Enhanced photocurrent in thin-film amorphous silicon solar cells via shape controlled three-dimensional nanostructures

Thin-film solar cells (TFSCs) have well-established advantages that allow them to make a significant contribution to electricity generation and to ultimately compete with traditional energy sources for powering the grid. These advantages include their lower manufacturing cost as well as their versatility in the types of applications [1]. Nevertheless, further reduction in the $/watt-peak cost is needed to achieve grid-parity. An attractive approach to achieve this is in TFSCs is by increasing the cell conversion efficiency via enhanced optical absorption without significantly adding to the manufacturing cost.

Amorphous Si is a technologically mature thin-film photovoltaic (PV) material, and Si is more abundant in nature than elements such as tellurium or indium that make up other photovoltaic materials such as CdTe and CIGS [2, 3]. However, the a-Si thin-film absorber layer would need to be thick enough to absorb as much of the solar radiation as possible while still being thin enough to lessen the Staebler–Wronski effect (SWE) that results in a degradation of the solar cell conversion efficiency due to increased defects of the hydrogenated amorphous Si i-layer with light exposure [4]. Thus, thin-film a-Si solar cells absorb fewer photons than conventional wafer-based Si solar cells and hence are less efficient. Light trapping photonic structures can be useful in keeping the absorber layer adequately thin to mitigate light-induced degradation, while at the same time attaining good optical absorption and high photocurrent [5]. In this paper we investigate structures in which the optical absorption volume is significantly increased while the carriers still have a small distance to travel to the collecting junction. In other words, the electrical performance is decoupled from the optical performance of the solar cell.

Since thin-film photovoltaic (PV) a-Si cells have an absorber layer thickness that is of the order of or less than the wavelengths of interest, the geometric Lambertian scattering approach for light trapping [6] would not be as effective in this case. This approach increases absorption by randomizing light based on statistical ray optics. Also, the typical surface texturing structures [7, 8] such as pyramidal structures used to enhance light absorption in wafer-based silicon solar cells due to wide angle light scattering and surface double-bounce are not suitable for use for thin-film solar cells because of their large-scale geometries. In this paper, we evaluate three-dimensional (3D)-wrapped nano-scale structures for enhancing the optical absorption of for thin-film single junction (TFSJ) a-Si solar cells. Recently, silicon nanostructures such as nano-cone and Si nano-pillar or nanowire array structures have suggested a potential for excellent antireflection and light trapping ability based on simulation [9] and optical characterization [10]. Incorporating these nano-features could provide a potential for thin-film Si based solar cells to have enhanced light absorption by increasing the optical path length. In this work device structures are simulated, characterized, and fabricated, to show that for a thin-film a-Si cell the photocurrent can be significantly enhanced demonstrating that these 3D nanostructures have substantial merit for thin-film solar cells.

The photocurrent enhancement of a-Si solar cells previously reported using nano-cones and nanowires [9, 11] were based on nano-patterning techniques that are not readily suited for cost-effective scale-up. To enable large-scale production, any nano-patterning technique should possess the following features:

• (1)  
  Low overall cost based on high throughput processes with low consumable cost.
• (2)  
  Ease of controlling the geometric features (shape, size, aspect ratio, pattern density, etc) of the pattern to optimize the subsequent deposition of a-Si and electrode materials, and the resulting the light trapping.

Low cost patterning. In this paper, we use jet and flash imprint lithography (J-FIL) as a fabrication approach for controlled 3D nano-patterning. J-FIL was developed in recent years for low cost and high throughput large-scale nanomanufacturing [13–15] for a cost-sensitive application, namely bit patterned media for hard disk drives. If J-FIL technology is appropriately scaled up to pattern large glass substrates, it has the potential to achieve a total cost of less than $2 m⁻² that could revolutionize thin-film PV fabrication (see table 1). J-FIL uses low viscosity resists and an ink-jet based resist deposition technique. The low viscosity materials lead to a high-speed patterning process, and the ink-jet dispense leads to very low resist consumption. Additionally, creation of template (mold) replicas can lead to low cost templates for J-FIL. These aspects of J-FIL—namely high throughput, low resist consumption, and use of template replicas—have the potential to create a very low cost structure as shown in table 1.

Controlled nano-patterning. The J-FIL process is also good at tailoring the geometric features of the nano-patterns. The 3D pattern on the template can be tailored by various lithographic methods. Figure 1 shows three SEM images of pyramids with different aspect ratios fabricated by J-FIL with the aspect ratio increasing from the first to the third figure.

