Iron pyrite nanocrystal film as a copper-free back contact for polycrystalline CdTe thin film solar cells

Abstract

The invention discloses nanocrystalline (NC) FeS.sub.2 thin films as the back contact for CdTe solar cells. In one example, the FeS.sub.2 NC layer is prepared from a solution directly on the CdTe surface using spin-casting and chemical treatment at ambient temperature and pressure, without a thermal treatment step. Solar cells prepared by applying the NC FeS.sub.2 back contact onto CdTe yield efficiencies of about 95% to 100% that of standard Cu/Au back contact devices. In another example, FeS.sub.2 is interposed between Cu and Au to form a Cu/FeS.sub.2 NC/Au back contact configuration yielding an efficiency improvement of 5 to 9 percent higher than standard Cu/Au devices.

Claims

1. A photovoltaic cell comprising: a front window comprising an n-type semiconductor material; a semiconductor layer comprising a p-type semiconductor material; and a back contact layer configured as an ohmic contact having an FeS.sub.2 portion; wherein the FeS.sub.2 portion is a nanocrystalline portion comprising a film of FeS.sub.2 in a cubic phase; and wherein the FeS.sub.2 portion comprises nanocrystals having an average size in a range of from about 70 nm to about 151 nm.

2. The photovoltaic cell of claim 1 wherein the FeS.sub.2 portion is in direct contact with the semiconductor layer.

3. The photovoltaic cell of claim 1 wherein the semiconductor layer comprises CdTe and the front window comprises CdS.

4. The photovoltaic cell of claim 1 wherein a conductive material forms a portion of the back contact layer.

5. The photovoltaic cell of claim 4 wherein the FeS.sub.2 portion is disposed between the semiconductor layer and the conductive material portion of the back contact layer.

6. The photovoltaic cell of claim 5 wherein the conductive material is one of Au, Cu, Sb, Hg, Bi-telluride, and graphene.

7. The photovoltaic cell of claim 4 wherein the conductive material is a first conductive material disposed between the semiconductor layer and the FeS.sub.2 portion, and a second conductive material is disposed on an opposite side of the FeS.sub.2 portion from the first conductive material.

8. The photovoltaic cell of claim 7 wherein the FeS.sub.2 portion has a thickness in a range of about 5 nanometers to about 1500 nanometers.

9. The photovoltaic cell of claim 7 wherein one of the first and second conductive materials is Cu and other of the first and second conductive materials is Au.

10. The photovoltaic cell of claim 4 wherein the semiconductor layer includes at least one pinhole that permits a charge carrier dysfunction in an interface region of the semiconductor layer and the FeS.sub.2 portion blocks the charge carrier dysfunction.

11. The photovoltaic cell of claim 10 wherein the conductive material is one of Au, Cu, Sb, Hg, Bi-telluride, and graphene.

12. The photovoltaic cell of claim 1 wherein the semiconductor layer includes one of CdTe, copper indium gallium di-selenide (CIGS), copper zinc tin sulfide (CZTS), copper zinc tin sulfur selenium alloy (CZTSSe), copper antimony sulfide, and tin sulfide.

13. The photovoltaic cell of claim 12 wherein a conductive material forms a portion of the back contact layer.

14. The photovoltaic cell of claim 13 wherein the conductive material is one of Au, Cu, Sb, Hg, Bi-telluride, and graphene.

15. The photovoltaic cell of claim 14 wherein the FeS.sub.2 portion directly contacts the semiconductor layer, the semiconductor layer including at least one pinhole that permits a shunting effect at an interface region, wherein the FeS.sub.2 portion blocks the shunting effect at the interface region.

16. A photovoltaic cell comprising: a transparent conductive layer permitting a portion of visible light to pass through and forming a first ohmic contact; a first active semiconductor layer having an n-type characteristic; a second active semiconductor layer having a p-type characteristic; and a back contact layer configured as a second ohmic contact having an FeS.sub.2 portion, the back contact layer applied onto the second active semiconductor layer; wherein the FeS.sub.2 portion is a nanocrystalline portion comprising a film of FeS.sub.2 in a cubic phase; and wherein the FeS.sub.2 portion comprises nanocrystals having an average size in a range of from about 70 nm to about 151 nm.

17. The photovoltaic cell of claim 16 wherein the transparent conductive layer includes a resistive coating layer, the first active semiconductor layer is a CdS layer, the second active semiconductor layer is a CdTe layer, and the back contact layer includes a conductive material.

18. The photovoltaic cell of claim 17 wherein the conductive material is one of Au, Cu, Sb, Hg, Bi-telluride, and graphene.

19. A photovoltaic cell comprising: a transparent conductive layer permitting a portion of visible light to pass through and forming a first ohmic contact; a CdS semiconductor layer having an n-type characteristic; a CdTe semiconductor layer having a p-type characteristic; and a back contact layer configured as a second ohmic contact, the back contact layer including a first conductive material applied onto the CdTe semiconductor layer, an FeS.sub.2 nanocrystalline layer applied onto the first conductive material, and a second conductive material applied onto the FeS.sub.2 nanocrystalline layer; wherein the FeS.sub.2 nanocrystalline layer comprises a film of FeS.sub.2 in a cubic phase; and wherein the FeS.sub.2 nanocrystalline layer comprises nanocrystals having an average size in a range of from about 70 nm to about 151 nm.

20. The photovoltaic cell of claim 19 wherein the FeS.sub.2 nanocrystalline layer is a ligand-free layer.

21. The photovoltaic cell of claim 1, wherein the film of FeS.sub.2 has a thickness of greater than 1.5 microns.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The patent or application file may contain one or more drawings executed in color and/or one or more photographs. Copies of this patent or patent application publication with color drawing(s) and/or photograph(s) will be provided by the U.S. Patent and Trademark Office upon request and payment of the necessary fees.

