The authors report on systematic photoconductivity σph and ambipolar diffusion length Lamb measurements carried out on a series of lightly p- and n-doped samples and monitored throughout the entire light-induced degradation process. The results are interpreted based on a transport model involving free carriers (nf and pf), localized charge on bandtails, as well as dangling bond and dopant charges.

We have measured the transient photocurrent in “time-of-flight” structures of a-Si:H using bias voltages applied following photocarrier generation by a laser flash. The delayed field technique appears promising both as a probe of the deep level distribution and also of photocarrier thermalization. We describe a simplified theory for the delayed field transients including the effects of the delay time and the magnitude of the applied voltage. The theory accounts for the measurements adequately.

The room temperature non-radiative efficiency, defined as the ratio of the heat released per absorbed photon for doped and undoped hydrogenated amorphous silicon (a-Si:H) has been measured using photo-pyroelectric spectroscopy (PPES) for photon energies ranging from 2.5 to 1.6 eV. There is a fairly sharp minimum in the non-radiative efficiency when the a-Si:H is illuminated with near bandgap photons. We describe a model wherein this minimum arises from the variation in the amount of heat generated by free carrier thermalization as the incident photon energy is varied, and report measurements of the excitation kinetics of the non-radiative efficiency which support this proposal.

Under a dc light bias, the ac constant photocurrent method (CPM) yields an anomalous subgap optical absorption over that without the bias. When the intensity of the bias is high, an absorption band appears. However, absorption measured by ac photothermal deflection spectroscopy (PDS) is much less affected by the bias. A simple model qualitatively explains the effect on the PDS spectra and suggests that the anomalous CPM absorption is an apparent one which represents a variation of the recombination lifetime.

The electrical conductivity of a-Si:H films was investigated in the temperature range 300K - 450K under an alternative electric field whose frequency was in the range 100 Hz - 6×106 Hz. Loss peaks were detected, located at a temperature dependent frequency and superimposed to a nearly linear dependence on frequency. Their amplitude has found to be sensitive to Staebler-Wronski Effect. The observed photodegradation-sensitive loss peaks have been described in terms of a Simple Pair Hopping (SPH) mechanism, involving the electronic states associated to dangling bonds, superimposed to a broadband process described by Correlated Barrier Hopping (CBH) model.

The lifetime of carriers injected optically in the extended states of amorphous silicon-germanium alloys has been measured by time-resolved pump and probe optical techniques using either a femtosecond dye laser or a picosecond free electron laser. When Ninj > 1018 cm-3, the lifetime of the carriers is in the picosecond time domain. Our results are comparable to what we have observed previously in a-Si:H and very recently in a-Si,C:H. There are two lifetime regimes: at high densities, the recombination is bimolecular and nonradiative, whereas at lower densities, the recombination tends to be monomolecular but still nonradiative. The origin of these lifetimes is discussed.

We have applied the modulated photocurrent (MPC) method over a wide range of frequencies (5Hz-100kHz) and temperatures (120K-380K) to assess its ability to accurately deduce the mobility gap distribution in a-Si:H. We have also investigated the effects of moving both the Fermi level within some samples (by light soaking and partial annealing) and the quasi-Fermi level (by applying the bias light) to observe how such changes influence the deduced density of states (DOS). We then compared the MPC results directly with the DOS determined by junction capacitance measurements in the same sample devices. We have determined general conditions under which, we believe, the MPC results provide an accurate picture of the gap state distribution. However, we found that under other conditions, the appearance of the deep defect peaks and other features do not represent the actual defect distribution but, rather, are artifacts due to recombination processes.

The tunneling transmission coefficient as calculated by the WKB and the multi-potential approximation (MPA) is compared for tunneling through an a-Si:H/a-SiNx:H double barrier structures. The MPA method leads to shifts in the location of the quantum well states, which complicates the interpretation of quantum size effects in amorphous silicon multilayers. Experimental studies of tunneling currents through a-Si:H/a-SiNx:H single barrier structures are presented. At low temperatures the conductance data resembles tunneling data for metal/insulator/metal structures with an asymmetric barrier, while at room temperature a pronounced peak is observed in the conductance-volatge characteristics.

