Fine-grained Polycrystalline Silicon Research Articles

The influence of space charge regions on effective charge carrier lifetime in thin films and resulting opportunities for materials characterization

The analysis of injection-dependent charge carrier lifetimes is a well-established method to determine material and interface quality in crystalline silicon wafer-based device structures such as solar cells. However, for thin films, this method has rarely been used. One reason is that the physical interpretation of experimental data must rely on advanced theoretical models. In this study, we show by numerical simulations and analytical approximations that the effective charge carrier lifetime in thin films is heavily affected by space charge regions (SCR) over a wide range of injection levels. By analysis of the characteristic features in the injection-dependent effective charge carrier lifetime curves, qualitative information about SCRs that occur at grain boundaries or interfaces can be obtained. In contrast, information about the defect density can only be extracted in a very limited range of injection levels and the relationship between effective charge carrier lifetime and the quasi-Fermi level splitting, which is limiting the open circuit voltage of wafer-based solar cells, is not valid in thin films. On the basis of this theoretical study, we analyze measurements of effective charge carrier lifetime in 1.5 μm thin and 2 μm fine-grained polycrystalline silicon films with lifetimes of up to 100 μs and find experimental evidence for grain boundary potential barriers. Finally, we present guidelines for optimized photoconductance measurements and the evaluation of charge carrier lifetimes in thin films, in general.

Effective hydrogenation and surface damage induced by MW-ECR plasma of fine-grained polycrystalline silicon

This work reports the investigations on the effects of the hydrogenation process of thin film polycrystalline n+pp+ mesa silicon cells using MW-ECR plasma in a conventional PECVD system. Different operating parameters such as MW-ECR power, annealing temperature and the doping level of the emitter region were varied. The n+-type emitter regions were obtained by phosphorus diffusion in a conventional furnace using an oxide doping source containing phosphorus (P507 or P509 solutions, from Filmtronics Inc.). The MW hydrogenation was carried out at a sample temperature of 400°C for 60 min. Both types of emitters formed from P507 and P509 showed V oc of 155 mV and 206 mV, which increased linearly to 305 mV and 331 mV, respectively, after hydrogenation when the MW power varied from 200 to 650 W. However, the sheet resistances of the n+ emitter region showed a slight increase depending upon hydrogenation power because of its etching. In a further study, hydrogenated samples were annealed in neutral or forming gas (FG) and we observed interesting results on V oc in the presence of FG. The FG annealing temperature study revealed a strong dependence of V oc on MW power, which affected the etching level of emitter and emitter dopant concentration, which controls the diffusion of hydrogen ions during post-hydrogenation step. The results were explained in detail by combining the effects of MW power and dopant level of the emitter.

Ellipsometric characterization of nanocrystals in porous silicon

Porous silicon layers (PSLs) were prepared by electrochemical etching of p-type single-crystalline silicon (c-Si) wafers having different dopant concentrations to obtain systematically changing sizes of nanocrystals (walls). The microstructure of the porous material was characterized using spectroscopic ellipsometry with multi-layer effective medium approximation (EMA) models. The dielectric function of PSL is conventionally calculated using EMA mixtures of c-Si and voids. The porosity is described by the concentration of voids. Some PSL structures can be described only by adding fine-grained polycrystalline silicon (nc-Si) reference material to the EMA model. Modified model dielectric functions (MDF) of c-Si have been shown to fit composite materials containing nanocrystalline regions, either by fitting only the broadening parameter or also other parameters of the parametric oscillator in MDF. The broadening parameter correlates with the long-range order in the crystalline material, and, as a consequence, with the size of nanocrystals. EMA and MDF models were used to describe systematically changing nanostructure of PSLs. Volume fraction of nc-Si in EMA and broadening parameter in MDF provide information on the nanocrystal size. The longer-term goal of this work is to provide a method for the quantitative characterization of nanocrystal size using quick, sensitive and non-destructive optical techniques.

