Search

The hydrogen concentration in hydrogenated amorphous silicon (a-Si:H), produced by plasma-enhanced chemical vapor deposition, was measured by Raman and Fourier transform infrared spectroscopy before and after thermal annealing. The possibility of Raman spectroscopy to perform depth dependent analysis in adapting the analysis wavelength is used. The findings presented in this paper are supported by secondary ion mass Spectrometry depth profiles. Moreover, we show that Raman spectroscopy can be used to get depth-dependent information about the microstructure and the influence of annealing on the microstructure of a-Si:H.

Laser and oven annealing effects on hydrogen concentration, hydrogen diffusion and material microstructure in hydrogenated amorphous silicon films deposited on crystalline silicon substrates are compared. For laser annealing, a 6 W green (532 nm) continuous wave laser with 100 µm focus diameter was applied and samples of about 1 cm2were scanned in ambient with a line distance of 50 µm and at a speed of 1 – 100 mm/s. Hydrogen content and microstructure were measured by infrared spectroscopy, and hydrogen diffusion was investigated by secondary ion mass spectroscopy (SIMS) measurements of depth profiles of deuterium and hydrogen in layered structures of deuterated and hydrogenated material. The results show that in both annealing experiments hydrogen diffuses predominantly in form of atoms although some formation of H2molecules cannot be excluded. By comparison of laser and oven treatment, an effective temperature describing the laser treated state can be defined. Furthermore, the temperature of the thin silicon film during laser treatment is estimated.

The effects of postdeposition air exposure of microcrystalline silicon films, prepared at varied deposition rates, are investigated. The changes in the oxygen content, evaluated from Fourier transform infrared spectroscopy measurements, were studied over a period of time after deposition (up to 180 days) depending on deposition rate and Raman intensity ratio. Two types of behavior were identified: in the case of highly crystalline samples, an oxygen uptake increases with increasing Raman intensity ratio; while less crystalline samples were found to be more stable against oxygen incorporation. These observations are related to the film microstructure and porosity and are linked to the variations in Raman intensity ratio and growth rate of microcrystalline silicon films.

The influence of implanted hydrogen (up to a concentration level of 3 at. %) on the microstructure of silicon (Si) materials is investigated by Fourier transform infrared spectroscopy as well as by effusion of hydrogen and of (low dose) implanted helium. Three materials of low original hydrogen concentration, crystalline Si, electron beam evaporated amorphous Si, and plasma-deposited hydrogenated amorphous Si (using high deposition temperature) were investigated. Significant differences between crystalline and amorphous materials were observed. In crystalline Si, implanted hydrogen is found to generate multivacancies that trap diffusing helium while this is not the case in amorphous Si. Accordingly, cavities where hydrogen is located in amorphous Si must be smaller than divacancies. Those cavities in amorphous Si, present from the growth process, that trap helium tend to disappear when the implanted hydrogen concentration increases. Annealing of the materials up to temperatures of 575 °C was also studied. No significant change in the density of voids (trapping helium) occur but in case of crystalline Si the bonding sites of hydrogen as well as the diffusion paths of helium change.

Single-chamber processes for the deposition of high efficiency thin-film silicon tandem cells of an a-Si:H p-i-n (top cell)/μc-Si:H p-i-n (bottom cell) structure involving short fabrication time are reported. An industry relevant reactor and an excitation frequency of 13.56 MHz were used. The conversion efficiency is found to be highly sensitive to dopant cross contamination into the μc-Si:H i-layer of the bottom cell and within the n/p-interface of the tunnel recombination junction (TRJ). Different reactor treatments at the p/i-interfaces of the top and bottom cells and at the n/p-interface of the TRJ were applied, aiming to prevent dopant cross contamination. The phosphorus and the boron concentrations were evaluated by secondary ion mass spectrometry measurements. Phosphorus cross contamination after TRJ n-layer deposition is found to result in significant n-type doping of the μc-Si:H i-layer of the bottom cell if no reactor treatment is applied. In situ reactor treatment via an Ar flush and pumping step of 15 min applied at the n/p-interface of TRJ results in reduction of the μc-Si:H i-layer phosphorus concentration to values below 1017 cm− 3. A conversion efficiency of 11.8% for such tandem cells is demonstrated. Shorter interface treatment time with phosphorus concentrations in the μc-Si:H i-layer of about 5 × 1017 cm− 3 results in lower conversion efficiencies of 10.6%, mainly due to the decrease of open-circuit voltage and fill factor.

