Session Fr-B2

a-Si:H/c-Si Interface II

Chair: Nazir Kherani, University of Toronto

Fr-B2.1 10:30–10:50

Thin Microcrystalline Layers for Application in Silicon Heterojunction Solar Cells

Johannes P. Seif, Antoine Descoeudres, Zachary C. Holman, Stefaan De Wolf, and Christophe Ballif

Ecole Polytechnique Fédérale de Lausanne (EPFL), Institute of Microengineering (IMT), Photovoltaics and thin film electronics laboratory / Rue A.-L. Breguet 2, CH-2000 Neuchâtel, Switzerland

Current generation in amorphous/crystalline silicon heterojunction (SHJ) solar cells is limited by losses at the front of the device. Light absorbed in the transparent conductive oxide (TCO) layer at the front and the front doped and intrinsic hydrogenated amorphous silicon (a-Si:H) layers either generates carriers that tend to recombine before being collected or is lost to free carrier absorption. To mitigate this so-called parasitic absorption, materials with a lower absorption coefficient especially at shorter wavelengths could be used. Exhibiting an indirect band gap and thus showing a lower absorption in the range of interest, microcrystalline silicon (μc-Si:H) in contrast to a-Si:H is a promising candidate for this purpose. However, the deposition of a thin, doped yet highly crystalline μc-Si:H layer on an intrinsic (i) a-Si:H passivation layer, that is the fundamental ingredient to high-efficiency devices, is not straightforward. One reason is that on such substrates, prior to the onset of crystalline growth, the material is prone to show an a-Si:H incubation layer, whose thickness can amount to several tens of nanometers. For p-doped layers the latter is closely linked to the boron concentration in the plasma. It has been shown [1] that boron catalyzes the desorption of hydrogen—which is crucial for crystalline growth—from the growing film and thus leads to its amorphization. Various techniques have been developed to avoid the formation of this incubation layer, some of which will be reviewed in the present paper. To maximize the short circuit current, layers used in SHJ cells are typically in the order of approximately 10 nm. Hence, the main challenge is to impede the formation of the incubation layer and achieve an abrupt transition at the i a-Si:H, doped μc-Si:H interface, as well as a sufficiently high crystalline volume fraction. At the same time, a good passivation quality is crucial, since it enables elevated open-circuit voltages. Therefore, a degradation of the underlying i a-Si:H layer, upon μc-Si:H deposition, should be avoided. In order to find deposition conditions meeting these requirements, we explored regimes with plasma frequencies between 13.56 MHz and 80 MHz, and also tested the influence of the deposition temperature and different gas mixtures, i.e. disilane instead of silane and deuterium instead of hydrogen, on the crystalline volume fraction of the layer. Furthermore, carbon dioxide plasma treatments of the a-Si:H substrate as well as the deposition of seed layers for μc-Si:H were studied. Based on these results, we achieved a 12.2 nm thick p-doped μc-Si:H layer with a crystalline fraction of approximately 71% deposited on glass. In a next step, we will test these deposition conditions on an a-Si:H substrate and implement it into a SHJ device. With this, we expect to see an increase in short-circuit current density, potentially up to 0.4 mA/cm2 (determined by simulations with OPAL [2]), as well as an increase in fill factor as a result of a better contact to the TCO, which is due to the expected increased doping efficiency of a μc-Si:H in contrast to an a-Si:H p-layer [3].

Keywords: silicon, microcrystalline, incubation, heterojunction, passivation

[1] J. Perrin et al., Surf. Sci., 210 (1), pp. 114–128, 1989

[2] S. C. Baker-Finch and K. R. McIntosh, Proc. 35th IEEE PVSC, Honolulu, pp. 002184–002187, 2010

[3] M. Bivour et al., Sol. Energ. Mat. Sol. Cells, 106, pp. 11–16, 2012

Fr-B2.2 10:50–11:10

Temperature and Bias Dependence of Hydrogenated Amorphous Silicon/Crystalline Silicon Heterojunction Capacitance: The Link to Band Bending and Band Offsets

