Journal of Crystal Growth

Published by Elsevier
Print ISSN: 0022-0248
Bimetallic nanoparticles of Au-Pd find important applications in catalysis. Their catalytic performance is directly related to the structure, alloy formation and variation of composition in the structure. A standard idea is that bimetallic nanoparticles can be either an alloy or a core shell structure. Our group has investigated the structure and composition of Pd-Au nanoparticles by using aberration corrected high angle annular dark field scanning transmission electron microscopy (HAADF-STEM). We reported previously that the nanoparticles are composed of an evenly alloyed inner core, an Au-rich intermediate layer, and a Pd-rich outer shell. The structure is more complicated than what simple models can predict. In this paper we report additional studies of this system wherein by carrying out spectral and chemical analysis (STEM*-EDAX, STEM-EELS) the interface structure can now be better identified and understood. Apart from the three-layered core-shell structures we have also been able to observe in some cases a four-layered core-shell structure as well. The entire core-shell structure is not rigid and there is indeed intercalation of Au-Pd into the other layers as well. In addition we have been able to locate stacking faults present in the nanoparticles. We also address the problem of the interface structure between the layers. By using nanodiffraction we have found that the whole structure of the nanoparticles becomes hcp in contrast to the bulk structure of Au or Pd.
The nucleation processes involved in calcium phosphate formation in tooth enamel are not well understood but are believed to involve proteins in the extracellular matrix. The ability of one enamel protein, amelogenin, to promote the nucleation and growth of calcium phosphate was studied in an in vitro system involving metastable supersaturated solutions. It was found that recombinant amelogenin (rM179 and rp(H)M180) promoted the nucleation of calcium phosphate compared to solutions without protein. The amount of calcium phosphate increased with increasing supersaturation of the solutions and increasing protein concentrations up to 6.5 μg/mL. At higher protein concentrations, the amount of calcium phosphate decreased. The kinetics of nucleation was studied in situ and in real time using a quartz crystal microbalance (QCM) and showed that the protein reduced the induction time for nucleation compared to solutions without protein. This work shows a nucleation role for amelogenin in vitro which may be promoted by the association of amelogenin into nanosphere templates, exposing charged functionality at the surface.
Ferromagnetic Ge(1-x)Mn(x)Te is a promising candidate for diluted magnetic semiconductors because solid solutions exist over a wide range of compositions up to x(Mn)≈0.5, where a maximum in the total magnetization occurs. In this work, a systematic study of molecular beam epitaxy of GeMnTe on (1 1 1) BaF(2) substrates is presented, in which the Mn concentration as well as growth conditions were varied over a wide range. The results demonstrate that single phase growth of GeMnTe can be achieved only in a narrow window of growth conditions, whereas at low as well as high temperatures secondary phases or even phase separation occurs. The formation of secondary phases strongly reduces the layer magnetization as well as the Curie temperatures. Under optimized conditions, single phase GeMnTe layers are obtained with Curie temperatures as high as 200 K for Mn concentrations close to the solubility limit of x(Mn)=50%.
Calcium phosphate crystalline powders grown under terrestrial and space (EURECA 1992-1993 flight) conditions in the Solution Growth Facility are analyzed and compared by optical and electron microscopy (scanning and transmission), electron and X-ray microdiffraction and microanalyses. On earth, only small, micrometer size scale, spherolites of hydroxyapatite (HAP) grow. In space, the HAP spherolites reach hundreds of micrometer. Also, octacalcium phosphate (OCP) spherolites up to 3 mm have been obtained. Computer modelling of diffusion in a real chamber has been performed. It suggests high spatial supersaturation gradients at zero gravity which may provide much higher local supersaturations on earth, where convection takes place. The analyses suggest that the dramatic difference between the terrestrial and space samples should come from much lower supersaturation in space.