Passivation and etching of fine-grained polycrystalline silicon films by hydrogen treatment

Abstract Here we investigated the effects of hydrogen treatment on highly defected polycrystalline silicon solar cells in terms of defects passivation and surface etching. The poly-Si films were formed by high-temperature chemical vapour deposition. The hydrogen treatment was carried out through deposition of a-SiN x :H layer followed by a thermal treatment or by direct hydrogen plasma. The deposition of silicon nitride layers on polysilicon cells led to a slight increase in the open-circuit voltage without damage to the surface. In contrast, after plasma hydrogenation, the results revealed an etching process of the emitter simultaneously with an important increase of the measured open-circuit voltage by a factor 2, reaching 420 mV.

Defect passivation in chemical vapour deposited fine-grained polycrystalline silicon by plasma hydrogenation

This study investigates defect passivation in fine-grained polycrystalline silicon layers by means of hydrogenation. The polycrystalline silicon layers are deposited on oxidized Si wafers using high-temperature chemical vapour deposition. The hydrogenation treatment is performed in a direct plasma enhanced chemical vapour deposition system. The samples are characterized by resistivity vs. temperature measurements of p-type layers and by quasi-steady state open-circuit voltage measurements of p–n diodes made in the material. The results show a large increase of the measured open-circuit voltage by the hydrogenation treatment, with the open-circuit voltage of the samples at a light intensity of 1 Sun rising from 180 mV without hydrogenation to values up to 380 mV. This corresponds to a defect concentration decrease by a factor of three.

EBIC technique applied to polycrystalline silicon thin films: minority carrier diffusion length improvement by hydrogenation

Direct deposition of 10–15-μm-thick silicon layers onto foreign substrates yields a polycrystalline Si material with an average grain size in the range 1–5 μm. The strong electrical activation of such large density and small grains is detrimental to solar cell performances. Here we investigate the passivation of such defects by ECR plasma hydrogenation. The electron beam induced current (EBIC) technique is used to reveal the electrical activity of the defects and to extract the minority carrier diffusion length on hydrogenated fine-grained polycrystalline silicon thin films. It is shown that the minority carrier diffusion length is mainly ruled by the grain size (1–2 μm) in the as-grown polycrystalline silicon layer. Hydrogen plasma treatments are performed, taking into account the deep defect passivation and the doping rate control. The diffusion length is then improved up to 10 μm, the same order of magnitude than the probed silicon layer thickness and larger than the mean grain size.

A two-dimensional modeling of the fine-grained polycrystalline silicon thin-film solar cells

A two-dimensional device modeling for polycrystalline silicon thin-film solar cells was performed. A ThRee-dimensional Emitter Based on Locally Enhanced diffusion (TREBLE) concept was applied to evaluate the degree of leveling of the device efficiency. The model assumes a n +-p-p + junction device of a 3-μm grain size and 10-μm-thick polycrystalline Si columnar structure unit, where the n + regions extend along the grain boundaries. The analysis was carried out using ISE-DESSIS, a two-dimensional semiconductor device simulator. It has been found that open circuit voltage could be improved by increasing the base doping level to the optimum value of 10 17 cm −3. Preferential doping has a beneficial effect on short-circuit current of the cell and a slight influence on open circuit voltage. Conversion efficiency of ∼10% could be expected at the surface, recombination velocity of 10 4 cm/s and dopant diffusion depth along the grain boundary x gb=2.5 μm At base doping ∼10 17 cm −3 and well-passivated grain boundaries ( S gb=10 3 cm/s), an efficiency of approximately 12% may be obtained when x gb=3.7 μm.

