In this work we present results of a study performed on MIS diodes with the following structure: substrate (glass) / Cr (2000Å) / a-Si:H n+ (400Å) / a-Si:H i (5500Å) / oxide (0-40Å) / Au (100Å) to determine the influence of the oxide passivation layer grown by different techniques on the electrical performance of MIS devices. The results achieved show that the diodes with oxides grown using hydrogen peroxide present higher rectification factor (2×106)and signal to noise (S/N) ratio (1×107 at -1V) than the diodes with oxides obtained by the evaporation of SiO2, or by the chemical deposition of SiO2 by plasma of HMDSO (hexamethyldisiloxane), but in the case of deposited oxides, the breakdown voltage is higher, 30V instead of 3-10 V for grown oxides. The ideal oxide thickness, determined by spectroscopic ellipsometry, is dependent on the method used to grow the oxide layer and is in the range between 6 and 20 Å. The reason for this variation is related to the degree of compactation of the oxide produced, which is not relevant for applications of the diodes in the range of ± 1V, but is relevant when high breakdown voltages are required.

Amorphous silicon (a-Si:H) thin-film transistors (TFT) used in emerging, non-switch applications such as analog amplifiers or active loads, often have a bias at the drain terminal in addition to the gate that can alter their threshold voltage (VT) stability performance. At small gate voltages (0 ≤ VST ≤ 15 V) where the defect state creation instability mechanism is dominant, the presence of a bias at the TFT drain is found to decrease the shift in VT (ΔVT) compared to the ΔVT in the absence of a drain bias. In this paper, a ΔVT model accounting for TFT drain bias, is used to predict the performance of a-Si:H analog circuits in active pixel sensor (APS) medical xray imaging and active matrix, organic light emitting diode (AMOLED) display applications.

The non-reversible degradation without light soaking has been observed in both thin films and solar cells fabricated using the hot-wire CVD technique. For solar cells, a 9.8% initial efficiency became 9.2% when measured after a few weeks of being stored in the dark. For the intrinsic layers, the conductivity increased in the air but remained steady in vacuum up to 700 hours. Also, the conductivity increased and activation energy decreases from an initial ∼0.95 eV to ~∼0.85 eV after several thermal cycles, even in a vacuum. We suggest that the degradation in both materials and solar cells come from the same origin: the upward shift of the Fermi-energy position from its initial value. In other words, the as-grown material is slightly p-type and gradually became slightly n-type after sitting in the air or after thermal annealing cycles in a vacuum. The shift in the Fermi-energy made the fill factor of the solar cell decrease and changes in the electronic properties of the i-layer. It is likely that adsorption of oxygen from water vapor is driving this degradation.

High rate deposition of high quality and stable hydrogenated amorphous silicon (a-Si:H) films were performed near the threshold of amorphous to microcrystalline phase transition using a very high frequency plasma enhanced chemical vapor deposition (VHF-PECVD) method. The effect of hydrogen dilution on optic-electronic and structural properties of these films was investigated by Fourier-transform infrared (FTIR) spectroscopy, Raman scattering and constant photocurrent method (CPM). Experiment showed that although the phase transition was much influenced by hydrogen dilution, it also strongly depended on substrate temperature, working pressure and plasma power. With optimized condition high quality and high stable a-Si:H films, which exhibit σph/σd of 4.4×106 and deposition rate of 28.8Å/s, have been obtained.

Photoluminescence spectroscopy has been applied to investigate localized states in microcrystal-line silicon (μc-Si:H) films and to address the problem of the changes of the electronic properties of this material upon changes of the hydrogen dilution during film growth. By a comparison of photoluminescence and Raman spectra on device grade sample series prepared at different silane concentration in hydrogen (SC) by PE-CVD and HW-CVD a correlation between the micro-structure and the photoluminescence energy is found. It is proposed that the density of band tail states is reduced with increasing SC leading to the increase of the PL energy as well as to the increase of Voc of solar cells. The reason for the tails and their reduction is not clear but strain might play a crucial role and the amorphous hydrogenated phase might be effective for strain reduction.

