scispace - formally typeset
Search or ask a question

Showing papers in "Solid-state Electronics in 1971"


Journal ArticleDOI
S.M. Sze1, D.J. Coleman1, A. Loya1
TL;DR: In this paper, the thermionic emission theory has been used to study the currentvoltage characteristics of a metal-semiconductor-metal (MSM) structure, and the critical voltage at which the minority carrier injection increases rapidly can be varied by varying the semiconductor doping and thickness.
Abstract: The current-voltage characteristics of a metal-semiconductor-metal structure (essentially two metal-semiconductor contacts connected back to back) have been studied based on the thermionic emission theory. When a uniformly doped semiconductor is thin enough that it can be completely depleted before avalanche breakdown occurs, the structure can exhibit many novel transport behaviors. Two outstanding features of the structure are that (1) a wide range of high-level injection of minority carriers can be achieved by varying the barrier heights of the two contacts and (2) the critical voltage at which the minority carrier injection increases rapidly can be varied by varying the semiconductor doping and thickness. Experimental silicon MSM structures of PtSi-Si-PtSi have been made from n-type silicon with doping of 4×1014 cm−3 and thickness of 12 μm. The critical voltage at room temperature is about 30 V. The current increases over five orders of magnitude with only 10 per cent increase of the voltage. The above results and other measurements over wide temperature range do substantiate the theoretical predictions.

395 citations


Journal ArticleDOI
TL;DR: In this article, the specific contact resistance at zero bias, R c, is calculated for metal-Si and metal-GaAs barriers on p-type and n-type samples.
Abstract: The specific contact resistance at zero bias, R c , serves as a measure of the ohmic or rectifying behavior of a metal-semiconductor barrier under operating conditions. It is thus an important design parameter for semiconductor devices. The values of R c have been calculated for Metal-Si and metal-GaAs barriers on p -type and n -type samples. The theoretical calculation is based on the generalized transport study of metal-semiconductor systems. The results, which are presented graphically, show the dependence of R c on temperature over the range 50°K–500°K, the barrier height from 0.2 to 1.0 eV, and the ionized impurity concentration from 10 14 to 10 21 cm −3 . Generally R c decreases exponentially with increasing temperature and with decreasing barrier height. For samples with lower dopings where the thermionic emission dominates, R c is essentially independent of doping; for higher dopings where the tunneling dominates, R c decreases rapidly with increasing doping. The experimental results of R c for various metals on silicon samples are in good agreement with the predictions.

309 citations


Journal ArticleDOI
TL;DR: In this article, the effect of surface generation on the C-t response and Zerbst plot is demonstrated and values for s0 are correlated with the density of fast surface states.
Abstract: The interpretation of lifetimes and surface generation velocities determined from pulsed MIS-C measurements is examined. The effect of surface generation on the C-t response and ‘Zerbst’ plot is demonstrated and values for s0 are correlated with the density of fast surface states. Different kinds of surface dominated responses are analyzed and we discuss how the bulk characteristics can be retrieved. It is also shown that pulsing from inversion significantly lowers s0, but for high densities of surface states this technique does not suffice to eliminate surface effects. Finally, an approximate analysis of the C-t data is shown to agree quite well with the more exact analysis and it is then applied to lifetime maps of Si wafers.

158 citations


Journal ArticleDOI
TL;DR: In this paper, a detailed theoretical treatment is given of the effects of a thermionic velocity boundary condition on diffusion transport of majority carriers and the spatial variation of the majority carrier quasi-Fermi potential (imref) in metal-semiconductor (Schottky) diodes.
Abstract: A detailed theoretical treatment is given of the effects of a thermionic velocity boundary condition on diffusion transport of majority carriers and the spatial variation of the majority carrier quasi-Fermi potential (imref) in metal-semiconductor (Schottky) diodes. Diode parameters such as n(={kT/q[dlnJ/dV]}−1) and Js, the reverse saturation current are calculated for the case of parabolic band bending. These results are expressed in terms of the Dawson function, the solution to the differential equation d D(y) d y = 1 − 2yD(y) . In contrast to recent estimates of ≅ 1·06 for typical conditions, the parameter n is shown to be slightly bias dependent, to have a value of less than 1·01 for typical operating conditions and a maximum value of 1·28 for forward bias ≅ 2kT/q from flat band. A calculation of spatial variation of imref in the same case shows that under moderate forward bias the electron imref is very nearly constant through the depletion region. For reverse bias the imref rapidly deviates from constancy for applied bias in excess of kT/q. Diffusion effects on T-F emission are shown to be very small.