Epitaxial, graphitic carbon thin films were directly grown on C-face/ (0001̄) SiC and (0001) sapphire by chemical vapor deposition (CVD), using propane as a carbon source and without any catalytic metal on the substrate surface. Raman spectroscopy shows the signature of multilayer graphene/graphite growth on both the SiC and sapphire. Raman 2D-peaks have Lorentzian lineshapes with FWHM of ~60 cm(-1) and the ratio of the D-peak to G-peak intensity (I(D)/I(G)) linearly decreases (down to 0.06) as growth temperature is increased. The epitaxial relationship between film and substrates were determined by x-ray diffraction. On both substrates, graphitic layers are oriented parallel to the substrate, but exhibit significant rotational disorder about the surface normal, and predominantly rhombohedral stacking. Film thicknesses were determined to be a function of growth time, growth temperature, and propane flow rate.
Schematics of pattern creation and nanowire growth. (a) Si substrate with the oxide layer. (b) Openings in the PMMA after EBL exposure and development. (c) Substrate with openings in the oxide layer after the PMMA is removed. (d) Position controlled InAs nanowires after growth.  
Demonstration of well controlled nanowire growth with 100% yield. The wires are 3.38 in length and 231±36 nm in diameter. As growth parameters a growth temperature of 550 °C, for TMI and AsH 3 molar fractions of 3.9×10 −6 and 2.8×10 −4 , respectively, and a growth time of 4 min has been used. The sample is tilted 20° in the SEM image.  
Position controlled growth of InAs nanowires on a Si (111) substrate. The opening sizes in these images are (a) 180 nm, (b) 165 nm, (c) 140 nm, (d) 130 nm. The sample is tilted by 30° in the image.  
SEM image of InAs growth following a 1 s (a), and a 16 s (b) indium deposition prior to growth, leading to large crystallites for 16 s indium pre deposition. The samples are seen under 45° tilt.
In this work, the nucleation and growth of InAs nanowires on patterned SiO(2)/Si(111) substrates is studied. It is found that the nanowire yield is strongly dependent on the size of the etched holes in the SiO(2), where openings smaller than 180 nm lead to a substantial decrease in nucleation yield, while openings larger than ≈500nm promote nucleation of crystallites rather than nanowires. We propose that this is a result of indium particle formation prior to nanowire growth, where the size of the indium particles, under constant growth parameters, is strongly influenced by the size of the openings in the SiO(2) film. Nanowires overgrowing the etched holes, eventually leading to a merging of neighboring nanowires, shed light into the growth mechanism.
Expecting that protein crystal growth in space would show us a new aspect of its crystallization process and might result in the improvement of crystal quality, we performed a series of growth experiments for bovine pancreatic ribonuclease S crystals under microgravity. For this purpose, the COSIMA2 system launched in September 1989 was utilized. In one of the eight crystallization chamber setups in the COSIMA2 experiment, a large "Z-form" single crystal appeared and we found that (a) the space-grown single crystal exhibited a higher fraction of significant X-ray diffraction spots than ground-grown crystals of approximately the same size and (b) the corresponding Wilson plots also tended to show a slightly lower overall temperature factor for the space-grown crystals, although the quality of the plot was not sufficient for quantitative evaluation. Thus, mainly based on the above observation (a), it is most likely that the internal order of the ribonuclease S crystal was in fact improved through the crystallization under microgravity.
Two-step crystallization experiments were conducted in low gravity employing a liquid-liquid diffusion method in an effort to eliminate problems associated with protein crystal growth under the supersaturating conditions required for nucleation. Experiments were performed in diffusion cells formed by the sliding of blocks on orbit. Step gradient diffusion experiments consisted of first exposing protein solutions in diffusion half-wells for brief periods to initiating buffer solutions of high precipitant concentrations to induce nucleation followed by expoure of the same protein solutions to solutions of lower precepitant concentration to promote growth of induced nuclei into crystals. To avoid convective disturbances that occur when solutions of discrepant densities are interfaced at normal gravity, crystallization of hen egg-white lysozyme and rabbit skeletal muscle aldolase by step gradient diffusion was investigated in low gravity on four NASA space shuttle flights. In general, the largest ctystals of both proteins formed at the highest initiating precipitant concentration used, which is consistent with nuclei formation upon brief exposure to high precipitant concentration, and that these nuclei are competent for sustained growth at lower precipitant concentration. The two-step approach dissociates nucleation events from crystal growth allowing parameters affecting nucleation kinetics such as time, precipitant concentration and temperature of nucleation to be varied separately from conditions used for post-nucleation growth.