O. Maslova (1,2), A. Brézard-Oudot (1), M. E. Gueunier-Farret (1), J. Alvarez (1), W. Favre (3), D. Muñoz (3), A. S. Gudovskikh (2), E. Terukov (4), and J. P. Kleider (1)

1. LGEP; CNRS UMR 8507; SUPELEC; Univ Paris-Sud; UPMC Univ Paris 06; 11 rue Joliot-Curie, Plateau de Moulon, F-91192 Gif-sur-Yvette Cedex, France

2. St.Petersburg Academic University - Nanotechnology Research and Education Centre of the Russian Academy of Sciences, Hlopina str. 8/3, 194021, St.-Petersburg, Russia

3. INES-CEA, 50 Avenue du Lac Léman, BP332, F-73310 Le Bourget du Lac, France

4. A.F. Ioffe Physico-technical Institute, Polytechnicheskaya str. 26, 194021, St.-Petersburg, Russia

Hydrogenated amorphous silicon/crystalline silicon (a-Si:H/c-Si) heterojunctions have proved to be promising candidates for solar cell fabrication while combining low cost manufacturing and high efficiency. Recently a lot of attention has been paid to the possibilities of capacitance measurements for studying a-Si:H/c-Si heterojunctions and determining interface parameters.

While most previous studies were focused on the low temperature behaviour, in this work we present the temperature dependence of the capacitance-voltage data (C-V-T) in a wide temperature range, up to 400K. Two approaches for the calculation of C-V-T data are compared: depletion approximation and extended analytical calculation. In both approaches we take into account the peculiar space charge distribution in a-Si:H. However in the extended analytical calculation, both types of carriers are taken into account in the space charge density in c-Si, while their contribution within the space charge layer is neglected in the depletion approximation. We show that the temperature dependence of 1/C2 versus voltage plots and the intercept with the voltage axis, Vint, of the linear extrapolation of these plots depend on whether strong inversion occurs at the c-Si surface or not. If a strong inversion layer exists, a larger decrease of Vint with increasing temperature is obtained in the extended calculation compared to the one obtained in the depletion approximation, leading to room temperature values that are considerably smaller than the diffusion potential Vd expected from the depletion approximation model. As a consequence, band offsets determined at room temperature from this technique assuming that Vint=Vd are also strongly underestimated. Due to the temperature shift of the Fermi level in c-Si the surface strong inversion layer may disappear at low temperature, depending on the band offset value. Interestingly, this can be revealed in the temperature dependence of Vint, thus providing a way to deduce this parameter from capacitance measurements of the heterojunction.

Experimental results on high efficiency (p) a-Si:H/ (n) c-Si solar cells (η > 21%) are then compared to the calculations. It is shown that the experimental temperature dependence of C-V curves cannot be reproduced by the depletion approximation calculation. On the contrary, they are very well reproduced by the full analytical calculation suggesting that strong inversion does exist at the c-Si surface in the whole temperature range (100K–400K), which implies that the valence band offset be large enough (ΔEV> 0.3 eV), in agreement with previous results obtained on dedicated devices from other types of measurements.

Keywords: capacitance spectroscopy, a-Si:H/c-Si heterojunction, strong inversion, band offsets

Fr-B2.3 11:10–11:30

Hydrogenated Amorphous Silicon and Quasimorphous Silicon Thin Film for Solar Cells Application

