PCE degradation is the dramatic increase in the Auger and surface recombination of the cells due to the high Si-nanostructure surface-to-volume ratio. Compared to conventional solar cells with microscale pyramidal structures, nanostructured Si solar cells have much larger p−n junction areas, in which the relatively highly doped Si (>1018 cm−3) is necessarily formed via the thermal doping process. Thus, the Auger recombination intensified by the larger area of the highly doped Si layer induces degradation of the nanostructured Si solar cell photovoltaic performance.6,8 In addition, severe surface recombination occurs, as the high dopant concentration at the Si surface negates the electrical field effect of the surface passivation layer, due to the formation of a weak inversion layer at the Si/passivation layer interface.9 Integration of all-back-contact (ABC) technology into nanostructured Si solar cells can prevent severe recombination, as the planar emitter layer and Si nanostructures are placed at the rear and front device surfaces, respectively.10,11 Because the © XXXX American Chemical Society
Received: September 29, 2015 Revised: January 7, 2016
A
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
respectively. This inversion layer acted as a p-type Si layer; thus, the photogenerated electron−hole pairs in the nanostructured Si solar cells are separated by the MoOx/n-Si interface rather than a p−n junction, as in conventional solar cells.21 To evaluate the MoOx/n-Si junction, planar front-backcontact Si solar cells were fabricated by depositing MoOx films with various thicknesses (Figure 1a). A heavily doped region
An alternative low-cost fabrication strategy is to apply dopant-free, selective carrier contacts (SCCs) to the Si solar cells, which yields selective hole and electron collections through the negative- and positive-polarity contacts, respectively. SCCs, such as organic polymers,2,13−16 alkaline metal compounds,17−19 and transition metal oxides20−22 have been successfully introduced to conventional Si solar cells. Among them, molybdenum oxide (MoOx) and lithium fluoride (LiF) films are very attractive for hole and electron contacts, respectively, because of their simple, direct deposition processes at room temperature. Herein, we fabricate low-temperature processed ABC solar cells by selectively depositing MoOx and LiF films as dopant-free SCCs between the Si absorber rear surfaces and Al electrodes in nanostructured Si solar cells. Our nanostructured Si solar cells achieve a PCE of up to 15.4%, with an open circuit voltage (VOC) of 561 mV, a short circuit current density (JSC) of 36.8 mA/cm2, and a fill factor (FF) of 74.6%. These results indicate that our dopant-free ABC solar cells can effectively separate the photocarriers at the MoOx/Si junction and collect electrons through the LiF/Al contact. This efficiency is higher than those of planar solar cells both with (13.9%) and without antireflection layers (10.2%), because of the strong broadband absorption of the Si nanoholes. In particular, the overall advantages provided by dopant-free ABC solar cells can be observed by comparing with the process chart of the conventional ABC solar cell (see Supporting Information, Figure S1). In total, six process steps are bypassed in the fabrication of the dopant-free ABC solar cells, including the SiO2 deposition, photolithography, the SiO2 etch-back process, and the thermal doping process, which are required for fabrication of the conventional ABC solar cell. This simplification of the fabrication process corresponds to a significant reduction in fabrication cost compared to that of conventional ABC solar cells. The key concept underlying this work is the application of MoOx and LiF thin films as the emitter and BSF layers at the ABC solar-cell rear surfaces, respectively, rather than the thermally formed heavily boron-doped (p+) or phosphorusdoped (n+) Si layers used in conventional ABC solar cells. First, we investigated the effect of varying MoOx film thickness (as the hole contact layer) on the photovoltaic behavior of n-type Si solar cells. In order to collect holes selectively, a substoichiometric MoOx film was obtained by thermally evaporating a solid MoO3 source (an insulating material). The deposited MoOx film could be employed as a semiconducting metal oxide with a metallic defect band, having 104 S/cm electrical conductivity due to the band gap defect states originating from the oxygen vacancies.23,24 Note that Battaglia et al. have reported that the work function of MoOx film without carbon contamination is approximately 6.6 eV.20 When the MoOx film was deposited onto the n-type Si substrate, a barrier formed at the MoOx/Si interface because of the larger work function gap between MoOx and Si. As the barrier height at the interface (∼1.6 eV) was larger than half the Si band gap, the holes (minority carriers) outnumbered the electrons (majority carriers) in the Si region close to the interface, leading to the formation of a hole inversion layer25 (see the MoOx/Si junction energy band diagram in Figure S2). The built-in field in the region between the inversion layer and n-Si allowed the device to function as a solar cell, despite the absence of the conventional p−n junction. The field in the depletion layer separated the electron−hole pairs and caused drifting of the holes and electrons toward the MoOx and n-Si,
Figure 1. (a) Schematic of MoOx/n-Si junction solar cell. (b) MoOx/ n-Si junction solar-cell J−V curves without (black solid line with filled squares) and with MoOx films of 2.5 (green solid line with filled triangles), 5 (red solid line with filled circles), 10 (blue solid line with filled inverse triangles), and 20 nm (magenta solid line with filled diamonds) thickness. (c) Schematic of planar p−n junction solar cell with LiF/Al electrode. (d) J−V curves of p−n junction solar cells without (black solid line with filled squares) and with LiF films of 0.5 (green solid line with filled triangles), 1 (red solid line with filled circles), and 2 nm (blue solid line with filled inverse triangles) thickness.
