Open Access
Add to BibSonomyAbstract
Thin-film hydrogenated amorphous silicon (a-Si:H) solar cells that are free-standing over a 2x2 mm area have been fabricated with thicknesses of 150 nm, 100 nm, and 60 nm. Silver nanoparticles (NPs) created on the front and/or back surfaces of the solar cells led to improvement in performance measures such as current density, overall efficiency, and external quantum efficiency. The effect of changing silver nanoparticle size and incident light angle was tested. Finite-Difference Time-Domain simulations are presented as a way to understand the experimental results as well as guide future research efforts.
© 2015 Optical Society of America
1. Introduction
One way to decrease the cost of solar cells is to reduce the amount of material used to create them. This is problematic in crystalline silicon (c-Si) based photovoltaics because of its indirect band gap. Reducing the thickness of the silicon (Si) in turn reduces the amount of sunlight absorbed. This decrease in physical thickness can be counteracted by increasing the optical path length inside a solar cell via a light trapping structure. c-Si commonly uses pyramidal structures for light trapping, which are created using a chemical etch [1]. This etch removes a relatively large amount of high quality c-Si and creates pyramidal structures several microns deep. In order to reduce materials costs, Si solar cell thicknesses need to be several microns or less [2], making this subtractive etching process inapplicable. Pyramidal structuring of the active layer also has the disadvantage of increasing the surface recombination rate, leading to a decrease in electrical performance. Current light trapping structures for a-Si:H based solar cells are based on textured substrates. This leads to deposited a-Si:H being a lower quality and negatively impacts electrical performance. Achieving better light trapping with this method requires increasing the aspect ratio of the textures, which will further reduce the a-Si:H quality [3]. This has led to an interest in creating light trapping structures that are not dependent on texturing [4]. Stuart and Hall first introduced a non-destructive method of increasing the photocurrent in silicon-on-insulator waveguides by the addition of plasmonic structures in the form of metal island films (MIF) at the front surface of the device in the mid-1990s [5]. In more recent years, interest in front-side plasmonic structures for enhanced solar cell efficiency has continued [6–12] while expanding to the backside [13–20] of the solar cell and even the inside [21–23]. These studies have produced results with varying success. The location of the NPs plays a large role in how they will effect a solar cell’s performance. NPs on the front of the cell are used as anti-reflection coatings that scatter light at high angles to improve light trapping, but they increase reflection below the NPs resonance. Rear located NPs avoid the reflection loss and still provide a high angle scattering structure. Several studies have compared the performance of front side and rear side NPs [24–27]. In more recent approaches to light management in thin-film solar cells, using photonic crystals or ordered plasmonic structures [28–32], researchers have demonstrated the ability to reach or exceed the light trapping limit of 4n2, where n is the refractive index of the absorber layer, as calculated by Yablonovitch [33].
In this study we use thin, suspended a-Si:H solar cells to study the impact of adding Ag NPs to the front and back of ultrathin solar cells. Because the cells are freestanding, the NPs can be added to front, back, or both sides of the cell without affecting the morphology of the active layer. By creating ultra-thin active layers of 60 nm, the interparticle spacing between the front and backside NPs can be reduced and the solar cell performance can benefit from their interaction. Research has shown that a region of increased electric field intensity is produced between two metallic nanostructures with interparticle spacing of a few hundred nanometers or less [34, 35]. Modeling has also shown that plasmonic structures on the surface and embedded in an absorber material produce enhanced absorption [36]. This work provides simulations and experimental data showing an increase in efficiency with the addition of Ag NPs on the front and back of the cell. The device fabrication is also discussed in detail.
2. Theoretical simulations
Finite-Difference Time-Domain (FDTD) simulations were performed using Lumerical’s FDTD Solutions [37]. Simulations were run to determine the reflection and transmission from thin-film a-Si:H solar cells with front and/or back NPs. Figure 1(a) is a typical simulation model for devices used in this study and was based on the experimental device geometry. The absorber layer was a-Si:H and the contact material directly above and below the a-Si:H is indium-tin-oxide (ITO) with thicknesses of 25 nm and 45 nm, respectively. These thicknesses for the ITO were chosen for reasons that will be mentioned in the fabrication sections. The Ag NP hemispheres were positioned above and below the cell as desired and a 10 nm SiO2 buffer layer was placed between the NP and ITO. Unless stated otherwise, simulations were run with a 60 nm a-Si layer and 120nm diameter NPs. The sidewalls were periodic boundaries, meaning the simulation acted as a regularly spaced array of NPs. The distance between the sides was 200 nm in order to minimize particle-particle interactions while maximizing surface coverage. The top and bottom boundaries of the simulation space were Perfectly Matched Layers (PML).
Fig. 1 (a) A 2D drawing of a typical 3D model used in FDTD simulations. Light is incident from above and the monitors at the top and bottom record the total power reflected and transmitted. (b) Plot showing the absorptance for cells with the three designs: front only NPs, back only NPs, and both sides NPs. (c) Plot showing the calculated absorptance from simulation with monitors placed at the interfaces of each constituent film. The plot contains the amount of light reflected from the top and transmitted through the bottom of the simulation space. Absorptance of the front and rear NPs, ITO films, and SiO2 films are also included. The absorptance in the a-Si:H layer is separated into three parts to represent the n-, i-, and p- layers with thicknesses of 20 nm, 25 nm, and 15nm respectively. The dashed black line represents the absorptance of the i-a-Si:H of a cell without Ag NPs. (d) Plot showing the effect of various sizes for NPs located on the front and back of a solar cell.
