CIGS solar cells were fabricated on a hybrid back contact comprised of a TCO layer (ZnO:Ga (GZO)) and Mo layers. It was discovered that an additional Mo layer introduced underneath the TCO layer promotes sodium diffusion through the TCO back contact into the upper CIGS absorber layer. Improvement in VOC and JSC values relative to those of sodium-free solar cells was achieved with the Mo/GZO/Mo hybrid back contact as a result of the enhanced sodium diffusion.

We have grown device-grade Cu(In,Ga)Se2 (CIGS) thin films using a RF-cracked Se-radical beam source. A unique combination of film properties: smooth surface, large grain size and high photovoltaic performance are shown. A competitive energy conversion efficiency of 17 % has been demonstrated from a solar cell fabricated using a CIGS absorber grown with a Se-radical source. In addition to the unique combination of film properties and high photovoltaic performance, a significant improvement in the use of Se source material in comparison with the conventional Se-evaporative sources has been demonstrated.

Dependence of band alignments at interfaces between CdS by chemical bath deposition and Cu(In1-xGax)Se2 by conventional 3-stage co-evaporation on Ga substitution ratio x from 0.2 to 1.0 has been systematically studied by means of photoemission spectroscopy (PES) and inverse photoemission spectroscopy (IPES). For the specimens of the In-rich CIGS, conduction band minimum (CBM) by CIGS was lower than that of CdS. Conduction band offset of them was positive about +0.3 ~ +0.4 eV. Almost flat conduction band alignment was realized at x = 0.4 ~ 0.5. On the other hand, at the interfaces over the Ga-rich CIGS, CBM of CIGS was higher than that of CdS, and CBO became negative. The present study reveals that the decrease of CBO with a rise of x presents over the wide rage of x, which results in the sign change of CBO around 0.4 ~ 0.45. In the Ga-rich interfaces, the minimum of band gap energy, which corresponded to energy spacing between CBM of CdS and valence band maximum of CIGS, was almost identical against the change of band gap energy of CIGS. Additionally, local accumulation of oxygen related impurities was observed at the Ga-rich samples, which might cause the local rise of band edges in central region of the interface.

In-situ characterization of composition, electronic structure and their depth profiles of surface of Cu(In1-xGax)Se2 (CIGS) film grown by three stage co-evaporation has been carried out by means of photoemission and inverse photoemission spectroscopy (PES/IPES), for the purpose of investigating the formation mechanism of the CIGS-side wide band-gap region adjacent to CBD-CdS/CIGS interface in cell structure. Sample-transportation in vacuum below 1 x 10-8 Torr yielded almost contamination-free feature of the CIGS surface. The as-transferred surface of Cu0.93(In0.65Ga0.35)Se2 grown at the identical condition for the high performance solar cell exhibited seriously Cu and Ga deficient composition. Chemical formula of this region was inbetween Cu : (In+Ga): Se = 1 : 3 : 5 and 1 : 5 : 8. In-situ UPS/IPES measurements CIGS showed that the as-grown surface region of the CIGS already had expanded band gap energy up to 1.4 eV and n-type character. A gradual decrease of band energy and a rise of valence band maximum as a function of depth from the original surface were observed. These results have revealed that the surface of CIGS by the three stage method already has the wide band gap, which might originate in so-called Cu-vacancy ordered phase.

For understanding the origin of the improvements of properties in the CIGS-based cells, of which the CIGS absorber has been fabricated by H2O-introduced co-evaporation [CIGS-H2O], band alignment at the interfaces between chemical bath deposited CdS and CIGS-H2O with Ga substitution ratio ~ 40 % has been studied by photoemission and inverse photoemission spectroscopy. The CdS layer over the CIGS-H2O showed an identical electronic structure with that of CdS on the conventionally grown CIGS; band gap energy of 2.3 ~ 2.4 and a location of conduction band minimum (CBM) and valence band maximum (VBM) relative to Fermi level were + 0.75 eV and -1.6 ~ -1.7 eV, respectively. In the interface region, decreases of CBM and a rise of VBM were observed. Total amount of the decrease of CBM over the interface was 0.2 ~ 0.3 eV. XPS measurements of the core-level signals over the interface showed a small upward bend bending of 0.1 ~ 0.2 eV. Consequently, the conduction band offset (CBO) and valence bad offset (VBO) at the CBD-interface over the CIGS-H2O (Ga~40%) are about +0.1, and 0.9 ~ 1.1 eV, respectively. This positive CBO is contrast with a slightly negative CBO at the interface between CBD-CdS/conventionally grown CIGS with Ga ~ 40 % measured previously. These results indicate that the H2O introduction is effective to extend the upper limit of the Ga substitution ratio where the Type-I conduction band alignment is maintained. The observed band alignments are consistent with the rise of Voc and efficiency in the CIGS-H2O based cells.

