A total area conversion efficiency of 10.2% has been achieved for a 1Ocm×1Ocm integrated-type single-junction amorphous silicon (a-Si) solar cell submodule. It is the highest conversion efficiency ever reported for an a-Si solar cell with an area of 100cm2, including multi-junction cells. The effective area conversion efficiency is as high as 11.3%. The high efficiency is obtained by improving the quality of the i-layer and the p/i buffer layer, as well as by utilizing a highly textured, high-quality transparent electrode. The quality of the i-layer plays a dominant role in the performance of a-Si solar cells, especially in high efficiency cells. Techniques that control the properties of the high-quality a-Si films for the i-layer are described. Electric conductivity, ESR spin density and the Raman spectra of high-quality a-Si:H films are investigated as well as their thickness-dependence and substrate-dependence.

A Through-Hole Contact (THC) integrated-type submodule has been developed as a new-type a-Si solar cell module structure. Numerical simulations on the output power of the structure show that the output power can be significantly improved by the THC structure.

A series of systematic experiments has been made on a bandgap profiling control using a-SiGe and a-SiC cells to clarify the cause for the improvement of the photovoltaic performance by the bandgap profiling in the photovoltaic active layer. It has been shown from the analysis that the most important contribution to the cell performance is an optimization of the ambipolar carrier collection by the bandgap profiling. On the basis of the results, an efficiency of 10.1% was obtained at the promising a-SiC/a-Si/ a-SiGe multi-bandgap stacked cell by the bandgap profiling.

The transport properties of intrinsic a-SiGe films are found to be limited by poor electron mobility. The relationship between electronic transport and composition (Si, Ge and H content) is established. The electron mobility decreases with increasing hydrogen content for a given Ge content. The relation between growth conditions and film composition is developed. Growth conditions that reduce the extent of gas phase polymerization as well as the flux of hydrogen radical to the growth surface yield the best electronic transport. The best a-SiGe:H alloys of 1.3 eV band gap are used to study the effects of band gap grading on solar cell performance. Solar cells analysis is used to determine both hole and electron transport. Cell designs that minimize performance loss due to poor electron mobility are considered.

We have studied the photoelectronic properties of pipi... multilayers made of a-SiC:H p-layers (∼ 100Å) alternating with a-Si:H i-layers (500Å to 4000Å). A large enhancement in d.c. coplanar photoconductivity (more than an order of magnitude) over that of bulk a-Si:H was observed with decreasing i-layer thickness of the multilayer. On the other hand, the ambipolar diffusion length was observed to decrease with decreasing i-layer thickness. These observations can be explained by the effect of charge separation due to the strong built-in electric field in the i-layer. The insertion of an ultrathin SiC layer at the p/i interface is shown to appreciably reduce the built-in field which may be directly related to the observed decrease in fill factor when such a buffer layer is used in a solar cell.

We investigated p-i-n solar cells in which the i-layer was grown from hydrogen-diluted silane. The deposition parameters which were varied include flow rates, deposition pressure and power, and the degree of hydrogen dilution. We found that high quality devices could be made in which the i-layer was grown under a wide range of deposition conditions but that over this range neither the initial performance nor the stability of the devices differed significantly from those of cells having the same structure, but in which the i-layer was deposited from SiH4 with no H2-dilution.

Data is presented on the field test of amorphous silicon photovoltaic modules. The modules have been obtained by SERI under the field test and performance evaluation program to track the changes in performance following the continuing development of improvements in their manufacture. The data have been gathered for up to five years. Procurement time intervals are used to distinguish between the sets of modules under test. The results show that more recent technogies have a more rapid initial loss of performance but that stability sets in at an earlier time than that observed for the earlier modules. Data are presented to show efficiency and fill factor degradation from both the date of deployment and a date selected for post initial degradation(pi). Performance models are presented and compared, including models which account for annual variations in insolation and temperature and, therefore, generation of light induced defects and annealing.

A two-layer MgF2/ITO antireflection (AR) coating has been used to reduce the reflection losses from the surface of a hydrogenated amorphous silicon alloy solar cell. This has resulted in a higher efficiency device primarily due to an improved blue response. The relative thicknesses of the MgF2 and ITO layers have been tailored to give the highest overall quantum efficiency (Q) values, which are higher than that obtained with a single-layer antireflection coating. Typically, the 0 value at 400 nm (Q400) has been increased from 0.58 to 0.68 for a single a:SiH cell. Incorporation of the double-layer AR coating in conjunction with μc-SiC p-layer has yielded Q400 value of 0.77. The total current density obtained by adding the individual contribution of the component cells of a dual bandgap triple amorphous silicon alloy solar cell has been increased from 21.90 to 23.27 mA/cm2 using the double-layer AR coating.