118 citations


Journal ArticleDOI
TL;DR: In this paper, the results of p-n junctions in GaP which emit green electroluminescence (EL) characteristic of recombination at the isoelectronic trap, N are presented.
Abstract: A summary is presented of the results that have been achieved in studies of p-n junctions in GaP which emit green electroluminescence (EL) characteristic of recombination at the isoelectronic trap, N. The junctions are formed by growing epitaxial layers onto GaP crystal substrates. The highest value achieved for the room temperature external quantum efficiency, η, of the green-emitting A line and its phonon replicas, measured under pulsed conditions, is 0·6 per cent at a current density, J, of 80 A·cm−2 using a solution grown substrate and 0·34 per cent at J = 200 A·cm−2 using double layer growth onto a pulled crystal substrate. These values of η are measured with the diode mounted on a header and covered by an epoxy dome to reduce reflection losses. At a d.c. current of 5 mA (J = 10 A·cm−2), η is reduced by a factor of ∼2 from these values. The variation of η with specific dopants and the level of shallow dopants in both p and n layer growth are described. The properties of junctions formed by p-layer growth onto n-substrates (p/n junctions) differ from those of n/p junctions but η increases approximately linearly with the nitrogen concentration N in both p and n layers. Recent experiments with double layer growth onto pulled crystal substrates show that η increases with the level of N in both conductivity regions of the p-n junction. Higher η are generally achieved in p/n junctions and in these junctions η is increased by a factor of 1·5–3 by annealing for 12 hr at 625°C. The integrated green EL varies with bias as exp qV/kT, but the diode current varies as exp (qV/nkT) with 1 < n < 2 so that η increases with I at low levels (J ⪝ 100 A·cm−2) and becomes constant at higher I. The high levels of N needed to generate the EL also significantly absorb the emitted radiation. Some suggestions are offered to modify the distribution of N in the junction structure to minimize absorption effects.

103 citations


Journal ArticleDOI
C.J. Kircher1
TL;DR: The fabrication, metallurgical properties, and electrical characteristics of palladium silicide (Pd 2 Si) contacts to n-type Si have been investigated in this paper, where Pd was found to react readily with Si at 200°C to form a silicide phase.
Abstract: The fabrication, metallurgical properties, and electrical characteristics of palladium silicide (Pd 2 Si) contacts to n -type Si have been investigated. Pd 2 Si/Si contacts are similar in electrical and metallurgical behavior to PtSi/Si contacts, but can be fabricated at much lower temperatures. Pd was found to react readily with Si at 200°C to form a silicide phase which was identified as Pd 2 Si from X-ray diffraction analysis. The electrical resistivity of the silicide is 40 × 10 −6 Ω cm as determined from sheet resistivity measurements. The barrier height at the Pd 2 Si/Si interface was determined from differential capacitance measurements to be 0.745±0.015 V. Current-voltage measurements and activation energy analysis gave barrier heights within this same range. Contact resistance measurements were made on contacts to Si surfaces with phosphorus doping levels of 2×10 20 /cm 3 . Resistance values obtained are comparable to both theoretical predictions and measurements reported on PtSi and Al contacts to Si.

101 citations


Journal ArticleDOI
TL;DR: In this article, the electrical characteristics of Schottky barriers made by evaporating films of various metals (Al, Pb, Ni, Au, Ag, Cu) onto p-type silicon were measured.
Abstract: Measurements have been made of the electrical characteristics of Schottky barriers made by evaporating films of various metals (Al, Pb, Ni, Au, Ag, Cu) onto p-type silicon. The barriers were generally lower than on n-type silicon, and in the case of Au the barrier was so low as to provide an effectively ohmic contact at room temperature. The truly exponential portion of the forward I–V characteristic was restricted to a comparatively small voltage range. Within this range ‘n’ values of about 1.10 were obtained. The reverse characteristics could be explained in terms of generation in the depletion region. The variation of barrier height with metal work-function indicates that the surface-state parameters (density of states and position of neutral level) are essentially the same for p-type as for n-type silicon. This is confirmed by the fact that, for a given metal, the sum of the barrier heights on n-type and p-type silicon is approximately equal to the band-gap.