(2) FIG. 1 is a graph of non-equilibrium band diagram showing conduction and valence band positions relative to the vacuum level;

(3) FIG. 2(a) is a graph of absorbance spectrum of as-obtained NC film dispersed in chloroform in accordance with the invention;

(4) FIG. 2(b) is a graph of absorbance spectrum of NC film deposited on soda lime glass in accordance with the invention;

(5) FIG. 2(c) is a graph of direct band gap determination for FeS.sub.2 material in accordance with the invention;

(6) FIG. 2(d) is a graph of indirect band gap determination for FeS.sub.2 material in accordance with the invention;

(7) FIG. 2(e) is a graph of absorbance spectra of as-synthesized and hydrazine treated FeS.sub.2 NC films taken from FTIR measurement;

(8) FIG. 2(f) is a graph of absorbance spectra of FeS.sub.2 NC thin films, prepared on soda-lime glass substrate, before and after hydrazine treatment;

(9) FIG. 3(a) is a graph of an XRD spectrum (focused beam) characterization of untreated and unsintered FeS.sub.2 NC film deposited by drop-cast using NCs of size 130 nm;

(10) FIG. 3(b) is a graph of a Raman scattering spectrum characterization of an untreated and unsintered FeS.sub.2 NC film deposited by drop-cast using NCs of size 130 nm;

(11) FIG. 3(c) is an SEM image (10 kV accelerating potential) of untreated and unsintered FeS.sub.2 NC film deposited by drop-cast using NCs of size 130 nm;

(12) FIG. 3(d) a plot of EDX measurements of seven FeS.sub.2 NC films fabricated from distinct NC syntheses;

(13) FIG. 4(a) is a graph of XRD pattern of oleylamine (OLA) and 1,2-hexanediol capped FeS.sub.2 NCs;

(14) FIG. 4(b) is a graph of Raman spectra of phonon vibrations of FeS.sub.2 for OLA capped and 1,2-hexanediol capped NCs;

(15) FIG. 4(c) is an SEM image for 1,2-hexanediol capped NCs;

(16) FIG. 4(d) is an SEM image for OLA capped FeS.sub.2 NCs;

(17) FIG. 5 is a table showing the effect of annealing temperature and time on the FWHM of the (200) XRD peaks measured for FeS.sub.2 NC films;

(18) FIG. 6(a) is a plot of an XRD spectral characterization of FeS.sub.2 NC films deposited by LbL drop-cast method using NCs of size 70 nm showing the effect of hydrazine and thermal annealing treatment;

(19) FIG. 6(b) is a plot of a Raman spectrum characterization of FeS.sub.2 NC films deposited by LbL drop-cast method using NC of size 70 nm for a film annealed at 540 C. for 1 hour;

(20) FIG. 6(c) is an SEM image for an as-deposited FeS.sub.2 NC film deposited by LbL drop-cast method using NC of size 70 nm film

(21) FIG. 6(d) is an SEM image for a FeS.sub.2 NC film deposited by LbL drop-cast method using NC of size 70 nm, the film treated in hydrazine and annealed at 540 C. for 3 hours;

(22) FIG. 7 is a table showing the sheet resistance for two pairs of two different FeS.sub.2 NC films on soda-lime glass, each in three different conditions;

(23) FIG. 8 is a table showing Hall Measurements; average thickness of the films=3.5 m;

(24) FIG. 9 is a table showing average parameters for 15 CdTe solar cells before and after heating;

(25) FIG. 10(a) is a plot of current density vs. bias voltage measurements of sputtered CdS/CdTe devices having Au, Cu/Au, and FeS.sub.2/Au back contacts evaluated under simulated AM1.5G solar spectrum conditions;

(26) FIG. 10(b) is a plot of external quantum efficiency for sputtered CdS/CdTe devices with Au, Cu/Au and FeS.sub.2/Au back contacts.

(27) FIG. 11(a) is a plot of J-V curves for CSS CdTe devices where the FeS.sub.2 thickness varies from 0.35 m to 1.4 m; comparison is made with standard device.

(28) FIG. 11(b) is a plot of EQE of CSS CdTe solar cells in three different FeS.sub.2 thicknesses along with EQE of standard device

(29) FIG. 12 is a plot of J-V curves for a sputtered CdTe photovoltaic cells having a back contact of untreated FeS.sub.2 NC film.

(30) FIG. 13a is a schematic representation of a conventional CdS/CdTe solar cell having a Cu/Au back contact.

(31) FIG. 13b is a schematic representation of an embodiment of a CdS/CdTe solar cell having an iron pyrite/metal (FeS.sub.2-NC/Au) back contact in accordance with an embodiment of the invention.

(32) FIG. 13c is a schematic representation of an alternative embodiment of a CdS/CdTe solar cell having a metal/iron pyrite/metal (eg. Cu/FeS.sub.2-NC/Au) back contact in accordance with another embodiment of the invention.

(33) FIG. 14 is a plot of current density-voltage (J-V) characteristics of CdS/CdTe solar cells using Cu in conjunction with a FeS.sub.2/Au back contact compared with standard back contact (Cu/Au).

(34) FIG. 15 is a comparative plot of current density-voltage (J-V) characteristics of CdS/CdTe solar cells using an FeS.sub.2/Au back contact and a standard back contact (Cu/Au).

(35) FIG. 16 is a comparative table of CdTe solar cell performance having different back contact configurations.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT(S)

(36) In a first aspect, there is provided herein an iron pyrite (iron persulfide, FeS.sub.2) nanocrystal-based thin film low barrier back contact to CdTe solar cells. While the semiconductor layers are described in the context of CdS and CdTe layers, other materials may be used. For example, the various semiconductor active layers may alternatively be formed from CIGS (copper indium zinc gallium di-selenide), CZTS (copper zinc tin sulfide), CZTSSe (copper zinc tin sulfur selenium alloy), tin sulfide, and copper antimony sulfide. In one embodiment, the FeS.sub.2-based back contact can be applied to the cell without pre-treatment of the CdTe p-type film. Creating a low-barrier, ohmic back contact to CdTe often involves addressing both the high work function and the low resistivity of the thin film. Standard back contact preparation often introduces copper, as a thin evaporated layer or as a CuCl.sub.2 solution deposition, followed by thermally-assisted diffusion to create a low-resistivity Cu.sub.xTe phase. Several studies have shown that Cu diffuses readily, and over time reaches the CdS/CdTe interface, reducing the operating voltage of the device. Copper diffusion therefore serves as a critical pathway to degradation, influencing device performance over the life of a PV systemultimately degrading the economic performance of the technology. Alleviating the degradation associated with Cu diffusion has advantages for CdTe-based photovoltaic systems.

(37) Iron pyrite exhibits a yellowish metallic luster reminiscent of gold, and is commonly known as Fool's Gold. As-deposited FeS.sub.2 NC films, evaluated at room-temperature using four-point probe and Hall-effect measurements, have shown low resistivity of about 100 -cm, high free carrier concentrations of about 10.sup.19 cm.sup.3, and a low electron mobility of about 10.sup.1 cm.sup.2V.sup.1s.sup.1. In addition, FeS.sub.2 NCs show strong sub-bandgap optical absorption, indicating the presence of sub-bandgap electronic states arising from core and/or surface defects. Thus, the inventors have concluded that a strong photovoltaic performance would not be expected when using FeS.sub.2 as an absorber layer, based on the combination of high free carrier density, low mobility, and mid-gap absorption leading to reduced photovoltage. While not wishing to be bound by theory, the inventors believe that prior poor cell performance, compared with CdTe-based cells, can be attributed, in part, to both strong Fermi level pinning associated with surface states and to bulk defects caused by sulfur deficiency and local variations in crystal structure.