(n+) was formed at the front of the Si solar cell via thermal doping, to create an ohmic contact between the Si and front electrode. We positioned the MoOx/Si junction on the device rear side to prevent optical-transmittance dependence on the MoOx film thickness (see Supporting Information, Figure S3). The current density−voltage (J−V) curves of the MoOx/Sijunction solar cells with MoOx films of varying thickness were analyzed under air mass (AM) 1.5G illumination (Figure 1b). The VOC, JSC, FF, and PCE photovoltaic parameters are summarized in Table 1. The MoOx-film-free device exhibited a Table 1. Planar Solar-Cell Photovoltaic Performance of MoOx/Al and LiF/Al Electrodes SCC materials
thickness (nm)
VOC (mV)
JSC (mA/cm2)
FF (%)
PCE (%)
MoOx
0 2.5 5 10 20 0 0.5 1 2
19 290 547 572 573 541 539 539 540
21.7 27.2 29.7 30.6 29.8 34.0 35.4 35.2 35.2
25.2 55.2 58.8 73.2 72.8 68.1 68.8 66.5 54.0
0.1 4.4 9.5 12.8 12.4 12.5 13.1 12.6 10.3
LiF
B
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters very low PCE of 0.1%, along with VOC, JSC, and FF values of 19 mV, 21.7 mA/cm2, and 25.2%, respectively. This result was obtained because the Schottky junction at the Si/Al interface did not effectively separate the photocarriers, as a result of the high density of the defect sites at the metal/Si interface.26,27 In contrast, the MoOx film between the Si and the Al electrode significantly improved the PCE, from 0.1−12.8%. For a MoOx thickness of 10 nm, the PCE value increased and then saturated. This behavior was observed because very thin MoOx film (less than 10 nm) cannot fully cover the Si surface (Figure S4a), diminishing the diode property of the MoOx/Si junction through the direct Al−Si contact Schottky junction. As the 10nm-thick MoOx film produced the highest front-back-contact solar-cell PCE (12.8%), it was subsequently applied in the ABC solar cells. The optimized LiF film thickness (electron contact layer) was also investigated using the front-back-contact solar cells, with the p−n junction being formed by diffusing the boron dopant into the n-Si. The LiF film’s function was to effectively collect the majority carriers (electrons), rather than to separate the photocarriers (Figure 1c). The solar-cell J−V characteristics and photovoltaic parameters for different LiF film thicknesses are summarized in Figure 1d and Table 1, respectively. The JSC increased from 34.0 to 35.4 mA/cm2 for LiF film inserted between the Si and Al electrode. The EQE of the longwavelength region (900−1100 nm) improved notably for the LiF-film-containing cells (Figure S5). In general, photocarriers generated by low-energy photons (i.e., long-wavelength light) deep within the Si absorber easily migrate toward the rear surface, where they are recombined with the opposite charge carriers. Here, the presence of LiF film between the Si and the Al electrode suppressed the surface recombination by effectively collecting the electrons through the Al electrode. The Al work function was decreased by up to 3.3 eV as a result of the large LiF dipole moment, leading to Schottky barrier reduction18,28,29 (see the LiF/Al contact energy band diagram in Figure S6). Despite the JSC enhancement, the PCEs were identical to or less than that of the LiF-film-free device for >1 nm LiF film thickness, because the FF value decreased significantly with increased LiF thickness. This behavior was observed because the electron transport from the Si to the Al electrode was impeded by the thick insulating LiF film. For increased LiF film thickness, the Si−Al area of direct contact decreased (Figure S7), leading to increased contact resistance and series resistance (Rs). As a result of the increased Rs, the VOC also decreased. In practical applications, high VOC can be achieved by forming a BSF layer of heavily doped Si (n+) between the n-Si and metal film, so as to reduce the contact resistance and prevent interface recombination. For