This work demonstrates the use of an amorphous silicon (a-Si:H) photodetector to measure the density of covalently-bound DNA molecules tagged with a fluorescent molecule. This device is based on the photoconductivity of a-Si:H in a coplanar electrode configuration. Excitation of a fluorescently-tagged biomolecule with near UV/blue light results in the emission of visible light. The emitted light is then converted into an electrical signal in the photodetector, thus allowing the detection of the presence of the tagged DNA molecules. The design, fabrication and characterization of this integrated a-Si:H-based bio-detector is described. The detection limit of the present device is of the order of 20 pmol/cm2. A surface density of ≤ 30 pmol/cm2 for DNA covalently-bound to an active silica layer was measured with the a-Si:H-based bio-detector.

Incomplete recrystallized junction defects of self-aligned, excimer laser annealed polycrystalline silicon (poly-Si) thin film transistor (TFT) was investigated by high-resolution transmission electron microscopy (HR-TEM). TEM observation and simulation result verify that the laser irradiation intensity decreased remarkably at the junction due to diffraction of laser beam at gate electrode edge. We proposed oblique-incidence excimer laser annealing method and successfully eliminated the residual junction defects.

Electron energy distribution function (EEDF) is directly proportional to the second derivative of the probe I-V characteristics. Because of an amplifying effect of unavoidable noises in the experimental probe I-V curves on the derivation process, the experimental I-V curves should be smoothed before performing the numerical derivation. This article investigates the effect of adjustable factors used in the smoothing process on the deduced second derivative of the I-V curves, and an optimum group of the adjustable factors is selected to make the rms deviation of the smoothed I-V curves from the measured ones less than 1 %. A simple differentiation circuit is designed and used to measure the EEDF parameter straightforwardly. It is the first time, so far as we know, to measure the EEDF parameters simultaneously by means of both numerical and circuit derivative methods under the same discharge conditions and on the same discharge equipment. The deviation between two groups of mean electron energy (E) and electron density (ne) obtained by the above different methods is within about 7 %. This apparently improves the reliability of the measurements on the EEDF parameters.

Dangling bond defects (DB) in hydrogenated microcrystalline silicon (μc-Si:H) have been studied by X-band (9 GHz) Q-band (33 GHz) and W-band (90 GHz) electron spin resonance (ESR) spectroscopy. In X-band ESR spectra, all the samples showed asymmetric dangling bond defect signal at g = 2.005 – 2.006. The DB signal shape shows little dependence on substrate temperature in the X-band electron spin resonance (ESR) spectra. In the Q-band and W-band ESR spectra, existence of two centers in DB signals is clearly indicated by the shape of the spectra. The Q-band ESR spectra shape reviles that the peak of one center is at g = 2.0055andthe other is around at g = 2.0060. In addition, the DB signal showed dependence on substrate temperature. The dependence of the DB signals can be explained by difference of intensity ratio of the peaks by these two centers. The signal at g = 2.0060 is consistent with the asymmetric ESR signal observed in the microcrystalline silicon embedded in SiO2. W-band ESR measurement indicates that the signal observed at g = 2.0060 is due to single inhomogeneous species and does not consist of plural species.

HCl was added to SiH4 containing plasmas to grow a-Si:H(Cl) films with dangling bonds terminated with Cl instead of H. Bulk and surface infrared spectra, film thickness and optical band gap were examined by in situ multiple total internal reflection Fourier transform infrared spectroscopy and in situ spectroscopic ellipsometry. SiH2Cl2 was also used as a conventional Cl source for reference a-Si:H(Cl) film deposition experiments. The introduction of HCl does not affect the deposition rate significantly, and the deposited a-Si:H(Cl) films contain over 1021cm-3 Cl atoms. HCl addition to the gas phase changes the surface compositions of the growing films drastically from higher silicon hydride to chlorinated lower hydride. The surface reaction control eliminates unfavorable hydride bonding structures such as SiH2 and/or SiH in voids in the deposited films. The a-Si:H(Cl) films deposited from mixtures of SiH4 and HCl do not show significant optical band gap widening in spite of containing over 1021cm-3 Cl atoms, a concentration that is comparable to that of hydrogen. In contrast, a conventional chlorine source of SiH2Cl2 increases the deposition rate significantly compared to HCl. The increase in the deposition rate results in monotonic decrease of the refractive index and the optical band gap widening.