(38) As stated above for CdS/CdTe solar cells, conventional back contacts are commonly made with Cu/Au or Cu/graphite. Copper introduced at the back contact diffuses atomically to the CdS/CdTe junction, resulting in shunting at the n-p junction. The inventors' development of forms of FeS.sub.2 as a back contact to CdTe provides a Cu-free and low-cost cell structure. Alternatively, these forms of FeS.sub.2 can be used in conjunction with Cu, or other metals, to form an alloyed back contact layer. Such a construction may include orientations where Cu is applied to the FeS.sub.2 material and positioned as an outermost back contact layer, Cu is positioned between the FeS.sub.2 material and the adjacent semiconductor layer, or the back contact layer is formed from three sub-layers as a metal interface with the semiconductor layer adjacent to the FeS.sub.2 material and a final metal layer. In addition, it is contemplated by the inventors that, for example, Co.sup.2+ or Ni.sup.2+ may substitute for Fe.sup.2+, and other Group VI elements such as, for example, Se or Te can substitute for S. Further, the inventors contemplate that preparation of the CdTe layer to accept and improve charge transport may involve exposure to an activation agent, such as FeCl.sub.2 or FeCl.sub.3, in a similar manner that CuCl.sub.2 is used, as described above. The inventors have observed that the work function of FeS.sub.2 (5.45 eV) compares favorably with that of Au (5.1 eV). As shown in the non-interacting band diagram of FIG. 1, the inventors have recognized the possibility of a barrier-less interface using an FeS.sub.2-based back contact. The inventors have demonstrated the applicability of such a barrier-less interface in spite of recognizing that charge polarization effects associated with Fermi level equilibration typically introduces band bending. The inventors' recognition of the high free carrier concentration and low resistivity have resulted in FeS.sub.2 NC development of a conductive back contact layer. Therefore, despite repeated, failed attempts to use FeS.sub.2 as a solar cell absorber layer, the inventors have realized the potential of FeS.sub.2 NCs as an ohmic or low barrier contact to CdTe. As used herein, the term ohmic, in the context of back contact electrical performance, is understood to include both ohmic and near ohmic contacts. As generally understood, ohmic contacts typically have no electronic barrier to the flow of holes at the back contact. Near ohmic contacts may exhibit some retained barrier to hole flow, as may be the case with some (though not all) embodiments of back contacts using FeS.sub.2.

(39) In one embodiment of a method of forming a photovoltaic cell having FeS.sub.2-based back contact, the inventors have created a solution-based synthesis and deposition process that offers a low-cost and scalable photovoltaic manufacturing method for large glass substrate processes and roll-to-roll processing on flexible substrates. In spite of other routes available to synthesize FeS.sub.2 NCs, the inventors have developed this solution-based method, in part due to the use of a thermal injection reaction of an iron salt solution with an elemental sulfur source. Whereas FeCl.sub.2 may be used as the iron source, the inventors have developed the method using FeBr.sub.2, which yields improved results regarding crystal structure and infrared absorption. In an embodiment of the method, the synthesis of FeS.sub.2 NCs, which may be performed in a Schlenk line under N.sub.2 environment, begins with about 1.49 mmol of FeBr.sub.2 (321 mg) and 3 mmol of trioctylphosphine oxide (TOPO) (1.16 g), which are mixed in 30 mL of oleylamine (OLA) in a three neck flask under constant stirring. The FeBr.sub.2 mixture is heated to 170 C. for 2 hours and 30 minutes using a heating mantle; during this time, the sulfur precursor solution is prepared. For this precursor, 8.98 mmol of elemental sulfur (288 mg) is dissolved in 15 mL of OLA. For complete dissolution of sulfur in OLA, 10 minutes of ultra-sonication is performed. The sulfur solution is kept in hot water bath at 90 C. Once the sulfur solution is ready, the temperature of the FeBr.sub.2 solution is raised toward 220 C., and once it exceeds 216 C., the sulfur solution is rapidly injected. Nucleation of FeS.sub.2 clusters initiates upon sulfur injection, and the growth of FeS.sub.2 NCs proceeds at a temperature of 220 C.

(40) Following two hours at 220 C., the NC solution is allowed to cool to room temperature, with continued stirring. Nanocrystals so obtained are washed, in one embodiment of the method they are washed a minimum of three times, using methanol as a non-solvent and toluene as solvent. In a first wash, methanol is added to the as-synthesized NC solution, followed by centrifugation for 10 minutes at 5000 rpm. After decanting the supernatant, the solvent is used to disperse the NCs with the assistance of sonication. Then, methanol is added to effect precipitation of the NCs, which allows for physical separation via centrifugation. In this embodiment, the washing procedure is repeated three times. Finally NCs so obtained are dried under nitrogen gas flow.

(41) In another embodiment of the synthesis process, TOPO may be used as the surfactant and OLA as a non-coordinating solvent. In this method, FeS.sub.2 NCs so obtained may be capped by TOPO. In an alternate embodiment, FeS.sub.2 NCs can be synthesized using OLA without the presence of TOPO. In yet another alternate embodiment, high-quality FeS.sub.2 NC can also be synthesized using 1,2-hexanediol as the surfactant and OLA as a non-coordinating solvent.

(42) As described below, the inventors have developed embodiments of a method to fabricate FeS.sub.2-based films suitable as back contact layers. Because of their large size (70 nm-140 nm), FeS.sub.2 NCs do not remain in stable suspension for long periods of time. A well-dispersed but unstirred solution will effectively change in concentration as the NCs settle to the bottom of the container. Thus, some impediments to conventional dip-coating methods and spin-coating have led to use of a drop-cast method for initial development. The inventors understand that these aforementioned impediments may be ameliorated using known production techniques to make NC films. For the sake of expediency, the inventors have utilized drop-cast films in a layer-by-layer (LbL) method. To fabricate the FeS.sub.2 NC films, an FeS.sub.2 NC solution was prepared in chloroform at a concentration of about 6 mg/ml. Film formation proceeded in an N.sub.2 environment. In an embodiment of the method, a layer of drop-cast NCs is deposited onto the chosen substrate, and allowed to dry. At this point, the film can optionally be treated with hydrazine for ligand removal, as described below. In the case of an untreated film, the film thickness may be increased by simply repeating the drop-cast process followed by the drying process. The inventors have found that preparation of an approximately 1 m film, in one embodiment of the method, typically relies on 2 drop-cast cycles.