A. R. Middya

Silicon Solar, Inc., Fremont, CA 94536

Since the invention of hydrogenated amorphous silicon (a-Si:H), the science and application of this tetrahydrally bonded disordered semiconductor advanced so much that anybody would imagine at the early period of amorphous silicon (a-Si:H). So far, a-Si:H thin-film is the most successful disordered semiconductors for application in microelectronics and photovoltaics. Unusally high photosensitivity (ratio of dark to photocurrent) ~106 makes it an interesting semiconductor for making good detector. The application expands from thin-film transistor (TFT) to amorphous silicon solar cells panel. Scientists discovered very soon the disadvantages of amorphous silicon that it degrades under intense light. Worldwide research started to control the light-induced degradation known as Staebler-Wronski effect and discussion on the role of hydrogen. In this report, I shall discuss the success of development of stress-free hydrogenated amorphous silicon (a-Si:H) using helium and argon dilution technology. Thick (~50 um) amorphous silicon can be easily deposited by plasma-enhanced chemical vapor deposition (PECVD) technique using helium and argon dilution technology. These class (stress-free) of a-Si:H thin-film have less strain bonds and substrate-to-film surface Si-Si network propagation close to ideal structure of a-Si, where few percentage (5 at.% to 20 at.%) of H relaxing stress within the lattice. I shall present single and tandem-junction a-Si:H solar cells performances based on these class of a-Si:H thin-film, where amorphous silicon (a-Si:H) and amorphous silicon-germanium (a-SiGe:H) layers have been developed under helium and argon dilution. We found single-junction a-Si:H solar cells is more stable if i-layer is deposited under argon dilution compared to hydrogen a-Si:H solar cells. This is beginning of the road, where the cost of a-Si:H solar cells can be decreased with time. I shall introduce a new class of amorphous silicon, where we do not find thickness dependence of phase transition yet they at the phase boundary of amorphous-to-nanocrystalline silicon thin-film. We call these class of amorphous silicon, quasimorphous silicon thin-film [1]. Astonishing properties are found in its electronic properties and stability under light. The mobility-lifetime (μτ) product is improved by a factor of 50 in case of quasimorphous silicon thin-film. We found using UV-VIS ellipsometry the evidence that new structure in Si lattice within a-Si:H in case of quasimorphous silicon thin-film. A comprehensive summary about the development of this class of amorphous silicon and its properties as an intrinsic (i) layer as well as solar cells performance will be presented.

[1] A. R. Middya, S. Hamma, S. Hazra, S. Ray and C. Longeaud, Mat. Res. Symp. Proc. 664 (2001) p. A9.5

Fr-B2.4 11:30–11:50

Stability of (n)c-Si Passivation Properties by a-Si:H Layers During Thermal Treatments

Wilfried Favre, Romain Champory, Renaud Varache, Thibaut Desrues, and Delfina Muñoz

CEA-INES, 50 avenue du lac leman, 73375 Le Bourget-du-lac, France

Silicon heterojunction (SHJ) are the basis of solar cells that have recently demonstrated the highest efficiency on large area [1]. An open-circuit voltage as high as 750 mV has been achieved using a crystalline silicon (c-Si) 90 μm thick wafer, passivated by intrinsic and doped hydrogenated amorphous silicon (a-Si:H) thin layers (<20 nm). Such excellent passivation properties can be obtained for high quality c-Si surfaces after cleaning, and a-Si:H and TCO deposition steps where each parameter has to be well controlled. Indeed, it was recently shown that passivation properties are not only sensitive to a-Si:H layers deposition conditions (temperature, doping, hydrogen content, etc.) but also to following steps: thermal and light treatments during TCO deposition, hydrogen plasma treatments [2–5].

In this paper, we study the evolution of passivation properties induced by thermal post-treatments. For this purpose several configuration of samples were prepared (monolayer and stacks of both doped and undoped layers, symmetric samples and SHJ cell precursor) using a semi-industrial PECVD/PVD cluster for both doped and undoped amorphous silicon layers, as described elsewhere [6]. We used an upgraded Quasi Steady-State Photoconductance tool whose chuck enables samples heating up to 180°C. It was thus possible to monitor the evolution of minority carrier injection level dependent passivation properties (effective lifetime and implied open-circuit voltage) as a function of time for several given temperatures. We find that the effective lifetime of samples passivated with undoped a-Si:H layers follows stretch exponential laws as proposed by DeWolf [7], which is not the case for samples passivated with doped layers and undoped/doped stacks that show a clear passivation degradation at temperatures higher than 160°C. The experimental results are coupled together with numerical and semi analytical simulations to determine the device parameters (interface, amorphous bulk layers properties) that are modified by the thermal treatments.