the LiFfilm-free cell, however, the Al film is thermally deposited onto the lightly doped n-Si substrate without a BSF layer, resulting in a large Schottky barrier due to Fermi-level pinning at the Si/Al interface.30 Thus, the LiF-film-free device exhibits a lower VOC of 541 mV compared to the MoOx/Si junction solar cell with the BSF layer (572−573 mV). Even though the Schottky barrier reduction of the LiF/Al contact caused JSC enhancement, VOC degradation of 1−2 mV was observed for the LiFfilm-containing cells. This result can be explained by the fact that a voltage drop is induced by a leakage current flowing through Rs.31 In this case, although the LiF film cannot reduce the contact resistance (unlike the BSF layer), the recombination of the minority carriers at the Si/Al interface may be moderate as a result of the effective electron collection through
the LiF/Al contact. Through the above analysis, the solar cell with 0.5 nm-thick LiF film was found to exhibit improved JSC with minimized VOC and FF degradation. Thus, the 0.5 nmthick LiF film was identified as being optimal for the purposes of this study. As the 10-nm-thick MoOx and 0.5 nm-thick LiF films produced the highest PCEs, they were subsequently applied to the ABC solar cells (Figure 2a). The planar ABC solar-cell front
Figure 2. (a) Schematic of planar ABC solar cell with selective carrier contacts. (b) Optical image of solar-cell rear side with MoOx/Al and LiF/Al-electrode interdigitated contacts. (c) Magnified image of redsquare region in part b showing uniform pitch between MoOx/Al and LiF/Al electrodes. (d) Planar ABC solar-cell J−V curves for pitch values between hole and electron contacts of: 225 (black solid line with filled squares), 375 (green solid line with filled circles), 575 (red solid line with filled triangles), and 1025 (blue solid line with filled inverse triangles) μm.
surface was covered by a thin Al2O3 passivation layer with no metal-electrode content, which was formed using atomic layer deposition (ALD). On the rear side of the cell, the hole and electron contacts were formed interdigitatedly using MoOx/Al and LiF/Al electrodes, respectively (Figure 2b and c). First, the 10-nm-thick MoOx film was deposited onto a part of the rear side, followed by the 500 nm-thick Al film. The exposed rearside part was then covered with 0.5-nm-thick LiF and 500-nmthick Al films to produce a uniform-pitch interdigitated pattern between the two contacts. In order to optimize the ABC solarcell PCEs, we controlled the intercontact pitch while maintaining the same MoOx/Al electrode area. For all devices, the areal ratio of the emitter was set to approximately 60%, based on a previous report,11 as the photocarrier separation and collection probabilities were potentially affected by the emitter area. Figure 2d shows the planar ABC solar-cell J−V characteristics for different pitch values. The JSC and FF values C
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters Table 2. Planar and Nanohole ABC Solar-Cell Photovoltaic Performance device
structure
dopant-free
planar
thermal
nanohole nanohole
antireflection
SiNx
pitch (μm) 225 375 575 1025 225 225 225
JSC (mA/cm2)
VOC (mV) 555 548 557 546 562 561 618
23.8 23.0 22.8 20.1 33.2 36.8 38.2
FF (%) 77.1 76.9 75.1 67.2 75.0 74.6 77.3
PCE (%)
shunt R (Ω·cm2)
Rs (Ω·cm2)
10.2 9.7 9.5 7.4 13.9 15.4 18.2
× × × × × × ×
1.27 2.05 2.82 6.43 2.17 2.48 1.13
3.16 2.87 2.34 1.29 3.32 2.85 6.87
3
10 103 103 103 103 103 103
Figure 3. Simulated data for: (a) VOC, (b) JSC, (c) FF, and (d) PCE as functions of pitch and J0 at MoOx/Si and LiF/Si interfaces. The J0 values are MoOx and LiF: 300 fA/cm2 (low; black solid line with filled squares); MoOx: 1200 fA/cm2 and LiF: 3500 fA/cm2 (experimental; red solid line with filled circles); and MoOx: 1.5 pA/cm2 and LiF: 35 pA/cm2 (high; blue solid line with filled triangles). The SARAH simulation tool developed by Fraunhofer ISE34 was used.
