The control of the barrier height at the Cu(In,Ga)Se2(CIGS)/CdS interface, via the chemical state of the CIGS surface, is a key to improve solar cell performances. In this paper chemical modifications have been achieved by the electrochemical method, by varying both the pH of the solution and the applied potential. It is shown that the potential for electrochemical oxidation depends on the pH, in good agreement with the thermodynamic predictions. The resulting surface composition of CIGS, analyzed by XPS, also depends on the pH. Contact Potential Differences (CPD) and Surface Photovoltage (SPV) measurements show that the electronic properties of the CIGS surface are modified. The important result is that the oxidation leads to the formation of permanent positive surface charges, which increase the band bending up to 0.20 V in a direction favorable for the solar cell performances.

The interface between a Cu(In,Ga)Se2 (CIGS) and an underlying Mo layer was studied by X-ray diffraction and high resolution transmission electron microscopy. The CIGS layer was deposited onto Mo coated soda-lime glass using the “3-stage” process. A MoSe2 layer found to form at the CIGS/Mo interface during the 2nd stage of the “3-stage” process. The thickness of the MoSe2 layer depended on the substrate temperature used for CIGS film deposition as well as the Na content of the CIGS and/or Mo layers. For higher substrate temperatures, thicker MoSe2 layers were observed. The Na in the CIGS and/or Mo layer is felt to assist in the formation of MoSe2. Current-Voltage measurements of the heterojunction formed by the CIGS/Mo interface were ohmic even at low temperature. The role of the MoSe2 layer in high efficiency CIGS solar cells is discussed.

Two CuGaSe2 films were grown by molecular beam epitaxy onto GaAs (001) substrates with varying Cu/Ga flux ratios under Se overpressure conditions. Growth was interrupted at predetermined times and the surface composition was measured using Auger electron spectroscopy after which growth was continued. After growth, the film composition was analyzed using voltage dependent electron microprobe spectroscopy. Film structure and morphology were also analyzed using x-ray diffraction and atomic force microscopy. The film with a Cu/Ga ratio larger than unity showed evidence of surface segregation of a second Cu-rich phase with a Cu/Se composition ratio slightly greater than unity. A second CuGaSe2 film with a Cu/Ga ratio of less than unity showed no change in surface composition with time and was also consistent with bulk composition measurements. Diffraction measurements indicated a high concentration of twins as well as the presence of domains with mixed c and a axes in the Ga-rich film. The Cu-rich films by contrast were single domain and had a narrower mosaics. High sensitivity scans along the [001] reciprocal axis did not exhibit any new peaks not attributable to either the substrate or the CuGaSe2 thin film.

The piezoelectric photoacoustic (PPA) spectra for Cu-rich CuGaSe2 (CGS)/GaAs epitaxial layers were successfully observed between liquid nitrogen and room temperatures for the first time. Bandgap energy of CGS (A band) is estimated to be 1.72 eV at liquid nitrogen temperature. The activation energies of three possible intrinsic defect levels are estimated to be about 80, 130 and 190 meV.

Thin film(s) of chalcopyrite CulnSe2 have been grown by low-pressure metal-organic chemical vapour deposition (LP-MOCVD) using the precursors In(Se2CNMenHexyl)3 and precursors Cu(Se2CNMenHexyl)2. The precursors were prepared from carbon diselenide. Films were grown on glass between 400 – 450 °C, and characterized by X-ray diffraction, optical spectroscopy (UV/Vis), EDAX and scanning electron microscopy.

CuInSe2 thin films have been synthesized from binary precursors by Rapid Thermal Processing (RTP) at a set-point temperature of 290°C for 70 s. With appropriate processing conditions no detrimental Cu2-xSe phase was detected in the CIS films. The novel binary precursor approach consisted of a bilayer structure of In-Se and Cu-Se compounds. This bilayer structure was deposited by migration enhanced physical vapor deposition at a low temperature (200°C) and the influence of deposition parameters on the precursor film composition was determined. The bilayer structure was then processed by RTP and characterized for constitution by X-ray diffraction and for composition by Wavelength Dispersive X-ray Spectroscopy.