The FDTD mesh size was set using Lumerical’s conformal meshing algorithm except for the mesh around the NPs. A mesh size of 0.7 nm was used around the Ag NPs and this dimension was determined by comparing the scattering of a lone sphere in vacuum with that of the analytical solution using Mie theory. The refractive index for Ag used in these simulations was based on the Johnson and Christy values available in the software’s database. The refractive indices used for a-Si:H were derived from ellipsometry measurements on deposited intrinsic material. Unless otherwise indicated, the following simulations were performed with a monitor placed above and below the entire structure to capture the reflected and transmitted intensities. Absorptance was used as a performance metric and it was calculated as 1-R-T, where R is reflectance measured by the top monitor and T is transmittance measured by the bottom monitor. The simulation results were used to help analyze experimental data as well as guide future work. A free-standing solar cell structure allows for testing of a variety of designs. The designs tested in this work were front side only NPs, back side only NPs, and both side NPs. Figure 1(b) shows simulation results for these three designs. The simulations showed that enhancement by using NPs, with NPs on both sides showing a strong peak in absorptance around 620 nm, is expected.
An important question regarding our simulation setup was whether the reflectance and transmittance being recorded by the monitors is an accurate measure of the light absorbed by a-Si:H, or if a significant fraction of light is absorbed by the Ag, ITO or SiO2. Figure 1(c) compares the power absorbed by each layer in the entire structure (with both side NPs). The Ag NPs used in this model were 120 nm diameter hemispheres placed on the front and back. The top two shaded regions in Fig. 1(c) represent the reflection and transmission losses. Figure 1(c) also shows ~15% absorptance is due to the Ag NPs, ITO, and SiO2 (front and back) combined. The a-Si:H contributes to a majority of the absorption. Due to absorption in the i-a-Si:H being the only contributor to solar cell performance, estimating the losses in the p- and n-type layers is important. These losses are estimated by separating the a-Si:H into three regions; a top 15 nm to represent the p-layer, a middle 25 nm to represent the i-layer, and a bottom 20 nm for the n-layer. The same refractive indices were used for each layer. The black dashed line in Fig. 1(c) is the absorptance of the i-a-Si:H from a solar cell simulated with no Ag NPs. The addition of the Ag NP results is slightly less absorption in the i- layer below ~590 nm but a larger increase at the higher wavelengths near the plasmon resonance. Approximately half of the absorptance above ~690 nm is parasitic losses in the ITO, SiO2, and NPs. Figure 1(c) shows that the absorptance increase found in Fig. 1(b) is not only due to the NPs, which indicates that the absorptance will serve as a good metric for analyzing optical simulations. In all the simulations with NPs on the front there is a noticeable drop in absorptance below the NP resonance. This is due to the destructive interference between the scattered light and the incident light, which reduces the transmission into the solar cell at the shorter wavelengths [24, 38, 39].
Figure 1(d) shows the effect of varying the size of the Ag NPs on the front and back of the solar cell. Plotted are the absorptance spectra with various Ag NP diameters ranging from 60 to 160 nm. As expected, the peak of the plasmon resonance red-shifts with increasing particle diameter. A realistic solar cell would be covered with a distribution of Ag NPs with sizes including all of those simulated. In addition, NPs created experimentally were not ideal hemispheres. The combined effects of the wide range of NP sizes and their unique shapes leads to a broad spectral response, instead of a response localized to the resonance of a single NP size and shape. From these simulations one expects an increase in absorptance in the 500 to 800 nm range with the addition of Ag NPs to the front and back of a the solar cell depicted in the schematic.
The preceding simulations studied Ag hemispherical NPs aligned on the front and back, as shown in Fig. 1(a). The NPs created in this study will be randomly distributed, making the simulations with aligned NPs an oversimplification. Figure 2 plots simulated absorptance for Ag NPs on both sides with various center-to-center offsets. Figure 2(a) is a plot showing light polarized perpendicular to the direction of offset and b) is for light polarized with the offset. Figure 2(b) shows the absorptance enhancement peak shifting and broadening as the NP offset increases. Notice that the plot of the 80 nm offset in Fig. 2(b) begins to show the individual response of the NPs, which is shown in Fig. 1b). The simulations show that the disorder inherent in our method of Ag NP creation leads to a broader enhancement region with a reduced maximum intensity if the incident light is polarized parallel to the direction of offset. Figure 2(a) shows that light polarized perpendicular to the direction of offset results in almost no change to the absorptance of the system. These simulations continue to suggest that the addition of Ag NPs to our freestanding solar cells will enhance the spectral response in the 500 to 800 nm range.
Fig. 2 Plots showing calculated absorptance from simulations investigating the effect of offsetting the center of a frontside NP from the center of the backside NP where the polarization is (a) perpendicular and (b) parallel to the direction of offset.
Our simulations are in reasonable agreement with other models used to explore the effect of plasmonic structures colinearly aligned in the direction of the propagation of incident light. Lin et. al. showed an increased absorbance in a-Si films with the addition of rectangular Ag structures to varied combinations of the front, middle, and back side [36]. The cause of enhanced absorption at 642 nm is supported by a plot showing the electric field profile with higher intensities near the Ag structures, especially those in the center of the a-Si:H.