(43) To improve the conductivity of the NC films, in one embodiment of the method, long chain hydrocarbon molecules (C.sub.24H.sub.51OP, TOPO) were removed from the NC surface in LbL process by cyclically depositing and then treating films with 1 M hydrazine in ethanol. An embodiment defining preparation of an FeS.sub.2 film treated with hydrazine to remove the surfactant is accomplished as follows. Subsequent to the first drop-cast layer deposition, the film is allowed to dry in the N.sub.2 environment. The film is subsequently submerged in a 1 M hydrazine solution in ethanol for about 2 minutes. The film is withdrawn from the hydrazine solution and immediately submerged into a pure ethanol solution to remove any residual surfactant or hydrazinei.e., as a rinse. The film is then allowed to dry. To attain a thicker film, the drop-cast/dry/hydrazine/rinse/dry process may be repeated as necessary for the desired film thickness.

(44) FeS.sub.2 NC films prepared by the LbL drop-casting method, above, may at times exhibit pin holes, i.e., microscopic areas of incomplete coverage by FeS.sub.2, which have been found to occur independent of film thickness up to even 1 m. The inventors have observed that, in some materials, an improvement in charge carrier mobility may be realized when neighboring NCs are brought into improved contact and/or surface defect states are removed. In an effort to ameliorate the presence of pinholes and to improve control over electronic properties, the NC films were sintered in the presence of sulfur vapor. The inventors believe that sintering may anneal the NCs together and/or promote grain growth resulting in a film exhibiting more uniform coverage and/or improved electronic properties.

(45) In one embodiment, sintering of the films is conducted in a cylindrical quartz tube furnace such that the film is heated uniformly. In the quartz tube, two heaters are arranged: one is being used for evaporating the elemental sulfur at 350 C. and the other is used to anneal the sample at various temperatures. Two ends of the quartz tube are sealed with flanges fitted with O-rings. Initially the tube is purged with a forming gas (95% argon, 5% hydrogen) for about 5-10 minutes and then low pressure argon gas (5 SCCM) is introduced during the sintering process. The sintering process proceeds with the substrate and film held at a temperature of about 500 C., within a sulfur vapor, for 1-3 hours. Alternatively, the temperature may be in a range of about 500 C. to about 600 C., and more particularly from about 500 C. to about 540 C. Sintering FeS.sub.2 NC films relies on process monitoring due to a phase change from FeS.sub.2 to FeS at high temperature. In other embodiments of the method and of the PV cells, FeS may also provide a suitable medium for Cu-free back contact formation. According to this particular embodiment of the method, a sulfur vapor is provided during sintering to protect pyrite FeS.sub.2 from phase change at elevated temperature.

(46) An embodiment of fabricating the semiconductor layers of a CdS/CdTe photovoltaic cell are describe. Cells described as sputtered and having CdS/CdTe layers were grown by RF magnetron sputtering. The cells consisted of a CdS film sputtered onto fluorine doped SnO.sub.2 transparent conducting layer deposited on soda lime glass substrate, on top of which was deposited CdTe by the method of closed spaced sublimation (CSS). The inventors believe that magnetron sputtering techniques provide films with better adhesion, smaller grain size and relatively smooth surface due to its slow deposition rate at low temperature (200-300 C.). The inventors have also found that CSS has been a productive method due to the high deposition rate, low material consumption and low cost of operation. As compared with all-sputtered CdS/CdTe films, CSS of CdTe yields films with larger grain size (1 m vs. 100 nm) due to the high deposition temperature (600 C.). Independent of the deposition method, following the CdTe deposition a CdCl.sub.2 treatment was carried out by dipping the CdTe layer in a solution of CdCl.sub.2-methanol and subsequently annealing at 378 C. in dry air to advance grain growth, release interfacial strain, and facilitate sulfur and tellurium mixing at the CdS/CdTe interface. The thickness of CdS films in both methods was 80 nm whereas the sputtered CdTe was 2.0 m and the CSS CdTe was 4 m. The inventors have found sputtered CdS/CdTe films enabling a typical conversion efficiency of 12%, and CSS CdTe on sputtered CdS films enabling typical devices of 14.0% efficiency. All device efficiency measurements were made under AM1.5G simulated solar spectrum at ambient laboratory temperature. For comparison purposes, standard devices were prepared on TEC15 TCO-coated glass (Pilkington N.A.) with a high-resistivity layer as part of the normal coating process. Following CdS/CdTe deposition, the standard back contact consists of a Cu/Au sequential deposition in which 3 nm of Cu followed by 30 nm of Au is evaporated onto the CdTe, and the film is then heated to 150 C. for about 45 minutes in air to drive Cu diffusion and improve the quality of the Au contact.

(47) FeS.sub.2 NC films were produced for evaluation of their structural and electrical characteristics. Absorbance spectrum of FeS.sub.2 NCs dispersed in chloroform as well as absorbance spectrum of FeS.sub.2 film deposited onto soda lime glass substrate are shown in FIGS. 2a and 2b. As shown in the graphs of FIGS. 2(a) and 2(b), most of the light absorption takes place throughout visible and near-infrared regions. In the region higher than 1200 nm, the film is more transmissive, though not completely. The absorption of light in the infrared region below the indirect band gap energy (0.95 eV, 1305 nm) has been ascribed to S vacancies in the FeS.sub.2 film. It has been found that FeS.sub.2 films, which show p-type defects and a high free carrier concentration, are prone to the formation of low-energy phases of Fe:S stoichiometry exceeding 0.5. Such phases include troilite (FeS) and pyrrhotite (FeS.sub.1+x, x=0- 1/7). Thus, the sub-bandgap infrared absorption may be ascribed to absorption caused by non-FeS.sub.2 iron sulfide phases. FIGS. 2(c) and 2(d) show the measured results of direct and indirect band gaps of FeS.sub.2 NCs in solution via optical absorbance spectroscopy. Based on the concentration of the NCs in solution together with the path length, an effective thickness of FeS.sub.2 is used to convert optical absorbance to absorption coefficient (). Then, a graph of (h).sup.2 vs. h is used to extract the direct band gap of 1.3 eV, and a graph of (h).sup.1/2 vs. h is used to extract the indirect band gap of 0.94 eV as shown in FIGS. 2c and 2d.

