High-efficiency, four-terminal tandem solar cells composed of thin GaAs films mechanically stacked onto interdigitated back contact silicon solar cells with a glass interlayer are demonstrated. The optimal thickness of the absorber layer of a rear heterojunction GaAs subcell for use in four terminal tandem solar cells was studied. GaAs top cells with absorber layer thicknesses of 1.5, 1.9, 2.3, 2.8, and 3.5 μm were fabricated on glass and mechanically stacked onto interdigitated back-contact Si bottom cells. All tandem cells were found to have efficiencies above 30% under the AM1.5 G spectrum demonstrating a relatively weak sensitivity to thickness in the four-terminal configuration. We found the 2.8 μm absorber layer cell to have the highest top cell and tandem cell efficiency at 26.38% and 32.57%, respectively. Optical modeling with transfer matrix method for the planar top cell and Lambertian light trapping in the textured Si subcell, along with drift-diffusion Hovel equations, were used to show photon recycling enhancement to the effective diffusion length and VOC of the top cell as a result of the low-index glass interlayer.

Most commercial photovoltaic systems use silicon cells, which are affordable and reliable but limited by their single junction design. Moreover, they are approaching a predicted practical efficiency limit of 27.1% for heterojunction cells with interdigitated back contacts (IBC).1 This type of cell has the highest recorded performance for single-junction silicon.2 Still, a higher efficiency is needed to address increasing energy demand, especially for areas with limited surface availability. One method to improve upon silicon solar technology is to incorporate it as the bottom cell of a tandem cell. To improve the overall tandem cell efficiency from that of silicon alone, the top cell material must have a higher spectral efficiency than silicon for the wavelengths of light within the top cell's bandgap.3 GaAs cells have a significantly higher spectral efficiency than that of silicon cells above the GaAs bandgap3 and are, therefore, a leading candidate for high-efficiency tandems on silicon. In addition, GaAs solar cells have shown the highest single junction efficiency to date.4,5 While the relatively high cost of GaAs fabrication may eventually preclude it from wide-spread use, this demonstration exemplifies the potential of such tandems on silicon as other potential material systems, such as perovskites, mature.

Tandem cells are commonly connected in series in a 2-terminal (2T) configuration, but can also be operated independently in a 4-terminal (4T) configuration. Unlike tandem cells in the 2T configuration, tandem cells in the 4T configuration are not limited by the cell with the lowest current.6 The non-reliance on current matching means 4T tandems have greater bandgap flexibility than 2T tandems. The GaAs is poorly current matched to Si, making 4T the best option for the tandem cell configuration.3 GaAs and Si subcells could be combined using wafer bonding or direct growth techniques,7 but these techniques require pristine flat surfaces or challenging heteroepitaxy. Previously,8,9 mechanically stacked 4T III–V//Si tandems utilized a silicon heterojunction bottom subcell contacted from the top and bottom, but here textured silicon IBC subcells with rear-only contacts are used to simplify the processing and contacting (see Fig. 1). This GaAs//Si IBC structure was recently introduced10 by the authors, demonstrating 29.6% efficiency using a GaInP window layer rather than the AlInP used here. The cells are grown and fabricated separately with access to both the front and rear sides, and the completed subcells are later combined together with transparent epoxy that is compatible with the textured silicon IBC cell. Thin epitaxial GaAs top subcells removed from the substrate utilized the rear heterojunction (RHJ) structure.11 In the RHJ structure, the absorber layer is a thick n-type layer at the top of the p–n junction rather than a thick p-type layer at the bottom of the p–n junction. The RHJ configuration is associated with increased cell voltage relative to a traditional front junction configuration11 and benefits greatly from optical designs that increase photon recycling.12 

FIG. 1.

Configuration of the GaAs//Si four-terminal, rear heterojunction tandem solar cells.

FIG. 1.

Configuration of the GaAs//Si four-terminal, rear heterojunction tandem solar cells.