Pb(In(0.5)Nb(0.5))O(3)-Pb(Mg(1/3)Nb(2/3))O(3)-PbTiO(3) (PIN-PMN-PT) ferroelectric crystals attracted extensive attentions in last couple years, due to their higher usage temperatures range (> 30°C) and coercive fields (~5kV/cm), meanwhile maintaining similar electromechanical couplings (k(33)> 90%) and piezoelectric coefficients (d(33)~1500pC/N), when compared to their binary counterpart Pb(Mg(1/3)Nb(2/3))O(3)-PbTiO(3). In this article, we reviewed recent developments on the PIN-PMN-PT single crystals, including the Bridgman crystal growth, dielectric, electromechanical, piezoelectric and ferroelectric behaviors as function of temperature and dc bias. Mechanical quality factor Q was studied as function of orientation and phase. Of particular interest is the dynamic strain, which related to the Q and d(33), was found to be improved when compared to binary system, exhibiting the potential usage of PIN-PMN-PT in high power application. Furthermore, PIN-PMN-PT crystals exhibit improved thickness dependent properties, due to their small domain size, being on the order of 1μm. Finally, the manganese acceptor dopant in the ternary crystals was investigated and discussed briefly in this paper.
This paper reports comparative characterizations of calcium phosphate crystals grown on earth and in space. At the CaCl2 and KH2PO4 + K2HPO4 solution concentrations and the pH used, only hydroxyapatite (HAP) crystals grow under terrestrial condition while both HAP and octacalcium phosphate (OCP) crystals grew during the space experiment. The space-grown OCP crystals reach 3 mm in size, the space-grown HAP crystals reach sizes up to 100 times larger than the earth-grown crystallites. It was found also that the space-grown crystallites are more perfect than the terrestrial ones, being more stable under electron beam during HRTEM examination. Spherolites of hydroxyapatite consist of small and thin HAP crystals with different orientations. Space-grown OCP crystals containing almost pure OCP phase show strong striations along the c direction due to thickness variations. Terrestrial OCP crystals grown at lowest supersaturation on earth may be almost as large as the space-grown ones, possess a regular habit and are homogeneous in thickness. However, they always contain substantial regions of HAP structure. Also, in these crystals electron irradiation induces phase transformation from crystalline to amorphous (disordered) state during transmission electron microscopy observations. In the space-grown crystals, such transformation needs longer radiation time. We believe that the differences described above come from much lower supersaturation and different pH for crystals nucleating and growing in space compared to those formed on earth.
sketch of the configurations: (a) no sedimentation allowed (b) Lysozyme sedimenting on the gel interface
This paper deals with experimental investigation, mathematical modelling and numerical simulation of the crystallization processes induced by counter diffusion method of a precipitant agent in a lysozyme protein solution. Novel mathematical strategies are introduced to simulate the experiments and in particular to take into account the kinetics of the growth process and the motion of the crystals due to the combined effect of gravitational force and viscous drag if the sedimenting process is allowed (protein chamber free of gel). Comparison between experimental observations and numerical simulations in the presence of convection and sedimentation and without them provides a validation of the model. The crystal formation in gel results modulated in space. If the gel matrix is not present, convective cells arise in the protein chamber due to local inversions in the density distribution associated to nucleation phenomena. As time passes, these vortex cells migrate towards the top of the protein chamber exhibiting a different wave number according to the distance from the gel interface. The sedimentating particles produce a wake due to depletion of protein from the surrounding liquid. The models and the experiments may represent a useful methodology for the determination of the parameters and conditions that may lead to protein crystallization.