CuInSe2 epitaxial films having superior luminescence properties were prepared by using (001) GaAs substrates and then by thermal annealing in the presence of a Se-beam flux. We also found that exciton emission became dominant when the film composition was very close to the stoichiometry.

Studies of the reaction path during annealing of Cu-In-Se thin films for solar cell absorbers have been limited up to now to ex-situ analyses of the phase composition by X-Ray Diffraction (XRD) after processing by a specific temperature-time program. As an indirect method, the application of ex-situ XRD is not sufficient for the determination of reaction temperatures and reaction times for setting up a general model of CIS-formation.

We show in this paper that the use of a calorimetric method (Thin Film Calorimetry, TFC) offers the advantage of a direct (in-situ) observation of thin film reactions. Special care is taken to use film thicknesses of practical interest for industrial application (1.5 – 3 μm). In a first step we show results of binary reactions in the Cu-In, In-Se and Cu-Se systems. Their knowledge is necessary for understanding the processes involved in the ternary CIS-layers. It turned out that thin Cu-In and Cu-Se films react already at room temperature and behave as predicted by the bulk equilibrium phase diagrams during heating. In-Se thin films show prominent exothermic reactions starting with the melting of In. The first phase to be formed is generally In2Se which is then converted to more Se-rich compounds. In ternary Cu-In-Se films (Cu/In = 1.00) we observe transitions of the Cu-Se-system which can be attributed to the decomposition of CuSe2 and CuSe. Consequences for the model of improved CIS-growth by a Cu-Se flux agent are discussed.

The incorporation of Na into Cu(Inl-xGax)Se2 thin-films is known to lead to an improvement in device performance. The authors use X-ray photoelectron and Auger electron spectroscopies to determine the chemical nature of Na in Cu(Inl-xGax)Se2 thin-films. The Na concentration is determined to be ∼ 0.1 atomic percent in the bulk of Cu(In1-xGax)Se2 thin-films. The Na is chemically bonded to Se. The authors propose a model invoking the replacement of column III elements by Na during the growth of Cu(Inl-xGax)Se2 thin-films. These Nain and NaGa defects act as acceptor states to increase the p-type conductivity of Cu(In1-xGax)Se2 thin-films.

Single and dual-laser ablation techniques have been used to grow conductive ZnO films at room temperature by ablating a Zn metal target in oxygen ambient. The emission spectroscopy of the material plumes shows a significant presence of oxygen ions and Zn ions in the dual-laser ablated plume. Furthermore, dual-laser ablated plumes expanded rapidly in the radial direction resulting in large-area uniform films. The electrical properties of the films deposited on glass substrates depend critically on the ambient oxygen pressure. Conductivities of the order of 103 (ω.cm)-1 have been obtained for films deposited at room temperature by the dual-laser ablation process.

The MOCVD growth of cadmium sulphide on indium tin oxide (ITO) coated glass is investigated in this work. Optimum transmission characteristics and grain size are observed for films grown at 290° C at VI/MI ratios of 1–1.5. It is postulated that the nucleation delay observed during in situ growth monitoring via laser interferometry s a critical step in acieving these results. The nucleation delay is noted to vary with temperature and also VI/II ratio, allowing the rationalisation of the resulting film quality in terms of this observation. Structural analysis indicates polycrystalline CdS of hexagonal phase of (100) preferred orientation. Results of preliminary doping studies indicate that CdS can be n-type doped during growth with nbutylchloride.