reducing the Rs loss. Furthermore, the photocarriers could be effectively diffused to the contacts before recombination with the majority carriers, because of the shorter diffusion length; this behavior resulted in improved JSC. In contrast, the photocarriers of the large-pitch solar cells were not effectively collected through the contacts, because of the severe recombination. This recombination loss primarily occurred at the MoOx/Si and LiF/Si interfaces, as we assumed that the recombination at the bulk Si and front surface was relatively suppressed by the selected simulation parameters (bulk lifetime: 1.5 ms, front-surface J0: 350 fA/cm2). For J0 values less than the experimental values (black solid line with filled squares, Figure 3b), the JSC values were improved regardless of pitch. The JSC degradation for a large pitch of 1025 μm was relatively low compared to the higher J0 values (red and blue solid lines, Figure 3b). This result indicates that photocarriers can be collected through the contacts by reducing the recombination at the MoOx/Si and LiF/Si interfaces. In other words, our devices are negatively affected by recombination at the MoOx/Si and LiF/Si interface defect sites. This behavior can be prevented using a passivation layer (e.g., amorphous Si film) between the SCC and Si, leading to an increased effective minority carrier lifetime (i.e., decreased J0 values at the MoOx/Si and LiF/Si interfaces). According to the simulation results obtained for decreased J0 values, a pitch of
were significantly increased, by 18.4% and 14.7%, respectively, for a decrease in pitch from 1025 to 225 μm, while the VOC values remained almost constant (Table 2). In general, the FF is the only photovoltaic parameter previously reported to be affected by the pitch of conventional ABC solar cells fabricated using thermal doping processes.32,33 Increased FF is primarily due to decreased Rs (Table 2), as the photocarrier diffusion length decreases with smaller pitch. Here, however, both the JSC and FF were affected by the pitch. To elucidate the experimental results, we performed a twodimensional model simulation (SARAH, Fraunhofer ISE) for a conventional ABC solar-cell design,34 with an emitter, ohmic contact regions, and a front-surface passivation layer. As the MoOx/Si junction and LiF/Al contact act as an emitter and ohmic contact, respectively, the simulation parameters, including the pitch (with the same area as the MoOx/Si) and the saturation current density (J0) at the MoOx/Si and LiF/Si interfaces, were considered. Experimental J0 values (MoOx: J0 = 1200 fA/cm2, LiF: J0 = 3500 fA/cm2) were obtained from lifetime measurements of samples in which both sides of the Si substrate were covered by MoOx or LiF films (Figure S8). The simulation results showed that the FF and JSC increased with decreasing pitch, which agrees with the experimental results (Figure 3). For smaller pitch, the photocarriers could be collected through the contacts with shorter diffusion lengths, D
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 4. (a) Schematic of Si nanohole ABC solar cell with selective carrier contacts. (b) Optical images of planar and nanohole solar cells. (c) Cross-sectional SEM image of Si nanoholes. (d) J−V curves of planar Si without (black solid line with filled squares), with SiNx layer (red solid line with filled circles), and nanohole (blue solid line with filled triangles) ABC solar cells. (e) Planar Si without (black solid line with squares), with SiNx layer (red solid line with circles), and nanohole (blue solid line with triangles) ABC solar-cell EQE (open symbols) and reflectance (filled symbols) spectra.