The electrical properties of single Pt (30 nm) and Pt (30 nm)/Ni (30nm)/Au(50nm) multilayer contacts on moderately doped p-ZnO (Na = 5.0 × 1017 /−3) were investigated. Although linear current-voltage characteristics were observed for all samples, a sample that was annealed for 1 min at a temperature above 500 °C resulted in an ohmic contact with good characteristics. The best ohmic contact to p-type ZnO was obtained using a Pt/Ni/Au multilayer contact that was annealed at 600 °C for 1 min under a N2 ambient, showing a specific contact resistance Rc of 1.97 × 10−5 Ω cm2. The fundamental mechanisms for the lower contact resistivity of Pt/Ni/Au contacts are discussed based on glancing angle x-ray diffraction results and Auger depth profile analysis of the multilayer alloying process. Furthermore, we fabricated a ZnO p-n homojunction using Pt/Ni/Au and Ti/Au as the p-type and n-type ohmic contact metal, respectively. The threshold voltage was determined to be about 3.7 V, comparable to the bandgap energy of ZnO.

The reasons why the open circuit voltage (Voc) of high-x CuIn1-xGaxSe2 (CIGS)/ZnO solar cells remain low are discussed. Here it is shown that the Voc ceiling can be interpreted simply on the basis of a model that the valence-band energy (Ev) of CIGS is almost immovable irrespective of x. When the conduction-band energy (Ec) of ZnO is lower than that of high-x CIGS (DEc<0), the built-in potential (Vbi) of a CIGS/ZnO junction is equivalent to the flat-band potential (Vbi) that arises from the separation between the Fermi energies of the two materials. If the Ev (and therefore the Fermi energy) of p-type CIGS is constant with increasing x, the Vbi and Voc that follows the Vbi remain unchanged since the Fermi energy of ZnO is constant. This unchangeable Voc reduces the conversion efficiency of high-x CIGS cells in cooperation with reduced photocurrents due to a larger bandgap. A positive offset, ΔEc>o gives rise to a photoelectrons barrier in the conduction-band that partially cancels Voc, thus the Voc of a low-x CIGS cell is governed by the Ec of CIGS. Based upon this concept, a material selection guideline is given for the windows and transparent electrodes appropriate for high-x CIGS absorbers-based solar cells.

Using EBIC and EDX measurements, CIGS solar cells prepared under several different conditions were observed and characterized. The results of EBIC and EDX measurements suggest that Cd plays an important role in the forming of a buried pn-junction in the CIGS layer via diffusion, and de-emphasize the possibility of the formation of the hetero pn-junction at the CdS/CIGS heterointerface. The correlation of the extent of the space charge region and the observed shift in the pn-junction location with the diffusion of the constituent elements in CIGS was investigated.

We have fabricated CIGS:Fe polycrystalline thin films using a standard three-stage method, and investigated the effects of Fe doping on cell performances. The Ga / (In+Ga) ratio was varied between 0.3 ˜ 1.0 (= CGS), and the Fe concentration was varied between 0.0 ˜ 1.2 mol%. The films were characterized by various means, including the cell performance. Increment of the grain size with higher Fe content was observed. Redshift with higher Fe content was observed in the absorbance spectra. The spectral response of the fabricated solar cells deteriorated with higher Fe content, from the long wavelength side.

Epitaxial CuGaSe2 films were grown on GaAs substrates under Cu-excess conditions to obtain stoichiometric compositions. The films were annealed in Ar, Sex or O2 ambients with or without a Cu or Cu-Se cap layer with the intention of changing the intrinsic defect concentrations. Samples were evaluated using low-temperature photoluminescence (PL) measurements. Annealing of the samples dramatically changed the PL spectra indicating that not only interdiffusion had occurred, but defect species and populations were changed. Comprehensive consideration of the changes led to the conclusion that the emissions at 1.62 eV, 1.66 eV and in the range from 1.2 to 1.4 eV are related to specific defects of Se vacancies, Cu vacancy-Se vacancy complexes and interstitial Cu, respectively.