3. Materials and methods
3.1 Fabrication of free-standing solar cell
Freestanding p-i-n a:Si:H solar cells were prepared using a method similar to the one discussed in [40]. A 200 µm thick double-side polished 5-10 Ω-cm n-type <100> wafer with 100 nm of thermally grown dry oxide was backside patterned with typical openings of 2 x 2 mm by standard contact lithography followed by a 10:1 Buffered Oxide Etch (BOE). The BOE etch opens the windows as well as strips the oxide from the front side of the wafer. The front side was then covered with an oxide film of approximately 45 nm using a GSI Plasma Enhanced Chemical Vapor Deposition (PECVD) tool. This layer acts as an etch stop when removing selected areas of the silicon substrate. Next, an a-Si:H p-i-n junction was deposited using the same GSI PECVD, starting with the n-layer. Silane (SiH4) was the source gas, and Phosphine (PH3) and diborane (B2H6) were used as the doping gases. Both the doping gases were diluted in helium at 1.5%. For each layer the power was 50W, the frequency used was 13.56 MHz, the H2/SiH4 ratio was 14, and pressure was 4 Torr. To reduce cross-contamination problems associated with a single chamber process, a simple hydrogen plasma etch was used after the n-layer deposition [41]. The wafer was stored in the pumped down load lock during the hydrogen etch. A final 30 nm SiO2 top layer was deposited to protect the front side during the following chemical etch. The patterned backside was then exposed to Ethylene Diamine Pyrocatechol (EDP) at 110þC using a specially designed single side etch cell. EDP anisotropically etches silicon in the <100> crystallographic orientation. The etch removes the exposed Si until reaching the SiO2 etch stop layer where the etch rate is dramatically reduced. The wafer was next dipped in 15:1 BOE for 2 minutes to remove both SiO2 layers, thereby exposing the a-Si:H p-i-n junction. These steps are outlined schematically in Fig. 3(a)-3(e).
Fig. 3 Fabrication steps: (a) The starting wafer, (b) patterning, (c) SiO2/a-Si:H/SiO2 stack deposited, (d) EDP etch, (e) BOE dip, (f) ITO deposition.
Indium Tin Oxide (ITO) was then sputtered onto the exposed p- and n-type layers to act as transparent contacts, shown schematically in Fig. 3(f). ITO was sputtered using a Kurt J Lesker PVD75 at 200 þC on a 20 rpm rotating platen with a working distance of 15 cm. The pressure was set to 3 mTorr with the Turbo pump set to 50%. Argon and oxygen flow rates automatically adjusted to maintain the pressure. Oxygen flow was set to 5% of the total gas flow. The target was a Kurt J. Lesker 3 inch diameter 90-10 In2O3-SnO2 and the power was 100 W. A typical deposition involved a complete coating of the entire back surface of a quarter wafer section. A shadow mask is used to avoid shorting at the edges. For the purposes of this work, the ITO was made as thin as possible. A 25 nm front-side and 45 nm back-side ITO layer provided predictable results. Depositing thinner layers led to inconsistent test results. The back-side is made thicker because of the sloped region. Due to the non-conformal nature of sputtering, the thickness is not uniform on the backside and the slopes create regions of low conductivity.
The solar cells were tested using an Oriel Arc Lamp with a xenon bulb. Light current density versus voltage (JV) measurements are taken at 1000W/m2 using an AM1.5G filter. External quantum efficiency (EQE) was done using an Oriel Cornerstone 130 1/8m monochromator coupled to our light source. The current was recorded at 10 nm wavelength increments. A c-Si Hamamatzu S1781-12 diode was used to determine the incident power at each wavelength. Test results show that the devices are functional, although efficiencies are low. This can be attributed to the significant transmission losses that are associated with these very thin cells with no rear reflector, which was an anticipated loss. The goal of this work was not to set a record efficiency, but to explore the effects of NPs by doing a relative comparison. For this work three wafers were processed to test three different solar cell thicknesses. The three basic cells created have total a-Si:H layers of 150, 100, and 60 nm. The layer thicknesses and solar cell parameters for each set of cells is listed in Table 1.
Table 1. A summary of the thicknesses and measurement results of cells before enhancement. The thicknesses are calculated using deposition rates obtained from interferometric measurements and analysis of SEM cross sectional images.
3.2 Ag NP addition
The Ag NPs were fabricated by depositing a thin-film of Ag, generally less than 30 nm, and then annealing it. Relatively low temperatures around 200þC are enough to cause the Ag thin-film to coalesce and form metal island films [42]. Before the Ag film is deposited, e-beam evaporation was used to deposit a 10 nm SiO2 buffer layer on top of the ITO. The SiO2 buffer layer is deposited at room temperature, a working pressure of 5e-6 Torr, an accelerating voltage of 6 kV, and a rate of ~0.5 Å/s. There are two reasons for adding this thin SiO2 layer. First is to electrically isolate the NPs from the solar cell. Earlier experiments showed reduced performance with the NPs in contact with the ITO, which has been observed by others [43]. Second, the SiO2 acts as an NP forming layer. The substrate affects the formation of the NPs [44]. The SiO2 layer is not expected to have an impact on the optical performance of the cell. FDTD simulations show that the SiO2 has a very small effect, which is apparent in Fig. 1(c).