Close modal

Optimizing the absorber layer thickness is different for a tandem cell vs a single junction cell: first, the light transmitted through the top cell due to absorber layer thinning can still be collected by the bottom cell. This means a 4T tandem is less sensitive to top cell thinning than a single junction would be. Second, there can be no metal back reflector behind the top cell to send normal light back into the top cell absorbing region. This might mean the top cell needs to be thicker to absorb the photons that would, otherwise, be transmitted. Finally, since each subcell in the 4T configuration can operate independently at its maximum power point, the current matching constraint of a 2T configuration need not be considered when optimizing thicknesses.

The purpose of this paper is to optimize the n-GaAs absorber layer thickness to maximize 4T RHJ GaAs//Si IBC tandem cell efficiency. If the absorber layer is too thin, transmission through the top cell will increase, and high-energy photons will be collected by the bottom cell at a lower voltage. If the absorber layer is too thick, approaching the minority carrier diffusion length of the absorber layer material, the carriers generated will recombine prematurely, and the photon energy will be lost as heat. If the top cell absorber layer thickness is optimized, energy losses due to light transmission to the bottom cell and recombination in the top cell will be minimized.

The GaAs cells were grown on GaAs substrates using Metal-Organic Vapor Phase Epitaxy (MOVPE), following the same procedure as in Ref. 9. The absorber layer thicknesses were varied from 1.5 to 3.5 μm to bracket an expected optimum of 2.4 μm based on preliminary Hovel modeling13–15 with a diffusion length of about 10 μm. The actual thickness was determined by in situ reflectivity of multi-beam optical sensors.16 Structures were grown on single-crystal, Si-doped, (001)-oriented GaAs wafers with a 2° offcut toward the (111)B orientation. The samples were processed to create 1 cm2 GaAs cells with the same grid pattern used previously.8,9 The samples were cleaned with acetone and isopropyl alcohol. Then, the back grid was patterned using UV contact-lithography and electroplated. An anti-reflective coating of 2 nm MgF2 followed by 105 nm ZnS was deposited. The cell was then inverted and stacked onto a piece of 0.6 mm glass with epoxy (LOCTITE ECCOBOND 931-1) and cured at 100 °C for 20 min. The substrate and stop-etch were etched off, and the front contacts were electroplated. Then, the cell was mesa etched, the exposed contact layer was etched off, and an anti-reflective coating of 2 nm MgF2, 54 nm ZnS, and 104 nm MgF2 was evaporated onto the front of the cell.

Interdigitated back contact (IBC) silicon cells were supplied by ISFH. Cells had a base resistivity of 1.5 Ω-cm, a thickness of 300 μm, and poly-silicon on oxide junctions of both polarities prepared by thermal oxidation, LPCVD deposition of amorphous silicon, ion-implantation, and inkjet patterning.20 The tandems were assembled by stacking the processed GaAs cells on top of a Si bottom cell with a thin layer of the same epoxy used for inversion in between. The resulting cells were then cured at room temperature for 24 h. Figure 1 shows the final structure of the GaAs//Si 4T tandem.

The tandem cells were measured by the independent NREL Cell and Module Performance (CMP) team. External quantum efficiency (EQE) and current density–voltage (J–V) measurements were obtained for each subcell of the mechanically stacked 4T device. The J–V curves were measured, while both subcells were illuminated at one-sun under the AM1.5G spectra. The bottom cells were measured, while the top cells were biased at the maximum power point to account for luminescent coupling.9 

The J–V curves and EQE spectra of the top and bottom cells are shown in Fig. 2 and summarized in Table I. The effect of increased light transmission through the top cell as the thickness of the absorber layer decreases can be observed in the EQE within the 700–850 nm range [Fig. 2(a)]. The largest change in EQE within this wavelength range is from 1.5 to 1.9 μm absorber layer thickness, where the top cell EQE decreases and the bottom cell EQE increases. The sum of the subcell EQE shown in the top of Fig. 2(a) is relatively constant at 93%–98%, indicating very little absorption or reflection losses in the glass and epoxy interlayers.

