The etching of a-Si:H, c-Si and SiO2 in glow discharges of SiF4 and H2 was studied. The order of the etch rates in a SiF4 plasma is a-Si:H > c-Si > SiO2. The etch rates of these materials are nearly independent of temperature. Only a-Si:H is etched in a H2 plasma.

Undoped a-Ge:H films have been made both by sputtering and by glow discharge under various combinations of deposition parameters, and characterized structurally, electrically, and optically. A detailed comparison of subband gap spectra of sputtered and glow discharge samples has been carried out using photothermal deflection spectroscopy and the constant photocurrent method. The results show subband gap absorption in a-Ge:H to be an order of magnitude larger than in device-quality a-Si:H. For some of the films the absorption spectra deduced from transmission and PDS fit very well onto the absorption spectra inferred from CPM measurements, while for others it is clear that the match is very poor. We infer that the photoconductivity-derived spectrum does not necessarily reflect the dominant absorption process, and that this is related to the existence of a two-phase structure clearly observed by transmission and scanning electron microscopy.

Results from structural measurements on r.f. glow discharge produced a-Ge:H films have been found to be substrate dependent. The variations in the results were found to depend on both the substrate temperature, Ts, and the substrate yield strength. Differential scanning calorimetry results were particularly affected by these parameters. For films prepared at Ts = 150°C, the DSC spectra contain two exothermic peaks when the films are deposited on low yield strength substrates while only one exothermic peak is present for films deposited on high yield strength substrates. One exothermic DSC peak is seen in spectra for all films prepared at Ts = 300°C no matter what substrates were used. This DSC spectral dependence is attributed to differences in the microstructure of films deposited at the two substrate temperatures, as seen in TEM micrographs. X-ray diffraction measurements performed on films annealed to various temperatures show that all of the exothermic DSC peaks described above are associated with the crystallization process. Thus, for the films prepared at low Ts, crystallization is either a one or two step process depending on the yield strength of the substrate.

A vacuum UV-photo-CVD reactor for direct decomposition of Si2H6, SiH4, GeH4, and B2H6 for a-Si:H, a-Ge:H, and a-SiGe:H deposition has been operated with a large area dielectric-barrier discharge lamp emitting excimer radiation of Xe (7.3eV), Kr (8.6eV), or Ar (9.8eV).

It is shown that the deposition process (Si2H6 or GeH4 at 0.5 mbar) is governed by the diffusion of long-lifetime radicals.

High quality undoped and p-type doped a-Si:H has been deposited from Si2H6 and B2H6 with the Xe lamp spectrum (σph/σ≈106 at Eg=1.9eV) and film growth rates of up to 5 nm/min have been achieved.

The remote plasma deposition process was studied by optical emission spectroscopy using Ne and N2 to detect He-metastables. a- Ge:H was prepared and its optoelectronic and structural properties were characterized. AM 1.5 photoconductivities around 10−6(Ωcm)−1 were obtained in intrinsic material with the Fermi level position lying near midgap. However, even in the best film the defect density is higher than 1017cm−3.

Amorphous hydrogenated germanium films have been deposited in a conventional glow discharge reactor (diode as well as triode configuration) from GeH4 with and without additional H2 dilution. In this study special emphasis was put on samples deposited at different substrate temperatures, different neutral gas densities and dilution ratios. The deposition rate has been varied from 0.2 to 2.5 Å/s. Optical, structural and electronic properties have been studied by VIS/NIR thin film spectroscopy, PDS, Raman backscattering, FTIR, as well as temperature dependent dark and photoconductivity. Hydrogen content was monitored by quantitative SIMS analysis and FTIR spectroscopy.

By ‘soft’ deposition, i.e. high pressure and H2 dilution ratio, Fermi level positions near midgap (Eact =0.48eV at Eg=1.0ev) are achieved. Nevertheless the optimum ratio σph/σd is restricted to 0.6 at room temperature (λ=632.8nm). To point out the influence of Fermi level position for transport properties, we investigated two samples with different Fermi level positions in detail. σd(T) is strongly dependent on Fermi level positions, as well as the absolute values of σph(T), whereas the temperature dependence of σph(T) does not differ greatly. This is an indication, that Fermi level position and recombination centers of a-Ge:H can be effectively influenced by changing deposition parameters.

