Be-implanted GaAs Research Articles

Improved Precision in a Resonance Ionization Mass Spectrometer by the Use of Stark-Shifted Spectral Lines as a Probe for Extraction Field

High-lying (Rydberg) electronic levels of sputtered Na atoms are spectrally shifted by the strong ion extracting electric field (⩾12 kV/cm) present in a resonance ionization mass spectrometer (Stark effect). The Stark-shifted lines are a beneficial diagnostic to probe the ion source's local electric field, which is related to sample alignment and hence ion transmission efficiency. Changes in electric field of 1% can be detected spectrally, so the sample position is controlled to ⩽ 10 μm by this technique. Results show that by the use of the Na spectrum as an alignment aid prior to sputter depth profiling, the precision of the RIMS instrument, as measured by the reproducibility of either the peak of the integrated signals from Be-implanted GaAs, is improved to about ±3%—at least a fourfold improvement over the best previous results. Spatial variations in the Be-ion implant dose over a GaAs wafer may be detected because of improved precision.

Diffusion and deactivation characteristics in Be-implanted GaAs

In this work, we have observed for the first time that irreversible changes in the sheet hole concentration of the rapid thermally annealed Be-implanted GaAs samples occurred during subsequent heat treatments. The hole concentration is found to decrease with annealing time. Also, the rate of decrease in the hole concentration increases with increasing annealing temperature. By determining the initial rate of change of the carrier concentration during subsequent heat treatments, an activation energy of around 1 eV has been identified. It is concluded that the rate dependent process for this irreversible reaction is the out-diffusion of beryllium atoms. The electrical profiles of the Be implanted samples have also been modified during the annealing processes. These modifications in the electrical properties could lead to deleterious effects to transistor operations if Be is used as a dopant.

Effect of F co-implant during annealing of Be-implanted GaAs

F+ co-implantation at different doses and energies was performed into GaAs already implanted with Be+ at high dose (1015 cm−2) and low energy (20 keV), in order to reduce the beryllium diffusion during post-implant annealing. The redistribution behavior of Be and associated electrical effects were studied by secondary-ion mass spectrometry, transmission electron microscopy (TEM), Hall effect measurements, and current-voltage profiling. Be outdiffusion was reduced by co-implantation of F; more than 80% of the implanted Be was retained during rapid thermal annealing up to 850 °C. The dose and energy of the F implant strongly influenced Be electrical activation efficiency. High activation, up to 48.5%, was obtained when F was co-implanted at high dose (1015 cm−2) and low energy (10 keV). Hole profiles shown reduced electrical activation in the region where F and Be profiles overlapped and TEM studies indicated the formation of {111} coherent plates, possibly BeF2 precipitates, in the same region. The reduction of Be outdiffusion in F co-implanted samples led to high activation after annealing, and was believed to be due to chemical interaction between Be and F.

Influence of rapid thermal annealing temperature on the electrical properties of Be-implanted GaAs p-n junctions

Planar, Be-implanted p-n junctions were fabricated in GaAs with rapid thermal annealing (RTA). Five second isochronal anneals over a temperature range of 600–1000 °C were studied with secondary ion mass spectrometry (SIMS), sheet resistance measurements, and variable diameter p-n junctions. Sheet resistance measurements indicate that a minimum RTA temperature of 600 °C is necessary for electrical activation of the implanted Be. SIMS analysis indicates that significant outdiffusion and surface evaporation of Be occur at all RTA temperatures in this range, while indiffusion of Be is insignificant for concentrations below 1×1018 cm−3. Forward bias current in diodes ranging in diameter from 10 to 1000 μm is dominated by surface recombination, rather than bulk space-charge recombination, over the entire 600–1000 °C temperature range. The magnitude of the surface recombination current is insensitive to the RTA temperature, which suggests that 600 °C RTA should be sufficient for the formation of satisfactory p-n junctions.