Zoom In Zoom Out Reset image size

Simulation results show that high aspect ratio nano-pillar structures are superior to other 3D structures in photocurrent enhancement. However, it is very difficult to obtain good conformal deposition of a-Si and electrode films on such pillar structures. For instance, the a-Si layer tends to be much thicker on the pillar top and thinner on the bottom and on the sides. The higher the aspect ratio of the pillars and the higher the density of the pillars, the more severe this problem is. Therefore, for this paper, we have adopted a pyramid structure, which can enable both good conformability and good light trapping. High aspect ratio pyramids can come close to achieving the photocurrent enhancement of nano-pillar structures, while are more suited for scalable fabrication. Therefore, we have chosen this particular 3D geometry for fabricating high-efficiency a-Si solar cells.

In summary, J-FIL enables high throughput and controlled 3D nano-patterning with the potential for the cost structure needed to make this approach viable for production of nano-patterned a-Si solar cells.

To obtain a better insight into the optical effects such as reflection, light trapping, and absorption as well as to have guidelines for optimal design choices, optical simulations are performed in addition to the experimental work.

Simulating or predicting the behavior of nanostructured thin-film a-Si PV cells becomes a non-trivial problem that requires the solution of Maxwell's equations for electromagnetic waves in order to obtain accurate and representative results. In the simulation part of this work the reflected and transmitted light waves need to be calculated from the incident field for these scattering problems. The patterned structures that we investigate in this study are periodic with horizontally periodic boundary conditions. This allows us to use a rigorous coupled wave analysis (RCWA) algorithm [16] where a periodic permittivity function is represented using Fourier harmonics, where each harmonic is related to a coupled wave. Recently, the RCWA technique has been used to model metal–dielectric stack-coated sub-wavelength metallic gratings for selective absorbers for solar thermal applications [17]. This technique or algorithm is available in the RSOFT photonics design software package used in simulations for this work. This analytical technique enables full vectorial Maxwell's equations to be efficiently solved in the Fourier domain and the spatial field distribution to be derived from the Fourier harmonics. An otherwise direct solution of Maxwell's equations with boundary conditions would be a computationally involved task. Using the periodicity property of the simulated structures the incident field in a periodic layer according to Bloch's theorem [18] could be expressed as follows (plane-wave incidence is assumed):

$$\begin{eqnarray} &&\boldsymbol{E}(x,y)=\hat {\boldsymbol{z}}{E}_{\Lambda }(x,y){\mathrm{e}}^{-\mathrm{i}({k}_{x,0}x+{k}_{y,0}y)} \end{eqnarray} \tag{ 1 }$$

where the electric field E(x,y) is expanded as a periodic function with period Λ_x in the x-direction and a propagation constant k_x,0 and with period Λ_y in the y-direction and a propagation constant k_y,0. E_Λ(x,y) is the periodic electric field that propagates in the xy-plane only and can be expressed as a double Fourier series expansion in x and y with coefficients a_p,q representing the dependence on the z-direction:

$$\begin{eqnarray} &&{E}_{\Lambda }(x,y)={\Sigma }_{p}{\Sigma }_{q}{a}_{p,q}{\mathrm{e}}^{-\mathrm{i}(\frac{2 \pi }{{\Lambda }_{x}}p x+\frac{2 \pi }{{\Lambda }_{y}}q y)}. \end{eqnarray} \tag{ 2 }$$

Similarly, the magnetic field can be expressed as

$$\begin{eqnarray} &&\boldsymbol{H}(x,y)=\hat {\boldsymbol{z}}{H}_{\Lambda }(x,y){\mathrm{e}}^{-\mathrm{i}({k}_{x,\alpha }x+{k}_{y,\alpha }y)} \end{eqnarray} \tag{ 3 }$$

where the periodic magnetic field H_Λ(x,y) can be expanded in a double Fourier series with coefficients b_p,q,

$$\begin{eqnarray} &&{H}_{\Lambda }(x,y)={\Sigma }_{p}{\Sigma }_{q}{b}_{p,q}{\mathrm{e}}^{-\mathrm{i}(\frac{2 \pi }{{\Lambda }_{x}}p x+\frac{2 \pi }{{\Lambda }_{y}}q y)}. \end{eqnarray} \tag{ 4 }$$

Using the above formulation in the transverse form of Maxwell's wave equations with the time harmonic factored out yields an eigenvalue problem formulation. Modal transmission line theory and methods are used to solve the boundary-value problem [19].