We present results on the growth, doping, and application to lasers of GaInP on GaAs(100) substrates using chemical beam epitaxy (CBE). The growth studies were performed in the substrate temperature range of 490-555-degrees-C. We were able to obtain lattice-matching with good surface morphology over the entire substrate range investigated. For a fixed triethylgallium (TEGa) flow, a sharp increase in the trimethylindium (TMIn) flow required to obtain lattice-matching for T(sub) above 520-degrees-C is observed. This can be attributed to an increase in GaP growth rate and a decrease in InP growth rate due to desorption of TMIn species. The p-type and n-type doping of Ga0.51In0.49P was investigated using diethylzinc (DEZn) and hydrogen sulfide (H2S), respectively. It was found that low substrate temperature (less than or similar to 510-degrees-C) was necessary to obtain high p-type doping. Separate confinement heterostructure (SCH) lasers with strained In0.2Ga0.8As/GaAs multiple-quantum-well (MQW) active layers and Ga0.51In0.49P cladding layers for operation at 0.98 mum were grown. Broad-area lasers show extremely low threshold current densities, J(th), of 70 A/cm2. Ridge waveguide lasers with 4 mum stripe width have a threshold of 7.8 mA and gave linear CW output powers upto 100 mW. High external quantum efficiency of 0.9 mW/mA and a very low internal waveguide loss of 2.5 cm-1 were obtained from these lasers.
The InAsN epilayers have been successfully grown on GaAs (0 0 1) substrates with a similar to 0.8-mum-thick InAs buffer layer by radio-frequency plasma-source molecular beam epitaxy. A series of samples obtained at different growth temperatures have been analyzed by high-resolution X-ray diffraction and atomic force microscopy to characterize the structural properties. The strain and the nitrogen concentration of the InAsN epilayers grown on InAs (0 0 1) substrates have also been evaluated for application to those on the InAs buffer layer on the GaAs substrates. The maximum nitrogen concentration of the epilayers grown on the GaAs substrate was estimated to be 1.62% when the growth temperature was 420degreesC.
Si-doped GaAs[001] 1° off <111>A (n = 2 × 10<sup>18</sup> cm<sup>-3</sup>) substrates were prepared by standard procedures in MBE and then annealed at 550°C without As overpressure in order to produce a (4 × 2) phase surface as determined by RHEED and RAS(or RDS). Under the same conditions as these measurements, STM images were obtained. The STM data strongly points to a co-existence of reconstructions. One set of candidates is predicted as the ζ(4 × 6), ζ(4 × 4), and ζ(4 × 4) reconstructions. All models are based on ζ(4 × 2) by Lee et al. and satisfy electron-counting heuristics. The models differ in the presence and location of Ga atoms. At elevated temperatures Ga adatoms can detach and diffuse to make Ga clusters. Mobile Ga would result in different surface reconstructions on different parts of the surface. Each of these reconstructions does not form large domains and is distributed randomly on the [001] surface. This reasonably explains why we do not observe the 1/4- and 1/6-order reflections in RHEED patterns obtained along the [1 - 10] direction. However, since all reconstructions are derived from the ζ(4 × 2), the surface dynamics associated with Ga motion will produce transient regions with this symmetry. Therefore, it is natural to observe the 1/2-order reflection along the [1 - 10] direction.
InSb quantum wells (QWs) with remotely doped Al<sub>x</sub>In<sub>1-x</sub>Sb barriers are candidates for several novel device structures that rely on a long electron mean free path. Mesoscopic magnetoresistors that take advantage of the high electron mobility in InSb QWs at room temperature are currently being developed for read-head applications. The promise of InSb QWs for spin-transistor applications has been shown recently by experiments that demonstrate a large zero-field spin splitting and ballistic transport at temperatures as high as 185 K. Since a semi-insulating substrate is required for electronic applications, the InSb/Al<sub>x</sub>In<sub>1-x</sub>Sb structures are grown on GaAs [001] substrates. The ∼14% lattice mismatch between the epilayers and the substrate results in a high defect density that partially limits the electron mean free path. We will present a detailed characterization of these defects and elucidate their role in limiting electron mobility.