Low temperature, high field properties of electron and hole transport are investigated by time-of-flight (TOF), steady-state and transient space charge limited current (SCLC) experiments on intrinsic a-Si:H. Charge collection and TOF experiments performed at T= 80 K reveal hole μτ-products of = 8 × 10-10 cm2/V and hole mobilities μ ≤ 9×10-3 cm2/Vs. The field effect on hole thermalization is demonstrated by evaluation of the post transit current decay. SCLC experiments on p+ -i-n+ (electron transport) and p + -i-n+ (hole transport) configurations are introduced and interpreted in terms of field enhanced conductivity and mobility. The experiments demonstrate the overwhelming field effect on carrier hopping in the band tail regions of a-Si:H. Considerably higher fields have to be applied in the case of hole transport than electron transport to achieve comparable conductivities; this is discussed on the basis of the different tail state distributions and localization lengths.

When a forward bias is applied to an amorphous silicon p-i-n diode, there is a delay in the onset of the current. We present experimental data and numerical calculations, showing that the effect is due to trapped space charge which inhibits the onset of a double injection current. The dependence of the delay time on defect density, bias, illumination and the location of the defects is examined.

Carrier transport properties, including mobility and carrier lifetimes, or the mobility-lifetime-product, are important parameters for the understanding of the electronic properties of amorphous silicon devices. We have attempted to study these parameters by the junction recovery method which is useful in crystal p-i-n devices, and by the decay of the forward and reverse bias current with varying periods of zero bias delays after removal of the forward bias. We found that the current decay by the standard reverse bias recovery is much faster than that due to the decay under zero bias. In this paper we present our experimental results. We conclude that the true decay of the stored charged due to forward bias is much longer and consequently, the stored charge is much larger than that suggested by standard reverse recovery experiments.

A series of experimental investigations has been made on the a-Si // poly-Si tandem solar cell which is one of the most promised candidate of high cost-performance photovoltaic cell, e.g., high efficiency, low cost with almost no light induced degradation. Employing high conductivity with wide optical band gap p type microcrystalline SiC (μ-SiC) as a window material together with a-SiC as an interface buffer layer and also n type μc-Si as a back ohmic contact layer in the poly-Si based bottom cell, the conversion efficiency of 17.2 % has been obtained. Combining an optically transparent a-Si p-i-n cell as a top cell with an optical coupler between the top and the poly-Si bottom cell, a total efficiency of 20.3 % has been obtained so far on the four-terminal stacked mode structure. A systematic technical data for the optimization of cell structure variation on the developed tandem solar cells are presented and further possibility to improving the performance are discussed.

We have obtained the world's highest total area conversion efficiency of 11.1% for a 100cm2 integrated-type single-junction a-Si solar cell submodule. This was achieved by the development of various advanced technologies, such as a new ultra-thin i/n interface layer and a new laser patterning method using an ablation phenomenon.

To acheive further improvement in the conversion efficiency of a-Si based solar cells, we focus on polycrystalline silicon (poly-Si) thin-film for a-Si/poly-Si tandem solar cells. As far as material technology is concerned, we have used a new solid phase crystallization (SPC) method from amorphous silicon (a-Si) films deposited by plasma-CVD. The maximum mobility of 623 cm2/V.s was achieved on textured substrates at a carrier concentration of 3.0 × 1015 cm-3. This film has been applied to the active layer of poly-Si solar cells on metal substrates and a conversion efficiency of 6.2% has been obtained with poly-Si film of 12 μm thickness made by SPC at 600°C.

In the field of device technology, we have developed new artificially constructed junction (ACJ) solar cells using p-type a-Si/i-type a-Si/n-type crystalline silicon (c-Si). We call this a HIT (Heterojunction with Intrinsic Ihin-layer) structure, and we have achieved a conversion efficiency of 18.1% for this type of solar cells. This is the highest reported value for a cell with a junction fabricated at low temperature (∼ 120°C).