FIG. 2.

(a) External quantum efficiency (EQE) spectra with the sum of the top + bottom EQE shown above. Solid lines show measured data and dashed lines show modeled EQE based on the transfer matrix method17 for the top junction and Lambertian light trapping18,19 (as filtered by the top cell) for the textured silicon bottom cell. (b) Measured current density–voltage curves. Insets zoomed in near the maximum power points are included in the current density–voltage plot.

FIG. 2.

(a) External quantum efficiency (EQE) spectra with the sum of the top + bottom EQE shown above. Solid lines show measured data and dashed lines show modeled EQE based on the transfer matrix method17 for the top junction and Lambertian light trapping18,19 (as filtered by the top cell) for the textured silicon bottom cell. (b) Measured current density–voltage curves. Insets zoomed in near the maximum power points are included in the current density–voltage plot.

Close modal
TABLE I.

Performance of the GaAs//Si cells. Top and bottom cells were measured at one-sun under the AM1.5G spectrum. The top cell was held at the maximum power point while the bottom cell was measured.

n-GaAs absorber layer thickness (μm)Voc (V)Jsc (mA/cm2)Fill factor (%)Single junction efficiency (%)Tandem efficiency (%)
1.5 Top cell 1.075 (± 0.01) 27.77 (± 0.36) 81.3 (± 0.4) 24.25 (± 0.1) 30.91 (± 0.1) 
Bottom cell 0.675 (± 0.01) 12.2 (± 0.16) 80.9 (± 0.4) 6.66 (± 0.01)  
1.9 Top cell 1.081 (± 0.01) 28.84 (± 0.37) 82.6 (± 0.41) 25.75 (± 0.1) 32.14 (± 0.11) 
Bottom cell 0.680 (± 0.01) 11.55 (± 0.15) 81.3 (± 0.41) 6.39 (± 0.01)  
2.3 Top cell 1.084 (± 0.01) 28.65 (± 0.37) 82.5 (± 0.41) 25.63 (± 0.1) 31.68 (± 0.11) 
Bottom cell 0.675 (± 0.01) 10.99 (± 0.14) 81.5 (± 0.41) 6.05 (± 0.01)  
2.8 Top cell 1.092 (± 0.01) 28.84 (± 0.37) 83.8 (± 0.42) 26.38 (± 0.11) 32.57 (± 0.11) 
Bottom cell 0.679 (± 0.01) 11.2 (± 0.15) 81.4 (± 0.41) 6.19 (± 0.01)  
3.5 Top cell 1.088 (± 0.01) 29.36 (± 0.38) 81.9 (± 0.41) 26.14 (± 0.1) 32.52 (± 0.11) 
Bottom cell 0.675 (± 0.01) 11.46 (± 0.15) 82.3 (± 0.41) 6.38 (± 0.01)  
n-GaAs absorber layer thickness (μm)Voc (V)Jsc (mA/cm2)Fill factor (%)Single junction efficiency (%)Tandem efficiency (%)
1.5 Top cell 1.075 (± 0.01) 27.77 (± 0.36) 81.3 (± 0.4) 24.25 (± 0.1) 30.91 (± 0.1) 
Bottom cell 0.675 (± 0.01) 12.2 (± 0.16) 80.9 (± 0.4) 6.66 (± 0.01)  
1.9 Top cell 1.081 (± 0.01) 28.84 (± 0.37) 82.6 (± 0.41) 25.75 (± 0.1) 32.14 (± 0.11) 
Bottom cell 0.680 (± 0.01) 11.55 (± 0.15) 81.3 (± 0.41) 6.39 (± 0.01)  
2.3 Top cell 1.084 (± 0.01) 28.65 (± 0.37) 82.5 (± 0.41) 25.63 (± 0.1) 31.68 (± 0.11) 
Bottom cell 0.675 (± 0.01) 10.99 (± 0.14) 81.5 (± 0.41) 6.05 (± 0.01)  
2.8 Top cell 1.092 (± 0.01) 28.84 (± 0.37) 83.8 (± 0.42) 26.38 (± 0.11) 32.57 (± 0.11) 
Bottom cell 0.679 (± 0.01) 11.2 (± 0.15) 81.4 (± 0.41) 6.19 (± 0.01)  
3.5 Top cell 1.088 (± 0.01) 29.36 (± 0.38) 81.9 (± 0.41) 26.14 (± 0.1) 32.52 (± 0.11) 
Bottom cell 0.675 (± 0.01) 11.46 (± 0.15) 82.3 (± 0.41) 6.38 (± 0.01)  