In situ measurement of the crystallization of amorphous silicon in a vertical furnace using spectroscopic ellipsometry

The in situ measurement of the crystallization of amorphous silicon at 600°C was carried out during the annealing of amorphous silicon-on-oxide samples inside a vertical furnace using spectroscopic ellipsometry. The ellipsometer arrangement was adapted to the furnace geometry using a special beam-guiding system to have a minimum impact on the furnace process performance. Modifications in the furnace geometry were restricted as far as possible, to show that a fast integration in an existing industrial equipment with minor costs can be done. The dielectric function of the changing structure was calculated using the Bruggeman-effective medium approximation (B-EMA). Two optical models were compared. In the first model, the crystallinity was described by the ‘mixture’ of single-crystalline silicon and amorphous silicon. A better fit was obtained using the second model, in which fine-grained polycrystalline silicon and amorphous silicon are used. The high-temperature reference data were obtained by a measurement on the annealed samples at the beginning and at the end of the annealing process. Using this method the crystallization can be monitored by the change of the ratio of amorphous and fine-grained polycrystalline silicon in the best-fit model.

Ellipsometric characterization of oxidized porous silicon layer structures

Electrochemically prepared porous silicon (PS) layers were oxidized thermally and investigated by spectroscopic ellipsometry (SE). The SE spectra were measured in the range of 270–850 nm with a rotating polarizer ellipsometer. The PS was modelled as a mixture of void and crystalline silicon or fine-grained polycrystalline silicon with enhanced absorption due to extensive grain-boundary regions, i.e. the complex refractive index of the layer was calculated by Bruggeman effective medium approximation. The dielectric function of the fine-grained polycrystalline silicon was taken from the work published by G.E. Jellison, Jr., M.F. Chisholm, S.M. Gorbatkin, Appl. Phys. Lett. 62 (1993) 3348. The porosity, the layer thickness and the composition of the oxidized PS layers were determined. Oxidation at 900°C was performed after a stabilizing heat treatment at 320°C. The oxidation at 900°C for 10 min generated only a few nm silicon dioxide on single crystalline Si while in the case of PS with 57% porosity nearly complete oxidation was found. For PS with 68% porosity complete oxidation was observed.

Critical issues in ion implantation of silicon below 5 keV: Defects and diffusion

Continued use of ion implantation for doping of silicon integrated circuits will soon require implantation energies below 5 keV in order to form electrical junctions < 50 nm deep. At such low energies, dopant diffusion and formation of extended defects may be modified by both the proximity of the surface and by the large volume concentrations of point defects and dopant atoms that will arise from reduced range straggling. This brief review summarizes our recent experiments measuring defect formation and evolution, as well as transient enhanced diffusion (TED), in silicon implanted with Si + and B + ions < 10 nm deep. The results have demonstrated that {311}-type extended defects are generated from Si implants even within 3 nm of the surface. However, when these defects eventually dissolve, the surface acts as a perfect sink to efficiently annihilate the released interstitials. As a result, the amount of TED measured at epitaxially-grown B markers decreases approximately linearly with decreasing ion energy (at least, for Si + implantation). For low energy B + implants typical doses required for source-drain doping lead to the formation of a fine-grain polycrystalline silicon boride phase during activation annealing.

Damage, defects and diffusion from ultra-low energy (0–5 keV) ion implantation of silicon

Continued use of ion implantation for doping of silicon integrated circuits will soon require implantation energies below 5 keV to form electrical junctions less than 50 nm deep. At such low energies, dopant diffusion and formation of extended defects is modified by both the proximity of the surface and by the large volume concentrations of point defects and dopant atoms that arise from reduced range straggling. This brief review summarizes our recent experiments which measured defect formation and evolution, as well as enhanced diffusion, in silicon implanted with Si + and B + ions at energies as low as 0.5 keV. The results have demonstrated that {311}-type extended defects are generated from Si + implants even within 3 nm of the surface. However, when these defects eventually dissolve, the surface acts as a perfect sink to efficiently annihilate the released interstitials. As a result, the amount of TED from Si + implantation measured by epitaxially-grown B markers decreases approximately linearly with decreasing ion energy. For sub-keV B + implants typical doses currently used for source-drain doping lead to a boron diffusion enhancement of 3–4× despite the proximity of the surface. Enhanced diffusion is also observed from molecular beam-deposited silicon layers containing a high boron concentration. This newly emerged diffusion enhancement mechanism, boron-enhanced-diffusion (BED), is associated with the formation of a fine-grain polycrystalline silicon boride phase in the implanted layer during activation annealing. These investigations of ultra-low energy (ULE) implantation have thus reinforced and validated our understanding of the role of implantation damage in enhancing dopant diffusion in silicon, while simultaneously revealing some important new materials issues which will impact semiconductor processing in coming device generations.