Irradiation of thin films and solar cells with 1.00 MeV protons has been investigated for a fluence of 5.0E14 cm−2. We have used photothermal deflection spectroscopy and light conductivity to characterize the effect of irradiation on thin films; current-voltage and quantum efficiency measurements have been used to determine the effects of irradiation on solar cells.

Irradiation introduces increases in the sub-band-gap optical absorption and decreases in the photoconductivity of thin films. The major effect on solar cells is to reduce the short-circuit current. The spectral dependence of the quantum efficiency is reduced in a somewhat uniform manner with irradiation and also recovers in a uniform manner with annealing. Investigations of the annealing behavior of both thin films and solar cells show similar behaviors in the measured properties. The properties of both the films and cells are essentially restored with a one-hour anneal at 200 °C. Attempts have been made to use simple models to calculate defect densities and carrier transport.

Textured TCO (Sn02:F) substrates were fabricated using atmospheric pressure CVD (APCVD). Various kinds of TCO developed were classified into three types (A, B, C) in terms of shape of texture, electrical and optical properties. TCO substrates of type A with low carrier concentration and pyramidal texture fulfill the conflicting properties of low absorption and low resistivity after hydrogen plasma treatment. By using type A TCO substrates, the highest conversion efficiencies of 12 % for a small area of 1cm×1cm as a-Si single junction solar cell were certified by several related NEDO research organizations.

The large amounts of methane commonly used to prepare wide-gap a-SiC:H films present a lot of problems. This study was aimed at investigating the properties of fluorinated silicon nitride as an alternative to silicon carbide and to study the thermal stability of the films.

The RF glow discharge of NF3 (an etchant gas) and pure SiH4 produced a-Si:H,N,F films. Films were deposited with a gas ratio of NF3 over (NF3 + SiH4) from 0% to 15%. The substrate temperature, chamber pressure, and power were kept constant; only flow rate was varied. Characterizations of the films included AES, FTIR, visible/UV spectroscopy, dark conductivity, and photoconductivity.

Deposition rate increased about eight times with an addition of 6% NF3. The deposition rate increased from 1.22 A/s to 10.5 A/s. The Tauc optical gap increased from 1.72 eV to about 2.06 eV, suggesting excellent N incorporation. The dark conductivity decreased one order of magnitude while the photoconductivity decreased by as much as five orders of magnitude with ELH light. FTIR spectra showed the nitrogen incorporation. The Si-N bonds replaced the Si-Si bonds with increasing NF3 concentration. Also with increasing NF3 content in the gas phase, the hydrogen concentration in the material decreased and the films became more thermally stable.

Hydrogenated amorphous SiC:H films prepared by rf glow discharge from a silane-methane mixture have been B-doped either from diborane or from trimethyl boron (TMB). In the latter a higher hydrogen content is found explaining a higher Tauc optical gap EG, however also a lower dark conductivity for the same boron concentration in the film. Thus the figure of merit of our p-layers doped from TMB is not improved. Further the observed narrowing of the Tauc gap could not be correlated with a decrease of the hydrogen content at high B-doping. The reduction of EG is mainly explained by a drastic broadening of the absorption edge caused by creation of boron induced defects and structural disorder but not due to a shift of the absorption edge.

Internal photoemission of both electrons and holes is used to investigate the movement of the mobility edges in high quality intrinsic, undoped hy-drogenated amorphous silicon (a-Si:H) with temperature and electrical field. The electron mobility edge is found to move up in energy by ∼40meV between 298K and 120K. On the other hand, the hole mobility edge remains essentially unchanged between 298K and 160K. The injection (and collection) of photoemitted holes is less efficient than that for electrons and in the films studied could not be measured below 160K.

We compare the experimentally observed correlation between the density of charged dangling–bond defects and the Fermi–level position in n– and p–type a–Si:H films with the predictions of two hydrogen–related models for the thermal equilibrium state at the glass–transition temperature of the hydrogen glass in a–Si:H. Reasonable agreement is obtained with a defect reaction in which the transfer of hydrogen from Si–H bonds to weak bonds is limited to distances comparable to one atomic spacing. The results for p–type (boron–doped) films suggest that the formation of positively charged dangling–bond defects is not only determined by electronic energy differences but also by a relaxational energy gain of approximately 0.2eV.