Figure 3 shows how the one-sun J–V parameters vary with GaAs absorber thickness. Surprisingly, the top cell Jsc continues to increase with absorber layer thickness up through 3.5 μm. Total power conversion efficiencies of the stacked 4T devices are around 32% for every tandem cell with a GaAs thickness above 1.5 μm. The 2.8 μm absorber layer cell had the highest top cell and tandem efficiency, 26.38% and 32.57%, respectively. This is only slightly less than the record RHJ GaAs//Si 4T tandem cell efficiency of 32.82%.8 While the GaAs top cell here had a slightly decreased fill factor (FF), the IBC bottom cell exhibited a slightly higher efficiency than the silicon heterojunction bottom cell used previously.

FIG. 3.

Short-circuit current density (JSC), open-circuit voltage (VOC), fill factor (FF), and efficiency of 4T RHJ GaAs//Si solar cells vs the n-GaAs absorber layer thickness. Large markers show data measurements from CMP group at NREL with error bars showing uncertainties. Lines show model predictions using various methods. The blue lines show the fitting of the silicon bottom cell using Lambertian light trapping with the Jo value shown. The gray lines show predictions of the optical modeling of Steiner et al.12 for various internal radiative efficiencies (IRE) of the GaAs absorber. The Hovel model of the GaAs subcell is shown for various effective diffusion lengths (Leffd) (purple = 5 μm, orange = 10 μm, green = 20 μm). The Hovel model also assumes emitter properties: surface recombination velocity = 100 cm/s, mobility = 300 cm2/V s, and a depletion width lifetime of 2 × 10−11 s.

FIG. 3.

Short-circuit current density (JSC), open-circuit voltage (VOC), fill factor (FF), and efficiency of 4T RHJ GaAs//Si solar cells vs the n-GaAs absorber layer thickness. Large markers show data measurements from CMP group at NREL with error bars showing uncertainties. Lines show model predictions using various methods. The blue lines show the fitting of the silicon bottom cell using Lambertian light trapping with the Jo value shown. The gray lines show predictions of the optical modeling of Steiner et al.12 for various internal radiative efficiencies (IRE) of the GaAs absorber. The Hovel model of the GaAs subcell is shown for various effective diffusion lengths (Leffd) (purple = 5 μm, orange = 10 μm, green = 20 μm). The Hovel model also assumes emitter properties: surface recombination velocity = 100 cm/s, mobility = 300 cm2/V s, and a depletion width lifetime of 2 × 10−11 s.