In situ spectroscopic ellipsometric investigation of vacuum annealed and oxidized porous silicon layers

Vacuum annealed and oxidized porous silicon layers (PSL) were investigated by in situ spectroscopic ellipsometry (SE). The nominal porosity of the layers was between 60 and 77% and the nominal thickness was 500 nm. The annealing was performed by direct ohmic heating (R.T.<450°C) in 5×10 −10 Torr vacuum. The oxidation was performed in two steps, the first step at 5×10 −5 Torr, the second at 10 Torr. Two optically different types of silicon compounds, a bulk-type silicon (c-Si) and a fine-grain polycrystalline silicon with enhanced absorption due to extensive grain-boundary regions (p-Si) were mixed with voids in the appropriate ratio to fit the spectra of as-prepared PSL. For the annealed PSL amorphous silicon (a-Si) was needed in conjunction with p-Si. The oxidized PSL could be fitted with a reduced a-Si content. We can interpret the annealing effect as a depassivation process of the inner surfaces of the PSL (a-Si fraction). At the same time, oxidation leads to a repassivation process.

Anomalously high collection probability in thin film polycrystalline silicon solar cells from numerical modeling

The impact of grain size on the average collection probability of fine grained (&amp;lt;10 μm) polycrystalline silicon has been investigated using a numerical device simulator. This model predicts that the collection probability can improve dramatically once the depletion regions around adjacent grain boundaries overlap. Thus, contrary to most analytical models, the collection probability is not monotonic with grain size. The collection probability has a local maximum at a grain size determined by the grain boundary trap state density. The magnitude of the improvement is dependent upon the grain boundary recombination velocity.

Optical lamp recrystallization of plasma-sprayed silicon deposits on different substrates

Fine-grained polycrystalline silicon (Si) was deposited using a plasma torch onto different substrates such as single crystal Si, polycrystalline Si, and ceramic alumina (Al 2O 3). These deposits were then recrystallized by zone-melting using optical (tungsten-halogen) lamps. The grain morphology of the deposits on Si were greatly improved, reaching that of the substrate. Those deposited on Al 2O 3 substrates were self-supporting and their grain size largely improved after recrystallization. Here we report the improvement in bulk and surface properties after zone-melting. The hole mobility, resistivity, and minority carrier diffusion length of the zone-melted samples were 225 cm 2/V-s, 0.3 ohm-cm, and 25 microns respectively. Notes on process-introduced impurities are also given.

Grain-boundary states and hydrogenation of fine-grained polycrystalline silicon films deposited by molecular beams

Abstract The energy distribution of grain-boundary states is determined for polycrystalline silicon films grown under ultra-high vacuum conditions. Conductivity and electron spin resonance measurements on n-type films reveal both exponential bandtails and deep gap states corresponding to disorder-induced gap states and dangling-bond defect levels (D° and D−). The latter are responsible for the pinning of the Fermi level observed at moderate doping. Both experimental techniques agree with a location of the D− level at E C-0.30 eV and the D° level at E C-0.65eV ± 0.05eV. It is shown that a hydrogen-plasma treatment at 500°C reduces the dangling-bond density by an order of magnitude and that it also yields a conduction bandtail twice as steep. The replacement of weak Si-Si bonds by more energetic Si-H bonds would explain the steeper bandtails. This view is supported by absolute measurements by nuclear reaction showing that the hydrogen content exceeds by two orders of magnitude the original dangling-bond density. A spin density as Iow as 5 × 1016cm−3 is measured for the first time in fine-grained polycrystalline silicon. The passivation of Si dangling bonds is accompanied by a shift of the pinned position of the Fermi level from 0.5 to 0.45 eV below E C, which could mean that the remaining deep defects have a slightly different energy distribution.

