Modeling and simulation of emissivity of silicon-related materials and structures

A brief review of the models that have been proposed in the literature to simulate the emissivity of silicon-related materials and structures is presented. The models discussed in this paper include ray tracing, numerical, phenomenological, and semi-quantitative approaches. A semi-empirical model, known as Multi-Rad, based on the matrix method of multilayers is used to evaluate the reflectance, transmittance, and emittance for Si, SiO₂/Si, Si₃N₄/SiO₂/Si/SiO₂/Si₃N₄ (Hotliner), and separation by implantation of oxygen (SIMOX) wafers. The influence of doping concentration and dopant type as well as the effect of the angle of incidence on the radiative properties of silicon is examined. The results of these simulations lead to the following conclusions: (1) at least within the limitations of the Multi-Rad model, near the absorption edge, the radiative properties of Si are not affected significantly by the angle of incidence unless the angle is very steep; (2) at low temperatures, the emissivity of silicon shows complex structure as a function of wavelength; (3) for SiO₂/Si, changes in emissivity are dominated by substrate effects; (4) Hotliner has peak transmittance at 1.25 μm, and its emissivity is almost temperature independent; and (5) SIMOX exhibits significant changes in emissivity in the wavelength range of 1–20 μm.     

Modeling emissivity of rough and textured silicon wafers

A method for calculating the emissivity of Si wafers with planar and nonplanar (such as rough or textured) surface morphologies is described. The technique is similar to that used in modeling of light trapping in solar cells and is also applicable to those cases when the wafer may have thin dielectric or metal layers. A software package is developed that uses this method. This package includes an approach for calculating the refractive index and absorption coefficient as a function of wavelength, for various temperatures and dopant concentrations. We present results for a number of cases to demonstrate the applications of this model.     

Formation of Boron‐doped region using spin‐on dopant: investigation on the impact of metallic impurities

Investigation on the electrical properties of p⁺‐doped regions formed by spin‐on‐dopant (SOD) technique was achieved. Using this technique, boron‐diffused regions were formed on both p‐type and n‐type float zone wafers. Homogeneous sheet resistances were obtained for both types of wafers. Bulk properties were investigated by measuring effective carrier lifetime. An iron contamination was observed after the boron diffusion step and interstitial iron concentrations were deduced from lifetime measurements. More investigations proved that the iron was initially present within the SOD film. A phosphorus diffusion allows to remove this bulk contamination, leading to an improvement of effective lifetime values. Nevertheless, the corresponding emitter saturation current density was estimated on n‐type wafers and presented a high value. It is likely that this poor electrical quality is the consequence of a high iron concentration which remains in the diffused region. Some possibilities are suggested to avoid or to limit this contamination. Copyright © 2008 John Wiley & Sons, Ltd.     

Correlation between photoluminescence data and device performance of p-channel strained-layer materials

The 4-K photoluminescence spectrum and room temperature transconductance for modulation dopedp-type GaAs/(In,Ga)As dual-channel strained-quantum-well field-effect transistors with comparable dopant and 2-D carrier concentrations were studied. All gate sizes were nominally 300 μm wide by 1 μm long. The best sample has a peak normalized extrinsic transconductanceg _moat room temperature of 31 mS/mm and a 4K photoluminescence linewidth of 6 meV. Depending upon the sample,g _movaried from about 0.5 to 31 mS/mm while the 4-K photoluminescence linewidth decreased from 26 to 6 meV. The low-temperature photoluminescence linewidth and room temperature transconductance were correlated. These results indicate that photoluminescence spectroscopy can be used for screening wafers for potential device peformance before processing.     