100 citations


Journal ArticleDOI
TL;DR: The electron thermal emission and capture rates and the electron impact ionization rate of trapped electrons at the sulfur donor centers in the depletion region of reverse biased silicon p-n junctions have been measured by the dark capacitance and current transient methods as mentioned in this paper.
Abstract: The electron thermal emission and capture rates and the electron impact ionization rate of trapped electrons at the sulfur donor centers in the depletion region of reverse biased silicon p-n junctions have been measured by the dark capacitance and current transient methods. Least square fits of the low field data give the following electron thermal emission rates: en0t = 1.64 × 1010(T/300°K)2exp [−276/kT] sec for the neutral center and en−1t = 1.03 × 1012 (T/300°K)2 exp [−528/kT] sec for the singly ionized center where kT is in meV. The thermal activation energies are then 276 and 528 meV respectively. The hole emission rates are much smaller and not determined. The electric field dependences of the thermal emission rates of electrons are considerably smaller than that predicted by the Poole-Frenkel theory applied to the ground state: en0t increased by 1.5 at 130°K from 0.2 to 1.0 × 105 V/cm and en−1t by 3 at 200°K. Electron capture coefficients are obtained from capacitance transient during steady state electron injection into the junction depletion layer either by transistor emitter or by optical generation at the surface next to the junction. The electron capture rate at the doubly charged center, cn−2t, decreases from 5 × 10−7 cm3/sec at 3 × 104 V/cm to 10−7 cm3/sec at 1.0 × 105 V/cm with essentially no temperature dependence. The electron capture rate at the singly ionized centers, cn−1t, obtained at 82°K was about two orders of magnitude smaller than cn−2t but had essentially the same electric field dependence. The electron impact ionization rate of trapped electrons at the neutral centers and its electric field dependences were also determined at 82°K.

95 citations


Journal ArticleDOI
TL;DR: In this article, an approach for describing the density of quantum states in a semiconductor as a function not only of energy but also of impurity concentration is presented. And the density and the Fermi level in n-type silicon are calculated.
Abstract: Utilizing an approach for describing the density of quantum states in a semiconductor as a function not only of energy but also of impurity concentration, we calculate the density of states and the Fermi level in n -type silicon. The calculated densities of states are shown to agree with pertinent experimental results.

73 citations


Journal ArticleDOI
TL;DR: In this article, the d.c. characteristics of the lateral transistor are examined in detail and the theory presented accounts for the observed dependences of collector and base currents on emitter-base bias.
Abstract: The d.c. characteristics of the lateral transistor are examined in detail. The theory presented accounts for the observed dependences of collector and base currents on emitter-base bias. Two-dimensional geometrical effects are included, and various components of base recombination current are discussed separately. For the experimental devices studied in this work, high-level injection in the base was found to be chiefly responsible for the decreasing current gain at high currents.

71 citations


Journal ArticleDOI
TL;DR: In this paper, the authors have calculated the thermoelastic stresses in an elastic plate of finite length bonded to a rigid substrate using the energy minimization method developed by Aleck.
Abstract: The thermoelastic stresses in an elastic plate of finite length bonded to a rigid substrate have been calculated using the energy minimization method developed by Aleck. The strain and displacement fields have been derived and the inclination of the plate surface with respect to the interface has been given for different ratios of coefficients of thermal expansion.

Journal ArticleDOI
Chung-Yu Ting1, Charles Y. Chen1
TL;DR: In this paper, the total resistance of a diffused resistor can be expressed in terms of bulk sheet resistance and three correction terms, namely, contact resistance, current crowding resistance in the contact region, and the contact width correction.
Abstract: The total resistance of a diffused resistor can be expressed in terms of bulk sheet resistance and three correction terms—namely, contact resistance, current crowding resistance in the contact region, and the contact width correction. The latter two terms have been arrived at theoretically. A new test structure has been designed to measure the metal-silicon contact sheet resistivity and the contact resistance of the resistor. The maximum current density in the contact region is also determined experimentally.

Journal ArticleDOI
TL;DR: In this paper, the thickness of the oxide layers which are built up by reaction with dry oxygen, wet oxygen and by a mixture of oxygen and nitrogen in the presence of phosphorus pentoxide, are given as a function of time and temperature.
Abstract: When thin films of silicon nitride are annealed in the presence of oxygen, thin films of silicon dioxide are formed on the nitride layers while silicon nitride is consumed. The thickness of the oxide layers which are built up by reaction with dry oxygen, wet oxygen and by a mixture of oxygen and nitrogen in the presence of phosphorus pentoxide, are given as a function of time and temperature. The results are compared with the thermal oxidation of silicon.