The influence of oxygen and nitrogen impurities on the material properties of a-Si:H and μc-Si:H films and on the corresponding solar cell performances was studied. For intentional contamination of the i-layer the impurities were inserted by two contamination sources: (i) directly into the plasma through a leak at the chamber wall or (ii) into the gas supply line. The critical oxygen and nitrogen concentrations for silicon solar cells were determined as the lowest concentration of these impurities in the i-layer causing a deterioration of the cell performance. Similar critical concentrations for a-Si:H and μc-Si:H cells in the range of 4–6 × 1018 cm− 3 for nitrogen and 1–5 × 1019 cm− 3 for oxygen by applying a chamber leak are observed. Similar increase of conductivity with increasing impurity concentration in the a-Si:H and μc-Si:H films is found. A more detailed study shows that the critical oxygen concentration depends on the contamination source and the deposition parameters. For a-Si:H cells, the application of the gas pipe leak leads to an increased critical oxygen concentration to 2 × 1020 cm− 3. Such an effect was not observed for nitrogen. For μc-Si:H, a new deposition regime with reduced discharge power was found where the application of the gas pipe leak can also result in an increase of the oxygen concentration to 1 × 1020 cm− 3.

For deposition of a-Si:H p–i–n solar cells, a single-chamber plasma enhanced chemical vapor deposition process at the frequency of 13.56 MHz is developed. A 40×40 cm² deposition chamber, which represents typical industry reactors equipped with showerhead electrodes is employed. Various methods are applied to reduce the boron-cross contamination from the boron-doped p-layer into the intrinsic layer, which is considered to reduce solar cell efficiency by losses especially in the short wavelength range. Three different device configurations and four different chamber treatment methods are studied, aimed to reach stable device efficiencies comparable to multi-chamber systems at minimum chamber treatment effort and treatment time. An ex-situ CO2–plasma treatment applied after deposition of the p-doped layer is found to be effective to reduce boron-cross contamination. However, this CO2-treatment is a time-consuming process step for production. We found a less time consuming treatment: by a chamber evacuation to 9×10−7 mbar subsequent to p-layer deposition. Initial and stable efficiencies of 10.2% and 7.7%, respectively, were obtained. This latter treatment results in a sharp drop of the boron concentration from ∼5×1020 cm−3 in the p-doped layer to ∼1017 cm−3 in the intrinsic layer. For comparison of different reactor geometries and their influence on the cross-contamination we used a small-area (10×10 cm2) lab-type reactor.

Microcrystalline Si:Cl:H films prepared by plasma deposition using mixtures of silicon tetrachloride and hydrogen as process gases were studied by effusion of implanted helium. The temperature of maximum (low temperature) effusion rate is used as a measure of material density and microstructure. The results suggest that chlorine incorporation at concentrations exceeding a few percent results in a void-rich structure. High p-type conductivities obtained for boron-doped microcrystalline Si:Cl:H are explained by the presence of rather dense material with high crystalline fraction.

The use of chlorosilanes (silicontetrachloride and dichlorosilane) for rf plasma-deposition of microcrystalline silicon is explored. Deposition conditions similar to those employed for high rate deposition of efficient microcrystalline silicon solar cells (frequency: 13.56 MHz, high rf power, high pressure and high hydrogen dilution of process gas) were used and high deposition rates were achieved. Silane-based materials, deposited under the same conditions, were studied for comparison. Lowest porosity is obtained for both chlorinated and unchlorinated material at a crystallinity (measured by Raman scattering) near 70%. Infrared absorption measurements indicate the presence of more void-related microstructure in chlorosilane-based material than in silane-based. Conductivity measurements give a Fermi level near midgap for an undoped chlorinated material while our silane-based samples show unintentional doping effects. By doping of chlorinated material with boron, high conductivities are obtained. Chlorinated material was successfully incorporated as p-layer in microcrystalline Si solar cells.