Using various simulation techniques, such as molecular dynamics and the activation-relaxation technique, we are slowly developing a consistent picture of the dynamical properties of amorphous silicon. For example, results of an extensive search for the activated events surrounding a single minimum, in a well-relaxed model represented by a modified Stillinger-Weber potential, confirm that barrier height at the transition point, for activated mechanisms, is determined essentially by the binding energy of a single bond and not the details of the mechanism. We will discuss these results in some detail as well as recent simulations of nucleation in liquid and amorphous silicon.

With an Expanding Thermal Plasma Chemical Vapor Deposition system (ETP-CVD), solar grade amorphous silicon (a-Si:H) can be deposited at high deposition rate (> 2 nm/s). We think that during the first stage of deposition, a material is grown with a higher defect density than the rest of the bulk creating a defect-rich layer (DRL). Therefore we analyzed, by the means of simulations, the influence of the position of the DRL on the performance of a p-i-n a-Si:H solar cell when moved from the p-i towards the i-n interface and as a function of its thickness. We investigate the effect of a buffer layer in between the p- and the i-layer on the external parameters of the solar cell. The presence of a buffer layer increases the electric field near the p-i interface, which leads to a higher collection of free charge carriers at the interface, although the electric field is then diminished deeper in the bulk. It appears that 10 nm thick buffer layer is sufficient to improve the performance. In case no buffer layer is applied, recombination losses at the p-i interface diminish the performance of the solar cell. We also observe that an increase of the DRL thickness results in a reduction of the solar-cell performance, which is more pronounced when the DRL is located in the region close to the p-i interface rather than close to the i-n interface.

A method for monitoring the evolution of the crystallinity during the deposition of thin hydrogenated silicon films by using in situ spectroscopic ellipsometry is presented. The crystallinity of the topmost 10-20 nm of a film is derived from the analysis of the shape of ellipsometric spectra in the UV range. The values are closely related to parameters of the deposition process and in good agreement with Raman scattering results. Examples of different kinds of microcrystalline silicon films are shown. Improvements of the time resolution and/or accuracy are discussed. The method turns out to be well suited for process control.

The sensitivity of transient photocurrent measurements to the spatial location of native and metastable electronic defects in hydrogenated amorphous silicon films is demonstrated. The technique utilises red and green laser excitation to generate excess carriers in the bulk and at the surfaces of the film, respectively. In annealed films the defect density is found to be higher in the surface regions. Following white light soaking, the metastable defect density at the surface at which the light is incident is greater than that in the bulk, which in turn is greater than that at the exit surface. This is attributed to the white light absorption profile within the film. Green light soaking creates metastable defects at the incident surface, with the bulk of the film largely unaffected.

Using thermally induced hydrosilylation, organic molecules were covalently bonded to H-terminated crystalline silicon (111) and hydrogenated amorphous silicon (a-Si:H) surfaces. The resulting chemical surface structure was analyzed by X-ray photoelectron spectroscopy (XPS) and compared to that of silicon surfaces covered by a native oxide or terminated with hydrogen. For both kinds of substrates, the presence of oxygen on the surface is found to hinder the hydrosilylation reaction. Stable H-termination as a starting point for a successful hydrosilylation can be obtained on a-Si:H surfaces with much less technological effort than on crystalline silicon surfaces. Photoconductivity measurements of the different a-Si:H surfaces at low intensity of illumination (monomolecular recombination regime) indicate that the hydrosilylated surface has less defects than the H-terminated surfaces or surfaces covered with native oxide. Spin-dependent photoconductivity measurements identify the dominant paramagnetic defect at the hydrosilylated a-Si:H surface to be the silicon dangling bond.