We have studied the improvement of the quality of undoped a-Si:H deposited by remote-plasma chemical vapour deposition. The effects of reactant gas concentration, rf power, substrate bias voltage on the electrical and optical properties have been investigated. Some hydrogen dilution of si lane improves the photoeletric property and a high rf power gives rise to the defect creation due to the ion bombardment on the growing surface. The positive substrate bias improves the quality of undoped a-Si:H.

The influence of deposition temperature on hydrogen incorporation in a-Si:H prepared by VHF glow discharge at 70 MHz is investigated using hydrogen evolution, elastic recoil detection analysis and infrared spectroscopy.The films were further characterized by dark- and photoconductivity and by photothermal deflection spectroscopy. While the electronic film properties deteriorate in the usual manner with decreasing substrate temperature it is found that the total hydrogen content CH and the degree of microstructure that can be directly correlated to CH increase only moderately. It is concluded that a higher flux of low energy ions in the VHF plasma plays a key role in this context, possibly by increasing the surface mobility of the H atoms and thereby preventing the build-in of a large amount of hydrogen at low substrate temperatures.

The dependence of chemical composition, structure and optoelectronic properties of sputtered a-SiC:H thin films on substrate temperature, Ts, and hydrogen flow rate has been studied. The films are amorphous for the growth conditions used in this work. The chemical composition of the alloys is very little influenced by the Ts, whereas the hydrogen content and the optical absorption coefficient depends strongly on Ts and hydrogen flow rate.

Double modulated beam mass spectrometry was used to obtain the first measurements of the identity, relative flux, and energy of film precursors in reactive dc magnetron sputtering. In these experiments, a 2″ diameter silicon target was sputtered in an argon plus deuterium plasma at an argon pressure of 1.5 mTorr. This system produces high quality a-Si:D. Energetic neutral Si and SiD were observed, while energetic neutral SiD2 and SiD3 were absent. The arrival rate of D in the form of SiD increased and then saturated at 19 percent of the total flux as the deuterium pressure was varied from 0 to 5 mTorr. The sputtered Si energy agreed qualitatively with the standard Thompson theory. Energetic ions located at mass 2 and 4 were measured with energies much greater than the sputtered Si. Thermal silane was also detected, raising the possibility that thermal silane radicals in addition to energetic Si and SiD contribute to film growth.

Preliminary results are presented for a new approach proposed by the present investigators to solve the problem of light-induced degradation in amorphous silicon semiconductors. The approach uses low-temperature metal-organic chemical vapor deposition (LTMOCVD) of tailored organometallic precursors. The precursors employed are non-toxic, non-hazardous and easy to handle. In the present paper, a-Si:H films were grown, using argon with various hydrogen concentrations as carrier gas, in a cold-wall CVD reactor at a reactor pressure of 1-10 torr and substrate temperature in the range 300–450°C. Characterization studies were performed using x-ray diffraction (XRD), x-ray photoelectron spectroscopy (XPS), and extended electron-energy-loss fine structure spectroscopy (EXELFS). The results of these studies showed that the films were uniform, continuous, adherent and highly pure--contaminant levels were below the detection limits of XPS. In addition, EXELFS results showed that short-range order (SRO), consisting of the same tetrahedral coordinated units found in crystalline silicon, does exist in all the amorphous samples, regardless of hydrogen concentration. However, the degree of stuctural disorder in the silicon local tetrahedral units decreased as hydrogen was added.

The deposition process of multiple layer structures of intrinsic and p-type hydrogenated amorphous silicon was followed by measuring the microwave detected transient photoconductivity (TRMC) during the film growth. In an i-p-i structure we can show that after deposition of an upper layer of about 500 nm, former deposited layers do not influence the TRMC-signal any more. In an i-p+-i-p structure we can clearly distinguish between p-layers of different doping concentration.

The thickness dependence of the sub-gap optical absorption in plasma-deposited hydrogenated amorphous silicon is carefully studied by photo-thermal deflection spectroscopy. The deep-level defect concentration decays from the top surface into the bulk where it approaches the thermal equilibrium defect density. This defect profile is interpreted in terms of the annealing, during growth, of growth-induced surface defects. It is also shown that the defect profile is compatible with the known growth-temperature dependence of the average defect density in amorphous silicon.