CuGaSe2 (CGS) is a promising material for high efficiency thin film solar cells though predicted device performance has not been realized. Understanding the difference in the chemical nature between CuInSe2 (CIS) and CGS is critical for improving Cu (In, Ga) Se2 solar cells with high Ga concentrations. In this work, we have investigated the effects of oxygen-annealing on Ga-rich CGS epitaxial films focusing on compositional changes and secondary phase formations. The photoluminescence (PL) spectrum of Ga-rich films after oxygen-annealing was observed to always change into a spectrum characteristic of CGS grown under Cu-excess conditions. Electron probe micro-analysis (EPMA) measurements indicate the formation of Ga-O after oxygen-annealing. Selective etching of the Ga-O phase showed the composition of the CGS phase became close to stoichiometric. The oxygen-annealed films showed multiple pits ∼ 100 nm in depth and ∼ 2.5 μm in width. The Ga-O phase is founded in a layer formed on the surface of the CGS phase and in a columnar form rising from the bottom of the pits to the film/substrate interface. The above results suggest that excess Ga in Ga-rich CGS tends to react with oxygen to form Ga-O, thus the composition of the remaining CGS approaches stoichiometry consistent with the changes observed in PL.

ZnO films were grown by radical-source molecular beam epitaxy (RS-MBE) on sapphire and glass substrates, and they were characterized in terms of Hall mobility and optical transmission. Undoped ZnO films exhibit a low intrinsic defect density and optical properties close to bulk ZnO. By Ga doping, a resistance ρ as low as 2×10−4 Ωcm could be achieved. Balancing high conductivity and low transmission losses due to free carrier absorption in the infrared, the optimum was obtained for ρ=3.4×10−4Ωcm, electron mobility μe=37 cm2/Vs and an average transmission T of 96% in the wavelength range 400-1100 nm. Polycrystalline growth on glass yields slightly reduced but still good film quality (μe=30 cm2/Vs, T=90%). By the incorporation of Mg, conducting Mg0.3Zn0.7O films with an increased band gap up to ∼ 4eV were realized.

ZnO films were grown directly on epitaxial CuInSe2 (001) (CIS) by radical-source molecular beam epitaxy (RS-MBE). The substrate-film interdiffusion was investigated dependent on the ZnO growth temperature. Secondary Ion Mass Spectroscopy (SIMS) profiles indicate the mutual temperature-activated diffusion of Zn and In at a growth temperature of 440°C which is absent at 250°C. Zn indiffusion into the CIS substrate leads to characteristic changes in the photoluminescence (PL) properties, whereas the In outdiffusion into the growing ZnO film causes an increased carrier concentration.

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.

Manganese ions implantation into ultrapure GaAs layers grown by molecular beam epitaxy were investigated by photoluminescence technique systematically in a wide range of manganese concentration up to 1×1020cm-3. Five shallow emission bands denoted by (Mn°, X), “G”, “G' “ “H” and (D, A)2 are formed in the implanted layers in addition to the well known Mn impurity related emission at ~880nm. With increasing manganese concentration to 1×1019cm-3, “G” exhibits no energy shift, suggesting that “G” is different from the behavior of [g-g] emission that is commonly formed in shallow acceptor (such as C) incorporated ultrapure GaAs. (Mn°, X), “G” and “G' “ present no energy shift with increasing excitation intensity, while “H” and (D, A)2 indicate peak energy shift greatly showing typical donor-acceptor pair characteristics. “G” and (Mn°, X) are found to hold similar radiative origin which is different from “G”. Temperature dependence measurement reveals that emission “G” has a thermal activation energy of 5.4meV.

Recently, we introduced various acceptor impurities into MBE-grown ultra-pure GaAs by conventional high-energy ion implantation and found many novel shallow emissions associated with acceptor-acceptor pairs. Most of these emissions were easily quenched by extremely small amount of residual donor atoms which were unintentionally introduced during doping processes. For the interpretation of impurity effects, the usage of mass-separated atom as dopant source was strongly suggested. Along this consideration, we developed combined ion beam and molecular beam epitaxy (CIBMBE) technology, in which damage-free doping with high mass purity (M/ΔM=100) is expected to be possible. We here present the results of low-energy (100 eV) carbon ion doping using CIBMBE method. Samples were prepared asa function of growth temperature (Tg=400-700°C) and ion beam current. Net hole concentration, |NA-ND| as high as ~1×1020 cm-3 was obtained in as-grown samples. In 2K photoluminescence spectra, emissions due to acceptor-acceptor pairs exhibit specific energy shift with growing |NA-ND|. Results indicate that carbon doping can be made efficiently even at Tg as low as 500°C without any post heat treatment. These results also tell that by CIBMBE method no serious radiation damages are produced and the undesired impurity contamination can be considerably suppressed.