In this paper we model these 3D nanostructures and we compare the experimental results and trends with our optical simulation. We also compare the simulated and experimental results of the 3D nanostructured solar cell with planar and Asahi glass TFSCs. The goal of this work is to illustrate nanostructure designs that demonstrate a manufacturable or feasible 3D structure for enhancing optical absorption in TF a-Si solar cells.

As discussed earlier, the 3D nano-features were fabricated using J-FIL, a high throughput nano-imprinting technique [13, 15, 20, 21]. Recently, imprinting or embossing has been proposed as a method for texturing or patterning thin-film solar cells (e.g. for patterning a back reflector or for texturing) for optical enhancement [22, 23]. A typical process sequence for patterning the back reflector (also employed in this study) is as follows: the glass substrates are first cleaned in piranha solution and are subsequently rinsed and dried. Then an adhesion promoter is deposited using either a vapor deposition process or spin-coating. The substrates are subsequently imprinted using nano-imprint lithography to pattern or emboss the resist with the required features. The residual resist layer is subsequently etched with oxygen plasma reactive ion etching (RIE) and then the glass is etched using plasma RIE. Finally the resist is removed using an oxygen plasma asher. The thin-film hydrogenated amorphous Si layer (a-Si:H) is deposited by plasma-enhanced chemical vapor deposition (PECVD) at low temperatures ( ∼ 140–230 °C), which is the most common technique used to deposit these thin-film absorbers [24]. The Ag metal reflector and transparent conductive oxide (TCO) layers [aluminum zinc oxide (AZO) and indium tin oxide (ITO)] are deposited by sputtering. For the 3D pyramid nanostructures the deposited a-Si thickness is 178 nm as measured on a planar surface, this is typically somewhat thinner for a nanostructured surface for the same deposition conditions; the actual a-Si thickness on the nano-pyramid facets is  ∼ 140 nm. The deposited ITO thickness is 100 nm, and the deposited AZO thickness is about 20 nm. Various pyramidal nanostructures with different aspect ratios were fabricated as illustrated by the sample nanostructures shown in figure 1. Also, amorphous Si nano-pyramid TFSCs with feature size base widths ranging from  ∼ 250 to 460 nm, feature height in the range of 220–600 nm, and a pitch in the range of 530–650 nm have been tested. However, we report on a nano-pyramid structure that consistently yielded good results with the following feature parameters: a nanostructure pitch of  ∼ 650 nm with a feature height of approximately 381 nm and a base width of 450 nm. In order to obtain controlled 3D pyramid patterns, two possible approaches are proposed. In either case the templates are created by using a combination of electron beam patterning followed by dry and wet etching techniques to achieve proper control of the desired 3D structures. Then the template pattern can be transferred into the PV glass substrate by one of the following approaches.

• (1)  
  A 3D patterning of the imprint resist followed by a controlled etch process that matches the etch rate of the resist and the underlying glass thereby transferring the shape of the resist into the glass while substantially retaining the 3D shape.
• (2)  
  A one step process where the imprint resist is engineered to be a functional material that is left behind on the substrate and the resist may then be encapsulated by a thin layer of oxide to prevent outgassing of organic materials during subsequent deposition.

The final choice of approach will be dictated by the cost and practical integration challenges associated with these two approaches. Here Approach 1 was chosen. All the fabricated thin-film solar cells investigated in this work had identical active areas. Electrical characterization of the fabricated solar cell devices was performed using current–voltage (I–V) measurements as well spectral response measurements. Optical absorption measurements were also conducted.

Three-dimensional photonic nanostructures have been investigated in this work, namely nano-pyramids for experimental and modeled performance results. In addition to their low surface reflectance loss [25–27] it has been recently shown that nanostructures such as Si nanowires or nano-pillars may greatly enhance the optical path length of the incident solar radiation, and hence, have the potential to demonstrate very good light trapping properties [26]. According to recent studies, a nano-coaxial pillar design for amorphous Si solar cell structures may have a good potential for the enhancement of conversion efficiency [11, 28–30] by enhancing the optical absorption and light coupling into the amorphous Si material. Regular pyramidal photonic nanostructures, which are the main structures studied in this paper, were investigated for these TFSJ a-Si solar cells experimentally as well as by using computer simulations; these nanostructures are also simpler to manufacture compared with nano-pillars or nano-cones by nano-imprint patterning and etching. Hence, these nanostructures form the core of our study in this paper. The studies conducted in this paper focus on optical absorption and photocurrent enhancements in the nanostructured solar cell device compared with more conventional a-Si devices, which should result in directly proportional enhancement in the solar cell conversion efficiency assuming that the electronic properties of the solar cell (i.e. recombination, resistance losses, etc) remain the same.