>300 μm can be used without PCE degradation. These large pitches are suitable for low-cost solar-cell fabrication, as costeffective processes such as screen-printing and shadow-mask methods can be used. Although the planar ABC solar-cell PCE was improved by minimizing the intercontact pitch, the overall PCE remained lower than those of conventional ABC solar cells because of the enormous light reflection of over 40% from the planar surface.10 To diminish the front-surface reflection, we applied nanohole arrays to the ABC solar-cell front surfaces (Figure 4a). The metal-assisted chemical etching (MACE) method was used, as it is a fast and inexpensive technique. Si nanohole arrays completely blacken the ABC solar-cell front surfaces (Figure 4b). The Si nanohole arrays used here had typical heights of 240 ± 5 nm and pore diameters of 17 ± 2 nm, as characterized using cross-sectional scanning electron microscopy (SEM). As the Si nanohole arrays improved the matching of the optical impedance between air and bulk Si, the reflectance (averaged over the main 400−1000 nm spectral range) decreased significantly from 38.5% to 4.3% without use of an additional antireflection layer (Figure 4e). Hence, the nanohole ABC solar-cell JSC increased dramatically (23.8−36.8 mA/cm2) compared to that of the planar ABC cell (J−V curves, Figure 4d). Thus, the superior broadband antireflection property of the cell with Si nanoholes compared to that of the planar ABC solar cell with the conventional antireflection layer was clearly observed. The 80 nm-thick SiNx layer was deposited by plasma-enhanced chemical vapor deposition (PECVD) onto the front surface of solar cell. As the Si nanohole arrays exhibit lower reflectance over the entire wavelength range of 400 to 1100 nm (Figure 4e), the JSC of the nanohole ABC solar-cell is higher than that of the planar ABC cell with the conventional antireflection layer (33.2 mA/cm2). This result indicates that the conventional antireflection layer can be replaced with the Si nanohole array.
Further, the increased JSC value was consistent with the enhanced EQE (Figure 4e), in accordance with q ΔJSC = ( SQ holeλ dλ − SQλ dλ) (1) hc
∫
∫
where q is the electron charge, h is Planck’s constant, c is the speed of light, S is the AM 1.5G solar spectrum, Qhole is the nanohole-cell EQE, Q is the planar-cell EQE, and λ is the light wavelength. Remarkably, the Qhole for λ = 400 nm was 83.4%, indicating that the photocarriers generated in the lightly doped nanoholes were very efficiently collected without Auger/surface recombination. In addition, the VOC improved slightly (555− 561 mV) via the significant increase in JSC, as VOC =
⎛J ⎞ k bT × ln⎜⎜ SC + 1⎟⎟ q ⎝ J0 ⎠
(2)
where kb, T, and q represent Boltzmann’s constant, the temperature, and the electron charge, respectively. Although the nanohole solar-cell FF decreased slightly according to the resistive loss originating from the increased JSC,35 we obtained a very high PCE of 15.4% for our optimally performing cell, along with VOC, JSC, FF values of 561 mV, 36.8 mA/cm2, and 74.6%, respectively, because of the suppressed Auger/surface recombination and increased light absorption. Our dopant-free ABC solar cell was compared with a conventional ABC solar cell fabricated via a high-temperature doping processes. The experimental details regarding the fabrication of the conventional ABC solar cell are described in the Supporting Information. Both solar cells were designed with Si nanoholes at the front surface and a 225 μm pitch at the rear surface. The J−V characteristic and photovoltaic parameters of the conventional ABC solar cell are summarized in Figure S9 and Table 2, respectively. The conventional ABC solar cell exhibits a PCE of 18.2%, as the electrical field in the highly doped layer causes drifting of the minority carriers into the bulk Si before E
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Technology (UNIST). It was also supported by NRF through Basic Science Research Program (2014R1A1A1004885) from the Ministry of Science, ICT & Future Planning.
recombination occurs at the Si/Al contact. Although no hightemperature process was applied, our dopant-free ABC solar cell exhibited a photovoltaic performance equivalent to 84.6% that of the thermally doped solar cell. Therefore, our devices demonstrate the potential of thermal-doping-free and simply deposited dopant-free contacts applied in low-cost and highly efficient ABC solar cells. Further FF and VOC improvements would be possible using a thicker Al electrode and inserting thin amorphous Si film between the SCC and Si. In summary, we have demonstrated highly efficient (15.4%) ABC solar cells through simple, direct deposition of MoOx and LiF films, yielding dopant-free Si solar cells without the use of a high-temperature thermal doping process. Further, 10-nm-thick MoOx film and 0.5-nm-thick LiF film were shown to be optimal for effective photocarrier separation through inversion-layer formation on n-type Si and enhanced carrier collection via recombination reduction at the Si/Al interface, respectively. The ABC structure was also optimized by controlling the intercontact pitch, yielding the highest JSC and FF with the smallest pitch (225 μm). For cost-effective solar cells, Si nanoholes were fabricated via MACE on the ABC solar-cell front surfaces, leading to significant light-reflection reduction with no additional antireflection layer. The superior light absorption and significant Auger/surface recombination reduction at the nanoholes improved the JSC value dramatically (23.8−36.8 mA/cm2), with an EQE of 83.4% for shortwavelength light (400 nm), resulting in 51% PCE improvement compared to a planar ABC solar cell (10.2%). This work is therefore believed to present a very promising design using dopant-free contacts to achieve cost-effective and highly efficient nanostructured Si solar cells.