Then a film of Ag with a typical thickness ranging from 14 to 18 nm was evaporated by e-beam using similar conditions, except for changing the accelerating voltage to 10 kV. An aluminum shadow mask is used to prevent shorting between the front and back side and also to provide a contact region of ITO free of SiO2 and Ag. The device was then annealed in a convection oven in an air atmosphere. The annealing parameters are discussed in detail later in this section. A schematic cross-section of a cell with front and back NPs is shown in Fig. 4(a) and photographs of the tops of the solar cells are presented in Fig. 4(b)-4(j). Stresses are introduced with each thin-film deposition, and due to the free standing design of the cell, these stresses result in variations of the sample’s appearance. This is clearly shown in the series of Fig. 4(e)-4(g). For example, the silver deposition in Fig. 4(f) introduced wrinkles and annealing the film removed the stress induced by the Ag layer, which eliminated the wrinkles shown in Fig. 4(g). While these wrinkles are clearly visible, we do not believe that they made significant contributions to the cell’s electrical or optical performance due to their relatively large size and random orientation. The thinner wrinkled solar cells had comparable standard deviations for cell performance (Table 1) to thicker cells without wrinkles.
Fig. 4 (a) Schematic showing the cross-section of a solar cell with SiO2 and Ag on the front and back side. Part of the ITO is masked off from the SiO2 and Ag depositions so that it remains exposed for uninhibited electrical contact. (b)-(d) Photographs of 150 nm solar cells with ITO, 10 nm silicon dioxide + 16 nm of Ag, and annealed, (e)-(g), 100 nm solar cells, and (h)-(j) 60 nm. (k) A SEM cross-section of a 150 nm a-Si:H solar cell with Ag NPs on the front and back.
Figure 4(k) is an SEM cross section of the freestanding portion of a 150 nm a-Si:H layer cell. There is 25 nm of ITO on the top, 45 nm of ITO on the back and 10 nm of SiO2 with 16 nm of Ag on the front and back. This sample had been ramped to 200 þC, showing agglomerated NPs with Ag NPs on the front and back. (The bottom of the micrograph is displaying NPs formed on the sloped region of the supporting Si substrate.)
The samples were annealed by placing the solar cells in the oven at an initial temperature of 50 þC and then ramping up to the final temperature setpoint. Once the final setpoint was reached then the solar cells were immediately removed and allowed to cool to room temperature. Various final temperatures were tested on a set of similar solar cells. The goal is to use an anneal temperature that promotes agglomeration of the Ag film without damaging the solar cell. The results are presented in Table 2. An important trend was the reduction in Voc with increasing temperature. The Voc degradation could be the result of damage to the front and back contacts or to the a-Si:H material itself [2]. A final temperature setpoint of 170 þC resulted in an increase in Jsc without reducing the Voc. Therefore 170 þC was the temperature used for the rest of the experiments. Figure 5(a) shows an SEM image of the top surface of the solar cell from that had been annealed up to 170 þC. The software ImageJ [45] was used to analyze the number of particles and determine the area of each particle. An effective diameter was then calculated from the area of each NP by assuming they are roughly circular. These results are shown in Fig. 5(b). The Ag NPs have a surface coverage of approximately 28%.
Table 2. A table of key performance parameters for solar cells annealed by ramping the temperature from 50 þC to final temperatures ranging from 160 to 200 þC.
Fig. 5 (a) An SEM image of Ag NPs formed on the 10 nm SiO2 layer of a solar cell. The Ag film was initially 16 nm and the sample was heated in a ramp from 50 to 170 þC. (b) A histogram showing the size distribution of the Ag NPs from (a).
We next investigated the possibility that annealing the Ag films in air could lead to oxidation, thereby affecting the optical properties of the particles. We used transmission electron microscopy (TEM), scanning tunneling electron microscopy (STEM) and electron dispersive spectrometry (EDS) to acquire a more detailed view of a Ag NP fabricated with our method. Ag NPs were directly formed on an a-Si TEM grid manufactured by SiMPore Inc. An EDS scan on a single NP shows a coating approximately 3 nm thick on the perimeter of the Ag NP. EDS images are shown in Fig. 6. As expected, Ag was seen only on the NP while Si exists only on the sample holder. Oxygen was found throughout the sample, with higher concentrations around the edges of the NP. This indicates that the coating on the NP was likely a layer of oxide. A relatively simple method to determine if this oxide coating effects the optical properties of the Ag NP on an actual solar cell was to add a similar coating in the simulation model. Figure 6(e) is a plot of calculated absorptance vs wavelength showing two plots of identical models except for a 3 nm Ag oxide shell surrounding the 120 nm diameter Ag NP hemisphere. The real and imaginary values of the simulated Ag2O refractive index were estimated to be 2.5 and 0.11, respectively [46]. According to these simulations the oxidation occurring during anneal induces an approximate 15 nm red shift in the resonance. There is also an increase in absorptance by as much as 10% across wavelengths below 520 nm and above 610 nm. According to simulations run by Akimov and Koh [47], Ag NPs of our size with an oxide shell provide reduced enhancement over the photoactive region. Once the oxide shell reaches 6 nm and above we should expect to see negative enhancement. The increased absorptance we are reporting in our data is most likely due to losses in the Ag2O shell itself. In this simulation, the shell is also increasing the overall size of the hemisphere, which would also effect a change in absorptance and shift the resonance peak location towards longer wavelengths. Other researchers have also found deleterious effects of oxidizing Ag NPs over time and have shown that a thin film of MgF2 coating particles used as a backside reflector not only removed oxidation concerns but further improved cell performance [48].
