High-efficiency AIGaAs/GaAs concentrator solar cells
`R. Sahai, D. D. Edwall, and J. S. Harris, Jr.
`Rockwell International Science Center, Thousand Oaks, California 91360
`(Received 26 October 1978; accepted for publication 9 November 1978)
`
`Efficiencies of 25% have been obtained with l-cm-diam AIGaAs/GaAs heteroface
`concentrator solar cells utilizing an ultrathin AIGaAs window layer design. A low
`specific resistance « 0.005 n cm2
`) Ohmic contact is achieved by direct contact to the p(cid:173)
`GaAs active layer. Liquid phase epitaxy has been developed to grow < 500-A thick
`window layers on large-area (3.3 X 3.3 cm) GaAs substrates. Four l-cm-diam cells are
`produced from each wafer and demonstrate the potential for larger-scale production.
`
`PACS numbers: 72.40.+w, 84.60.Jt
`
`System studies 1
`2 on concentrator photovoltaic systems
`•
`clearly demonstrate the importance of high cell efficiency in
`reducing the overall cost of solar-electric generation. GaAs
`is near the optimum band gap for optimum photovoltaic
`conversion and has demonstrated high conversion efficien(cid:173)
`cy. J The high cost of high-quality single-crystal concentrator
`cells requires that the cells operate under highly concentrat(cid:173)
`ed sunlight. One major obstacle facing the fabrication of con(cid:173)
`centrator solar cells has been the relatively high sheet resis(cid:173)
`tance for the parallel combination of the top Ohmic contact
`and semiconductor layer. Prior work on AIGaAs/GaAs
`concentrator solar cells utilized a relatively thick heavily
`doped AIGaAs window layer to minimize the sheet resis(cid:173)
`tance"" This thick window layer, however, reduces the cell
`quantum efficiency for wavelengths < 5000 A because of
`optical absorption in the window layer. In this paper, we
`report the design and successful demonstration of high-effi(cid:173)
`ciency AIGaAs/GaAs concentrator cells based upon a thin
`AIGaAs window design which achieves near-theoretical
`performance. Terrestrial measurements on these cells have
`shown conversion efficiencies as high as 24.7% at 178 suns
`(AMI) and 21.7% at 900 suns.
`The thin-window AIGaAs/GaAs solar cell design in(cid:173)
`corporates the following major design elements:
`(1) An ultrathin ( < 500 A) p-AIGaAs surface which
`reduces surface recombination losses with negligible optical
`absorption.
`(2) A relatively thick p-GaAs active epitaxial layer with
`good electron diffusion length ( > 5 11m) which reduces top
`layer sheet resistance yet still achieves high current collec(cid:173)
`tion efficiency.
`(3) An efficient front contact grid design which directly
`contacts the active p-GaAs layer and reduces the cell series
`resistance.
`(4) A broad-band two-layer antireflection coating
`which minimizes reflection losses.
`The major advantage of using an ultrathin AIGaAs
`window layer is illustrated by the data of Fig. 1. Figure 1
`shows the calculated internal collection efficiency versus
`wavelength for various AIGaAs layer thicknesses. Thick
`window layers result in loss of photoresponse at wavelengths
`< 0.5 11m due to photogeneration in the AIGaAs layer and
`
`poor collection due to high surface recombination losses. By
`using an AIGaAs layer only 500 A thick, there is negligible
`absorption in the AIGaAs and the near-perfect AIGaAs(cid:173)
`IGaAs heterojunction interface eliminates the high surface
`recombination loss for all carriers generated in the p-GaAs
`layer. This results in efficient collection of photogenerated
`carriers down to ultraviolet wavelengths. The measured re(cid:173)
`sponse on a liquid phase epitaxially grown cell structure is
`also shown in Fig. 1 and demonstrates that indeed an ul(cid:173)
`trathin AIGaAs epitaxial layer with a low-loss hetrojunction
`interface has been achieved.
`Si,N. is a good single-layer antireflection (AR) dielec(cid:173)
`tric for GaAs; however, this results in -12% reflection loss
`for a thin-window AIGaAs/GaAs cell. In order to take full
`advantage of the broadband internal photoresponse de(cid:173)
`scribed above, a two-layer AR-coating is required to reduce
`this external reflection loss. A two-layer coating consisting
`of 530 A of Ta20, followed by 760 A ofSi02 is utilized,3
`which reduces reflection losses to -4%.
`The value of the cell series resistance becomes critical
`under high sunlight concentration. Figure 2 shows the effect
`of series resistance on cell performance as a function of con-
`
`0.10
`
`o. ag. ':-: 30:--''-:OJ .• C:-o ~---::o-L. 50"--~-'--0.L60~-OJ. '-'--0 ~----CO.-L80-~---b~----J1. 00
`
`WAVELENGTH
`
`IN MICRONS
`
`FIG. I. Calculated internal collection efficiency curves for different GaA(cid:173)
`lAs layer thickness. Also shown are the measured response points after
`correction for reflection loss.