■
■
(1) Garnett, E. C.; Yang, P. D. J. Am. Chem. Soc. 2008, 130, 9224− 9225. (2) Gong, X.; Jiang, Y. R.; Li, M.; Liu, H. R.; Ma, H. RSC Adv. 2015, 5, 10310−10317. (3) Jia, G. B.; Steglich, M.; Sill, I.; Falk, F. Sol. Energy Mater. Sol. Cells 2012, 96, 226−230. (4) Jung, J. Y.; Um, H. D.; Jee, S. W.; Park, K. T.; Bang, J. H.; Lee, J. H. Sol. Energy Mater. Sol. Cells 2013, 112, 84−90. (5) Lu, Y. R.; Lal, A. Nano Lett. 2010, 10, 4651−4656. (6) Um, H. D.; Park, K. T.; Jung, J. Y.; Li, X. P.; Zhou, K.; Jee, S. W.; Lee, J. H. Nanoscale 2014, 6, 5193−5199. (7) Wang, J. X.; Wang, H.; Prakoso, A. B.; Togonal, A. S.; Hong, L.; Jiang, C. Y.; Rusli, N. A. Nanoscale 2015, 7, 4559−4565. (8) Oh, J.; Yuan, H. C.; Branz, H. M. Nat. Nanotechnol. 2012, 7, 743−748. (9) Richter, A.; Benick, J.; Kimmerle, A.; Hermle, M.; Glunz, S. W. J. Appl. Phys. 2014, 116, 243501. (10) Savin, H.; Repo, P.; von Gastrow, G.; Ortega, P.; Calle, E.; Garin, M.; Alcubilla, R. Nat. Nanotechnol. 2015, 10, 624−628. (11) Jeong, S.; McGehee, M. D.; Cui, Y. Nat. Commun. 2013, 4, 2950. (12) Sinton, R. A.; Swanson, R. M. IEEE Trans. Electron Devices 1990, 37, 348−352. (13) He, J.; Gao, P.; Liao, M.; Yang, X.; Ying, Z.; Zhou, S.; Ye, J.; Cui, Y. ACS Nano 2015, 9, 6522−6531. (14) Jeong, S.; Garnett, E. C.; Wang, S.; Yu, Z.; Fan, S.; Brongersma, M. L.; McGehee, M. D.; Cui, Y. Nano Lett. 2012, 12, 2971−2976. (15) Wang, J.; Wang, H.; Prakoso, A. B.; Togonal, A. S.; Hong, L.; Jiang, C.; Rusli. Nanoscale 2015, 7, 4559−4565. (16) Zhang, Y. F.; Cui, W.; Zhu, Y. W.; Zu, F. S.; Liao, L. S.; Lee, S. T.; Sun, B. Q. Energy Environ. Sci. 2015, 8, 297−302. (17) Hsu, F.-H.; Wang, N.-F.; Tsai, Y.-Z.; Wu, C.-Y.; Cheng, Y.-S.; Chien, M.-H.; Houng, M.-P. Thin Solid Films 2014, 573, 159−163. (18) Xu, D.; Yu, X.; Zuo, L.; Yang, D. RSC Adv. 2015, 5, 46480− 46484. (19) Zhang, Y.; Liu, R.; Lee, S.-T.; Sun, B. Appl. Phys. Lett. 2014, 104, 083514. (20) Battaglia, C.; Yin, X.; Zheng, M.; Sharp, I. D.; Chen, T.; McDonnell, S.; Azcatl, A.; Carraro, C.; Ma, B.; Maboudian, R.; Wallace, R. M.; Javey, A. Nano Lett. 2014, 14, 967−71. (21) Bullock, J.; Cuevas, A.; Allen, T.; Battaglia, C. Appl. Phys. Lett. 2014, 105, 232109. (22) Yang, J. H.; Jung, H. H.; Seo, J.; Kim, K. D.; Kim, D. H.; Lim, D. C.; Park, S. G.; Kang, J. W.; Song, M.; Choi, M. S.; Kwon, J. D.; Nam, K. S.; Jeong, Y.; Kwon, S. H.; Park, Y. C.; Kang, Y. C.; Chung, K. B.; Kim, C. S.; Lim, K. S.; Ryu, S. Y. J. Phys. Chem. C 2013, 117, 23459− 23468. (23) Chuang, S.; Battaglia, C.; Azcatl, A.; McDonnell, S.; Kang, J. S.; Yin, X. T.; Tosun, M.; Kapadia, R.; Fang, H.; Wallace, R. M.; Javey, A. Nano Lett. 2014, 14, 1337−1342. (24) Meyer, J.; Hamwi, S.; Kroger, M.; Kowalsky, W.; Riedl, T.; Kahn, A. Adv. Mater. 