This study addresses the electrical and optical properties of radio frequency magnetron sputtered aluminum doped zinc oxide (ZnO:Al) films. The main focus was on the improvement in carrier mobility mu to achieve simultaneously high transparency for visible and particularly near-infrared light and low resistivity. The influence of Al concentration in the target, film thickness, sputter power, deposition pressure, and substrate temperature on material properties was investigated. The structural, compositional, electrical and optical properties were studied using x-ray diffraction, secondary ion mass spectrometry (SIMS), room temperature Hall effect measurements and spectral photometry, respectively. All ZnO:Al films were polycrystalline and preferentially oriented along [002]. The grain size along the direction of growth increased with higher Al doping and with increasing film thickness. The SIMS measurements revealed that the Al concentration in the film was nearly the same as in the target. Carrier concentration N and mobility mu are determined by the target Al concentration. In addition mu is influenced by the film thickness and the sputter pressure. For each Al concentration, the highest mu was generally observed at low deposition pressures. By using a target with low Al2O3 concentration of 0.5 wt %, mu could be improved up to 44.2 cm(2)/V s while maintaining the electrical resistivity rho as low as 3.8x10(-4) Omega cm. For these films the transparency in the near-infrared wavelength range strongly improved which makes them particularly interesting for the application in optoelectronic devices like thin-film solar cells. The mu-N dependence for films deposited under diverse conditions was studied to identify a practical limit for mu. (C) 2004 American Institute of Physics.

The effusion of implanted helium was studied for doped and undoped radio frequency plasma-deposited a-Si:H films as well as for Si:H films deposited by other techniques. The results suggest that the observed helium diffusion and precipitation effects give information on the material microstructure.

The dependence of hydrogen effusion and diffusion on hydrogen concentration was studied for crystalline and amorphous Si:H samples which were hydrogenated by hydrogen ion implantation. The results are compared to data for a-Si:H and μc-Si:H films grown with various H concentrations. Although general trends of the H concentration dependence of H stability are similar, there are differences between c-Si:H and a-Si:H. These differences involve larger hydrogen-generated microstructural effects in c-Si:H and a hydrogen diffusion coefficient in compact material increasing in a-Si:H and decreasing in c-Si:H with increasing H concentration. A different flexibility of the atomic network in amorphous and crystalline Si could be the origin. Microcrystalline Si:H films show a similar concentration dependence of H diffusion as a-Si:H.

Epitaxial CoSi2 layers were grown on Si(1 0 0) using molecular beam allotaxy. Boron ion implantations and rapid thermal oxidation (RTO) were performed. During oxidation, SiO2 formed on the surface of the CoSi2 layers, and the silicides was pushed into the substrate. The diffusion of boron was slightly retarded during oxidation for the specimen with a 20 nm epitaxial CoSi2 capping layer as compared to the specimen without CoSi2 capping layer. The electrical measurements showed that the silicide has good Schottky contacts with the boron doped silicon layer after RTO. A nanometer silicide patterning process, based on local oxidation of silicide (LOCOSI) layer, was also investigated. It shows two back-to-back Schottky diodes between the two separated parts of the silicide.

Hydrogen diffusion and solubility effects in hydrogenated crystalline and amorphous silicon prepared by hydrogen implantation with approximately equal hydrogen concentrations are compared. Hydrogen diffusion is found to be limited by hydrogen solubility in both cases. Between 300 and 450°C, an agreement of the hydrogen diffusion coefficients for the two Si:H materials approximately within a factor of ten is observed. With increasing hydrogen concentration, hydrogen diffusion increases for amorphous Si:H but decreases for crystalline Si:H. This latter decrease is attributed to a hydrogen related microstructure which appears in crystalline Si:H at about a factor of ten smaller hydrogen concentration than in amorphous Si:H.

SIMS depth profiling during O2+ bombardment has been performed to analyse epitaxially grown Si p-n-p layers, which define the p-channel region in vertical Si-p MOS transistors, as well as to establish “on-chip” depth profiling of the functional vertical device. The SIMS detection limit of 31P in Si, phosphorus used as n-type dopant in the transistor, has been optimised as a function of the residual gas pressure in the SIMS analysis chamber and of the sputter erosion rate. We demonstrate that good vacuum during SIMS analysis combined with high erosion rates allows the simultaneous quantitative SIMS depth profiling of n- and p-type dopant concentrations in the vertical transistor. Small area “on-chip” SIMS depth profiling through the layered structure of Al-contact/TiSi2/Si(p-n-p)/Si-substrate has been performed. Factors influencing the depth resolution during “on-chip” analysis of the transistor are discussed especially in terms of sputtering induced ripple formation at the erosion crater bottom, which has been imaged with atomic force microscopy.