(48) Iron pyrite NC films are prepared, as described above, using LbL drop-cast method. The films, used for structural and electrical characterization, are deposited on soda-lime glass or onto a zero-background single-crystal Si substrate. The films may be hydrazine-treated, untreated, and sintered or unsintered. Hydrazine treatment of the films is found to remove the capping ligands, serving to modify the electronic properties of the film. As shown in FIG. 2(e), in FTIR spectra, red is the reference background spectrum obtained from the bare Si substrate, while the green and blue lines respectively show signals for the FeS.sub.2 NC film before and after ligand exchange. The CH stretch signatures near 3000 cm.sup.1 and at 1500 cm.sup.1 show quantitative removal and/or replacement of TOPO (or 1,2 hexanediol) through hydrazine treatment.

(49) FIGS. 3(a) to 3(c) show XRD, SEM and Raman spectra of as-deposited FeS.sub.2 NC films. For XRD, the NC film was prepared on zero background Si substrate whereas for SEM and Raman they were prepared on soda lime glass substrate. As shown in FIG. 3(d), energy-dispersive X-ray spectroscopy (EDX) was used to determine the stoichiometry of FeS.sub.2 NC films. The XRD image in FIG. 3(a) shows pure FeS.sub.2 cubic phase with no evidence of other crystal structures. The sharp peaks in the XRD pattern indicate excellent crystallinity of the as-synthesized FeS.sub.2 NCs. Three digit numbers in the bracket represent the Miller Indices for cubic crystal structures. XRD images for OLA capped and 1,2-hexanediol capped FeS.sub.2 NCs are shown in FIGS. 4(a) to 4(d).

(50) Referring again to FIG. 3(b), the Raman spectra show peaks at 343.7 cm.sup.1, 380 cm.sup.1 and 431 cm.sup.1 which are consistent with phonon vibrations for FeS.sub.2. These results rely on a value of =632.8 nm for excitation of FeS.sub.2 film, due to a belief that that iron pyrite absorbs more strongly at 532 nm wavelength than at 633 nm. The Raman peaks we observe are well separated from Raman peaks reported for troilote (FeS) which shows peaks at 210 cm.sup.1 and 280 cm.sup.1 Raman spectra of OLA capped and 1,2-hexanediol capped FeS.sub.2 NCs are shown in FIGS. 4(a) and 4(b). Uniform cubically-shaped FeS.sub.2 NCs synthesized with TOPO/OLA combinations are shown in FIG. 3(c). Similarly, SEM images of FeS.sub.2 NCs synthesized with 1,2-hexanediol/OLA combination and OLA alone are shown in FIGS. 4(c) and 4(d). The size of the NCs can be varied by varying the amount of surfactant. The average size of these large cubes is in a range of about 13318 nm. Reducing the amount of TOPO resulted in FeS.sub.2 NCs with average edge length <70 nm.

(51) Photovoltaic cells previously produced using iron pyrite as one of the light-absorbing semiconductor layers exhibited reduced open circuit voltage (Voc) performance. The observed limitation of V.sub.OC performance was thought to be due to the possible significant sulfur deficiency. However, significant changes in the crystalline structure, and therefore the electronic properties of compound semiconductors, can arise from formation of phases that may correspond to relatively small deviations in stoichiometry. After these early photo-electrochemical solar cells, efficient solar cells using iron pyrite as a semiconductor layer have been elusive. Many of these prior FeS.sub.2 semiconductor cells exhibited a significant decrease in S:Fe ratio in samples considered to be nominally iron pyrite. The S:Fe ratio was found to range from 2:1 to 1.74:1. Iron pyrite NCs used in the creation of FeS.sub.2 NC back contact cells, as synthesized by the inventors, exhibit an essentially stoichiometric ratio, as shown in FIG. 3(d). Energy Dispersive X-ray Spectroscopy (EDX) measurement for seven different batches of FeS.sub.2 NCs, synthesized with varying amounts of surfactant yielded an average S:Fe ratio of 2.01:1 representing iron pyrite as a stoichiometric compound. The inventors, however, note that even small amounts of phase impurities, especially those near the 2:1 S:Fe ratio, may noticeably alter the aggregate sample optical and electronic properties. As shown in FIG. 3(a), considering the intense peak at (200), the full-width-at-half-maxima (FWHM) and grain size were calculated for all XRD spectra taken at different temperatures. The table of FIG. 4 shows that with increased temperature and annealing time, the (200) peak FWHM decreases, which also corresponds to an increase in the average grain size.

(52) In order to confirm the limitations in performance of cells using FeS.sub.2 films as semiconductor layers, FeS.sub.2 films that were (1) as-synthesized, (2) hydrazine treated, and (3) hydrazine treated and annealed, for the fabrication of Schottky junction and heterojunction solar cells employing FeS.sub.2 as the absorber layer. In all cases, results showed no improvement in PV performance resulting from hydrazine or thermal annealing treatments. Performance of the devices was slightly better when ITO was used as a front contact instead of HRT/TEC15. However, in all cases the PV devices yielded effectively zero open circuit photovoltage and zero short-circuit current density (the J-V curve passed through the origin).

(53) Electrical properties of FeS.sub.2 NC films were studied using hot probe measurement, four point probe measurement, and Hall measurement methods, with results summarized in the tables of FIGS. 7 and 8. All hot probe measurements indicated that the FeS.sub.2 films used as semiconductor layers were p-type, in agreement with comparative data for other polycrystalline and NC-based films. This indicated that the majority of charge carriers in these pyrite films were holes. The table of FIG. 7 shows the sheet resistance for two pairs of two different FeS.sub.2 NCs on soda-lime glass, each in three or four different conditions, depending on the surfactant/solvent type. For the first sample type, FeS.sub.2 NCs were synthesized using 1,2-hexanediol/OLA. For the second case, FeS.sub.2 NCs were synthesized using TOPO/OLA. In the first case, three different films were prepared: the first as-synthesized NC film (condition: No, No), the second NC film treated with hydrazine but at room temperature (condition: Yes, No) and the third NC film treated with hydrazine and annealed in sulfur vapor for an hour at about 500 C. (condition: Yes, Yes). In the second case, the additional condition of an as-synthesized NC film annealed in sulfur vapor (No, Yes) was tested. The sheet resistance of the films decreases by a factor of about 10 for films in the as-synthesized condition treated with hydrazine. Sheet resistance also decreased by another factor of 10 for films hydrazine-treated and annealed versus only treated with hydrazine. Resistivity of the films in each case is obtained by multiplying sheet resistance by the average thickness of the films. These sheet resistances are very close to ones obtained from Hall measurement as given in the table of FIG. 8. The decrease in sheet resistance of the hydrazine treated films correlates with the removal of the organic molecules from the surface of the NCs which insulate neighboring NCs against electrical conduction. When the films are annealed at high temperature, any leftover organic molecules are evaporated, and also the grain size of the NCs increases slightly which reduces the density of grain boundaries within the film.