Raman scattering evaluation of lattice damage and electrical activity in Be-implanted GaAs

Systematic Raman scattering experiments were carried out on [001] GaAs implanted with 70-keV Be+ ions both after implantation for samples subjected to fluences spanning the 1013–1015 cm−2 range and after successive isochronal thermal annealing steps performed between 200 and 900 °C for a selected series of Be doses. The implantation-induced damages were observed via the modifications with increasing fluence, of the first-order Raman line parameters (mainly for the LO peak) classified in terms of mode energy softening and linewidth increase. These modifications were quantitatively accounted for within the spatial coherence-length reduction formalism, which makes it possible to evaluate the mean size of unperturbed regions in as-implanted GaAs. Upon successive thermal treatments, the implanted layers were found to recover their crystalline perfection at a temperature as low as 500 °C. Above this temperature the Be electrical activation is evidenced by the sudden and drastic modification of only the ‘‘longitudinal’’ response, which is interpreted as due to the coupling between the LO phonon and an overdamped plasmon. The detailed analysis of the coupled-mode behavior is presented in the case of p-doped GaAs. From the Raman line-shape analysis, including a near-surface free-carrier depleted zone, we have obtained information about the electrical properties of the layer in terms of carrier density and mobility. It is further shown that in the case of strong damping, the ratio ω2p/Γp, where ωp and Γp are, respectively, the plasmon frequency and damping, is a pertinent quantity since it is proportional to the electrical conductivity. All the reported Raman conclusions agree well with results found in the literature that were obtained from direct electrical techniques. Possible mechanisms involved in the electrical activation of Be in GaAs are briefly discussed.

Dominance of surface recombination current in planar, Be-implanted GaAs p-n junctions prepared by rapid thermal annealing

Rapid thermal annealing with incoherent light was used to fabricate planar, Be-implanted p-n junctions in GaAs. Diodes of varying diameter were used to determine if the residual implant damage would cause the space-charge recombination current to dominate the surface recombination current. Unpassivated diodes are dominated by surface recombination over the 20–150 μm range of diameters investigated. Passivation of diode structures with a surface layer of high-resistivity Al0.4Ga0.6As grown by metalorganic chemical vapor deposition resulted in a significant reduction of surface recombination current and permitted the measurement of the space-charge recombination current. The space-charge recombination current for these diodes was found to be similar in value to that previously reported for Zn-diffused GaAs junctions.

Electrical characteristics of Be-implanted GaAs activated by rapid thermal annealing

Be-implanted GaAs are annealed by rapid thermal annealing (RTA) using halogen lamps. Electrical properties of the annealed GaAs are investigated, emphasizing those at 77K for application to the p+-layer of Be-implanted WSi <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</inf> -gate self-aligned two-dimensional hole gas (2- DHG) FET. An electrical activation of 90 percent (for 2.0 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">13</sup> cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup> ) or 80 percent (for 2.2 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">14</sup> cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup> ) is obtained. An annealing temperature dependence of carrier freezing at 77K is observed for higher dose samples. The phenomenon is attributed to the redistribution of impurity atoms near the high-concentration peak.

Solid phase regrowth of low temperature Be-implanted GaAs

Infrared reflection and transmission measurements were used to study the thermally induced regrowth of (100) oriented, Be-implanted GaAs samples. The samples used in this study were implanted at low temperature (−100 °C) with 250-keV Be ions to a fluence of 6×1015 cm−2. The samples were postannealed at temperatures ranging from 100 to 550 °C. Isochronal and isothermal annealing at a series of temperatures between 180 and 240 °C were performed. Infrared reflection spectra were analyzed by using a three or four layer dielectric model. Analysis of the annealing data suggests that an amorphous layer first anneals to a second metastable amorphous state and then becomes a damaged crystalline layer after annealing at 220 °C for 12 h. The observed regrowth is not by a simple epitaxial process. After annealing at 400 °C for 1 h, the damage in the layer is reduced sufficiently for the refractive index to recover almost to the preimplantation value. On annealing at 450 °C free carriers are observed. From the measured average regrowth rate for the amorphous layer at various anneal temperature, an effective activation energy is estimated to be about 1.45 eV. This compares with activation energies of 2.3 eV for Si and 2.0 eV for Ge.