Microcrystalline silicon carbide (μc-SiC:H) deposited by hot wire chemical vapor deposition (HWCVD) and plasma-enhanced chemical vapor deposition (PECVD) provide advantageous opto-electronic properties, making it attractive as a window layer material in silicon thin-film and silicon heterojunction solar cells. However, it is still not clear which electrical transport mechanisms yield dark conductivities up to 10−3 S/cm without the active use of any doping gas and how the transport mechanisms are related to the morphology of μc-SiC:H. To investigate these open questions systematically, we investigated HWCVD and PECVD grown layers that provide a very extensive range of dark conductivity values from 10−12 S/cm to 10−3 S/cm. We found out by secondary ion mass spectrometry measurements that no direct correlation exists between oxygen or nitrogen concentrations and high dark conductivity σd, high charge carrier density n, and low activation energy Ea. Higher σd seems to rise from lower hydrogen concentrations or/and larger coherent domain sizes LSiC. On the one hand, the decrease of σd with increasing hydrogen concentration might be due to the inactivation of donors by hydrogen passivation that gives rise to decreased n. On the other hand, qualitatively consistent with the Seto model, the lower σd and lower n might be caused by smaller LSiC, since the fraction of depleted grain boundaries with higher Ea increases accordingly.

SIMS depth profiling, using an oxygen primary beam at close to normal incidence, was applied to study lithium diffusion in thin films of hydrogenated amorphous silicon (a-Si:H) as well as in layered structures of doped and undoped silicon-based alloys. A strongly increasing decay length of the lithium profiles was observed for increasing primary beam energies and is attributed to the Li accumulation at the SiOx/Si interface. The stability of the lithium incorporation in a-Si:H is found to depend on the presence of charged acceptors or defects and of doping gradient related electrical fields.

For microcrystalline silicon based p–i–n solar cells the effect of deposition conditions on the critical oxygen concentration was investigated. All solar cells were prepared by 13.56 MHz plasma-enhanced chemical vapour deposition. The critical oxygen concentration defines the lowest oxygen concentration in the intrinsic absorber layer causing a deterioration of the solar cell performance. For intentional contamination of ∼1.2–1.3 µm thick i-layers, the oxygen was inserted by a controllable leak at the process gases supply line, i.e. by a gas pipe leak. For µc-Si:H deposited at a discharge power of 0.53 W/cm2 we find a critical oxygen concentration of 1–2 × 1019 cm–3 in agreement with values commonly reported in literature. However, changing the deposition conditions, we find that the critical oxygen concentration in µc-Si:H cells is not fixed. At reduced power of 0.20 W/cm2 a much higher value for the critical oxygen concentration of 1 × 1020 cm–3 is observed. (© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Using secondary ion mass spectroscopy (SIMS), the stability of lithium doping by in-diffusion and ion implantation is studied for doped and undoped a-Si:H films. The results show a higher stability of implanted compared to in-diffused samples attributed to interaction of Li with implantation defects. Furthermore, for highly boron-doped material the in-diffusion is greatly enhanced while the out-diffusion is strongly inhibited. These latter effects are attributed to field enhanced diffusion and ion pairing.

An improved version of a data processing and control system has been used with state-of-the-art AES and SIMS instruments. It consists of a mainframe with CPU, buffer and IEEE 488 interface to a host computer and a range of plug-in submodules. Its features include three-dimensional analysis, imaging, multichannel spectra and sample bias control. The performance of the system has been tested by AES and SIMS depth profile measurements of Si-Ge multilayers.

For three types of hydrogenated silicon films, amorphous, microcrystalline and crystalline hydrogenated silicon, hydrogen diffusion was studied as a function of doping level employing depth profiling by secondary ion mass spectrometry. Hydrogen implantation was used to control the hydrogen concentration. All three materials show a similar doping dependence of H diffusion, namely a strong increase upon boron (p-type) doping and a much lesser increase for n- type (P, As) doping. In a band model of H diffusion, the effect is related to a decrease in energy of the hydrogen diffusion path. Possible explanations are a different charge state of diffusing hydrogen or an effect of the Fermi energy on the release energy of neutral hydrogen.