3.1. Global optimization of 3D photonic structures

3.2. Simulation results for optical absorption and photocurrent

The quality and scattering properties of the light-facing broadband transparent conductive oxide (TCO) in thin-film single junction amorphous Si solar cells can result in a good light confinement capability allowing for enhancement of the conversion efficiency [39, 40]. Hence, in this work we also fabricated thin-film a-Si solar cells on rough Asahi-type TCO-coated glass for comparison with our 3D photonic structures. The finite-difference time domain (FDTD) method [41] is used to simulate the optical absorption of the thin-film a-Si solar cell with this random textured TCO on glass solar cell device structure. A cross-sectional schematic of the pyramidal nano-patterned thin-film a-Si solar cell is shown in figure 2. The corresponding cross-sectional SEM micrograph of the nano-patterned etched glass forming nano-pyramids before and after the deposited back reflector, TCO, and amorphous Si is shown in figures 3(a) and (b), respectively.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In order to investigate and understand the potential of these nanostructured thin-film a-Si solar cells 3D electromagnetic simulations described previously in section 1 of this paper were employed. As shown in figure 4, these nano-pyramid structures showed significant photocurrent enhancement over the planar or flat TCO glass a-Si TFSC as well as an enhancement over Asahi textured TCO glass a-Si TFSC. The simulated photocurrent enhancement for the nano-pyramid a-Si solar cells was  ∼ 90% and  ∼ 40% compared with the simulated photocurrent for the planar and Asahi textured glass a-Si solar cells, respectively. A slightly greater a-Si thickness was used for the planar and Asahi glass TFSCs of  ∼ 230 nm and  ∼ 205 nm, respectively, compared with  ∼ 140 nm for our 3D nanostructures to match our experimental parameters showing thinner deposited a-Si layers for the nanostructured TFSCs. The nanostructured and planar a-Si TFSCs have a glass substrate design while the Asahi glass a-Si TFSC has a superstrate design.

Zoom In Zoom Out Reset image size

To obtain the photocurrent the number of electron–hole pairs generated at a certain wavelength and collected at the junction electrodes is calculated, this can be given by:

$$\begin{eqnarray} &&{n}_{\mathrm{ph}}(\lambda )=\frac{\lambda }{h c}\mathop{\sum }\limits _{i}{\eta }_{c}s(\lambda ){\alpha }_{i}(\lambda ) \end{eqnarray} \tag{ 5 }$$

where h is the Planck constant, c is the speed of light, λ is the wavelength, S(λ) is the incident spectrum (for the simulations performed in this work the AM1.5G spectrum was used), and α_I is the absorption spectrum in layer i. In these simulations η_c, the carrier collection efficiency, is assumed to be equal to 1 or 100% (i.e. the internal quantum efficiency is 100%) as in this work only optical phenomena in the solar cell device are investigated/considered and not recombination issues. The total number of electron–hole pairs collected can be expressed as the following:

$$\begin{eqnarray} &&{N}_{\mathrm{ph}}=\int {n}_{\mathrm{ph}}(\lambda )\hspace{0.167em} \mathrm{d}\lambda , \end{eqnarray} \tag{ 6 }$$

and the photocurrent can then be computed as:

$$\begin{eqnarray} &&{J}_{\mathrm{ph}}=e{N}_{\mathrm{ph}} \end{eqnarray} \tag{ 7 }$$

where e is the electron charge density.

The simulation results show a significant improvement in absorption due to the 3D patterned nano-pyramid structures over the planar a-Si cell and even show an improvement over the Asahi textured glass a-Si cell. As shown by the absorption simulations (figure 5) this enhancement is broadband. Our results are in good agreement with the recent findings in [42] where it is reported that the surface reflection can be lowered due to the enhanced absorption in a broad spectral range (400–1000 nm) by forming sub-wavelength nano-patterned Si structures that were formed in that case by nano-sphere lithography for hetero-junction crystalline Si solar cells. From the simulated absorption results, it can be concluded that the merit of the 3D nanostructures is mainly in enhancing the optical absorption in the 590–750 nm wavelength range, which is the critical part of the spectrum for improving the performance of thin-film a-Si solar cells.