A parameter which determine internal stress of hydrogenated amorphous silicon (a-Si:H) prepared by plasma CVD method has been investigated to prevent a peeling off or crack forming problem. It is clarified that the internal stress changes from tension to compression as a function of supplied rf power density during deposition process. Supplied rf power density is closely connected to H o and SiH* optical emission intensity ratio (H α /SiH* ) of CVD plasma. Ve find that the internal stress can be reduced remarkably by controlling this ratio to 0.3. This H α/SiH* parameter has a superiority that the internal stress can be in situ controlled.

Microwave excited electron cyclotron resonance (ECR) plasmas have been used to fabricate a-Si:H,Cl and a-SiC:H films. By varying the absorbed microwave power, highly photoconductive a-Si:H,Cl films were obtained. The a-SiC:H films do not have appreciable photoconductivity, but they show a bright, broad continuous wave (cw) photoluminescence (PL) in the visible region under near-uv excitation. We have measured the photoluminescence decay in the picosecond range and have resolved the spectra into four distinct components. The four luminescence components may correspond to various recombination mechanisms of photogenerated electron-hole pairs.

We propose to apply a new method to model the optical response of amorphous silicon thin films. This method presents the advantage of having a good physical insight. On the other hand, although the model has been originally tested on different materials like a-Si, a-Ge and a-GaAs, we show that it is also sensitive to small differences like those that can exist between intrinsic and doped a-Si:H.

A set of samples grown under different substrate temperature has been characterised by complex optical, photoelectrical and transport measurements. A very sensitive and simple method for detection of optical scattering in a device quality amorphous silicon has been demonstrated. The spectral dependence of the optical scattering coefficient αs(E) can give us information on the nature and concentration of the scattering centers in a-Si:H thin films.

The influence of diluting SiH4 with H2 on the optical and structural properties of a-Si:H has been investigated. The major effects of the dilution are a pronounced decrease of the amount of (Si-H2)n-chains, an increase of the macroscopic density, and a decrease of the total amount of incorporated hydrogen in the films. Different models for the structure as a function of the hydrogen concentration are deduced. We also present new data on the dependence of the optical band gap Eg on the hydrogen concentration and density of the films. An important result from this work is that Eg appears to correlate with the distance between the silicon atoms, rather than with the overall hydrogen concentration.

The results of transient photoconductivity and photoinduced absorption measurements in lightly phosphorus and boron doped (0-10 ppm ) a-Si:H films at T = 300 K are presented. From the observed increase of minority carrier trapping with doping it is concluded that even light doping leads to a strong increase of the density of deep traps. Furthermore an abrupt change of the behaviour of the transient photoconductivity between 2 ppm and 3 ppm B2H6 is observed. This change may be interpreted as the transition to a p-type semiconductor.

The cw proton NMR spectra, spin lattice relaxation times Tx and free induction decay times T*2 were measured in various a-Si:H films deposited by rf-sputtering. All spectra exhibited broad and narrow components, of T*2 ∼ 20 and 45 μsec, respectively. In samples deposited at high rf power, the ratio of the integrated intensities of the broad and narrow lines was ∼ 1.5. These samples, which apparently contain mainly bulk monohydride bonds and a low volume fraction of microvoids, exhibit long-range H motion. In a sample deposited at low rf power, this ratio was ∼ 3. These samples, which contain significant dihydride and microvoid content, show a dramatic suppression of the long-range H motion. The T1 of both types of samples exhibits a clear minimum at ∼ 30 K, indicative of relaxation by H2 trapped in microvoids. The absolute values of T1, however, are smaller in the sample deposited at low rf power (0.1 sec at the minimum as opposed to 0.3 to 1.0 sec at the minimum), apparently due to a larger ratio of molecular hydrogen to bonded hydrogen. After annealing for 24 hours at 294°C, the ratio of broad-to-narrow components increases to > 10 in all samples, and the magnitudes of the T1 minima increase to about 1 sec in all samples.