Journal ArticleDOI
TL;DR: In this article, a d.c. tunnel current saturation under negative bias is observed for N -type MOS devices when average oxide thicknesses of d 0 A and non-degenerate (10Ω cm) semiconductors are employed.
Abstract: A d.c. tunnel current saturation under negative bias is observed for N -type MOS devices when average oxide thicknesses of d 0 A and non-degenerate (10Ω cm) semiconductors are employed. A model is proposed that allows the voltage distribution across devices in such a non-equilibrium situation to be calculated. Theoretical tunnel current expressions are derived and a comparison with experimental results made. The dependence of the existence of this saturation on semiconductor doping concentration, oxide thickness, and minority carrier injection is investigated. Consideration of band tunnel currents, and the effects of surface state charge, oxide charge, and work function differences on these currents, is shown to be sufficient to explain much of the observed current-voltage characteristic.

Journal ArticleDOI
A.Y. Cho1, I. Hayashi1
TL;DR: In this article, surface structures were observed during molecular beam epitaxial growth of GaAs in situ in an ultrahigh vacuum high energy electron diffraction (HEED) system.
Abstract: Surface structures were observed during molecular beam epitaxial growth of GaAs in situ in an ultrahigh vacuum high energy electron diffraction (HEED) system. The surface structures were related to the relative populations of gallium and arsenic on the substrate surface when the intensities of gallium and arsenic in the molecular beam were varied during growth. On the GaAs ( 111 ) face a ( 111 )-2 surface structure is interpreted in terms of an arsenic-rich surface condition and a ( 111 )-√ 19 structure is interpreted in terms of a gallium-rich condition. Only the (111)-2 surface structure was observed on the (111) face. The presence of these different surface structures under different growth conditions is interpreted in terms of rearrangements of the unshared bonds of the atoms on the crystal surface. Photoluminescence measurements were used to investigate the radiative recombination characteristics of the films grown with either a ( 111 )-2 or a ( 111 )-√ 19 structure present on the surface. The results show that the near gap photoluminescence intensity from a film grown under the latter condition is much higher than that from the former. A broad photoluminescence peak ∼ 0.1 eV below the band gap energy is observed on all epitaxial layers grown with the ( 111 )-2 structure present on the surface. The peak may be reduced by annealing and may be attributed to centers associated with Ga vacancies.

Journal ArticleDOI
TL;DR: In this paper, the mechanism of threshold switching in thin chalcogenide glass systems is shown to be essentially non-thermal in character by means of a double pulse technique.
Abstract: By means of a ‘rare double pulse technique’, it is shown that the mechanism of threshold switching in thin chalcogenide glass systems is essentially non-thermal in character. The switches were prepared by evaporating layers of approximately 1μ thickness on to polished graphite electrodes, using Te 40 As 35 Ge 7 Si 18 as starting material. Polarization effects and polarization reversals on switching have been observed which are in qualitative harmony with a space charge theory of threshold switching previously advanced, at any rate in the low-temperature range (−78°C to +20°C). A non-polar electronic process is evidently operative above room temperature. The magnitude of the perturbation caused by heat dissipation in these systems is also assessed.

Journal ArticleDOI
TL;DR: In this paper, a theoretical study of the wave-guiding properties of GaAs-Al x Ga 1−x As heterostructure injection lasers has been made and the electromagnetic field distributions and far-field radiation patterns of these devices have been calculated by solving Maxwell's equations for a three-layer waveguide with complex dielectric constants in each region.
Abstract: A theoretical study has been made of the wave-guiding properties of GaAs-Al x Ga 1−x As heterostructure injection lasers. The electromagnetic field distributions and far-field radiation patterns of these devices have been calculated by solving Maxwell's equations for a three-layer wave-guide with complex dielectric constants in each region. The theory also yields threshold currents and external quantum efficiencies as functions of the principal laser parameters—temperature, doping, alloy composition x in the Al x Ga 1−x As regions, active region thickness, and cavity length. This permits detailed comparison of the theoretical results with experiment, and hence further estimates of optimal laser structures. In particular the variation of laser threshold with active region thickness (and the existence of an optimum thickness) in single heterostructures is sucessfully accounted for by the theory. For double heterostructures with the same alloy composition in each passive region, there exists no optimum active region thickness, and hence the lowest thresholds can be obtained by making these structures as thin as possible.