Several efficiency influencing factors in MOCVD-grown CdTe/CdS solar cells, including preferential crystal orientation of CdTe layers, CdTe grain size and surface roughness, interfacial mixing, and surface and interface geometrical morphology, are studied. X-ray diffraction (XRD) shows that polycrystalline CdTe/CdS solar cells with higher efficiencies tend to have more (111) planes of CdTe parallel to the macro-surface. Scanning electron microscopy (SEM) and atomic force microscopy (AFM) analysis reveal the relationship between the grain size/surface roughness and cell efficiency. Secondary ion mass spectroscopy (SIMS) and Auger electron spectroscopy (AES) depth profiling show that the interfacial geometrical morphology has a significant influence on the efficiency of CdTe/CdS solar cells. Finally it is shown that interfacial mixing reduces the number of interfacial states and recombination centers and the energy loss due to internal reflectance, enhancing the performance of the solar cells.

X-ray diffraction and photoluminescence measurements have been used to characterize the diffusion of S into CdTe during post growth annealing of CdTe solar cells. For anneals at 410°C in the presence of CdCl2, evidence that both a CdTe1 xSx phase and nearly-pure CdTe are present near the back contact is observed. The ternary phase becomes more prominent and the S concentration increases with depth reaching roughly 4–5% near the CdS interface. Much less diffusion is observed at 350°C while for a 460°C anneal, CdTe1-xSx with a S concentration near 5% is found throughout the layer. The presence of CdC12 during the anneal enhances the interdiffusion.

Device performance in CdTe/CdS solar cells is critically linked to the quality of the back contact. In this paper we report on a phenomenon wherein application of the back contact significantly alters the electro-optical properties of the absorber near the junction. We have studied the photoluminescence (PL) spectrum of the near-junction CdTe region in CdTe/CdS solar cells before and after contact application. It is found that the elemental Tellurium layer formed on the CdTe surface by the standard nitric-phosphoric etch process results in a dramatic qualitative change in the junction PL spectrum. Prior to NP etch, the spectrum has two peaks at energies of 1.50 eV and 1.45 eV, corresponding to recombination in bulk CdTe, and in a CdTeS alloy with 9% sulfur content, respectively. After NP etch, the relative intensity of the low-energy peak is reduced, and the peak shifts towards higher energy. These changes are consistent with a model of increased band-bending at the grain boundaries near the CdTe/CdS heterojunction.

Currently, the best performing CdS/CdTe solar cells use a superstrate structure in which CdTe is deposited on a heated CdS/SnO2/Glass substrate. In the close-spaced-sublimation (CSS) process, substrate temperatures in the range 550°C to 620°C are common. Understanding how these high processing temperatures impact reactions at the CdS/CdTe interface in addition to reactions between previously deposited layers is critical. At the SnO2/CdS interface we have determined that SnO2 can be susceptible to reduction, particularly in H2 ambients. Room-temperature sputtered SnO2 shows the most susceptibility. In contrast, higher growth temperature chemical vapor deposited (CVD) SnO2 appears to be much more stable. Elimination of unstable SnO2 layers, and the substitution of thermal treatments for H2 anneals has produced total-area solar conversion efficiencies of 13.6% using non-optimized SnO2 substrates and chemical-bath deposited (CBD) CdS. Alloying and interdiffusion at the CdS/CdTe interface was studied using a new lift-off approach which allows enhanced compositional and structural analysis at the interface. Small-grained CdS, grown by a low-temperature CBD process, results in more CdTe1-x.Sx alloying (x=12–13%) relative to larger-grained CdS grown by high-temperature CSS (x7sim;2–3%). Interdiffusion of S and Te at the interface, measured with lift-off samples, appears to be inversely proportional to the amount of oxygen used during the CSS CdTe deposition. Our highest efficiency to date using CSS-grown CdS is 10.7% and was accomplished by eliminating oxygen during the CdTe deposition.