(54) Free carrier concentrations of the FeS.sub.2 NC films increased by a factor of about 2 following hydrazine treatments and by another factor of about 2 to about 5 following sulfur vapor thermal annealing. The maximum carrier concentration of treated and annealed film was on the order of 10.sup.19 cm.sup.3 or higher in some cases. Surprisingly, the inventors have found that these FeS.sub.2 NCs, even in their apparent pure phase, possess very high carrier concentrations. Also surprising is that when the films are annealed at high temperature (>500 C.), carrier concentrations are found to increase. As obtained by Hall measurement, in all different conditions, the mobility of the carriers in the NC film is found to be <<1 cm.sup.2 V.sup.1 s.sup.1. Conductivity is proportional to the product of carrier concentration and mobility. Very low mobility within an absorber layer strongly affects the workings of solar cells because poor transport under an electric field (drift) may lead to substantial charge at interfaces under solar illumination. In addition, high carrier concentration leads to very short depletion widths, and one would have to rely on relatively long diffusion lengths.

(55) The built-in electric field in the depletion region of a heterojunction solar cell serves to separate the photogenerated charge carriers across the interface. If the carrier concentration on p-region is significantly higher than for the n-region, width of window layer is depleted by the majority charge carriers of p-type region. However, the built-in electric field in n-type material does not contribute to the collection of photogenerated charge carriers as most of the incident photons pass through this region because of its wide band gap. In our efforts to create FeS.sub.2 absorber layer solar cells, the window layer transmitted most of the solar spectrum, and those photons absorbed within the n-type window layer did not generate meaningful photocurrent. In addition, the photogenerated e-h pairs within the FeS.sub.2 primarily recombine prior to diffusing to the (small) depletion region where they could be separated.

(56) As mentioned above, FeS.sub.2 NC films have made a high-performance back contact to CdTe solar cells without the use of Cu. Representative current density/voltage (J-V) characteristics for sputtered CdTe devices prepared with different back contacts are shown in FIG. 10(a), and average device parameters are presented in the table of FIG. 9. These solar cells were prepared on 11 substrates and mechanically scribed to produce smaller cell areas of approximately 0.085 cm.sup.2.

(57) The table of FIG. 9 demonstrates average and standard deviations of 15 cells of one kind prepared on a substrate. Each J-V curve in FIG. 10(a) shows representatives of these 15 cells. The solid lines represent light response and broken lines represent dark response of the devices. The red line represents the J-V curve when 30 nm Au was evaporated as a back contact, without any Cu diffusion layer, showing a short circuit current density of 20.3 mA cm.sup.2. Although the Jsc does not drop significantly compared to the other back contact types, Voc and efficiency () are poor when Au is used as a back contact. Because of the high electron affinity of CdTe, 4.5 eV, a high work function material is desirable to form a zero barrier height (ohmic) contact with p-type CdTe. With available metals used in typical back contact layers, a Schottky barrier is formed with CdTe at its typical free hole concentration. The resulting barrier causes a significant limitation to hole transport from the CdTe to metal. The Schottky junction acts as a diode, the direction of which opposes the main diode formed between CdS/CdTe interfaces. The diode at the back contact causes the current-voltage curve to roll-over at a forward bias, decreasing both the fill factor as well as the open-circuit voltage.

(58) The above-mentioned limitations associated with using a metal-only (e.g. Au) back contact are minimized by depositing about 3 nm Cu and about 30 nm of Au onto CdTe and annealing the film at 150 C. for about 45 minutes. The interdiffused Cu can form a distinct Cu.sub.2-xTe layer and increase the effective doping level of a thin layer of the CdTe through inter-diffusion. The Cu.sub.2-xTe layer increases the conductivity and reduces the barriers, thus allowing ohmic behavior at this interface. Referring now to FIG. 10(a), when a 3 nm thick layer of Cu was deposited before Au (blue curve), the Voc improved from 0.679 V to 0.794 V and efficiency improved from 8.7% to 11.4%. When 3 nm Cu was replaced by a 1.5 m thick FeS.sub.2 NC layer at room temperature and 30 nm thick Au atop of it (green curve), performance of the device was similar to the Cu/Au standard back contact. The FeS.sub.2 NC back contact, which consisted of an unheated hydrazine-treated film, showed a Voc=0.774 V, J.sub.SC=21.7 mA-cm.sup.2, and a 10.6% efficiency. The fill factor did decrease due to a residual barrier, and there was a slight increase in the apparent series resistance at the back contact. The inventors have observed an increase in the Jsc several times. While not wishing to be bound by theory, the inventors believe that sources of the increased Jsc are one or more of (a) Reabsorption of transmitted light by back-reflection, (b) a reduced interfacial recombination velocity at the back contact, and (c) a systematic error in the device area. The inventors believe that source (c) has been preliminarily addressed by QE measurements, which do not depend on the device area. As discussed below, the QE measurements shown in FIG. 10(b) show that the FeS.sub.2 back contact improves carrier collection in the CdTe and reduces the collection efficiency for photons absorbed in the CdS.

(59) The structure of solar cells formed having a conventional metal/metal (Cu/Au) back contact, an iron pyrite/metal (FeS.sub.2-NC/Au) back contact, and a metal/iron pyrite/metal (Cu/FeS.sub.2-NC/Au) are shown in FIGS. 13(a), 13(b), and 13(c). While the inventors have produced CdTe solar cells exhibiting high-performance back contacts utilizing FeS.sub.2 NC films made without Cu, to mitigate semiconductor layer shunting issues, Cu can provide benefits to cell performance. The thin layer of Cu utilized in the standard Cu/Au back contact improves performance by increasing the CdTe free hole concentration, one effect of which is a narrowing of the residual back barrier and improved hole tunneling efficiency. The inventors have realized an added benefit to performance by the inclusion of a metal, such as Cu, between the FeS.sub.2-NC/Au layer and the adjacent CdTe semiconductor layer. The demonstrated beneficial effects of a Cu/FeS.sub.2-NC/Au back contact configuration are shown in FIG. 14. Devices were fabricated where three-layer back contacts were deposited as 3 nm Cu/1 Pm FeS.sub.2/45 nm Au. The devices were formed as both sputtered and CSS CdTe devices. As shown in FIG. 14, the presence of the FeS.sub.2 NC layer improved the performance of the devices compared with that of the standard Cu/Au back contact devices. While not completely making the barrier ohmic, when the FeS.sub.2-NC layer is deposited onto Cu and the back contact is completed with Au, the performance of the device increases. While not wishing to be bound by theory, the inventors believe that copper narrows the barrier width, and FeS.sub.2 reduces the series resistance and increases V.sub.OC. This may be an effect of the FeS.sub.2 NC layer serving as a buffer layer with a relatively high work function. The device efficiency improves by 5-9% when incorporating an FeS.sub.2 NC layer between the Cu-treated CdTe and the Au external contact layer.