Close modal

The performance was modeled as shown by the lines in Fig. 3. The GaAs top cell reflectance, layer absorption, and transmission into the glass were modeled by the transfer matrix method.17 The top cell EQE was well fit with 98% of the absorbance in the GaAs layer plus 30% of the absorbance in the AlInP window layer as shown by the dashed lines in Fig. 2(a). The top cell EQE was also modeled using the Hovel model as a function of the emitter minority carrier diffusion length [not shown in Fig. 2(a)]. The Si EQE was modeled, assuming Lambertian light trapping18 attenuated by the reflection and absorption of the top cell.19 The JSC is calculated as an integration of the modeled EQE and AM1.5G spectrum as shown in Fig. 3. The Hovel model indicates that these top cells have an extremely long effective diffusion length. Lumb et al.21 have shown that the effective diffusion length for a RHJ GaAs absorber can be 2.2× longer as the result of photon recycling enhancement from a metal back surface reflector rather than an absorbing substrate. Furthermore, the radiative lifetime of GaAs has been shown to be enhanced due to its optical environment.22 We therefore conclude that this top cell performance is benefiting from the total internal reflection from the low-index glass interlayer.23 In fact, the high VOC of the top cell cannot be explained without enhanced photon recycling.12 We have calculated the photon recycling factor21 or the probability of photon reabsorption (Pabs) of this device as a function of thickness.12 The angle averaged reflectance from the glass back toward the GaAs absorber is 82.9%, giving Pabs = 95.4% for a GaAs thickness of 1.5 μm. This is considerably higher than if the top cell were grown directly on a planar Si cell without a low index interlayer. Coupled with extremely high internal radiative efficiency (IRE) as modeled by the gray lines in Fig. 3, the measured VOCs above 1.09 V can be explained.

Eventually, even thicker GaAs absorbers must result in decreased performance due to imperfect carrier transport, but we did not observe this falloff in JSC up through 3.5 μm. We did, however, see a slight drop in the FF of this thickest top cell. The fill factor (FF) can be very difficult to model but can play an important role in the final efficiency. Here, the top cell FF was modeled assuming an ideality factor of n = 2 from the calculated VOC and JSC in good agreement with the data, but assuming n = 1 would have vastly overestimated the data. We have also confirmed the reproducibility of this drop in FF for the 3.5 μm device but have not yet developed models that can completely explain the FF loss.

In conclusion, we have demonstrated over 32% efficiency in 4T tandems with a RHJ GaAs subcell on an IBC silicon subcell for a wide range of GaAs absorber thickness from 1.9 to 3.5 μm. The low-index glass and epoxy interlayer of this design enhance photon recycling to increase the GaAs top cell effective diffusion length and VOC leading to a range of absorber layer thicknesses that will produce high efficiency cells rather than only one specific optimal thickness.