Zoom In Zoom Out Reset image size

In figure 6, showing the simulated external quantum efficiency for various cell structures, the nanostructured thin-film a-Si solar cells as well as the standard Asahi glass a-Si solar cell show enhancement in the long-wavelength response as compared with the planar a-Si solar cell structure. The 3D nano-patterned pyramidal structure a-Si TFSC shows a slight enhancement in the long-wavelength response compared with the Asahi textured glass a-Si TFSC as well as some enhancement in the mid-wavelength response  ∼ 590 nm. Moreover, because of the thick TCO for the Asahi glass superstrate structure there are parasitic absorption losses in the short-wavelength response; hence, the 3D nano-patterned a-Si solar cell shows a significant enhancement in the short-wavelength response as well.

Zoom In Zoom Out Reset image size

Based on the simulation results above, it can be concluded that 3D-wrapped pyramidal nanostructure a-Si solar cells yield a better enhancement in photocurrent compared with plasmonic nanostructures for a-Si TFSC, e.g. those reported in [37, 43, 44]. Single-pitch-design plasmonic nanostructures give more narrowband improvements [43], and they typically show a photocurrent enhancement that is similar to that of the Asahi U-type textured glass a-Si TFSC [44]. However, as shown by the simulation results, our 3D nano-patterned pyramidal structures yield higher photocurrents compared with the Asahi textured glass a-Si solar cells. In the simulated E-field in figure 7, it is shown that the field is dampened as it goes through the 3D nanostructured a-Si device indicating enhanced optical absorption in the nano-patterned thin-film solar cell. The simulation results shown helped guide the design and optimization for the experimental work involving the fabrication of these 3D nanostructured a-Si TFSCs.

Zoom In Zoom Out Reset image size

3.3. Experimental optical and electrical results

In this section the fabricated a-Si TFSC structures are investigated based on the characterization results. It has recently been shown that the morphology of textured TCO (LP CVD ZnO) on glass substrates strongly affects the quality of the microcrystalline Si absorber material and the solar cell performance [45]. V-shaped morphologies were found to strongly enhance light trapping but at the same time degrade the growth and quality of the absorber film by helping to induce micro-cracks. Hence, V-shaped morphologies were smoothed to a more U-shaped form by optimizing the plasma treatment process and time. In our work amorphous Si is used, and hence micro-cracking should be less of an issue. Nevertheless, we optimized the nano-pyramid surface morphology in order to avoid or mitigate the degradation of the amorphous Si P–I–N and N–I–P solar cell devices. This type of morphological performance degradation mainly shows up in the open-circuit voltage (V_oc) and fill factor (FF) of the device and not in the absorbed photocurrent. We generally see an improvement in photocurrent for an N–I–P a-Si cell design (N-layer deposited first) on a substrate configuration where the P-type a-Si layer is facing the incident light. We do not see this significant photocurrent enhancement with a P–I–N superstrate configuration where the P-layer is deposited first, on the TCO on the light-facing side of the glass superstrate. This is probably due to stresses and delaminations that we observe in the ITO layer when deposited directly on these pyramidal structures. This is not the case for the substrate configuration, as the metal back reflector is deposited first, which seems to be less prone to the disadvantageous stress and CTE mismatch effects that may take place during the a-Si deposition on these patterned structures.

The measured absorption curves in figure 8 show an 88% average weighted absorption (AWA) for the 3D pyramidal nanostructure a-Si TFSC followed by the Asahi textured glass a-Si TFSC with an AWA of about 54%. The simulated absorption plots shown in the previous section and experimental absorption plots below show good agreement in the short wavelength and visible parts of the spectrum.