Fig. 6 (a) A TEM image of a single Ag NP showing a thin film around the perimeter. (b), (c), and (d) STEM EDS image analyses for Ag, oxygen, and Si, respectively. Note that the detector is located to the right of these images. The left side of the NP appears to have less oxygen, but this is due to the location of the detector and the signal being blocked by the NP. Carbon also appeared in the scan, but this image is not shown because it is a known contaminate. (e) A plot of simulated absorptance of 120 nm diameter Ag NP with and without a 3 nm coating of Ag oxide.
4. Results and discussion
4.1 Solar cell Jsc, Voc, FF, and efficiency
The relative enhancement due to the addition of NPs was then analyzed with light JV and EQE measurements. Figure 7 shows external quantum efficiency (EQE) measurements for solar cells with total a-Si:H thicknesses of 150, 100, and 60 nm. For each solar cell data is presented for four configurations: no NPs, front only NPs, back only NPs, and both side NPs. The sides with NPs had a 10 nm film of SiO2 and 16 nm of Ag, while the sides without NPs only had the ITO. All solar cells underwent the annealing process, including the cell with no NPs. We observe several trends. First, cells with Ag NPs on the front show increased absorptance towards longer wavelengths and reduced absorptance towards shorter wavelengths as compared with the control cell and the cell with Ag NPs on the back only. This is expected from simulation and, as was mentioned earlier, is due to the scattered light below the NPs resonance destructively interfering with the incident light. A second trend is that when Ag NPs are present on the front and back, the cells perform better than when NPs are on either the front or the back alone. The EQE of the Ag NPs on front and back closely resemble NPs on the front alone. Third, both Jsc and efficiency increase more with the addition of Ag NPs for thinner absorber layers. Fourth, there is a difference in performance between front side only and back side only NPs. The back side NPs resulted in relatively poor performance at longer wavelengths. This may be due to the spacer layer thickness between the NPs and the a-Si. Many have found that back side NPs will perform better when they are located very close to the high index absorber and front side NPs perform better with a thicker spacer [16, 24, 25, 27, 38, 49]. This work did not investigate the effects of the spacer layer thickness, but it is important to understand that the thickness used does have an impact on the results.
Fig. 7 A plot of EQE measurements for (a) 150 nm, (b) 100 nm, and (c) 60 nm thick a-Si:H absorber layers. Each cell has EQE data for ITO contact only, front oxide and Ag NPs, front and back oxide with Ag NPs and back only oxide and Ag NPs.
Lastly, we attribute any decrease in performance to reflectance or parasitic absorption caused by the Ag NPs and their oxide. The majority of the enhancement is most likely a result of increased scattering. Any near-field effects from the Ag NPs would be predominantly confined to the SiO2 and ITO films and therefore not have a significant impact on absorption in the a-Si:H region. Even when the NP is in direct contact with the a-Si, the near field has been shown to have little effect due to the relatively long absorption length of a-Si:H [50]. While the change in Jsc is the main interest in this paper, there are variations in the Voc and FF that should be discussed. Finding a trend in the Voc is difficult due to the multiple variables that contributed to the value of the Voc, and one of those variables is light trapping [29, 51] In theory, a planar cell will have a lower Voc than a cell with light trapping. The light trapping factor in the Voc is interdependent on other variables and therefore the change in Voc cannot be described with a single factor. The variability in the performance of each cell also contributes to the changes in the Voc and FF. Table 1 provides standard deviations for cells without NP coating and before annealing. Some of the changes shown in Fig. 7 are greater than the reported standard deviations. The Ag films added additional variation to the samples, especially when comparing the front side and back side NPs. While all fabrication steps were kept the same for all samples shown in Fig. 7, the samples themselves are asymmetric at the perimeter of the free-standing section. As is shown in Fig. 4(a), the back side has sloped sidewalls. This leads to the same silver film and anneal process behaving slightly differently between these front and back side interfaces. Future work will have to be done to further understand the fabrication limits and tolerances of this free-standing structure.
Our solar cells were designed to minimize the distance between the Ag NPs on the front and back surfaces, and therefore our front side ITO was kept to 25 nm. A more traditional solar cell using an ITO front contact uses 80 nm thick ITO as an optimum thickness to reduce reflectance over the visible spectrum. In future work, we should compare our thin solar cells with Ag NPs to solar cells with 80 nm thick ITO. Additionally, a more fair comparison to conventional solar cells could be made if we had added a version with a thick Ag mirror functioning both as a backside reflector and a more conductive contact. Unfortunately, depositing thick Ag films on the backside of the solar cells always resulted in drastically reduced electrical performance. We attribute the degraded performance to stress induced by the Ag on the thin dielectric films.
4.2 Off-axis performance improvement
An important aspect of solar cell performance is their performance when sunlight is not normally incident on the front surface. Figure 8 plots Jsc enhancement as a percent change between the sample and control versus the angle of incidence of the light source. These samples only had front side NPs and were measured at incidence angles varying from 0 to 60 degrees. Cells with Ag NPs on the front show a large improvement in Jsc versus the control cells at incidence angles greater than 30þ and larger. This indicates that the NPs improve the coupling of non-normal incident light when compared to using ITO alone. The data is in good agreement with other work that has shown increased off-axis performance with Au NPs on a planar silicon solar cell [52] and with metallic gratings on the surface of a Si absorber with an SiO2 spacer [53]. The transmittance of a planar interface is dependent on the angle of incidence of the incoming light. Adding plasmonic resonators such as Ag NPs reduces this angle dependence in transmittance because the localized surface plasmon generated from any angle of incident light is more likely to scatter into the higher index absorber material than not.