Rapid thermal processing of silicon wafers with emissivity patterns

Fabrication of devices and circuits on silicon wafers creates patterns in optical properties, particularly the thermal emissivity and absorptivity, that lead to temperature nonuniformity during rapid thermal processing (RTP) by infrared heating methods. The work reported in this paper compares the effect of emissivity test patterns on wafers heated by two RTP methods: (1) a steadystate furnace or (2) arrays of incandescent lamps. Method I was found to yield reduced temperature variability, attributable to smaller temperature differences between the wafer and heat source. The temperature was determined by monitoring test processes involving either the device side or the reverse side of the wafer. These include electrical activiation of implanted dopants after rapid thermal annealing (RTA) or growth of oxide films by rapid thermal oxidation (RTO). Temperature variation data are compared with a model of radiant heating of patterned wafers in RTP systems.     

Insitu incorporation of Al and N andp-n junction diode fabrication in alpha(6H)-SiC thin films

Aluminum and nitrogen have been introduced asp- andn-type dopants, respectively, during chemical vapor deposition (CVD) ofα(6H)-SiC films on 6H-SiC substrates. The atomic concentration of each dopant in the films showed a linear dependence on partial pressure of the dopant source gas. The Al species exhibited ideal behavior based on a dilute solution model. Thus elemental Al and/or a complex containing only one Al atom were the principal species which contributed to the incorporation of this constituent. The incorporation of N was greater than expected from the dilute solution theory which implied interaction between N and Si in the SiC. The relationship between ionized dopant concentration (carrier concentration) and the concentration in each dopant source gas was also linear and parallel to its atomic concentration. The ratios of the carrier concentration to the atomic concentration for Al and N were 0.02 and 0.06, respectively.P-n junction diodes were fabricated which exhibited rectification and reverse leakage currents at 100 V of 0.19, 0.75, and 1.3 ΜA at 298, 523, and 623 K, respectively. The turn-on voltage decreased from 2.2 to 2.1 and 1.9 V with each incremental increase in temperature.     

Full Field Temperature Measurement of Specular Wafers During Rapid Thermal Processing

Many of the processes involved in the creation of semiconductor devices involve high-temperature processing of silicon wafers. The benefits of reduced thermal budget and faster cycle time make rapid thermal processing (RTP) a possible key technology for semiconductor manufacturing. However, the problem of nonuniform wafer temperature has prevented it from further spread among the industry. The first step in developing controls to maintain a uniform wafer temperature is accurate temperature measurement during processing. In this paper, a system was developed to exploit the specular reflectivity of silicon wafers and obtain a measurement of the wafer temperature profile. The spectral reflectivity is determined by measuring the intensity of an incident beam and the beam reflected from the wafer surface. With this measured reflectivity value the spectral-directional wafer emissivity was determined using Kirchhoff's law. The obtained emissivity then was used to calculate the wafer temperature profile from an image obtained with an infrared camera. An experimental study of the transmittance of an undoped silicon calibration wafer at an elevated temperature is also discussed     

Towards thinner and low bowing silicon solar cells: form the boron and aluminum co‐doped back surface field with thinner metallization film

Using thinner wafers can largely reduce the cost of silicon solar cells. One obstacle of using thinner wafers is that few methods can provide good dopant concentration for the back surface field (BSF) and good ohmic contact while generated only in low bowing. In this paper, we have demonstrated the screening–printing B and Al (B/Al) mixture metallization film technique, making use of the screen‐printing technique and the higher solubility of B in silicon to form a B/Al‐BSF. This technique can raise the carrier concentration in the BSF by more than one order of magnitude and reduce the back surface recombination at a low firing temperature (≤800 °C). We have also shown that through the new technique, the metallization paste thickness at the rear could be reduced largely, which however did not degrade the solar cell efficiency. All these efforts are aiming for pushing forward the application of thinner wafers. Copyright © 2011 John Wiley & Sons, Ltd.     