Journal ArticleDOI
TL;DR: The relative dielectric constant ϵ_r of β-Ga_2O_3, in the direction perpendicular to the (100) plane, was found to be 10.2 ± 0.3.
Abstract: The relative dielectric constant ϵ_r of β-Ga_2O_3, in the direction perpendicular to the (100) plane is found to be 10.2 ±0.3. Within the measurement error ϵ_r is the same at 297°K and at 77°K, and is independent of frequency from 5 kHz to 500 kHz.

Journal ArticleDOI
TL;DR: In this article, hot electron transport and emission have been studied for Au-SiO-Au thin film cold cathodes with oxide thicknesses in the range 100-6400 A. The devices were observed to undergo an electro-forming process on application of a d.c. bias.
Abstract: Hot electron transport and emission have been studied for Au-SiO-Au thin film cold cathodes with oxide thicknesses in the range 100–6400 A. The devices were observed to undergo an electro-forming process on application of a d.c. bias, as described previously. Electron attenuation lengths in the gold layer were found to be in the range 1000–3000 A and to increase with applied bias, in strong contrast with tunnel cathode structures. Attenuation lengths in the oxide layer were ∼ 700 A. Electrons were emitted with energies up to the full applied bias and it is estimated that these electrons can acquire excess energies of up to 4·7 eV by excitation in the oxide layer. No correlation was observed between the areas of the devices yielding electroluminescence and electron emission.

Journal ArticleDOI
J.M. Andrews1, F.B. Koch1
TL;DR: In this article, X-ray diffraction was used to identify the orthorhombic (B31) monosilicide phase of the NiSi Schottky barrier diodes.
Abstract: The metallic compound NiSi has been produced by solid-solid metallurgical reaction in both n- and p-type silicon by sputtering Ni onto freshly back-sputtered silicon wafers and sintering for 5 min at 550°C. Identification of the orthorhombic (B31) monosilicide phase was made by X-ray diffraction, utilizing a glancing-angle Debye-Scherrer photograph of the 1300 A thin-film phase on the silicon substrate. NiSi Schottky barrier diodes, formed by this technique through holes in a silica mask and containing diffused guard rings around the periphery of the metallization, yield nearly ideal reverse I–V characteristics. The data have been fitted to a thermionic emission model for current transport, with barrier heights of 0.66 and 0.45 (eV) to n- and p-type silicon, respectively. The nature of the current transport characteristics strongly indicates the presence of a dipolar charge layer at the metal-semiconductor interface caused by quantum mechanical penetration of metallic electrons into the forbidden gap of the semiconductor.

Journal ArticleDOI
J.M. Shannon1
TL;DR: In this article, an analysis of the MOS Transistor with bias between the source and substrate has shown that when the surface is weakly inverted, the silicon space charge capacitance over a wide range of temperature and bias can be obtained from the change in gate voltage required to maintain a constant channel current.
Abstract: An analysis of the MOS Transistor with bias between the source and substrate has shown that when the surface is weakly inverted, the silicon space charge capacitance over a wide range of temperature and bias can be obtained from the change in gate voltage required to maintain a constant channel current. The substrate impurity profile beneath the gate oxide can then be calculated from capacitance-bias measurements. Measurements made on n -channell and p -channel transistors following the growth of a thick gate oxide indicate a segregation coefficient of ⋍ 0·3 and ⋍ 100 for boron and phosphorus respectively.

Journal ArticleDOI
S.T. Hsu1
TL;DR: In this article, two types of bistable noise, burst noise and microplasma noise, have been investigated with particular emphasis on that under reverse-bias operations, and the physical models for them are reviewed and generalized.
Abstract: Two types of bistable noise, burst noise and microplasma noise, have been investigated with particular emphasis on that under reverse-bias operations. The physical models for them are reviewed and generalized, and the distinction between burst noise and microplasma noise is discussed. The experimental results indicate that the mechanism responsible for burst noise in a reverse-biased p - n junction is the same as that in a forward-biased p - n junction. It is also shown that the fluctuation amplitude of the effective potential barrier can be determined by the measurements of the amplitude of burst noise pulses as functions of temperature.