Fig. 8 A plot of percent change in Jsc between the control sample and the sample with front-side NPs as a function of the light source’s angle of incidence. The percentage was calculated by taking the difference between the sample and control Jsc, then dividing that difference by the control Jsc to obtain a percent change.
4.3 Ag NP size effects
In order to better understand the effect of particle size on solar cell performance, the Ag thickness was varied on the front side of cell, which corresponds to a change in NP size. The mean particle diameter for the 14, 16, and 18 nm films are 100, 146 and 158 nm, respectively. The 150 nm solar cells used in this experiment came from a separate batch ITO deposition which is why there is a difference in performance from those referenced previously in Fig. 7. While deposition conditions were kept the same, the performance of each film varied slightly. EQE data is presented in Fig. 9 along with Jsc. In general, increased Ag NP size leads to an overall increase in photon capture. Larger Ag NPs sizes also resulted in increased cell efficiencies, as seen in the inset tables. As the NPs increase in size, their plasmon resonance frequency shifts to correspondingly longer wavelengths. The larger particles therefore direct more light into the absorber layer and the smaller particles tend to either lose absorbed light directly to heat or scatter light at wavelengths that were already fully absorbed without assistance.
Fig. 9 A plot showing the effect of different Ag NP size distributions on the EQE of (a) 150 nm and (b) 100 nm thick solar cells.
5. Conclusion
This has demonstrated that the addition of Ag NPs to thin film freestanding a-Si:H solar cell improves their overall performance. FDTD, EQE, and JV measurements show improved performance for longer wavelengths. Jsc and efficiency improved more with thinner cells with an increase in overall efficiency of 50% for the 60 nm cell. Cells generally showed little enhancement with Ag NPs on the back side, more with NPs on the front and the most when NPs were on both the front and back. Solar cells of all three thicknesses studied showed enhancement at all angles of incidence. All aspects of the enhancement from Ag NP addition to the solar cells fabricated in this work will need to be tested in future work with solar cells of higher quality. Repeating the experiments with higher efficiency cells will better examine the effects of losses to the Ag NPs. Additionally, the parasitic effects of the oxidization of Ag over time should be mitigated by evaporating a thin film of MgF2 over the Ag NPs to protect them from the environment. Lastly, all light trapping structures have certain trade-offs. Current texturing techniques must sacrifice material quality for better light trapping. So while NPs have inherent losses, the ultimate goal is to determine if the increased light trapping with NPs is greater than those possible with traditional techniques.
Acknowledgments
The authors would like to thank the staff at the Semiconductor Microfabrication Laboratory at the Rochester Institute of Technology, the staff at the Center for Nanoscale Science and Technology at Cornell University, Brian McIntyre and the staff at URnano at the University of Rochester and the funding sources for this research under New York State Energy Research and Development Authority Award (18809) and a National Science Foundation IGERT Award (DGE 0966089).
References and links
1. P. Campbell and M. A. Green, “Light trapping properties of pyramidally textured surfaces,” J. Appl. Phys. 62(1), 243–249 (1987). [CrossRef]
2. A. Luque and S. Hegedus, Handbook of Photovoltaic Science and Engineering (Wiley, 2011).
3. H. B. T. Li, R. H. Franken, J. K. Rath, and R. E. I. Schropp, “Structural defects caused by a rough substrate and their influence on the performance of hydrogenated nano-crystalline silicon n-i-p solar cells,” Sol. Energy Mater. Sol. Cells 93(3), 338–349 (2009). [CrossRef]
4. K. R. Catchpole and S. Pillai, “Absorption enhancement due to scattering by dipoles into silicon waveguides,” J. Appl. Phys. 100(4), 044504 (2006). [CrossRef]
5. H. Stuart and D. Hall, “Absorption enhancement in silicon-on-insulator waveguides using metal island films,” Appl. Phys. Lett. 69(16), 2327–2329 (1996). [CrossRef]
6. S. Pillai, K. R. Catchpole, T. Trupke, and M. A. Green, “Surface plasmon enhanced silicon solar cells,” J. Appl. Phys. 101(9), 093105 (2007). [CrossRef]