Wafer-level correlations of EL2, dislocation density,and FET saturation current at various processing stages

The neutral deep-donor density [EL2]⁰, and dislocation density,ρ _D, are measured on adjacent, semi-insulating GaAs wafers, grown by both high-pressure (HP) and low-pres-sure (LP) liquid-encapsulated Czochralski (LEC) techniques; also, other nearby wafers from each boule are used for low-noise, field-effect-transistor (FET) fabrication. Dense data maps (at least 3500 points per wafer per parameter) are then visually and math-ematically compared for [EL2]⁰,ρ _D,I _u, I_r, and I_g where the latter three quantities rep-resent the unrecessed-ungated, recessed-ungated, and gated saturation currents, re-spectively, for ion-implanted, 0.5 ]smm × 300 μm FET’s. For theparticular wafers and processing used in this study, the following conclusions can be drawn: (1) onall of the wafers, materials (EL2 andρ _D) non-uniformities are correlated with at least some of theI _u non-uniformities; (2) onsome of the wafers, materials non-uniformities follow all the way through toI _g, but on others, the gate-recess step itself introduces much stronger non-uniformities; (3) the HP-LEC wafers give slightly higherI _u’s than the LP-LEC waf-ers; and (4) [EL2]0 is a better predictor ofI _u than isρ _D.     

The dopant density and temperature dependence of electron mobility and resistivity in n-type silicon

Traditional analysis of electron mobility in n-type silicon neglects the effect of electron-electron scattering in the mobility calculations. As a result, theory fails to conform with experiment when dopant density exceeds 2 × 10¹⁶ cm^?3. In this work, an improved theoretical model for computing mobility and resistivity as functions of dopant density and temperature has been developed for n-type silicon. The model has been applied to phosphorus-doped silicon for dopant densities from 10¹³ to 10¹⁹ cm^?3, and temperatures between 100 and 500 K. The mobility was calculated analytically by appropriately combining lattice, ionized impurity and neutral impurity scattering contributions. The effect of electron-electron scattering was incorporated empirically for dopant densities greater than 2 × 10¹⁶ cm^?3. Additionally, the anisotropic scattering effect was included in the mobility formulations. Resistivity measurements on seven phosphorus-doped silicon wafers with dopant densities from 1.2 × 10¹⁴ to 2.5 × 10¹⁸ cm^?3 were carried out for temperatures from 100 to 500 K. Electron mobility at 300 K was deduced from resistivity and junction C-V measurements for dopant densities from 10¹⁴ to 10¹⁸ cm^?3. Agreement between theoretical calculations and experimental data for both electron mobility and resistivity of phosphorus-doped silicon was within ±7% in the range of dopant densities and temperatures studied.     

Radiation thermometry of silicon wafers in a diffusion furnace forfabrication of LSI

A radiation thermometry technique suitable for measuring the temperature of silicon wafers in a diffusion furnace has been developed. A principal feature of this technique is that it measures the temperature of wafers that are not in the line of sight of a conventional pyrometer. An optical guide, consisting of two quartz prisms, gives optical access to interior wafers in the load. A measuring wavelength of 0.9 μm is selected since a silicon wafer is opaque and its emissivity does not depend on temperature at this wavelength. The accuracy of the thermometry is examined by comparing the measured value of the pyrometer with that of a thermocouple. The two measured values agree within ±2°C in a steady state. When wafers are being inserted into or drawn out from the furnace, however, an error is caused by the veiling glare at the optical guide and the wafer     

Activation level in boron-doped thin germanium-on-insulator (GeOI): Extraction method and impact of mobility

We hereby present a non-destructive method for extracting the activation level on boron-doped germanium-on-insulator (GeOI) wafers, with a discussion on the impact of the hole mobility model. This method combines Monte Carlo boron profile simulations with optical Ge layer thickness T_Ge and electrical sheet resistance R_sh measurements. As B atoms are known not to diffuse in Ge for the usual activation temperatures (<800 °C), we can assume that the as-implanted dopant profile remains unchanged after annealing (no modelling of boron diffusion required). We highlight that the knowledge of the hole mobility dependence on activated impurities concentration in Ge is of paramount importance. Several experimental and theoretical models are available in the literature. After relative validity assessments, all of them have been implemented for extraction and unfortunately yield different values scattered over nearly one decade. Still, the lower-bound concentration 2.7×10¹⁹ cm⁻³ is in the range of the state-of-the-art values for B-implanted crystalline Ge and has proven suitable for functional GeOI pMOSFET demonstration.     