We have developed new procedures for determining the microstructure as well as the index of refraction and extinction coefficient spectra {n(E), k(E)} for textured SnO2 thin films on glass used as the top contact layers and superstrate for amorphous silicon (a-Si:H) p-i-n solar cells. These procedures combine (i) multichannel Mueller matrix spectroscopy using a dual rotating-compensator spectroscopic ellipsometer in reflection from the surface of the SnO2, a measurement that is most sensitive to microstructure and n(E), and (ii) transmission spectroscopy through a double-thick SnO2 sandwich contacted with index-matching fluid, a measurement that is most sensitive to k(E). An important optical loss in a-Si:H p-i-n solar cells is reflection from the SnO2/p-layer interface. In this paper, we characterize this optical loss through modeling the solar cell optical quantum efficiency and demonstrate the extent to which microscopic roughness at this interface can serve as an anti-reflection layer for enhanced collection.

In this work, we study the pulsed laser crystallization of hydrogen-free amorphous germanium-nitrogen alloys (a-GeN). We discuss the role of nitrogen during phase transitions and the possible application of the resulting structure as an optical diffraction grating. The crystallized region results of pure microcrystalline germanium (μc-Ge). An indication that Ge-N bonds have broken and nitrogen outdiffused of the film is obtained from infrared spectroscopy and confirmed by Raman spectra. A pattern of alternating a-GeN and μc-Ge lines with a period of about 4 μm acts as an optical diffraction grating due to the difference in optical properties between the two materials, and the three dimensional surface profile, caused by N2 effusion, that is formed on the sample.

We report on the area dependence of switching in both Cr/p+a-Si:H/Ag(Al) and Cr/p+μc-Si/Ag(Al) filament switches. The doped amorphous (a-Si:H) or microcrystalline (μc-Si) thin Si layers are made by hot-wire chemical vapor deposition. The device active region area (A) is varied over 5 orders of magnitude, from 10-7 to 10-2 cm2, using photolithographically defined Ag and Al top contacts. Before switching, the resistance of 100-μm2 devices is normally about 100 kΩ for μc-Si and 10 GΩ for a-Si:H. After switching with applied current ramps, the resistance decreases to a few hundred ohms in all a-Si devices and to a few thousands ohms in μc-Si devices. In both μc-Si and a-Si:H devices, the switching voltage (Vsw) decreases with increasing device area according to Vsw ~ V0-αln(A/A0) with α=0.3V for a-Si:H and α=0.04V for μc-Si. For both materials, the switching current roughly obeys the power law Isw ∞ Aβ with β~1. A statistical model is proposed to explain the area scaling of the switching voltage and relate the parameters to the material properties.

The influence of the density of gap states and the band gap width of the intrinsic a-Si:H active layer on the characteristics of a-Si PIN/OLED coupling pair was analyzed by a-Si:H PIN/OLED CAD simulation model. The CAD simulation model was carried out based on a-Si PIN Hack & Shur model and OLED TCL transport model. At the same band gap width, for the intrinsic a-Si:H active layer with the higher density of gap states, the reverse current of a-Si PIN trended to be saturated at the higher reverse bias voltage. As a result, I-V curve of a-Si PIN/OLED around the turn point Vt became smoother with the increase of the density of gap states. At the same state density, the light induced current of a-Si PIN increased against the band gap width, assuming the input light had the same spectrum as AM1.5 solar light. Thus the luminance emitted from OLED increased with the decrease of the band gap width because OLED belongs to the light-emitting device controlled by current. The simulation results also showed that the influence of the state density intensified with the increase of the band gap of a-Si:H.