Journal ArticleDOI
TL;DR: In this paper, the relationship between carrier densities at a p-n junction and the junction potential is discussed, and an extension of the normal boundary conditions to account for current flow within the junction space charge region is presented.
Abstract: A discussion is presented of the relationships between carrier densities at a p-n junction and the junction potential. The major purpose of the work is to discuss the relationship between the Fletcher and Misawa boundary conditions. In addition, a discussion is presented of an extension of the normal boundary conditions to account for current flow within the junction space charge region.

Journal ArticleDOI
TL;DR: In this paper, a simple experimental method is presented for determining absolute values of photoionization cross sections by combining steady state and transient measurements of the extrinsic photovoltaic effect.
Abstract: A simple experimental method is presented for determining absolute values of photoionization cross sections by combining steady state and transient measurements of the extrinsic photovoltaic effect. It is shown that, under certain conditions, the short circuit current caused by a two step excitation process via localized impurity levels, is mainly generated in the transition region of the junction. The equationms for the photovoltaic effect reduce then to a simple form and the photoionization cross sections of the impurity levels involved can be expressed in experimental quantities, where a constant factor is also involved, as a function of energy. To determine this factor, the decay of optically excited stored charge carriers was investigated. The concentration of deep impurity levels in the depletion region of the junction together with capacitance measurements was finally calculated from this constant factor. All experiments were performed with zinc and oxygen doped GaP diodes at room temperature. At 1·78 eV the photo ionization cross sections of an impurity level with binding energy 0·9 eV probably caused by oxygen are found to be 1 × 10−16cm2 and 1·5 × 10−16cm2.

Journal ArticleDOI
TL;DR: In this article, the density of quantum states in a semiconductor depends on the impurity concentration, and a pragmatic approach to the problem of describing this dependence was developed and justified.
Abstract: The density of quantum states in a semiconductor depends on the impurity concentration; in this paper, we develop and justify a pragmatic approach to the problem of describing this dependence. The approach combines heretofore separate theories concerning the various contributors to the dependence, utilizing an approximation that enables application of the superposed theories over the range of impurity concentrations of technological interest. To demonstrate the practical utility of the approach, the prediction yielded by its application to p-type gallium arsenide is compared to an experimentally determined density of states. The agreement is satisfactory.

Journal ArticleDOI
TL;DR: Capacitance voltage measurements on interface of pyrolytically deposited n-type silicon dioxide-InAs MOS diodes as function of admittance at room and 77 K temperatures were taken.
Abstract: Capacitance voltage measurements on interface of pyrolytically deposited n-type silicon dioxide- InAs MOS diodes as function of admittance at room and 77 K temperatures

Journal ArticleDOI
TL;DR: In this paper, the orientation dependence of boron in silicon has been studied in both inert and oxidizing ambients at 1100°C and it is concluded that, for (111) material, the diffusion coefficient is independent of the nature of the ambient.
Abstract: The orientation dependence of boron in silicon has been studied in both inert and oxidizing ambients at 1100°C. It is concluded that, for (111) material, the diffusion coefficient is independent of the nature of the ambient. For (100) material diffusion in an inert ambient is the same as (111), and is twice as fast in an oxidizing ambient. Mathematical analysis and experimental evidence suggest that this is unlikely to be due to any difference in the segregation coefficient at the SiSiO 2 interface, or in the oxidation rate. It is proposed that the enhanced diffusion may be due to defects generated at the interface and caused by a mismatch stress or a vacancy unbalance during oxidation. Either the density of these defects is higher in (100) material than in (111), or they are more mobile in the former.


Journal ArticleDOI
TL;DR: In this article, a method of determining the distribution of injected carriers within the pinch-off region of an insulated-gate field effect transistor was described, based on two-dimensional solutions of both Poisson's Equation and the current continuity equation for minority carriers, within a small region adjacent to the drain junction.
Abstract: A method of determining the distribution of injected carriers within the pinch-off region of an insulated-gate field-effect transistor is described. The analysis is based on two-dimensional solutions of both Poisson's Equation and the current continuity equation for minority carriers, within a small region adjacent to the drain junction. The effect of carrier velocity saturation in the pinch-off region, on the carrier density distribution is assessed. Regions of maximum field are determined. The magnitude of the maximum field within the device is shown to be dependent on the concentration gradient of impurities at the edge of the drain diffusion.