C-doped GaAs films were prepared by novely a developed, combined ion beam and molecular beam method (CIBMBE) as a function of hyperthermal (30–500 eV) energies (EC+) of carbon ion (C+) beam. Ion beams of a fixed beam current density were impinged during molecular beam epitaxy growth of GaAs at substrate temperature of 550 °C. Low temperature (2 K) photoluminescence (PL) has been used to characterize the samples together with Hall effects measurements at room temperature. Through the spectral evolution of an emission denoted by [g-g]β which is a specific emission relevant to acceptor-acceptor pairs, the activation rate was confirmed to increase with increasing EC+ for EC+ lower than 170 eV. It was explicitly demonstrated that the most effective Ec+ to establish highest activation rate is located at ~170 eV. This growing activation rate was suggested to be attributed to the enhanced migration of both impinged C and host constituent atoms with increasing EC+. This surmise was supported also by Hall effect measurements which revealed the maximum net hole concentration ( |NA-ND| ) for EC+=170 eV. For EC+ higher than ~170 eV, increasing EC+ was found to induce the reduction of activation rate. It was suggested that this observation is ascribed not to the formation of C donors but to the enhanced sputtering effect of impinged C+ ions with increasing EC+.

Low temperature (2° K) photoluminescence (PL) measurements were performed for the wavelength (λ) region from 800 to 1600 nm on Be–doped GaAs grown by molecular beam epitaxy (MBE) for wide range of net hole concentration at room temperature (p) up to 2 x 1020 cm−3 after annealing at 850 °C for 20 minutes. The annealing procedure formed several new PL emission bands associated with deep levels appeared between 0.4 and 1.4 eV. The most intensive new PL band appeared at around 1.2 eV with a band width of about 0.2 eV for all samples investigated. Its peak position and band width were observed to shift slightly toward lower energy side and increase, respectively with increasing p. An additional two sharp bands appeared overlapping the main band at around 1.31 and 1.35 eV for p less than 1 x 1018 cm−3. New broad band formation was also observed at around 1.0 eV in the same samples. In addition, a new prominent PL band was found at about 1.35 eV in the as-grown samples with p = 5 ∼ 6 x 1017 cm−3, which disappeared entirely after annealing. The formation and annihilation mechanism of these deep levels after annealing can be presumably attributed to Be redistribution from Ga sites to As sites and interstitial sites to Ga or As sites, driven by the formation of As vacancies (V As) due to As evaporation from samples during annealing.

The structural effects of low-energy (30-500 eV), mass separated C12 Ion doping of GaAs simultaneous with conventional solid source MBE growth have been studied using room-temperature raman scattering, Hall-effect, transmission electron microscopy and 2K photoluminesence measurements for GaAs epitaxy temperatures of 550 ºC. Results indicate good acceptor activation without detectable residual damage is achieved for ion energies ≤ 240 eV, while at EIon = 500 eV, residual damage is present with a corresponding reduction in electrical activation. Low-energy TRIM calculations indicate that the damage is related to the increased depth distribution of vacancies and interstitials created during the higher (500 eV) implantation process which can not be annealed out at growth temperatures. Constant energy (100 eV) film growth experiments for a range of implantation currents (45 pA/cm2 - 45 nA/cm2) and growth temperatures of 550 and 550 ºC, show LO Raman peak broadening and mode hardening for currents ≥15 nA while maintaining very high C acceptor activation. This is interpreted as residual stress due to small amounts of interstitial C in the highest doped films. Both Hall mobility measurements and photoluminesence show no evidence of C dopant compensation.

Low energy (100 eV) impinging of carbon (C+) ions was made during molecular beam epitaxy (MBE) of GaAs using combined ion beam and molecular beam epitaxy (CIBMBE) technologies for the growth temperature ( Tg ) between 500 °C and 590 °C. 2 K photoluminescence (PL), Raman scattering and Hall effect measurements were made for the samples. In the PL spectra two specific emissions, “g” and [g-g], were observed which are closely associated with acceptor impurities. PL and Hall effect measurements indicate that C atoms were very efficiently introduced during MBE growth by CIBMBE and were both optically and electrically well activated as acceptors even at Tg=500 °C. The results reveal that defect-free impurity doping without subsequent annealing can be achieved by CIBMBE method.