This paper describes new results on the relationship between the grain size, mobility, and Si (111) x-ray peak intensity of laser crystallized amorphous silicon as a function of the laser fluence, shot density, substrate temperature, and film thickness. These observations include an unexpected narrow peak found in the silicon (111) x- ray peak intensity, which occurs at a specific laser fluence for a given film thickness and substrate temperature. Amorphous silicon materials processed at laser energy densities defined by this peak exhibit exceptionally large grain sizes and electron mobilities that cannot be obtained at any other energy and shot density combination above or below the energy at which the Si (111) x-ray peak intensity maximum occurs.

Time Resolved Reflectivity during XeCl pulsed laser irradiation of amorphous silicon films deposited on glass was measured. Simulation of the process by a Heat Flow Calculation in which explosive crystallization was not forced to occur, predicts the coexistence of partial bulk nucleation and a traveling molten layer. Optical simulation of Time Resolved Reflectivity was used to critically examine the Heat Flow Calculation results, substantially confirming the existence of a mixture of thermodynamical phases.

Crystallization of amorphous silicon thin films by various methods has fostered enhancements in the electrical characteristics over their amorphous counterparts. For example, carrier mobilities ranging from 10 to >100 cm2/V-sec have been reported for laser crystallized films. The rather large variability of the transport characteristics with crystallization processing conditions is not well understood and, as a result, greatly complicates device process debugging. In addition, while it is generally believed that defects inherent in the grain boundaries provide the primary barriers degrading transport properties relative to single crystal silicon, the specific nature of these defects is not known. In this paper, we present data on the temperature dependence of the Hall mobility of thin silicon films crystallized by thermal and excimer laser processing. Hall data for the laser-crystallized phosphorus-doped material show a temperature dependence which differs dramatically from that for thermally crystallized materials, while the effects of hydrogenation are similar, reducing the barriers at the grain boundaries.

Metastable defect creation by illumination and by a forward current in p-i-n devices are compared using CPM and reverse current measurements of the defect density. The data show that the same defects are formed by the two mechanisms, but with different spatial profiles. Numerical modelling shows how the spatial profile influences the reverse bias current.

Thin three-layered a-Si:H photoconductive structures have been studied for application in high resolution electrophotography. The resolution of prints is determined by the a-Si:C:H top layer. It is observed that the resolution of prints decreases from 100 to 30 lines/mm after about 25 charge cycles of corona discharge. The effect of corona charging on the surface properties of a-Si:C:H films with different carbon content is studied. Before and after corona exposure Fourier transform infrared absorption and lateral dark conductivity were measured. We observe that modes around 3300 cm-1, ascribed to H2O and -OH related modes, arise in combination with an increase of the lateral dark conductivity. This effect is not dependent on the film thickness, indicating that only the surface is affected by the corona discharge. The increase of the dark conductivity is interpreted as due to band bending at the surface, induced by the adsorption of H2O or the growth of a thin Si-OH layer at the surface.

In this paper we present a method to determine the annealable defect density(ΔNann) present in hydrogenated amorphous silicon(a-Si:H). The effects of the annealable defects on the light-induced defect generation rate, saturated defect density (Nsat) and the change of defect density in the light-induced saturated state(ΔNsat) have been studied. Annealable defect density was varied by depositing samples at various substrate temperatures or by post-growth anneals of samples grown at low substrate temperatures. It is found that the generation rate, N satand ΔNsat are well correlated with ΔNann. In particular, the ΔNsat is found to follow a relation ΔNsat ≈ ΔNann. These results suggest that defect-related microscopic models are appropriate for light-induced metastability.

We extended the kinetics study of light-induced creation of metastable defects in undoped a-Si:H down to T =4.2K. Using band gap light with photocarrier generation rate G0 =3.8x1021cm-3s-1 we find that the change in metastable defect concentration ΔND depends on exposure time te as tm e with m = 0.35 ±0.02 at 4.2K, 80K and 300K. The dependence of ΔND on generation rate is G0.44 at 4.2K. The relation between ND and photoconductivity at 300K is nonlinear and not single-valued.

Measurements of the steady-state defect density (Nst) in hydrogenated amorphous silicon under illumination of pulse-laser light, as well as of continuous light, were carried out; and the dependence of Nst on the effective rate of carrier generation (G) is presented. The values of G ranged from 8 x 1021 to 2.4 × 1023 cm-3 s-1, while the illumination temperature was kept at 30 °C or at 105 °C. The results showed trends of Nst increasing with G similarly to the trends in the literature, but covered a higher and wider G range, and fitted a defect model which assumes a limited number of possible defect states.

Photoconductivity decay during monochromatic illumination has been measured on an ensemble of a-Si:H films deposited at different substrate temperatures. Degradation behaviour has been modelled within the framework of the bond-breaking model (dN/dt = Csw np). Simmons and Taylor recombination kinetics has been assumed, taking into account the divalent nature of dangling bonds and their three possible conditions of occupancy. The Staebler-Wronski susceptibility (Csw) has been extracted through a fitting procedure. As a result, a correlation between the obtained Csw and the measured electronic, optical and structural properties of as deposited a-Si:H films can be inferred.

The results of a study of the kinetics of the light-induced annealing of the deep-level defects in hydrogenated amorphous silicon (a-Si:H) are presented. They show that at elevated temperatures illumination increases the rate of annealing compared to annealing in the dark. We also detected light-induced annealing at room temperature. On the basis of a model in which the defects are generated by electron-hole recombination and annealing occurs through the action of a single carrier, we found values of 0.86 eV for the activation energy of the light-inducedgeneration coefficient γ-1, and 1.23 eV for the light-induced annealing prefactor λ

Metastable light-induced changes in hydrogenated amoiphous silicon (a-Si:H) have been monitored by stationary single photocarrier charge collection in p-i-n diodes under reverse bias. The influence of the following parameters has been systematically investigated: i-layer thickness, applied voltage and illumination time and energy. A strong dependence of cell degradation on i-layer thickness is observed. In the case of degraded diodes superlinear current vs. voltage (Iph αVm) in the low voltage regime were observed. These results can be related to space-charge-limited photocurrents. The results from variation of photon energy and single and/or double carrier injection indicate that recombination rather than single-carrier trapping plays a key role in metastable defect creation.