`
`147
`
`Appl. Phys. Lett. 34(2), 15 January 1979
`
`0003-6951/79/020147-03$00.50
`
`© 1979 American Institute of Physics
`
`147
`
` Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 18.9.61.112 On: Mon, 30 May 2016
`14:00:35
`
`Sony Corp. v. Raytheon Co.
`IPR2016-00209
`
`Raytheon2020-0001
`
`

`
`.001
`
`20.00
`
`~
`
`~
`
`10.00
`
`SERIES RESISTANCE VALUES SHOWN IN OHMS- CM 2
`o.or.':::OO:--~~--'-:I~O.':::OO:--~~--'-;I~OO'-:.O:::-O ~-~'-:I~OO':-O.::OO~-~~lo,J.OOE4
`CONCENTRATION
`
`FIG. 2. Calculated GaAs concentrator cell efficiency as a function of con(cid:173)
`centration for a number of series resistance values.
`
`centration level. For high-concentration high-efficiency op(cid:173)
`eration (-I000suns), theR,XA product for the cell must be
`< 10 mil cm2
`• Our approach to lower the series resistance is
`to use an efficient front contact grid design with a relatively
`thickp-GaAs active layer to which a direct Ohmic contact
`can be made. We have developed efficient grid designs for
`both the circular and rectangular concentrator cells. 6 Figure
`3(a) shows the radial contact pattern for l-cm-diam concen(cid:173)
`trator cells with 256 tapered grid lines leading to the outer
`contact ring. The average finger width is 6 11m with an aver(cid:173)
`age spacing of 92 11m. Figure 3(b) shows the details of the
`direct Ohmic contact to the p-GaAs active region. This con(cid:173)
`tact requires photolithographic definition of the grid pattern
`and selective etching ofthe thin AIGaAs layer before the Ag(cid:173)
`Zn contact metallization is evaporated. After contact alloy(cid:173)
`ing, the pattern is electroplated with Ag to a total thickness
`of 5 11m. This technique results in low specific contact resis(cid:173)
`tance to the active cell region.
`
`We have fabricated and tested a large number of con(cid:173)
`centrator cells with a l-cm-diam illuminated area. The ini(cid:173)
`tial cell design had a contact pattern similar to that shown in
`Fig. 3(a) with 128 radial grid lines, while the most recent
`cells have 256 lines. Both front contact grid designs cover
`- 10% of the illuminated cell area. Several of these cells
`were calibrated for AMO by NASA Lewis Research Center?
`on a high-altitude F-106 test flight. The best of these cali(cid:173)
`brated cells has short-circuit current of 23.49 rnA, open(cid:173)
`circuit voltage of 0.996 V, and a fill factor of 0.857, leading to
`an AMO efficiency of 18.8% for a l-cm-diam cell. These
`calibrated cells are used as standards for all of our test
`measurements.
`High-concentration measurements on similar l-cm(cid:173)
`diam cells were made in natural sunlight at the Jet Propul(cid:173)
`sion Laboratory's test facility on Table Mountain, Califor(cid:173)
`nia. A portable cassagrain concentrator mounted on an
`equatorial mount in the tracking mode was used to obtain
`concentration levels up to 440 suns. The concentrator and
`measurement techniques are described in Ref. 6. Figure 4
`
`1,1
`
`p-GAAs
`
`N-GAAs
`SUBSTRATE
`
`Ibl
`
`FIG. 3. (a) The front grid pattern for l-cm-diam concentrator cell. (b) A
`section of the solar cell showing the Ohmic contact technique for the front
`grid.
`
`shows the measured current-voltage (IV) characteristics for
`one of the best 128 grid line cells (No. 976). The cell tempera(cid:173)
`ture, as indicated by a thermocouple just under the center of
`the cell, was maintained at 50°C with the cell operating at
`the maximum power point. Although the cell could have
`been operated at lower temperatures at the lower concentra(cid:173)
`tions, a constant temperature was maintained so that cell
`performance as a function of concentration could be mea(cid:173)
`sured and the cell series resistance estimated from the set of
`1- V curves. The concentration ratio, the power conversion
`efficiency, and fill factor are indicated for each curve in Fig.
`
`CR 0 440
`
`'1020.32",
`
`INSOLATION
`AMl
`CELL #976
`
`270
`
`178
`
`24.69
`
`114
`
`74
`
`°0~L-~.2-L-~.4~~.L6~-.L8~-l.LO~-1~.2
`
`VOL TS
`
`FIG. 4. J- V curves for concentrator cell No. 976 measured at Table Moun(cid:173)
`tain, Calif. The cell temperature at the maximum power point was main(cid:173)
`tained at 50 'C for each curve.