This work was supported in part by the U.S. Department of Energy, Office of Science, Office of Workforce Development for Teachers and Scientists (WDTS) under the Science Undergraduate Laboratory Internship (SULI) program and the U.S. Department of Energy under Contract No. DE-AC36-08GO28308 with Alliance for Sustainable Energy, LLC, the Manager and Operator of the National Renewable Energy Laboratory, with funding provided by the Department of Energy Office of Energy Efficiency and Renewable Energy Solar Energy Technologies Office under Contract No. SETP DE-EE00034911. For ISFH, the funding was provided by the German State of Lower Saxony and the German Federal Ministry for Economics and Energy (BMWi) within the research project “EASi” (No. FKZ0324040) and “27Plus6” (No. FKZ03EE1056A). Scientific support also came from Dana Sulas, Steven Johnston, Tao Song, Waldo Olavarria, Michelle Young, Daniel Lu, and Jerry Pineau at NREL. We also thank H. Kohlenberg and the team at ISFH and Leibniz Universität Hannover for processing the POLO-IBC bottom cells, and H. Schulte-Huxel and F. Haase for fruitful discussions.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
K.
Yoshikawa
,
H.
Kawasaki
,
W.
Yoshida
,
T.
Irie
,
K.
Konishi
,
K.
Nakano
,
T.
Uto
,
D.
Adachi
,
M.
Kanematsu
,
H.
Uzu
 et al, “
Silicon heterojunction solar cell with interdigitated back contacts for a photoconversion efficiency over 26%
,”
Nat. Energy
2
,
17032
(
2017
).
2.
K.
Yamamoto
,
K.
Yoshikawa
,
H.
Uzu
, and
D.
Adachi
, “
High-efficiency heterojunction crystalline Si solar cells
,”
Jpn. J. Appl. Phys., Part 1
57
,
08RB20
(
2018
).
3.
Z.
Yu
,
M.
Leilaeioun
, and
Z.
Holman
, “
Selecting tandem partners for silicon solar cells
,”
Nat. Energy
1
,
16137
(
2016
).
4.
M. A.
Green
,
Y.
Hishikawa
,
E. D.
Dunlop
,
D. H.
Levi
,
J.
Hohl–Ebinger
, and
A. W. Y.
Ho–Baillie
, “
Solar cell efficiency tables (version 52)
,”
Prog. Photovoltaics
26
,
427
436
(
2018
).
5.
B. M.
Kayes
,
H.
Nie
,
R.
Twist
,
S. G.
Spruytte
,
F.
Reinhardt
,
I. C.
Kizilyalli
, and
G. S.
Higashi
, “
27.6% conversion efficiency, a new record for single- junction solar cells under 1 sun illumination
,” in Proc. 37th IEEE Photovoltaic Spec. Conf.,
2011
, pp.
4
8
.
6.
I.
Almansouri
,
A.
Ho-Baillie
,
S. P.
Bremner
, and
M. A.
Green
, “
Supercharging silicon solar cell performance by means of multijunction concept
,”
IEEE J. Photovoltaics
5
,
968
976
(
2015
).
7.
F.
Dimroth
,
S.
Roesener
,
T.
Essig
,
C.
Weuffen
,
A.
Wekkeli
,
E.
Oliva
,
S. G. K.
Volz
,
T.
Hannappel
,
D.
Häussler
,
W.
Jager
, and
A.
Bett
, “
Comparison of direct growth and wafer bonding for the fabrication of GaInP/GaAs dual-junction solar cells on silicon
,”
IEEE J. Photovoltaics
4
,
620
625
(
2014
).
8.
S.
Essig
,
C.
Allebé
,
T.
Remo
,
J. F.
Geisz
,
M. A.
Steiner
,
K.
Horowitz
,
L.
Barraud
,
J. S.
Ward
,
M.
Schnabel
,
A.
Descoeudres
,
D. L.
Young
,
M.
Woodhouse
,
M.
Despeisse
,
C.
Ballif
, and
A.
Tamboli
, “
Raising the one-sun conversion efficiency of III–V/Si solar cells to 32.8% for two junctions and 35.9% for three junctions
,”
Nat. Energy
2
,
17144
(
2017
).
9.
S.
Essig
,
M. A.
Steiner
,
C.
Allebé
,
J. F.
Geisz
,
B.
Paviet-Salomon
,
S.
Ward
,
A.
Descoeudres
,
V.
LaSalvia
,
L.
Barraud
,
N.
Badel
,
A.
Faes
,
J.
Levrat
,
M.
Despeisse
,
C.
Ballif
,
P.
Stradins
, and
D. L.
Young
, “