Our previous results of light-induced electron spin resonance (LESR) indicate that, in hydrogenated amorphous silicon (a-Si:H), light-induced defects differ from those formed during deposition or high-temperature annealing. A plausible interpretation, in which light-induced defects occupy higher-energy states, was proposed to explain these differences. In this study, the constant photocurrent method (CPM), dark conductivity and steady-state (SS) LESR are used to supply new evidence for the difference and conduct two important tests on our hypothesis. In striking agreement with our predictions, we find that the light-induced changes in the SS-LESR lineshape (a decrease in the narrow component relative to the broad one upon light exposure) become indeed more dramatic as the demarcation energies move closer to the midgap by increasing temperature or decreasing bias-light intensity for SS-LESR.

Dispersive description of defect generation in a-Si:H that leads to stretched-exponential transients is extended by relaxing the assumption that light-induced processes and thermally induced processes have the same dispersive character. This is done by separating the rate equation for the defect density into two parts, one thermal and one optical, each with its own dispersion parameter. The solutions of this new equation — which must be obtained numerically — generally have two distinct parts: there may be a two-part rise or a peak, depending on the relative values of the two stretch parameters. Using this formulation we have readily simulated the recently observed peak in relaxation of a previously heavily degraded solar cell while exposed to a weak light. We find no way to explain other reports in similar two-part experiments that relaxation is faster under weak excitation than without.

Amorphous Silicon p-i-nphotodiodes were obtained by PECVD in a reactor with a rotating substrate holder. Reverse currents as low as 5 × 10−11 A/cm2 at a bias of − 2V were measured using a guard ring electrode to minimize lateral edge currents. The devices were degraded by a Xenon flash lamp in open circuit conditions. The kinetics of the degradation process was evaluated by studying the long time dark current transient under reverse bias.

A two-step light soaking experiment at high and low intensities provided convincing evidence that defect generation and annealing in a-Si:H are controlled by defect states of different characteristics. We point out that the total defect density by itself cannot uniquely determine the state of material or be described by a single rate equation, even though it might be the only quantity that is experimentally measurable. A system of rate equations for all defect components, therefore, must be established in order to accurately describe the defect kinetics. A simple two-component model in which defects are categorized as “fast” or “slow” is shown to be adequate to explain a variety of experimental results in a consistent fashion.

Light soaking of a-Si:H solar cells gives rise to a decrease in the spatially resolved collection efficiency. In the internal collection efficiency distributions measured under bias light conditions two crossing points appear during light soaking. The existence and the shift of these points could be explained by a combined bulk i-layer light soaking effect and a more pronounced collection loss at the p+/i-interface. This spatially non-uniform degradation was confirmed by the Dynamic Inner Collection (DICE) method.

Based on results of a-Si:H solar cell computer simulation solving numerically the Poisson, transport and continuity equations, using a rather complex density of states distribution being in agreement with experimental observations, shows that light-soaking degradation can be attributed to both, light induced effects in heterojunction region and to changes in the i-layer, which together tend to reduce the built-in electric field inside of the i-region. Measurements of spectral sensitivity using small intensity monochromatic illumination in combination with an accurate computer analysis provides the possibility of ascertaining which of the degrading effects is the dominating one in a particular a-Si:H solar cell structure.

PIN solar cells were light soaked up to 60 hours. The cell characteristics, the optoelectronic properties and the microstructure parameter (R=I2100/I2100+I2000) as well as the hydrogen content (CH) and density of states (g(Ef)) of the active i-layer were monitored throughout the entire light induced degradation process and compared with the correspondents μτ product (for both carriers) inferred through steady photoconductivity and FST measurements.

Data show a strong correlation between the decrease of μτ product for electron and the increase of the fraction of hydrogen bonded on internal surfaces (R increases from 0.1 to 0.4) suggesting structural changes during the light induced defects' formation. For holes, the μτ product remains approximately constant and only dependent on the initial hydrogen content. As g(Ef) increases, μτ presents an asymmetrical decrease showing that electrons are more sensitive to defects' growth than holes.

We also observe that the rate of degradation is faster for samples having the lowest defect densities, R and CH, showing that the amount of degradation is not a simple function of the photon exposure (Gt product) but also depends on the material microstructure.

We report on a detailed study of EDMR in pin-type solar cells. Like in films the signals are dominated by the contribution of the e-db resonance. It is found that the spectra depend on the applied bias and photon energy of the exciting light. The data suggest that the dark current is controlled by recombination in the bulk of the i-layer. The sign of the signal depends sensitively on the sign of the internal field. At high forward bias and illumination recombination at the pi-interface plays an important role. Degradation by both light exposure and high forward current results predominantly from an increase of the bulk defect density.

We have studied the light-induced degradation and thermal annealing recovery of amorphous silicon alloy solar cells with different microvoid density in the intrinsic layer. The microvoid density was changed by altering the deposition rate. The experiments show that cells with higher microvoid density need longer annealing time to recover after prolonged light-soaking. As a consequence, cells with high density of microvoids do not seem to saturate even after long duration of light exposure. The cells with high microvoid density also show much lower degraded efficiency. A careful comparison between degradations caused by accelerated and one-sun light soaking and subsequent annealing recovery indicates that the defects created in the two cases have different nature.