`
`148
`
`Appl. Phys. Lett., Vol. 34, No.2, 15 January 1979
`
`Sahai, Edwall, and Harris, Jr.
`
`148
`
` Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 18.9.61.112 On: Mon, 30 May 2016
`14:00:35
`
`Raytheon2020-0002
`
`

`
`20
`
`\\
`
`,
`
`,
`
`CONCENTRATION RATIO
`
`FIG. 5. Measured efficiency-vs-concentration data for cell No. 1208 with
`the 256-line grid pattern of Fig. 3(a) and cell No. 976 with an earlier 128-
`line grid pattern.
`
`4. The highest measured cell efficiency was 24.7% at 178
`suns (AMI). To our knowledge, this is the highest efficiency
`ever measured for any solar cell. At higher concentrations,
`the efficiency decreases because offill-factor degradation
`due to a relatively high series resistance. The series resistance
`R s value estimatedS from this set of /- V curves is between 20
`and 30 mfl, giving a series resistance and cell area product
`• In order to achieve high perfor(cid:173)
`(RsXA) of -20 mfl cm2
`mance at higher concentration levels, the new 256-line pat(cid:173)
`tern shown in Fig. 3(a) was generated and cells fabricated.
`Cells with this pattern were tested in our new solar-cell test
`facility9 capable of reaching up to 2000 suns on I-cm-diam
`cells. Figure 5 shows the efficiency-vs-concentration data
`obtained on one of these new cells (No. 1208) tested up to
`- 900 suns and compares it with the performance of the
`earlier cell shown in Fig. 4. Table I lists the measured values
`of the conversion efficiency and the fill factor at various con(cid:173)
`centrations for these two cells. The Rs value estimated from
`the /- V curves for cell No. 1208 with the new contact pattern
`is -6mfl, giving aRsXA value of -4.7 mfl cm2
`• The rela(cid:173)
`tively high efficiency (- 22%) and fill factor for this cell do
`not show a significant fall off at the higher concentrations
`and demonstrate the improvement in series resistance real(cid:173)
`ized with the new contact pattern. The efficiency-vs-concen
`
`TABLE I. Concentration dependence of the cell efficiency and fill factor.
`
`Cell No.
`
`Concentration
`
`Efficiency (%)
`
`Fill factor
`
`976
`(AMI,50°C)
`
`1208
`(AM2, 65°C)
`
`74
`114
`178
`270
`440
`
`365
`596
`899
`
`24
`24
`24.7
`21.2
`20.3
`22.5
`22.3
`21.7
`
`0.855
`0.85
`0.822
`0.77
`0.71
`0.856
`0.835
`0.803
`
`tration behavior of Fig. 5 compares well with the predicted
`behavior shown in Fig. 2.
`In summary, we have demonstrated that the thin-win(cid:173)
`dow AIGaAs/GaAs solar-cell design leads to high conver(cid:173)
`sion efficiencies ( - 25%) and that an efficient grid design
`with direct Ohmic contact to the p-GaAs active layer leads
`to low series resistance (RsXA < 5 mil cm2
`) yielding effi(cid:173)
`cient cell operation at very high concentration levels.
`The assistance of L. Pearce and J. Elliot of Rockwell
`Space Division in making the concentrator measurements at
`Table Mountain and of J. Cape in setting up the high-con(cid:173)
`centration test facility at the Science Center is gratefully
`acknow ledged.
`
`'E.A. DeMeo and P.B. Bos, EPRI Special Report ER-589-SR, 1978
`(unpublished).
`'E.L Burgess, 20th Technical Symposium of Soc. of Photo-Optical In(cid:173)
`strum. Engineers, 1976 (unpublished).
`JR. Sahai, D.O. Edwall, E.S. Cory, and 1.S. Harris, 12th IEEE Photovoltaic
`Specialists Conference (IEEE, New York, 1976), p. 989.
`'LW. James R.L Moon, 11th IEEE Photovoltaic Specialists Conference
`(IEEE, New York, 1975), p. 402.
`'LW. James and R.L Moon, App!. Phys. Lett. 26, 476 (1975).
`'R. Sahai, D.O. Edwall, and J.S. Harris, 13th IEEE Photovoltaic Specialists
`Conference (IEEE, New York, 1978).
`'Courtesy of Dr. Henry W. Brandhorst, NASA-Lewis Research Center.
`SM. Wolf and H. Ranschenback, Adv. Energy Converso 3, 455 (1963).
`'J.A. Cape, J.S. Harris, and R. Sahai, Ref. 6.
`
`149
`
`Appl. Phys. lett., Vol. 34, No.2, 15 January 1979
`
`Sahai, Edwall, and Harris, Jr.
`
`149
`
` Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 18.9.61.112 On: Mon, 30 May 2016
`14:00:35
`
`Raytheon2020-0003