Self-assembled InAs quantum dots by utilizing islanding in a Stranski-Krastanov mode are expected to have tremendous potential for electronic and optical applications. So far, shapes of InAs islands (dots) formed on GaAs[001] have been intensively examined with the use of reflection high-energy electron diffraction (RHEED), atomic force microscopy (AFM), scanning tunneling microscopy (STM), and cross-sectional STM. There are some discrepancies between STM and XSTM results, i.e., considerable differences in shape between before and after GaAs overgrowth on islands. In this paper, STM and RHEED have been used to investigate the overgrowth of GaAs capping layers on InAs islands formed on GaAs[001] substrates. In particular, we focus on how the overgrowth of GaAs on InAs islands proceeds and whether or not the overgrowth of GaAs brings about changes in the shape of InAs islands.
The formation of self-assembled quantum dots (SAQDs) during the growth of epitaxial materials is of great interest for optical applications due to the high density of dots which can be achieved in comparison with lithographic patterning techniques. Much of the research in this field has been directed towards the systems which display a type-I band alignment such as In(Ga)As/GaAs. In these structures both the electrons and holes are confined within the dots themselves.
The need for long-wavelength cost-efficient GaAs-based lasers is currently a powerful driving force in the development of new GaAs-based heterostructures. 1.3-μm lasers based on self-organized InAs quantum dots (QDs) embedded into InGaAs quantum well (QW) have demonstrated very low threshold current density (16 A/cm<sup>2</sup>). On the other hand, low surface density of such QDs that provides the low transparency current density may also result in a small maximum optical gain on QD ground state. This motivated the use of low-loss cavity design (very ong cavities and/or HR/HR facet coatings), which is characterized by low external differential efficiency, η<sub>D</sub>, and output power. Previously the use of triply-stacked array of long-wavelength QDs allowed us to achieve η<sub>D</sub> of 57% for edge-emitting lasers and realize 1.3-μm VCSEL utilizing highly-reflective AlO/GaAs distributed Bragg reflectors. Further enhancement of optical gain of such QDs should improve efficiency of laser diodes as well as open new possibilities for their VCSEL application. However, increase in a number of QD planes, that results in corresponding increase of the saturated gain, typically led to degradation of threshold current and internal quantum efficiency of 1.3-μm-QD lasers.
Metalorganic chemical vapor deposition (MOCVD) has been used to grow low threshold 1.3 mum lattice matched and strained quantum well lasers. By careful optimization of growth conditions and growth interruptions between different layers, we have been able to grow good quality lattice matched and strained InxGa1-xAsyP1-y layers and quantum wells. This is demonstrated by achieving broad area lasers with threshold current densities as low as 190 A/cm2 for a two quantum well device with 0.85% compressive strain.
In recent years, self assembled quantum dots (QDs) have attracted much attention, because of the great potentialities expected from their zero dimensional confinement properties, especially for the realization of opto-electronic devices such as lasers. Indeed, for a QD based laser, a lower temperature dependence (high T<sub>0</sub>), a lower chirp (α<sub>H</sub>) and a higher modulation bandwidth are predicted compared to the conventional quantum well lasers. Numerous studies have been performed for the growth of InAs/GaAs QDs, and lasers have been realized, presenting improved performances (T<sub>0</sub>=161 K @ 350 K [1], α<sub>H</sub>=0 @ 1 GHz [2]).
A lot of knowledge about the formation of InAs or InGaAs quantum dots on GaAs based materials was obtained within the last decade and high quality quantum dot lasers could be realized with emission wavelengths between I and 1.3 μm. But the realization of quantum dot lasers well beyond 1.3 μm is still difficult. Very recently first device results of self-assembled InAs nanostructures on InP based materials with emission wavelengths between 1.4 and 1.8 μm were reported. This new type of quantum dot gain material offers the opportunity to take advantage of dot specific properties like broad band amplification, low threshold current and reduced chirp factor within the long communication wavelength range of 1.45-1.65 μm. In this paper the influence of the surface composition and of growth parameters on the formation of InAs quantum dashes and their optical properties is studied. Device data of corresponding quantum dash lasers is presented.