[1] http://panasonic.co.jp/corp/news/official.data/data.dir/2013/02/en130212-7/en130212-7.html

[2] S. De Wolf, B. Demaurex, A. Descoeudres, C. Ballif, Physical Review B 83, 233301 (2012). DOI: 10.1103/PhysRevB.83.233301

[3] B. Demaurex, S. De Wolf, A. Descoeudres, Z. C. Holman, and C. Ballif, Appl. Phys. Lett. 101, 171604 (2012). DOI: 10.1063/1.4764529

[4] W. Favre, J. Coignus, N. Nguyen, R. Lachaume, R. Cabal, D. Muñoz, accepted for publication in Appl. Phys. Lett (2013)

[5] M. Mews, T. F. Schulze, N. Mingirulli, and L. Korte, Appl. Phys. Lett. 102, 122106 (2013). DOI: 10.1063/1.4798292

[6] S. Martín de Nicolás, J. Coignus, W. Favre, J. P. Kleider, D. Muñoz, accepted for publication in Solar Materials & Solar cells (2013). DOI : 10.1016/j.solmat.2013.03.010

[7] S. De Wolf, S. Olibet, and C. Ballif, Appl. Phys. Lett. 93, 032101 (2008). DOI: 10.1063/1.2956668

Keywords : a-Si:H/c-Si heterojunction solar cells, passivation properties, thermal annealing

Fr-B2.5 11:50–12:10

Optical Enhancement in a-Si:H/a-SiGe:H Tandem and a-SiGe:H Single-junction Solar Cells

Hung-Jung Hsu, Shin-Wei Liang, Cheng-Hang Hsu, and Chuang-Chuang Tsai

Department of Photonics, National Chiao Tung University, 1001 University Road, Hsinchu, Taiwan

Optical management is important in high-efficiency thin-film solar cells. Yan et al. reported an world record efficiency of 16.3% by using a-Si:H/a-SiGe:H/μc-Si:H triple-junction configuration with efficient light management scheme [1]. Several issues such as thickness and opto-electrical properties of doped layers, bandgap grading in a-SiGe:H absorbers, intermediate reflector and back reflecting structure are critical to photon harvesting in solar cells.

Typically, a TCO layer was employed between n-layer and Ag contact to mitigate the optical loss in long-wavelength region [2]. Disadvantages of this approach are the inevitable ex-situ sputtering step and the damaged interface by ion bombardment. The n-type microcrystalline silicon oxide (μc-SiOx:H(n)) was reported to be a promising alternative for back TCO and n-layer in a-Si:H single-junction solar cells [3]. Moreover, the μc-SiOx:H(n) can serve as intermediate reflecting layer in multi-junction cells, which enhanced the utilization of solar spectrum [4]. The major challenge would be the substantially decreased conductivity of μc-SiOx:H(n) as oxygen content increased.

In this work, multiple light management schemes were employed in a-Si:H/a-SiGe:H tandem solar cells. More details on the effect of μc-SiOx:H(n), opto-electrical properties of doped layers and back TCO on the optical enhancement in a-Si:H/a-SiGe:H tandem cells will be discussed.

The deposition system is a 27 MHz single-chamber PECVD system with NF3 in-situ plasma cleaning and a load-lock chamber. The CO2, SiH4, PH3 and H2 are the source gases for preparing μc-SiOx:H(n). Solar cells were deposited on textured SnO2:F coated glasses in the superstrate configuration. The cells were characterized by AM1.5G illuminated J-V measurements and the external quantum efficiency, respectively.

The μc-SiOx:H(n) was served as part of tunnel recombination junction (TRJ) in a-Si:H/a-SiGe:H tandem solar cells. The lower refractive index nature of μc-SiOx:H(n) than adjacent Si layers created internal reflection. The cell employed μc-SiOx:H(n) as TRJ exhibited an increase in current density of a-Si:H top cell from 8.1 to 8.6 mA/cm2. Furthermore, the current density of the bottom cell increased from 9.3 to 10.5 mA/cm2 as the μc-SiOx:H(n) was employed. Presumably the μc-SiOx:H(n) facilitated the crystallization of the subsequent μc-Si:H(n) and thus improved carrier recombination at the TRJ.