A hydrogen-plasma treatment has been used for the first time to fabricate wide-gap, high-quality a-Si:H films. The hydrogen content (CH) of a-Si:H films substantially increases by the hydrogen-plasma treatment after deposition, without deteriorating the opto-electric properties of the films. The photoconductivity (σph) of ≥ 10-5 ο-1 cm-1, photosensitivity ( σ ph/σ d) of > 106 and SiH2/SiH of <0.2 are achieved for a film with CH of ∼25 atomic >%. The optical gap of the film is > 1.70 eV by the (α h ν )1/3 plot, and is >2 eV by the Tauc's plot. The open circuit voltage of a-Si solar cells exceeds 1 V conserving the fill factor of > 0.7 when the wide-gap a∼Si:H films are used as the i-layer, which proves the wide band gap and low defect density.

Hydrogenated amorphous silicon oxide (a-SiO:H) films were prepared by rf glow discharge decomposition of SiH4, CO2 and H2 gas mixture. These films showed better properties than a-SiC:H films. By applying the a-SiO:H to the p-layer, we attained an efficiency of 12.5% for 1cm2 single junction solar cells and a total-area efficiency of 10.1% for 30cm × 40cm tandem submodules.

We report a study of the correlation between solar cell parameters and electron transport parameters in a series of a-Si1-xGex:H alloys with bandgaps spanning the range 1.42 – 1.72 eV. Fill factors and open circuit voltages were measured in p-i-n solar cells with 0.32 μm thick i-layers. The electron drift mobility and deep-trapping mobility-lifetime product were measured using transient photocurrent techniques on p-i-n cells with thicker i-layers. The open circuit voltage tracked the bandgap accurately. Both the electron drift mobility and the deep-trapping mobility lifetime product correlated reasonably well with the fill factor measured with illumination absorbed near the p-i interface. We present a discussion of the relationship of the fill factor to these transport parameters.

This work reports on attempts to tailor the electric field of a-Si:H solar cells by the graded low-level doping of the intrinsic layer to optimize conversion efficiency in the degraded state. Based on wavelength dependent collection measurements and numerical modeling, the degradation behavior of doped and undoped cells is explained in terms of the interaction of dopants and the light-induced space-charge. Low level doping is shown to shift the electric field away from the p/i interface towards the bulk of the i-layer. This results in a better carrier collection from the back part of the solar cell, and solar cells with improved stabilized red light conversion efficiency can be realized. These cells can be readily applied as bottom cells of stacked solar cells.

High intensity light-soaking and post-degradation room temperature recovery of single junction a-Si:H solar cells were investigated. After high intensity short term light-soaking, it was found that the degraded state was itself not stable. In the dark, fast degraded samples improved with time, even at room temperature. Two different sets of samples showed recovery rates in the range of 1 % gain of relative efficiency per hour. Recovery kinetics were the same in the dark and under low intensity carrier generation. Continued high intensity light soaking resulted in a steady decrease of the recovery rate with time. To explain this behavior, we suggest the presence of a non-electronic, slow process related to lattice reconfigurations that lock-in the light-induced defect states.

We report the CPM defect density of a-Si:H material and the performance characteristics of pin solar cells during high-intensity light-soaking. In one group of experiments we compared the effects of soaking with monochromatic light from a Kr+ laser to white light from a Xe arc lamp. The effects are identical for the same electron-hole pair generation rate. In a second group of experiments we light-soaked at the two different temperatures of 50°C and 90°C. At 90°C the defect density saturates at a lower value than at 50°C, and correspondingly the cell performance parameters saturate at higher values.

Single junction a-Si p-i-n solar cells have been deposited by an Ultra High Vacuum (UHV) Multichamber PECVD system reaching an efficiency of 10.1% over 0.1 cm2 and 9.7% over 1 cm2. The effect of hydrogen treatments on the performance of the solar cells was studied on two different types of SnO2 coated substrates and was correlated with the I-V characteristics under AM1.5 (100 mW cm-2) illumination and the spectral response of the devices. The results show that modifications at the TCO/p-layer interface due to the hydrogen plasma treatments are strongly dependent on the initial characteristics of the TCO.