Spectroscopic ellipsometry and Raman scattering study of the annealing behavior of Be-implanted GaAs

The annealing of the lattice damage induced by Be-ion implantation in GaAs has been studied by spectroscopic ellipsometry and Raman scattering after each step of an isochronal thermal treatment. These two optical (i.e., nondestructive) techniques are shown to be very sensitive both to the lattice recovery and to the electrical activation of the implants. It has been observed that the latter process occurs after the lattice perfection recovery at T=550 °C, while for the ultimate annealing temperatures (T&amp;gt;750 °C) a significant Be redistribution takes place resulting in a decrease of the electrical activity which is confirmed by differential Hall effect measurements.

Electrical characteristics of Be-implanted GaAs diodes annealed with an ultrahigh power argon arc lamp

The potential of arc lamp annealing techniques in GaAs device processing is demonstrated by the fabrication of Be-implanted mesa pin diodes. Implants were done at 50 and 120 keV with doses of 4.4×1014 and 5.1×1014 cm−2, respectively (total dose =9.5×1014 cm−2) into a 14-μm-thick undoped (ND−NA≊7.5×1014 cm−3) GaAs epitaxial layer grown by vapor phase epitaxy. Ten-second annealing cycles with peak temperatures of 950° and 1050 °C have been studied. The electrical characteristics of these diodes are superior to published furnace-annealed, Be-implanted GaAs diodes.

Low temperature annealing of Be-implanted GaAs

Infrared reflection and transmission measurements of GaAs implanted with large fluences of Be ions have been made as a function of post-implantation annealing temperature. Annealing up to 400 °C resulted in decreases in the lattice disorder responsible for changes in the dielectric constant. This decrease appears to take place uniformly throughout the disordered layer rather than by an epitaxial process. Annealing at 400 °C for 2 h returns the dielectric constant to essentially the preimplantation value with no significant carrier activation being observed. Carrier activation is observed only after prolonged annealing at 400 °C, i.e., ≳50 h. Annealing for 1 h at 450 °C also produces measurable carrier activation. These results are in general accord with prior electrical data and transmission electron microscope measurements and suggest that the removal of the disorder-induced changes in the dielectric constant and the activation of free carriers might involve different annealing processes.

Application of thermal pulse annealing to ion-implanted GaAlAs/GaAs heterojunction bipolar transistors

Post-implant anneals of short (10-15 s) duration are compared with extended (10 min) furnace anneals for the fabrication of Be-implanted MBE-grown GaAlAs/GaAs heterojunction bipolar transistors. It is shown that the thermal pulse anneals can lead to improved Be-implant electrical activation in Ga 0.7 Al 0.3 As and GaAs, to improved contact resistance to Be-implanted GaAs, and to improved transistor characteristics. The transistor improvement is inferred to be due to a reduction in diffusion of the grown-in dopant profile.

Optical studies of Be-implanted GaAs

Infrared reflection and transmission measurements of Be-implanted GaAs have been made as functions of Be-ion fluence and annealing temperatures from 500 to 800 °C. Semi-insulating material was implanted with 280-keV ions with fluences of 1×1015, 3×1015, 6×1015, and 10×1015 ions/cm2. Hole carrier concentrations calculated from the optical measurements as functions of fluence and annealing temperature compare favorably with those determined from sheet Hall-effect data. The absorption cross section used was obtained from measurements of acceptor-doped GaAs with concentrations similar to those of the implanted and annealed samples. After annealing at 500 °C a free-hole plasma layer contribution to the electric susceptibility was observed by measurements of reflection interference effects. Except for the highest-fluence case the disorder effect on the dielectric constant was largely removed by this anneal. For the highest-fluence case the annealing results were more complicated. There are regions which still have significantly different optical properties due to residual disorder. The present results are consistent with redistribution effects produced by thermal aging at the higher temperatures.