Exchange reaction of Group V atoms is in situ monitored on a submonolayer scale during the MOVPE growth of pseudomorphic InAs/InP heterostructures on (001) InP substrate by the surface photo-absorption (SPA) method. When arsine was supplied onto the In surface of InP above 360-degrees-C, the substitution between As atoms of impinging AsH3 molecules and P atoms of InP surface was detected as a change in surface reflectivity. The reaction proceeded to about 0.4 monolayer exchange at 400-degrees-C, and was suppressed to less than 0.1 monolayer at 350-degrees-C. In contrast, the exchange between P atoms of impinging PH3 and As atoms of InAs surface was less than 0.1 monolayer below 400-degrees-C. To verify the formation of a metallurgically abrupt heterointerface, InAs/InP single quantum wells with 1-5 monolayers were grown at 350-degrees-C, and were characterized by transmission electron microscopy and photoluminescence. Atomic layer epitaxy was used to control precisely the InAs well thickness with an accuracy of one monolayer. A sharp and an intense quantum well photoluminescence was observed with a reasonable shift of emission energy due to the quantum size effect up to five InAs monolayers, indicating that metallurgically abrupt and atomically flat heterointerfaces were fabricated as designed.
Iron redistribution phenomena were studied in adjacent p-type Zn-, Cd- and Be-doped layers. We showed that iron diffusion behaviour occurs, independent of the acceptor and of the material (InP or GaInAs). In order to explain our observations, we propose an exchange mechanism involving an acceptor interstitial out-diffusion from p-type to semiinsulating material. As an application of our study, we present a SI-BH laser structure with a very low RC product (< 2 ps) and an optical response with bandwidth value up to 22 GHz.
Reflection high-energy electron diffraction (RHEED) oscillations are widely used in molecular beam epitaxy (MBE) as a technique to calibrate material growth rates. The growth rates are used to predict the composition of the following growth run. For many applications, the predicted composition uncertainties of a few percent are adequate, but some applications, like vertical cavity lasers (VCSELs) and distributed Bragg reflectors, demand greater accuracy. To improve the accuracy in determining the composition of MBE-grown films, it is obvious we need to understand the uncertainties and limitations associated with RHEED as an MBE tool. In this study, we investigate several aspects of RHEED intensity oscillations of the specular spot during growth of AlAs, GaAs, and AlGaAs on GaAs substrates, including the effects of beam positioning, substrate size, different growth rates, and incident beam along the [0(-1)1] (corresponding to the 4x reconstruction) and [011] (corresponding to the 2x reconstruction) direction. Additionally, we examine "beat" phenomena in the RHEED oscillations (due to nonuniformity in growth rate across the sample) and beam flux transients and their implications on composition. For the two largest factors, electron beam positioning and flux transients, the overall uncertainty can be reduced with careful experimental technique. The lower end of the range corresponds to technique that minimizes the error, while the upper number corresponds to allowing these factors to be essentially uncontrolled. We also present a procedure that uses the measured variance in the growth rates to calculate the composition with the smallest mean square error.
Nanoscale device technology is driving intense study of thin dielectric layers on semiconductors. The aggressive scaling of Si CMOS technology calls for identifying high κ dielectrics to replace SiO<sub>2</sub> and oxynitrides in gate related applications. The material, requirements for the alternative gate dielectric are very challenging in order to achieve performance comparable to SiO<sub>2</sub>. Furthermore, there are demanding issues for process integration compatibility. Among several binary oxides proposed the rare earth oxides are attractive candidates based on thermodynamic energy considerations and a high conduction band offset over 2eV. The interest in the rare earth oxide stems from our earlier work on GaAs passivation. The Ga<sub>2-x</sub>Gd<sub>x</sub>O<sub>3</sub> mixed oxides (κ =12) and the Gd<sub>2</sub>O<sub>3</sub> oxides (κ =14) films grown by ultrahigh vacuum deposition from an oxide source formed an excellent insulating barrier with low interfacial state density D<sub>it</sub> on the GaAs surface. This discovery has led to the first GaAs based inversion channel MOSFET devices. These dielectrics were also successfully applied to other III-V semiconductors including InGaAs, AlGaAs, InP, and GaN producing MOS diodes and MOSFETs.