(60) Contrary to Cu/Au back contacts, when Au was deposited after FeS.sub.2 layer, the device was not heated. A very thick layer of >1.5 m of FeS.sub.2 NCs was deposited onto the CdTe layer to prevent the final Au layer from reaching the CdTe layer through FeS.sub.2 layer pin holes. While a thick layer of FeS.sub.2 NCs may decrease fill factor of the solar cells because of a series resistance increase at the back contact, the FeS.sub.2/Au samples compared favorably over conventional Au and Cu/Au back contacts. To test the stability of the back contacts (Au, Cu/Au and FeS.sub.2-NC/Au), the completed devices were tested at room temperature, and then again after heating to 100 C. in an N.sub.2 environment for several hours. The table of FIG. 9 shows performance data for all three devices before and after the heat treatment. Any decreases in Voc and Jsc is comparatively less when FeS.sub.2/Au combination was used as a back contact. After a 23 hour period, cell efficiency of those having FeS.sub.2/Au back contacts is similar to the Cu/Au cells. The inventors have found better performance when these devices are heated to about 200 C. for a short period of time.

(61) Carrier collection efficiency of the devices used for FIG. 10(a) were examined by comparing their external quantum efficiency (EQE) measurements of devices with Au, Cu/Au and FeS.sub.2/Au back contacts and given in FIG. 10(b). In the wavelength region from 500 nm to 850 nm, carrier collection when Au was used as a back contact was less than when the back contacts were Cu/Au and FeS.sub.2/Au based. While not wishing to be bound by theory, the inventors believe that the observed deep penetration loss seen at wavelengths higher than 700 nm is caused by insufficient thickness of CdTe layer. In other words, the penetration depth of light at this wavelength region is higher than the thickness of the CdTe film. Higher carrier collection in long wavelength region, when FeS.sub.2 NC was used as a back contact, suggests a slower recombination velocity of minority charge carrier generated near the back contact. The EQE in the region from 375 nm to 500 nm depends on the CdS thickness and there is a slight variation of CdS thickness in each sample. In the case of FeS.sub.2/Au back contacts, an increase in band gap of the CdTe by about 10 nm can be observed.

(62) The dependence of device performance on the FeS.sub.2 NC thickness as a back contact is illustrated in the graphs of FIGS. 11(a) and 11(b). For comparison purposes, a CdTe device having an efficiency of about 14.5%, under STC, with Cu/Au as the back contact was used as a baseline (black curve). The Cu/Au back contact was replaced with varying thicknesses of FeS.sub.2 NC layers (from 0.35 m to 1.5 m), fixing the other device architecture and processing parameters. An about 30 nm Au layer was deposited following fabrication of the hydrazine-treated FeS.sub.2 NC film. The current-density voltage curves in all four different cases are shown in FIG. 11(a) and corresponding QE curves are shown in FIG. 11(b). The inventors found that the thin FeS.sub.2 layers (0.35 m) tended to yield lower performance, perhaps due to an increased occurrence of pinholes which may allow Au to contact the CdTe directly, thus, lowering the Voc. The thickest FeS.sub.2 film (1.4 m) showed an effectively increased series resistance and thus a decreased Voc. Comparatively low photo-conversion efficiency of the device, when the back contact was 0.35 m of FeS.sub.2 and 30 nm of Au, is due to the reduced Voc even though other parameters are considerably higher. When the FeS.sub.2 thickness was at the maximum, both Jsc and fill factor were lower ultimately decreasing the efficiency of the solar cells. Optimized performance of the solar cell was found when FeS.sub.2 thickness was 0.7 m in which Voc, Jsc, FF and are only 2.9%, 0.4%, 10% and 13% less than standard device.

(63) The inventors have found that the electrical properties of the FeS.sub.2 NC layer are significantly different from those of bulk metal or semiconductor layers. During the synthesis of the FeS.sub.2 NCs, a long chain CH molecule (TOPO) has been used. Trioctylphosphine oxide controls the growth rate, determines the size of the NCs, and caps the NC surface. The length of these TOPO molecules is about 1.1 nm with the outer part of the molecule being hydrophobic. An estimate of the typical separation between two neighboring NCs in the as-deposited film is about 2.2 nm. This separation creates an insulating film and inhibits charge transport through the contact. However, when the FeS.sub.2 films are treated with hydrazine, these long molecules are removed, and the NC surfaces are capped by hydrogen atoms. Since hydrazine tends to remove oxygen from the reaction medium, reaction with an oxygen containing compound leaves hydrogen in the place of oxygen. The hydrazine reacts with TOPO to remove the oxygen, and as such, the association of TOPO with the iron pyrite NCs ceases because this association takes place through oxygen atoms on TOPO. As a result, the TOPO is removed. In addition, the inventors believe that hydrazine may remove a substantial amount (or even all) of the dissolved oxygen from the suspension where FeS.sub.2 NCs are present, thus improving the resistance of the film to further oxidation. As an aspect of the method of the invention, in certain embodiments it may be advantageous to vary the hydrazine treatment time depending on the concentration of the NC solution and the resulting single-cycle layer thickness. FIG. 12 illustrates the effects on cell current-voltage curves in the case where TOPO remains in the FeS.sub.2 NC film. The untreated film shows a clear S-shape signature, indicative of a transport barrier formed at the back contact. The S or kink shape has been previously observed under illumination in CIS and also organic solar cells when a charge buildup occurs at one contact. The inflection points in the J-V curves, also called S-kinks, are mostly seen in organic solar cells which are attributed to charge transport layers energetically misaligned to the energy levels of the active materials in planer heterojunction solar cells. For the FeS.sub.2-NC back contact cells, a charging is likely occurring, associated with reduced transport through the FeS.sub.2 contacting layer, resulting in the S shape J-V curves.

(64) In FIG. 12, the J-V curve is obtained from sputtered CdTe solar cells. Thickness of the FeS.sub.2 NC and Au are shown in FIG. 12. Carrier concentrations of FeS.sub.2 NC films are about 10.sup.19 cm.sup.3, lower than the carrier concentrations of metals. In certain embodiment, the FeS.sub.2, used in conjunction with another metal, such as Au, eliminates the reliance on Cu and provides for long term stability of the CdTe devices. Thus, in certain embodiments, a thin layer of Au applied on the FeS.sub.2 layer facilitates good electrical contact. In FIG. 12, Voc and Jsc remain relatively high but the efficiency drops significantly due to reduced fill factor. The inventors believe that the low fill factor may result from the S-shape J-V curve, which shows d.sup.2J/dV.sup.2<0 for much of the power-generating quadrant.