Reduced lateral diffusion and reverse leakage in Be-implanted GaAs 1− xP x diodes : P. K. Chatterjee and B. G. Streetman Solid State Electronics,2, p.305 (1977)

Reduced lateral diffusion and reverse leakage in Be-implanted GaAs 1− xP x diodes : P. K. Chatterjee and B. G. Streetman Solid State Electronics,2, p.305 (1977)

Annealing studies of Be-implanted GaAs0.6P0.4

Differential resistivity and Hall effect measurements and secondary ion mass spectrometry (SIMS) are used to study the annealing behavior of Be-implanted GaAs0.6P0.4. Results similar to that previously reported for Be-implanted GaAs are observed, including outdiffusion of Be into the Si3N4 encapsulant during 900‡C annealing of high dose implants. Nearly all (85–100%) of the Be remaining after a 900‡C, 1/2 hr anneal is electrically active. However, the electrical activation at low annealing temperatures (600–700‡C) is much lower in GaAs0.6P0.4 than in GaAs. A substantial amount of diffusion is observed even for the low fluence Be implants in GaAs0.6P0.4 annealed at 900‡C, indicating a greater dependence of the diffusion on defect-related effects in the ternary.

Diffusion studies of Be-implanted GaAs by SIMS and electrical profiling

Secondary ion mass spectrometry (SIMS) has been used with differential resistivity and Hall effect measurements to study the 900°C diffusion of implanted Be in GaAs. Some outdiffusion of Be into the Si3N4 encapsulant occurs for surface Be concentrations above 1 × 1018cm−3. However, excellent agreement between the electrical and atomic profiles indicates that 85–100% of the Be remaining after annealing is electrically active. The concentration-dependent diffusion observed for implanted Be in GaAs was not significantly altered in experiments using hot substrate implants, two-step anneals, or annealing with Ga and As overpressure.

Diffusion studies of Be-implanted GaAs by SIMS and electrical profiling : W. V. McLevige, K. V. Vaidyanathan, B. G. Streetman, J. Comas and L. Plew. Solid St. Commun.25, 1003 (1978)

Diffusion studies of Be-implanted GaAs by SIMS and electrical profiling : W. V. McLevige, K. V. Vaidyanathan, B. G. Streetman, J. Comas and L. Plew. Solid St. Commun.25, 1003 (1978)

Electrical profiling and optical activation studies of Be-implanted GaAs

Differential resistivity and Hall-effect measurements have been utilized to study the annealing behavior and electrical carrier-distribution profiles of Be-implanted GaAs. A maximum of 90–100% electrical activation occurs during 900 °C anneals for implanted Be concentrations less than ∼5×1018 cm−3. For higher fluences, however, a heavily concentration-dependent diffusion is observed, and the measured electrical activation is complicated by outdiffusion of Be into the Si3N4 encapsulant. In these cases, a maximum in the electrical activation appears for annealing near 700 °C. Low-temperature (5 °K) photoluminescence substantiates previous findings that 900 °C annealing results in maximum optical activation and lattice recovery.

Reduced lateral diffusion and reverse leakage in Be-implanted GaAs 1−xP xdiodes

Lateral Zn diffusion under the Si 3N 4 mask encountered in standard red GaAs 0.6P 0.4 light emitting diodes can be eliminated in Be-implanted junctions. Typical reverse leakage current for implanted devices is ∼ 5nA compared with ∼ 250nA for Zn-diffused diodes.

Versatile double AC Hall effect system for profiling impurities in semiconductors

A versatile system is described for measuring electrically active impurity distributions in semiconductors. The system uses an improved double AC Hall system for sensitive measurement of the Hall effect in conjunction with chemical etching for removal of successive layers. All processing is done with samples mounted on standardized discs, increasing the convenience and speed of the technique for repetitive measurements. A measured profile of Be-implanted GaAs obtained with this system is presented as an example.