7. H. A. Atwater and A. Polman, “Plasmonics for improved photovoltaic devices,” Nat. Mater. 9(3), 205–213 (2010). [CrossRef] [PubMed]
8. M. A. Green and S. Pillai, “Harnessing plasmonics for solar cells,” Nat. Photonics 6(3), 130–132 (2012). [CrossRef]
9. D. Derkacs, S. H. Lim, P. Matheu, W. Mar, and E. T. Yu, “Improved performance of amorphous silicon solar cells via scattering from surface plasmon polaritons in nearby metallic nanoparticles,” Appl. Phys. Lett. 89(9), 093103 (2006). [CrossRef]
10. M. van Lare, F. Lenzmann, M. A. Verschuuren, and A. Polman, “Mode coupling by plasmonic surface scatterers in thin-film silicon solar cells,” Appl. Phys. Lett. 101(22), 221110 (2012). [CrossRef]
11. D. M. Schaadt, B. Feng, and E. T. Yu, “Enhanced semiconductor optical absorption via surface plasmon excitation in metal nanoparticles,” Appl. Phys. Lett. 86(6), 063106 (2005). [CrossRef]
12. K. Nakayama, K. Tanabe, and H. A. Atwater, “Plasmonic nanoparticle enhanced light absorption in gaas solar cells,” Appl. Phys. Lett. 93(12), 121904 (2008). [CrossRef]
13. V. E. Ferry, M. A. Verschuuren, H. B. T. Li, R. E. I. Schropp, H. A. Atwater, and A. Polman, “Improved red- response in thin film a-si:H solar cells with soft-imprinted plasmonic back reflectors,” Appl. Phys. Lett. 95(18), 183503 (2009). [CrossRef]
14. J. A. Schuller, E. S. Barnard, W. Cai, Y. C. Jun, J. S. White, and M. L. Brongersma, “Plasmonics for extreme light concentration and manipulation,” Nat. Mater. 9(3), 193–204 (2010). [CrossRef] [PubMed]
15. R. A. Pala, J. White, E. Barnard, J. Liu, and M. L. Brongersma, “Design of plasmonic thin-film solar cells with broadband absorption enhancements,” Adv. Mater. 21(34), 3504–3509 (2009). [CrossRef]
16. F. J. Beck, S. Mokkapati, and K. R. Catchpole, “Plasmonic light-trapping for si solar cells using self-assembled, ag nanoparticles,” Prog. Photovolt. Res. Appl. 18(7), 500–504 (2010). [CrossRef]
17. R. Santbergen, T. L. Temple, R. Liang, A. H. M. Smets, R. A. C. M. M. Swaaij, and M. Zeman, “Application of plasmonic silver island films in thin-film silicon solar cells,” J. Opt. 14(2), 024010 (2012). [CrossRef]
18. C. Eminian, F. J. Haug, O. Cubero, X. Niquille, and C. Ballif, “Photocurrent enhancement in thin film amorphous silicon solar cells with silver nanoparticles,” Prog. Photovolt. Res. Appl. 19(3), 260–265 (2011). [CrossRef]
19. A. Basch, F. J. Beck, T. Soderstrom, S. Varlamov, and K. R. Catchpole, “Combined plasmonic and dielectric rear reflectors for enhanced photocurrent in solar cells,” Appl. Phys. Lett. 100(24), 243903 (2012). [CrossRef]
20. H. R. Tan, R. Santbergen, A. H. M. Smets, and M. Zeman, “Plasmonic light trapping in thin-film silicon solar cells with improved self-assembled silver nanoparticles,” Nano Lett. 12(8), 4070–4076 (2012). [CrossRef] [PubMed]
21. H. Stiebig, C. Haase, S. Jorke, P. Obermeyer, E. Moulin, and M. Schulte, “New light management concepts for thin-film silicon solar cells,” Mater. Res. Soc. Symp. Proc. 1101 1101–KK06–03 (2011).
22. E. Moulin, J. Sukmanowski, M. Schulte, A. Gordijn, F. Royer, and H. Stiebig, “Thin-film silicon solar cells with integrated silver nanoparticles,” Thin Solid Films 516(20), 6813–6817 (2008). [CrossRef]
23. R. Santbergen, R. Liang, and M. Zeman, “A-si:H solar cells with embedded silver nanoparticles,” in Photovoltaic Specialists Conference (PVSC), 2010 35th IEEE (2010), pp. 748–753. [CrossRef]
24. F. J. Beck, S. Mokkapati, and K. R. Catchpole, “Light trapping with plasmonic particles: beyond the dipole model,” Opt. Express 19(25), 25230–25241 (2011). [CrossRef] [PubMed]
25. F. J. Beck, S. Mokkapati, A. Polman, and K. R. Catchpole, “Asymmetry in photocurrent enhancement by plasmonic nanoparticle arrays located on the front or on the rear of solar cells,” Appl. Phys. Lett. 96(3), 033113 (2010). [CrossRef]
26. Z. Ouyang, X. Zhao, S. Varlamov, Y. G. Tao, J. Wong, and S. Pillai, “Nanoparticle-enhanced light trapping in thin-film silicon solar cells,” Prog. Photovolt. Res. Appl. 19(8), 917–926 (2011). [CrossRef]
27. S. Pillai, F. J. Beck, K. R. Catchpole, Z. Ouyang, and M. A. Green, “The effect of dielectric spacer thickness on surface plasmon enhanced solar cells for front and rear side depositions,” J. Appl. Phys. 109(7), 073105 (2011). [CrossRef]
28. D. M. Callahan, J. N. Munday, and H. A. Atwater, “Solar cell light trapping beyond the ray optic limit,” Nano Lett. 12(1), 214–218 (2012). [CrossRef] [PubMed]