(65) For the FeS.sub.2-NC material, this S-shaped behavior is seen at or near the open circuit voltage region. Under open circuit conditions, the total current at every location inside the solar cell is generally zero. For this condition to transpire, the internal current densities for electrons and holes are equal and of opposite sign. In solar cells, an infinite minority surface recombination always creates a steady diffusive recombination current towards the back contact. This current is neutralized by an equivalent majority surface recombination current. Because of the long chain hydrocarbon molecules in an untreated FeS.sub.2 NC film, transport of light induced charge carriers is reduced thus reducing the surface recombination velocity. To satisfy a zero net-current condition, the majority charges will be accumulated at the surface. This situation creates a space charge region creating an electric field at the interface and thus generating S-shaped J-V curves around Voc. This situation lowers fill factor, and possibly Voc to some extent, which may decrease the efficiency of the solar cells. This double diode behavior has been appeared in J-V curves when the TOPO molecules are not removed or only partially removed from the NC surface.

(66) The FeS.sub.2 NC films used in the invention have shown p-type conductivity, as determined by hot probe measurement. When FeS.sub.2 films exhibit p-type conductivity, the valence and conduction band positions are approximately 4.3 eV and 5.6 eV from the vacuum level, respectively. Since the electron affinity of the FeS.sub.2-NCs (p-type) is about 4.3 eV and considering a direct band gap of 1.3 eV, the valence band edge is about 5.6 eV from the vacuum level. This is close to the valence band edge of CdTe, thus providing a close match to the CdTe layer. Since FeS.sub.2 is highly conductive, the position of the Fermi level is very close to the valence band edge, which is higher than the work function of CdTe. In such a case, FeS.sub.2 can form a zero barrier height contact with CdTe.

(67) When FeS.sub.2 NC thin films are prepared by solution based methods, some pinholes were encountered. These pinholes are not completely removed by just increasing the thickness of the films. The inventors believe that removing these pinholes will increase cell performance by eliminating the potential for shunts within the semiconductor-to-contact interface. Several techniques have been used to reduce pinholes, which have been thought to be formed in response to the strain at the substrate-film interface, as well as the cooling action of the spray droplets and the differences in thermal expansion between pyrite and glass materials. The inventors believe, generally, that spray forming the films at a low rate or spray forming the films on the metal/glass substrate, instead of just the glass, may have a beneficial effect. In an embodiment of the method of forming photovoltaic cells having FeS.sub.2-NC based back contacts, reducing pinhole effects begins with removing long chain CH molecules from the surface of the NCs with hydrazine treatment. The NC film is sintered at a relatively high temperature. Though FeS.sub.2 is thermodynamically unstable (FeS.sub.2 converts to FeS) at high temperature over longer times, these NCs are stable when heated at low temperature for a short time interval. For higher temperature and longer time, the FeS.sub.2-NC films are heated in sulfur vapor and in an argon atmosphere to prevent sulfur evaporation from FeS.sub.2 and thus maintaining S/Fe ratio before and after heating. FIGS. 6(a)-6(b) show XRD spectra and SEM images of FeS.sub.2 NC films before and after annealing. Raman spectroscopy is more sensitive than conventional XRD to explain the purity of the NC films before and after sulfur annealing. Raman spectra in FIG. 6(b) are sharper and more intense because of improved crystallinity of the film after annealing. Heating at 540 C. from one to three hours did not yield noticeable or appreciable grain growth of surface NCs on the films, as shown in FIG. 6(d). The XRD spectra in FIG. 6a show improvement in crystallinity after sintering. EDX spectral analysis further indicates that atomic percentage of S/Fe before and after annealing the film is maintained, indicating that the films were thermodynamically stable in the thermal sulfurization treatment.

(68) In another aspect of thin film photovoltaic cells using iron pyrite based back contacts, the inventors have found that the these back contacts provide passivation of pinholes present in the semiconductor layers, similar to the pinholes of the back contact described above. Referring now to FIG. 14, there is illustrated a comparative plot of electrical performance of photovoltaic cells having semiconductor layers formed with pinholes utilizing standard and iron pyrite-based back contacts. Two sets of CdTe devices were prepared using standard (Cu/Au) and iron pyrite/gold (FeS.sub.2NC/Au) back contacts and were left untreated for pinholes (passivation). Before back contact deposition, the CdS/CdTe thin films were treated normally with cadmium chloride. After the deposition of 3 nm Cu/40 nm Au, the standard sample was annealed at 150 C. for 45 minutes to diffuse copper (Cu) into the CdTe layer. In contrast, the iron pyrite back contact was prepared by depositing an FeS.sub.2 NC film at room temperature and then thermally evaporating 40 nm Au to complete the device. For the standard back contact deposition, a shadow mask was used with an area of 0.08 cm.sup.2. For the FeS.sub.2 NC/Au sample, the back contact layer was scribed manually making an effective area of cell of 0.085 cm.sup.2. Current density-voltage measurements were performed in 1 Sun illumination. As shown in FIG. 14, a plot of the strongest devices from each back contact configuration are presented. A summary of performance values for these samples and average values over a given population of devices are also shown.

(69) As shown in the plot, the Cu/Au back contact device performed poorly, due to pin holes. These pinholes or areas of structural nonuniformity produce deleterious effects on cell performance by creating regions of weak diode behavior or regions of shunting. The pinholes or structural nonuniformities can result from either defects within various semiconductor layers of the device or from morphological irregularities in the deposition surface of the substrate material. In one instance, the pinholes may permit the Cu/Au contact material to migrate through to the CdS/CdTe semiconductor junction creating shunting pathways. This decreases VOC and the FF of the device even though JSC is still reasonably high. In contrast to the Cu/Au cell performance, when a 1 m thick FeS.sub.2 NC film was deposited by a solution based method, the film functioned both as a back contact and as a pinhole passivation or barrier. The increased VOC, FF and efficiency show the effectiveness of iron pyrite-based back contacts in passivating semiconductor pinholes.

(70) While the invention has been described with reference to various embodiments, it should be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the essential scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof.

(71) Therefore, it is intended that the invention not be limited to the particular embodiment disclosed herein contemplated for carrying out this invention, but that the invention will include all embodiments falling within the scope of the claims.