2012, 24, 5408−5427. (25) Nelson, J. The Physics of Solar Cells; Imperial College Press: London, 2004; pp 125−130. (26) Fonash, S. J. J. Appl. Phys. 1975, 46, 1286−1289. (27) Pulfrey, D. L. IEEE Trans. Electron Devices 1978, 25, 1308− 1317. (28) Shaheen, S. E.; Jabbour, G. E.; Morrell, M. M.; Kawabe, Y.; Kippelen, B.; Peyghambarian, N.; Nabor, M. F.; Schlaf, R.; Mash, E. A.; Armstrong, N. R. J. Appl. Phys. 1998, 84, 2324−2327. (29) Wang, X. J.; Zhao, J. M.; Zhou, Y. C.; Wang, X. Z.; Zhang, S. T.; Zhan, Y. Q.; Xu, Z.; Ding, H. J.; Zhong, G. Y.; Shi, H. Z.; Xiong, Z. H.;
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b03955. Experimental section; Process charts of conventional and dopant-free ABC solar cells; Energy band diagram of MoOx/Si junction; Transmittance spectra of MoOx films with different thicknesses; AFM images of MoOx films; EQE spectra of the front-back contact solar cells without and with LiF film; Energy band diagram of LiF/Al contact; AFM images of LiF films; Inverse effective carrier lifetime reduced by inverse Auger carrier lifetime versus excess carrier density for MoOx/Si and LiF/Si contacts; J−V curve of the thermally doped ABC solar cell (PDF)
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Author Contributions
H.D.U. and K.S. conceived and designed the research study. H.D.U., N.K., K.L., I.H., and J.H.S. performed the experiments and analyzed the data. H.D.U. and K.S. wrote the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Future Strategic Fund (1.150030.01) of Ulsan National Institute of Science and F
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters Liu, Y.; Wang, Z. J.; Obbard, E. G.; Ding, X. M.; Huang, W.; Hou, X. Y. J. Appl. Phys. 2004, 95, 3828−3830. (30) Cowley, A. M.; Sze, S. M. J. Appl. Phys. 1965, 36, 3212−3220. (31) Cuevas, A.; Lopez-Romero, S. Sol. Cells 1984, 11, 163−173. (32) Kim, D. S.; Meemongkolkiat, V.; Ebong, A.; Rounsaville, B.; Upadhyaya, V.; Das, A.; Rohatgi, A. Conference Record of the 2006 IEEE 4th World Conference on Photovoltaic Energy Conversion, 2006; pp 1417−1420. (33) Swanson, R. M. Sol. Cells 1986, 17, 85−118. (34) Saint-Cast, P.; Padilla, M.; Kimmerle, A.; Reichel, C. IEEE J. Photovolt. 2014, 4, 114−121. (35) Fahrenbruch, A. L.; Bube, R. H. Fundamentals of Solar Cells: Photovoltaic Solar Energy Conversion; Academic Press: New York, 1983; pp 220−220.
G
DOI: 10.1021/acs.nanolett.5b03955 Nano Lett. XXXX, XXX, XXX−XXX