29. A. Polman and H. A. Atwater, “Photonic design principles for ultrahigh-efficiency photovoltaics,” Nat. Mater. 11(3), 174–177 (2012). [CrossRef] [PubMed]
30. E. Schiff, “Thermodynamic limit to photonic-plasmonic light-trapping in thin films on metals,” J. Appl. Phys. 110(10), 104501 (2011). [CrossRef]
31. Z. Yu, A. Raman, and S. Fan, “Fundamental limit of nanophotonic light trapping in solar cells,” Proc. Natl. Acad. Sci. U.S.A. 107(41), 17491–17496 (2010). [CrossRef] [PubMed]
32. Z. Yu, A. Raman, and S. Fan, “Nanophotonic light-trapping theory for solar cells,” Appl. Phys., A Mater. Sci. Process. 105(2), 329–339 (2011). [CrossRef]
33. E. Yablonovitch, “Statistical ray optics,” J. Opt. Soc. Am. 72(7), 899–907 (1982). [CrossRef]
34. C. Tabor, R. Murali, M. Mahmoud, and M. A. El-Sayed, “On the use of plasmonic nanoparticle pairs as a plasmon ruler: The dependence of the near-field dipole plasmon coupling on nanoparticle size and shape,” J. Phys. Chem. A 113(10), 1946–1953 (2009). [CrossRef] [PubMed]
35. J. P. Clarkson, J. Winans, and P. M. Fauchet, “On the scaling behavior of dipole and quadrupole modes in coupled plasmonic nanoparticle pairs,” Opt. Mater. Express 1(5), 970–979 (2011). [CrossRef]
36. A. Lin, S. M. Fu, Y. K. Chung, S. Y. Lai, and C. W. Tseng, “An optimized surface plasmon photovoltaic structure using energy transfer between discrete nano-particles,” Opt. Express 21(S1Suppl 1), A131–A145 (2013). [CrossRef] [PubMed]
38. K. R. Catchpole and A. Polman, “Design principles for particle plasmon enhanced solar cells,” Appl. Phys. Lett. 93(19), 191113 (2008). [CrossRef]
39. S. H. Lim, W. Mar, P. Matheu, D. Derkacs, and E. T. Yu, “Photocurrent spectroscopy of optical absorption enhancement in silicon photodiodes via scattering from surface plasmon polaritons in gold nanoparticles,” J. Appl. Phys. 101(10), 104309 (2007). [CrossRef]
40. C. C. Striemer, T. R. Gaborski, J. L. McGrath, and P. M. Fauchet, “Charge- and size-based separation of macromolecules using ultrathin silicon membranes,” Nature 445(7129), 749–753 (2007). [CrossRef] [PubMed]
41. O. Cubero, F. J. Haug, Y. Ziegler, L. Sansonnens, P. Couty, D. Fischer, and C. Ballif, “Reduction of the phosphorous cross-contamination in n–i–p solar cells prepared in a single-chamber pecvd reactor,” Sol. Energy Mater. Sol. Cells 95(2), 606–610 (2011). [CrossRef]
42. M. Ohring, The Materials Science of Thin Films (Academic Press, 1992).
43. H. Zhao, B. Ozturk, E. A. Schiff, B. Yan, J. Yang, and S. Guha, “Plasmonic light-trapping and quantum efficiency measurements on nanocrystalline silicon solar cells and silicon-on-insulator devices,” MRS Proceedings 1245, 1245–A03–02 (2010). [CrossRef]
44. C. H. Fosli, A. Thøgersen, S. Karazhanov, and E. S. Marstein, “Plasmonics for light trapping in silicon solar cells,” Energy Procedia 10, 287–291 (2011). [CrossRef]
45. W. Rasband, U.S. National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, “Imagej,” (1997–2014).
46. A. A. Schmidt, J. Offermann, and R. Anton, “The role of neutral oxygen radicals in the oxidation of ag films,” Thin Solid Films 282, 105–107 (1996). [CrossRef]
47. Y. A. Akimov and W. S. Koh, “Resonant and nonresonant plasmonic nanoparticle enhancement for thin-film silicon solar cells,” Nanotechnology 21(23), 235201 (2010). [CrossRef] [PubMed]
48. Y. Yang, S. Pillai, H. Mehrvarz, and M. A. Green, “Plasmonic degradation and the importance of over-coating metal nanoparticles for a plasmonic solar cell,” Sol. Energy Mater. Sol. Cells 122, 208–216 (2014). [CrossRef]
49. F. J. Beck, E. Verhagen, S. Mokkapati, A. Polman, and K. R. Catchpole, “Resonant spp modes supported by discrete metal nanoparticles on high-index substrates,” Opt. Express 19(S2Suppl 2), A146–A156 (2011). [CrossRef] [PubMed]
50. P. Spinelli and A. Polman, “Prospects of near-field plasmonic absorption enhancement in semiconductor materials using embedded ag nanoparticles,” Opt. Express 20(S5Suppl 5), A641–A654 (2012). [CrossRef] [PubMed]
51. U. Rau and T. Kirchartz, “On the thermodynamics of light trapping in solar cells,” Nat. Mater. 13(2), 103–104 (2014). [CrossRef] [PubMed]
52. D. Wan, H.-L. Chen, T.-C. Tseng, C.-Y. Fang, Y.-S. Lai, and F.-Y. Yeh, “Antireflective nanoparticle arrays enhance the efficiency of silicon solar cells,” Adv. Funct. Mater. 20(18), 3064–3075 (2010). [CrossRef]
53. J. N. Munday and H. A. Atwater, “Large integrated absorption enhancement in plasmonic solar cells by combining metallic gratings and antireflection coatings,” Nano Lett. 11(6), 2195–2201 (2011). [CrossRef] [PubMed]