Zoom In Zoom Out Reset image size

We also characterized the electrical characteristics of the solar cell. The I–V curves in figure 9 show a clear advantage of the 3D pyramid nanostructured a-Si TFSC over the Asahi textured glass a-Si TFSC and a significant enhancement over the planar or flat glass a-Si TFSC. A substantial short-circuit current percentage enhancement of  ∼ 80% was achieved for our nanostructured a-Si TFSC— typically enhancing J_sc from  ∼ 7 mA cm⁻² for the flat a-Si TFSC to  ∼ 13 mA cm⁻² for the nano-pyramid wrapped a-Si TFSC structure. The highest measured J_sc was 13.8 mA cm⁻² for these nano-pyramid thin film a-Si solar cells. We have repeated this experiment several times and this photocurrent enhancement is consistently achieved. This photocurrent enhancement is similar to that recently obtained in the literature [35, 36] for nano-coax-based solar cells for a similar a-Si thickness deposited on the sides of the nano-pyramids. However, in addition to the scalable manufacturability of the fabrication process used in this study our nano-pyramids only have a feature height of about 380 nm compared with 1–2 μm for the nano-pillars/nano-coax solar cells reported in [35, 36]. The very thin a-Si of  ∼ 140 nm deposited on the nano-pyramids is much thinner than the typical 500 nm thickness used for planar a-Si solar cells, which should mitigate the SWE that degrades the a-Si TFSC performance. We have also achieved conversion efficiencies approaching 6% for these nano-pyramid wrapped a-Si solar cells, which is also comparable to the result reported in [36], and which also shows a  ∼ 50% enhancement in relative conversion efficiency compared with their planar a-Si solar cell counterparts. Further process improvements are needed, particularly for the a-Si and TCO deposition parameters and conditions to improve solar cell performance.

Zoom In Zoom Out Reset image size

In order to get a better understanding and further investigate the photocurrent enhancement of the pyramidal nanostructured a-Si solar TFSCs, the external quantum efficiency (EQE) response was measured for all three types of cell (nanostructured, Asahi textured glass, and flat glass a-Si TFSCs) as shown in figure 10. The results show that the EQE of the 3D nano-patterned cell is the highest for the nano-patterned cells and has a better broadband wavelength response than the Asahi glass a-Si TFSC. Even though the simulated and experimentally measured optical absorption (figures 5 and 8, respectively) show good agreement, the simulated EQE of figure 6 is significantly higher than the experimental EQE (figure 10). This can be attributed to the fact that the experimental fabricated solar cells exhibit some shunting behavior as shown by the I–V curves of figure 9. In addition, there are also recombination losses in the a-Si bulk and at the a-Si surface/interfaces, which will inevitably result in a lower internal quantum efficiency (IQE), and consequently, this will result in a lower experimental EQE compared with the simulated counterpart (recall that the IQE is assumed to be 100% in the simulations, and no shunting effects were assumed in the EQE simulation). The results demonstrated in this work can be further improved by optimizing the back metal, TCO, and a-Si deposition parameters to avoid recombination losses in the a-Si i-layer and to minimize interface states in order to achieve open-circuit voltages closer to 0.9–1 V.

Zoom In Zoom Out Reset image size

In summary, controlled 3D nano-patterns were fabricated on glass substrates using the jet and flash imprint lithography technique for thin-film a-Si solar cells. This patterning technique provides high level of control of 3D nanostructures and has the potential to provide cost-effective solutions that are suited for fabrication of a-Si photovoltaic cells. Our nanostructured a-Si TFSC gave similar photocurrent enhancement to the best results previously reported in the literature but with almost one-third the nano-feature height, using a scalable nano-patterning process, and only  ∼ 140 nm a-Si absorber thickness. This can greatly mitigate the SWE which degrades a-Si module performance in the field. The chosen pyramidal nanostructures exhibit good optical absorption enhancement characteristics in a broadband response, and are suitable from the perspective of scalable fabrication including patterning and subsequent deposition techniques. The average weighted absorption of these nano-patterned thin-film amorphous Si solar cells has reached ∼90% for some of our fabricated pyramidal nanostructures, which is similar to the average weighted absorption of textured thick uncoated mono-crystalline Si wafer solar cells. Both simulated and experimental external quantum efficiency results show a significant improvement in both long- and short-wavelength responses by employing 3D nano-patterned pyramidal structures. This results in a 50% enhanced short-circuit current over planar a-Si solar cells and about 20% enhancement over Asahi textured glass a-Si solar cells. The difference between the simulated photocurrent enhancement and experimental short-circuit current enhancement as well as the difference between the simulated EQE and the measured EQE can be attributed to defects, interface states, and some shunting behavior in the deposited a-Si thin-film region, which requires further process optimization. Nevertheless, our results show that these controlled 3D nanostructures have the potential to be useful in low cost substrate designs for thin-film single junction a-Si solar cells.

This work was partially funded by the National Science Foundation Scalable Nanomanufacturing Program (NSF contract no. ECCS-1120823), the DARPA Tip-Based Nanofabrication Program (DARPA contract no. N66001-08-C-2040), the Texas Emerging Technology Funds Program, and by a research grant from Molecular Imprints, Inc. The authors would also like to thank the Institute of Electronics and Nanotechnology at Georgia Institute of Technology for their help and support with the amorphous Si deposition process.