Wafer temperature measurement in a rapid thermal processor with modulated lamp power

Pyrometry methods utilizing modulated lamp power (“ripple”) were used to improve wafer temperature measurement and control in rapid thermal processing (RTP) for silicon integrated circuit production. Data from a manufacturing line where ripple pyrometers have been tested show significantly reduced wafer to wafer and lot to lot variations in final test electrical measurements and increased yields of good chips per wafer. The pyrometers, an outgrowth of Accufiber’s ripple technique, are used to compensate for ordinary production variations in the emissivities of the backsides of wafers, which face the pyrometers. Power to the heating lamps is modulated with oscillatory functions of time at either the power line frequency or under software control. Fluctuating and quasi-steady components in detected radiation are analyzed to suppress background reflections from the lamps and to correct for effective wafer emissivity. Sheet resistances of annealed wafers with high dose shallow As implants were used to infer temperature measurement capability over a range in backside emissivity. Emissivities are varied when depositing or growing one or more layers of silicon dioxide, silicon nitride, or polycrystalline silicon on the backsides of the wafers.     

Photothermal Activation of Shallow Dopants Implanted in Silicon

Dopant impurities were implanted at high dose and low energy (10¹⁵ cm⁻², 0.5–2.2 keV) into double-side polished 200 mm diameter silicon wafers and electrically activated to form p–n junctions by 10 s anneals at temperatures of 1,025, 1,050, and 1,075°C by optical heating with tungsten incandescent lamps. Activation was studied for P, As, B, and BF₂ species implanted on one wafer side and for P and BF₂ implanted on both sides of the wafer. Measurements included electrical sheet resistance (Rs) and oxide film thickness. A heavily boron-doped wafer, which is optically opaque, was used as a hot shield to prevent direct exposure to lamp radiation on the adjacent side of the test wafer. Two wafers with opposing orientations with respect to the shield wafer were annealed for comparison of exposure to, or shielding from, direct lamp illumination. Differences in sheet resistance for the two wafer orientations ranged from 4% to 60%. n-Type dopants implanted in p-type wafers yielded higher Rs when the implanted surface was exposed to the lamps, as though the effective temperature had been reduced. p-Type dopants implanted in n-type wafers yielded lower Rs when the implanted surface was exposed to the lamps, as though the effective temperature had been increased. Effective temperature differences larger than 5°C, which were observed for the P, B, and BF₂ implants, exceeded experimental uncertainty in temperature control.     

Quantitative oxygen measurements in OMVPE Al x Ga1−x As grown by methyl precursors

Oxygen has always been considered to be a major contaminant in the organo-metallic vapor phase epitaxy (OMVPE) of Al _x Ga_1−x As. Oxygen incorporation has been invoked as a contributor to low luminescence efficiency, dopant compensation and degradation of surface morphology among other deleterious effects. This study presents quantitative measurements of oxygen concentration in nominally high purity Al _x Ga_1−x As. The oxygen concentration was measured as a function of alloy composition, growth temperature, andV/III ratio. Quantitative secondary ion mass spectroscopy (SIMS) measurements were used to determine the oxygen content as well as the carbon concentration in the film. The oxygen concentration increases with decreased growth temperature and V/III ratio while increasing superlinearly with Al content in the epitaxial layer.     