A monolithic interconnected module (MIM) structure has been developed for thermophotovoltaic (TPV) applications. The MIM consists of many individual InGaAs cells series-connected on a single semi-insulating (S.I.) InP substrate. An infrared (IR) back surface reflector (BSR), placed on the rear surface of the substrate, returns the unused portion of the TPV radiator output spectrum back to the radiator for recuperation, thereby providing for high system efficiencies. MIMs were fabricated with an active area of 0.9×l cm, and with 15 cells monolithically connected in series. Both lattice-matched and lattice-mismatched InGaAs/InP devices were fabricated, with bandgaps of 0.74 and 0.55 eV, respectively. The 0.74 eV MIMs demonstrated an open-circuit voltage (Voc) of 6.16 V and a fill factor of 74.2% at a short-circuit current (Jsc) of 0.84 A/cm2, under flashlamp testing. The 0.55 eV MIMs demonstrated a Voc of 4.85 V and a fill factor of 57.87percnt; at a Jsc of 3.87 A/cm2 Electrical performance results for these MIMs are presented.

The deposition kinetics of InP in MOCVD reactors is presented. The proposed chemical mechanism involves both gas phase and surface reactions. The fundamental hypothesis adopted in deriving the mechanism was a dual site dissociative adsorption of the precursors on the growing surface. In any case, all the rate constants either were taken from the literature or estimated through thermochemical methods. In addition, the deposition reactor was simulated by means of a monodimensional model that accounts for the main reactor features through the boundary layer theory.

The bandgap of InP is near ideal for the high theoretical conversion efficiency of the air mass zero (AM0) light spectrum experienced in space. This combined with the well known radiation resistance makes InP an ideal candidate for space solar cell use. The p/n configuration has several advantages over the n/p configuration, however, the conversion efficiency of p/n InP cells has been low compared to the n/p configuration, mostly do to poor performance of the Zn doped emitter. We have recently achieved a 17.6% AMO conversion efficiency for a p/n InP cell on an InP substrate and have achieved this record efficiency through understanding of the role of Zn interstitial defects and the hydrogen interactions with these defects. The effect of the Zn defect on device performance will be described.

New sources have been used to grow the first carbon doped, highly p-type InxGa1-xAs (0.2>x>0.7) epilayers on InP substrates by atmospheric pressure organometallic vapor phase epitaxy (APOMVPE). Excellent morphology was obtained simultaneously with high hole concentrations at growth temperatures near 450 °C. High hole concentrations of 1.6×1019 – 8.7×1019 cm−3 (the highest reported to date for APOMVPE), and the corresponding room temperature hole mobilities of 65 – 25 cm2/s, respectively, have been obtained from Hall measurements. X-ray diffraction is consistent with excellent crystal quality. Annealing at temperatures of T=400–500 °C in the presence of either nitrogen or hydrogen was found to change the carrier concentration by only 0–15%. However, after annealing at T=650 °C, irreversible changes occurred in the InGaAs. After a high temperature anneal, reversible order of magnitude changes in the hole concentration was obtained upon further annealing at low temperatures, depending upon the ambient. These results conclusively show that hydrogen does not passivate C acceptor ions in InGaAs. Since changes in the carrier concentration become substantial and reversible only after high temperature annealing, the results strongly suggest that a structural change occurred in the crystal at high temperatures. We consider it likely that this structural change is the precipitation of carbon out of a supersaturated solid solution, and that hydrogen atoms associated with these precipitates act as donors which compensate the hole concentration.

Germanium (100) substrates, 9° off axis towards the [011] direction, were heated in 2.0 Torr tertiarybutylarsine and 97.0 Torr hydrogen for 30 min at 650°C, then transferred to an ultrahigh vacuum system and annealed at several temperatures between 400 and 700°C. The arsenic coverage remained constant at 1.0 monolayer during heating in vacuum to 400°C. Further heating to 700°C caused the As coverage to gradually fall to zero. A mixed domain (l×2)+(2×l) dimer-terminated surface was observed by scanning tunneling microscopy after annealing at 400 to 510°C. Between 510 and 580°C, a single domain (l×2) reconstruction was recorded that contained equally spaced terraces separated by steps four atomic layers in height. Above 600°C, a (2×l) Ge(100) surface was produced that contained uniform terraces separated by steps two atomic layers in height.