Picosecond photoconductive response of polycrystalline silicon thin films

Using autocorrelation measurement, we have studied the picosecond photoconductivity of as-deposited, fine-grain polycrystalline silicon thin films. A strong correlation between photoconductive decay time and polycrystalline silicon deposition temperature has been observed. The fastest autocorrelated photoconductive response is 9 ps full width at half maximum, and is obtained from a polycrystalline silicon sample deposited at 590 °C. We estimated the carrier mobility from the peak autocorrelated current, as well as from the average photocurrent of a single gap. These two values show a large difference for samples deposited at relatively high temperatures, which can be explained by imperfect metal-semiconductor ohmic contacts.

Explosive crystallization of amorphous silicon: triggering and propagation

Amorphous silicon may be transformed into crystalline silicon via a self-sustained process driven by the latent heat released upon crystallization. This is called explosive crystallization and is found to occur under rapid-heating conditions such as laser annealing. In this paper we compare different kinds of explosive crystallization, with emphasis on pulsed-laser induced explosive crystallization of ion-implanted amorphous silicon. It is shown that explosive crystallization of amorphous surface layers yields randomly oriented fine-grain polycrystalline silicon while explosive crystallization of amorphous layers buried beneath a crystalline top layer results in epitaxially aligned single-crystal silicon. The results are used to discuss ultra-rapid crystal nucleation and growth.

Leakage current mechanisms in Hydrogen-passivated fine-grain polycrystalline Silicon on insulator MOSFET's

In this paper we identify various sources of leakage current in thin-film silicon on insulator (SOI) MOSFET's made in hydrogen-passivated small-grain polycrystalline silicon. The action of a parasitic bipolar transistor that can amplify the leakage current due to the thermally generated carriers has been confirmed and characterized. A current gain (β) of more than 6 for the parasitic bipolar transistor has been experimentally measured in accumulation-mode devices, in spite of the presence of a large number of defects. This high gain is attributed to the presence of the vertical electric field, which separates the carriers, thus reducing the probability of recombination. The presence of field-enhanced generation is shown to be the cause of the observed increase in the leakage current with positive front- or back-gate bias for p-channel accumulation-mode devices. Reasonable agreement has been obtained between experimental data and theory based on field-enhanced generation due to Poole-Frenkel barrier lowering.

Structural changes produced in silicon by intense 1-μm ps pulses

The numerous bulk and surface structural changes observed in c-Si following melting with 1-μm pulses that range from 4 to 260 ps in duration and fluences from about 0.6 to 2.8 J cm−2 are examined by Nomarski and transmission electron microscope techniques. For melting pulse widths 30 ps or longer, recrystallization from the melt was observed. By contrast, for the shorter pulses (∼7 ps), the steep temperature gradients that accompany the onset of two-photon absorption associated with pulses of this width produce an undercooled melt. Under these conditions, the resolidification velocities are evidently too high to allow epitaxial regrowth from the crystalline substrate and, for the first time, regions of amorphous and large- and fine-grain polycrystalline silicon are observed to form directly on a crystalline underlayer. In addition, alternate stripes of amorphous and crystalline material are produced by these short pulses. These are associated with localized melting, demonstrating that uniform surface melting is not always required before a spatial modulation of the absorbed surface energy can occur.

Solid-phase epitaxial regrowth of fine-grain polycrystalline silicon

Fine-grain polycrystalline silicon has been produced by low-energy pulsed-laser irradiation of copper-implanted amorphous silicon. This fine-grained material can be regrown epitaxially on the (100) substrate using thermal annealing at temperatures ranging from 800°–1000 °C.