Realization of GaInP/Si dual-junction solar cells with 29.8% 1-sun efficiency
,”
IEEE J. Photovoltaics
6
,
1012
1019
(
2016
).
10.
K. T.
VanSant
,
J.
Simon
,
J. F.
Geisz
,
E. L.
Warren
,
K. L.
Schulte
,
A. J.
Ptak
,
M. S.
Young
,
M.
Rienäcker
,
H.
Schulte-Huxel
,
R.
Peibst
, and
A. C.
Tamboli
, “
Toward low-cost 4–terminal GaAs//Si tandem solar cells
,”
ACS Appl. Energy Mater.
2
,
2375
2380
(
2019
).
11.
J. F.
Geisz
,
M. A.
Steiner
,
I.
García
,
S. R.
Kurtz
, and
D. J.
Friedman
, “
Enhanced external radiative efficiency for 20.8% efficient single-junction GaInP solar cells
,”
Appl. Phys. Lett.
103
,
041118
(
2013
).
12.
M. A.
Steiner
,
J. F.
Geisz
,
I.
García
,
D. J.
Friedman
,
A.
Duda
, and
S. R.
Kurtz
, “
Optical enhancement of the open-circuit voltage in high quality GaAs solar cells
,”
J. Appl. Phys.
113
,
123109
(
2013
).
13.
S. R.
Kurtz
,
J. M.
Olson
,
D. J.
Friedman
,
J. F.
Geisz
,
K. A.
Bertness
, and
A. E.
Kibbler
, “
Passivation of interfaces in high-efficiency photovoltaic devices
,”
MRS Proc.
573
,
95
(
1999
).
14.
A. L.
Fahrenbruch
,
R. H.
Bube
, and
R. V.
D'Aiello
, “
Fundamentals of solar cells (photovoltaic solar energy conversion)
,”
J. Sol. Energy Eng.
106
,
497
498
(
1984
).
15.
H. J.
Hovel
,
Semiconductors and Semimetals. Volume 11. Solar Cells
(
Academic Press
,
New York
,
1975
).
16.
W. G.
Breiland
and
K. P.
Killeen
, “
A virtual interface method for extracting growth rates and high temperature optical constants from thin semiconductor films using in situ normal incidence reflectance
,”
J. Appl. Phys.
78
,
6726
6736
(
1995
).
17.
E.
Centurioni
, “
Generalized matrix method for calculation of internal light energy flux in mixed coherent and incoherent multilayers
,”
Appl. Opt.
44
,
7532
7539
(
2005
).
18.
M. A.
Green
, “
Lambertian light trapping in textured solar cells and light-emitting diodes: Analytical solutions
,”
Prog. Photovoltaics
10
,
235
241
(
2002
).
19.
T. P.
White
,
N. N.
Lal
, and
K. R.
Catchpole
, “
Tandem solar cells based on high-efficiency c-Si bottom cells: Top cell requirements for > 30% efficiency
,”
IEEE J. Photovoltaics
4
,
208
214
(
2014
).
20.
M.
Rienäcker
,
M.
Schnabel
,
E.
Warren
,
A.
Merkle
,
H.
Schulte-Huxel
,
T. R.
Klein
,
M. F. A. M.
van Hest
,
M.
Steiner
,
J.
Geisz
,
S.
Kajari-Schröder
,
R.
Niepelt
,
J.
Schmidt
,
R.
Brendel
,
P.
Stradins
,
A.
Tamboli
, and
R.
Peibst
, “
Mechanically stacked dual-junction and triple-junction III-V/Si-IBC cells with efficiencies of 31.5% and 35.4%
,” in
33rd European Photovoltaic Solar Energy Conference and Exhibition
(
2017
).
21.
M. P.
Lumb
,
M. A.
Steiner
,
J. F.
Geisz
, and
R. J.
Walters
, “
Incorporating photon recycling into the analytical drift-diffusion model of high efficiency solar cells
,”
J. Appl. Phys.
116
,
194504
(
2014
).
22.
M. A.
Steiner
,
J. A.
Geisz
,
I.
García
,
D. J.
Friedman
,
A.
Duda
,
W. J.
Olavarria
,
M.
Young
,
D.
Kuciauskas
, and
S. R.
Kurtz
, “
Effects of internal luminescence and internal optics on Voc and Jsc of III–V solar cells
,”
IEEE J. Photovoltaics
3
,
1437
1442
(
2013
).
23.
M. A.
Steiner
,
J. F.
Geisz
,
J. S.
Ward
,
I.
García
,
D. J.
Friedman
,
R. R.
King
,
P. T.
Chiu
,
R. M.
France
,
A.
Duda
,
W. J.
Olavarria
,
M.
Young
, and
S. R.
Kurtz
, “
Optically enhanced photon recycling in mechanically stacked multijunction solar cells
,”
IEEE J. Photovoltaics
6
,
358
365
(
2016
).