The concentration dependence of hydrogen diffusion was studied in hydrogenated crystalline and amorphous silicon prepared by hydrogen implantation into crystalline Si wafers and into amorphous silicon of low hydrogen concentration. The results are compared with data for plasma-grown a-Si:H and µc-Si:H films. The increase of the diffusion coefficient with rising hydrogen concentration in a-Si:H is explained by an (equilibrium) energy band model of hydrogen diffusion whereas the decrease of the diffusion coefficient in c-Si:H is explained by a trapping model. The different behavior is attributed to a greater flexibility of the amorphous Si network compared to the crystalline Si lattice which is also visible in a difference in hydrogen-related microstructure formation.

The diffusion and effusion of hydrogen in hydrogenated microcrystalline silicon films deposited in an electron cyclotron resonance reactor were studied for various deposition temperatures Ts. For deposition temperatures below 250°C, hydrogen effusion is found to be dominated by desorption of hydrogen from internal surfaces followed by rapid out-diffusion of H2. Higher substrate temperatures result in an increased hydrogen stability suggesting the growth of a more compact material. For this latter type of samples, a hydrogen diffusion coefficient similar as in compact plasma-grown a-Si:H films is found despite a different predominant bonding of hydrogen according to infrared absorption.

SIMS measurements of implanted Li in (doped and undoped) a-Si:H, a-Si and a-Si:H alloys are discussed. The results suggest basically the same Li diffusion process in p- and n-type a-Si:H with negatively charged acceptors or dangling bonds acting as traps for the positively charged diffusing Li ions. Li is less stable in a-Si than in a-Si:H; Li stability in a-Si:H is enhanced at high concentrations of implanted Li. Both diffusion and stability are modified by doping gradient-related electric fields.

The influence of atmospheric contaminants oxygen and nitrogen on the performance of thin-film hydrogenated amorphous silicon (a-Si:H) solar cells grown by plasma-enhanced chemical vapor deposition at 13.56 MHz was systematically investigated. The question is addressed as to what degree of high base pressures (up to 10(-4) Torr) are compatible with the preparation of good quality amorphous silicon based solar cells. The data show that for the intrinsic a-Si: H absorber layer exists critical oxygen and nitrogen contamination levels (about 2 x 10(19) atoms/cm(3) and 4 x 10(18) atoms/cm(3), respectively). These levels define the minimum impurity concentration that causes a deterioration in solar cell performance. This critical concentration is found to depend little on the applied deposition regime. By enhancing, for example, the flow of process gases, a higher base pressure (and leak rate) can be tolerated before reaching the critical contamination level. The electrical properties of the corresponding films show that increasing oxygen and nitrogen contamination results in an increase in dark conductivity and photoconductivity, while activation energy and photosensitivity are decreased. These effects are attributed to nitrogen and oxygen induced donor states, which cause a shift of the Fermi level toward the conduction band and presumably deteriorate the built-in electric field in the solar cells. Higher doping efficiencies are observed for nitrogen compared to oxygen. Alloying effects (formation of SiOx) are observed for oxygen contaminations above 10(20) atoms/cm(3), leading to an increase in the band gap. (C) 2008 American Institute of Physics. [DOI: 10.1063/1.3009384]

Hydrogen diffusion in zinc oxide thin films was studied by secondary ion mass spectrometry (SIMS) measurements, investigating the spreading of implanted deuterium profiles by annealing. By effusion measurements of implanted rare gases He and Ne the microstructure of the material was characterized. While for material prepared by low pressure chemical vapour deposition an interconnected void structure and a predominant diffusion of molecular hydrogen was found, sputter-deposited ZnO films showed a more compact structure and long range diffusion of atomic hydrogen. Hydrogen diffusion energies of 1.8 – 2 eV, i.e. higher than reported in literature were found. The results are discussed in terms of a H diffusion model analogous to the model applied for hydrogen diffusion in hydrogenated amorphous and microcrystalline silicon.

The concentration profiles of B, P, Zn, and O in the active intrinsic (i) νc-Si:H layer and across the interfaces in p-i-n and n-i-p structures have been measured with SIMS. For the ZnO/νc-p/νc-i sequences, an apparent B and O profile extends over several hundred nanometers into the i-layer, and high levels of Zn can be found well above the ZnO substrate layer. These profiles are not affected by annealing at the deposition temperature. Much lower impurity concentrations are measured for n-i-p deposition sequences, or when the p-layer or the i-layer is amorphous. Ruling out diffusion and intermixing processes, evidence for the presence of pinholes in the material is presented, which explain most of the experimental findings.