Doping of trench capacitors by rapid thermal diffusion

A new rapid thermal diffusion process for shallow, heavily doped trench junctions in high density dynamic RAMs is described. Planar dopant sources are formed by spin-coating rigid substrates, such as silicon wafers or solid dopant sources, with liquid dopants. Diffusion takes place at high temperatures when the source, placed in proximity to the silicon wafer, releases dopant via evaporation followed by diffusion to the silicon surface. Well-controlled, heavily doped shallow junctions are readily obtained for B, P, and As. The doping process is shown to provide uniform doping of high-aspect-ratio trenches. Process control is achieved by controlling the wafer temperature and duration of the process. Junction depths near 0.1 μm have been demonstrated over the entire surface of trenches 0.7 μm in diameter and 6 μm in depth     

Annealing kinetics of boron-containing centers in electron-irradiated silicon

The annealing kinetics of B _i O _i pairs created by fast-electron irradiation in Si wafers is studied. The wafers are grown by the Czochralski method and doped with boron to different levels. It is found that, at a particular temperature, the annealing rate steadily increases with increasing boron concentration. The results are described with a simple model that takes into consideration the interaction of interstitial boron atoms with oxygen atoms and substitutional boron atoms. In the context of the model, the temperature dependence of the dissociation rate of the B _i O _i complex is calculated.     

Tin-doping ofn + InP OMVPE layers

Tin as a donor dopant in OMVPE epilayers of InP has been studied for its dependence on growth parameters to assess the nature of the dopant incorporation from a tetraethyltin (TESn) source and to establish reproducible conditions for tailoring carrier concentrations. Free carrier concentrations depend linearly on Sn/In ratio and are independent of PH₃ concentration consistent with the impurity incorporation on In-lattice sites. The carrier concentration saturates at 3 × 10¹⁹ cm³ and is accompanied by excess Sn accumulation on the layer surface. X-ray diffraction shows an expansion of the lattice proportional both to the measured free carrier concentration and to the TESn gas concentration up to solid saturation. The lattice expansion is larger than expected from Sn for In radius substitution. SIMS profiles for abrupt turn-on and turn-off of the dopant source show transient changes of Sn concentration consistent with a Sn surface layer buildup.     

Electrical Properties of Halogen-Doped CdTe Layers on Si Substrates Grown by Metalorganic Vapor-Phase Epitaxy

Electrical properties of halogen-doped CdTe layers grown on Si substrates using iodine and chlorine dopants are presented. No change in electrical properties of the layers was observed with chlorine as a dopant. However, doping with iodine resulted in highly conductive n-type layers or highly resistive p-type layers depending upon the growth conditions, even though a similar amount of dopant was introduced into the growth chamber. Layers grown at 560°C, with a vapor-phase Te/Cd precursor ratio of 3.0, were p-type. The resistivity of the layers remained unchanged for low dopant supply rates, but increased abruptly when the dopant supply rate was increased beyond a certain value. On the other hand, layers grown at 325°C with Te/Cd ratios from 0.1 to 0.25 were n-type. A maximum free electron concentration of 1.3 × 10¹⁷ cm⁻³ was obtained at room temperature. The types and conductivities of the grown layers were strongly dependent on the growth conditions.     

Influence of precursor layer ablation on laser doping of silicon

This paper presents a numerical model, which quantitatively demonstrates that ablation and partial recondensation of the dopant precursor layer are some of the dominating physical processes in laser doping (LD) of crystalline silicon. Our pulsed LD process uses a line focused laser beam, enabling the creation of solar cell emitters without the generation of dislocations, if the width w of the short axis of the line focus is w < 10 μm. The concentration profiles of the dopant atoms strongly depend on the pulse energy density E_p, the pulse to pulse separation Δx and the number of laser scans N_s. By comparing measured with modeled concentration profiles, we are able to evaluate the ablation width as well as the amount of the ablated precursor layer. In case of a sputtered phosphorus precursor layer, the ablation width w_a is w_a = 6 μm, whereas the width of the molten silicon layer w_m is w_m = 5 μm. The model also explains the dependence of experimental dopant concentration profiles on the number of subsequent laser scans N_s and pulse to pulse separation Δ_x. Copyright © 2010 John Wiley & Sons, Ltd.