Moreover, the μc-SiOx:H(n)/Ag back reflector (BR) was employed to improve light management in a-Si:H/a-SiGe:H tandem cells. The μc-SiOx:H(n) served as the replacement of n-layer of bottom cell and back TCO simultaneously, which contributed all PECVD process and process simplification due to non-necessity of TCO sputtering. The cell employed μc-SiOx:H(n)/Ag as BR exhibited comparable cell performance of 8.9% than the cell employed a-Si:H(n)/ITO/Ag as BR having efficiency of 9.1%. Further optimization on the oxygen composition in μc-SiOx:H(n) film and the current matching in tandem cells resulted in an efficiency of 10.03%.

[1] B. Yan, et al. Jiang, Appl. Phys. Lett. 99 (2011) 113512

[2] F.-J. Haug, et al. J. Appl. Phys. 104 (2008) 064509

[3] P. Delli Veneri, et al. Appl. Phys. Lett. 97 (2010) 023512

[4] A. Lambertz, et al. J. Appl. Phys. 109 (2011) 113109

Keywords: microcrystalline silicon oxide, amorphous silicon germanium, tandem solar cells, back reflector, light management

Fr-B2.6 (invited) 12:10–12:40

High-efficiency Amorphous/Crystalline Silicon Heterojunction Solar Cells

Stefaan De Wolf, Bénédicte Demaurex, Antoine Descoeudres, Jonas Geissbuehler, Zachary C. Holman, Johannes Seif, and Christophe Ballif

Ecole Polytechnique Fédérale de Lausanne (EPFL), Institute of Microengineering (IMT), Photovoltaics and thin film electronics laboratory, Rue A.-L. Breguet 2, CH-2000 Neuchâtel, Switzerland

Silicon heterojunction solar cells consist of thin hydrogenated amorphous silicon (a-Si:H) layers deposited on crystalline silicon wafers. This design enables energy conversion efficiencies well above 20% at the industrial production level, and is being pursued by an increasing number of groups in recent years. The key feature of this technology is that the metal contacts, which are highly recombination active in traditional, diffused-junction cells, are electronically separated from the absorber by insertion of a wider bandgap surface passivation layer. This enables the record open-circuit voltages typically associated with heterojunction devices without the need for expensive patterning techniques. In this presentation we discuss the key points to be considered when aiming for such high-efficiency devices.

First, we briefly elucidate device characteristics. This is followed by a discussion of each processing step, with a special focus on the properties of hydrogenated amorphous silicon films of only a few nanometer thin as such wider bandgap layers. We find that the ideal film should feature an atomically sharp interface with the wafer underneath, and that the bulk of the film should consist of material as close as possible to the amorphous-to-crystalline transition. We show these possibly conflicting deposition conditions can be resolved by careful monitoring of the plasma conditions, and interrupting the silane plasma to insert brief hydrogen plasma etching steps. With this sequence high-grade passivation layers can be obtained of only a few nanometer thin. We also show that from defect perspective this a-Si:H/c-Si interface has no unique features compared to the a-Si:H bulk, which allows to use such interfaces as sensitive probe for a-Si:H bulk defects. Next, we discuss the effects of film doping, and subsequent transparent conductive oxide (TCO) deposition on the electronic passivation properties of the a-Si:H/c-Si interface. These steps are required to fabricate respectively electron and hole collecting layers and external contacts. Here, we show that on the one hand film doping can lead to Fermi-level induced defect generation. On the other hand, the latter TCO sputtering can also lead to increased defect levels, but now more due to UV light exposure during processing. Despite this, by careful layer optimization and using floatzone wafers, we show independently confirmed conversion efficiencies of respectively 22.1% (n-type wafers) and 21.4% (p-type wafers) for 4 cm2 devices. These devices are fabricated with fully industrially compatible processing techniques. Finally, future trends are pointed out, where we also outline our strategy to lower parasitic absorption in both short- and long-wavelength regime by development of new window layers and transparent conductive oxides with increased transparency.

Keywords: silicon, high-efficiency, heterojunction, passivation, solar cells

ICANS 25