Time resolved and steady state electrololuminescence (EL) measurements have been performed in a-Si:H p-i-n junctions at 10K < T < 300K in order to study the recombination in these devices. At low temperatures the rise of the EL following a rectangular forward current pulse consists of a fast component with a time constant of a few μs and a slow component with a time constant of about a ms. The data suggest that the time response of the EL is determined not only by the recombination but also by the time to build up the carrier density for bimolecular radiative recombination. The fast component of the EL is attributed to the field enhanced recombination of carriers which survived from the preceding pulse. Steady state measurements reveal characteristics of bimolecular recombination. The EL spectrum is shifted by 50meV to lower energies with respect to the photoluminescence spectrum.

Photoluminescence excitation (PLE) spectra at 77 K have been measured over the range 1.201.75 eV using the Ti sapphire cw tunable laser as the excitation source. Two undoped a-Si:H samples on rough substrates have been investigated. The first sample has a very high dangling bond (DB) density; the second one has low DB density. The PLE spectrum of photoluminescence (PL) at 0.8 eV for the first sample follows the shape of the absorption spectrum measured by photothermal deflection spectroscopy (PDS) at room temperature. This behavior can be understood within the context of the existing models as due to recombination through defects which produces PL centered around 0.8 eV. However the PLE spectrum of PL at 0.8 eV for the second sample drops very rapidly with decreasing energy for energies less than about 1.3 eV. This behavior, which differs dramatically from that of the absorption spectrum, is consistent with earlier results and suggests that the PL measured at 0.8 eV for the second sample may be largely due to a contribution of the tail of the PL band which peaks near 1.3 eV. The PLE spectra for PL at 1.0 eV and 1.1 eV for the second sample approach the PLE spectrum previously obtained using the integrated PL intensities.

Optical phenomena associated with the absorption edge of crystalline silicon (c-Si) and amorphous silicon hydride (a-Si:H) are presented and compared. The optical properties discussed include the energy dependence, dipole matrix element and density of states associated with the absorption edge; the temperature dependence of the relevant optical energy gaps, and finally the magnitude and temperature dependence of the slope of the Urbach edge for each material. The comparison suggests that the optical properties of the two materials are closely related and that the absorption edge of a-Si:H may be derived from the effect of site disorder on the zone center direct gap of c-Si.

Detailed structural investigations have been performed on a-Si1−xGex:H in 0 ≤ × ≤ l namely Raman, X-ray Photoelectron, and Photothermal Deflection Spectroscopies, which reveal the silicon sub-matrix to be considerably more rigid than the germanium one. In Si-Ge dominated networks distinct deviations from an overall random distribution of Si and Ge atoms can be deduced from a semi-quantitative analysis of the Raman data.

Nuclear magnetic resonance (NMR) has provided essential information on the local atomic bonding and microstructure of hydrogen in hydrogenated amorphous silicon (a-Si:H). Here we describe results from NMR and Raman spectroscopy on the hydrogen distribution and bonding in a-Si:H prepared by remote hydrogen plasma (RHP) deposition and contrast the results with those from a-Si:H prepared by conventional glow discharge (GD) deposition. The films prepared by the two techniques have similar H bonding except for the presence in the RHP sample of about 1 atomic % molecular hydrogen, a factor of ten higher than in GD material. For RHP samples prepared from a deuterium plasma rather than a hydrogen plasma, substantial differences in the hydrogen NMR spectra, hydrogen spin lattice relaxation time and Raman spectra are observed. The hydrogen which necessarily originates from the silane has a dramatically altered spectra.

Phosphorus, boron and compensation doped hydrogenated amorphous silicon films were deposited in a glow discharge at different substrate temperatures in the range 50–330°C. Gas phase doping levels were 1%. At the lower temperatures the hydrogen concentration in the B doped and compensated doped films is larger than in the P and undoped films. For higher deposition temperatures the H concentration of the B doped films appeared to be smaller than in the other materials. The difference in hydrogen content of the doped and undoped material, deposited at various temperatures, is considered as a function of the measured activation energy for conduction in these films. This difference varies in much the same way with the activation energy as the hydrogen content in films deposited at one substrate temperature, but with varying gas phase dopant levels. This represents strong evidence that, apart from the deposition temperature, the hydrogen concentration in glow discharge a-Si:H is determined by the position of the Fermi level.

Two hydrogen-mediated models are applied to predicting the density of deep levels in hydrogenated amorphous silicon (a-Si:H) under variations in the material’s temperature and total hydrogen content. Both models depart from the assumption that hydrogen is bonded at two classes of sites, dangling bonds and weak bonds. The predictions of the two models for the temperature-dependence of the deep-level density are similar. The models differ in their treatment of the origins of the weak bonds; a model associating the weak bonds with the clustered phase of bonded hydrogen observed by nuclear magnetic resonance also appears to be in agreement with hydrogen evolution experiments.