We report here the low-pressure metalorganic chemical vapor deposition growth of AlGaAs-GaAs quantum well heterostructures having low-temperature (4.2 K) photoluminescence spectra with full width at half-maximum (FWHM) values ranging from approximately 6 to 4 meV for quantum wells having 6-28 monolayer (ML) widths, respectively. These linewidths are compared to those measured for quantum wells grown by molecular beam epitaxy, flow-rate modulation epitaxy, and atomic layer epitaxy. We find that the FWHM values for the thinnest quantum wells grown in the present study (approximately 6 ML) are equal to or narrower than those observed for comparable structures produced by other technologies.
The formation of vertical AlGaAs quantum wells during MOVPE growth of AlGaAs layers on submicron gratings has been investigated. The influence of the growth temperature, growth velocity, V/III ratio and the overall Al content has been examined and explained using surface diffusion effects of the group III atoms (or reactant species). 77 K PL measurements show a polarization anisotropy and indicate lateral quantum confinement. Also the growth of QWWs on submicron gratings is reported, showing very large PL intensities and polarization anisotropy. Finally the realization of QWWs has been proposed by growing AlxGa1-xAs/AlyGa1-yAs QWs on gratings and based on the different surface mobilities of Ga and Al.
Short-period superlattices of AlP-GaP with different periods have been grown by metalorganic vapor phase epitaxy using triethylaluminum, triethylgallium and phosphine as the sources, and they have been characterized by X-ray diffraction and photoluminescence measurements. As a result, it has been found that the samples of the present study are better in heterointerface abruptness than those of the previous study, where trimethylaluminum and trimethylgallium were used as Al and Ga sources. Light-emitting diodes consisting of the AlP-GaP superlattices have been fabricated as an application of this superlattice system.
InAlAs/InGaAs was grown by low-pressure MOVPE using trimethylamine alane as an aluminum source. Although there is a parasitic reaction between trimethylindium and trimethylamine alane, the growth rate and lattice-matching are well controlled. The photoluminescence intensity for an InAlAs/InGaAs multi-quantum well structure is approximately 3 times higher than for one grown with trimethylaluminum. This is due to the lower oxygen incorporation in InAlAs layer. The crystal quality of InAlAs/InGaAs grown with trimethylamine alane was demonstrated by results for InGaAlAs/InGaAs multi-quantum well lasers and InAlAs/InGaAs high electron mobility transistors. Low threshold current density and high transconductance were realized for each device. These results indicate the usefulness of trimethylamine alane for the MOVPE growth of InAlAs/InGaAs.
The study of interface morphology is of key importance for the optical as well as the transport properties of low-dimensional structures and can also lead to detailed understanding of growth processes. In this paper we report the results of systematic studies of the influence of the growth temperature on the morphology of GaAs/(AlGa)As quantum well (QW) interfaces. A combination of highly selective etching and subsequent atomic force microscopy (AFM) was used to determine the interface morphology. This technique enables three-dimensional mapping of the interfaces with an atomic height resolution on lateral scales from 10 nm up to 50 μm.
Photoluminescence mapping of 2 inch and 3 inch wafers with AlGaAs/GaAs heterojunction bipolar transistor (HBT) layer sequences is reported. Photoluminescence spectra taken at 77 K on these layers reveal four well distinct emission lines from cap, emitter, collector and base, respectively. The mapping of line intensities and spectral positions allows the analysis of doping, composition and thickness homogeneities influenced by epitaxy or device processing. The characteristics of the luminescence light emitted from GaAs: C base layers and the DC current gain of HBT structures after processing exhibit the same lateral distribution over the wafer. The analysis of the spectra showed that the intensity variations can partially be explained by an inhomogeneous base doping, Luminescence mapping of the HBT base related luminescence line was also performed on partially and fully processed wafers. This measurement allowed a comparison of luminescence data with high frequency device results. We found that the ratio between maximum oscillation frequency f(max) and transit frequency f(T) of the HBTs clearly reflect the variation of the base hole concentration.
The use of triethylgallium (TEGa) for the CBE growth of GaAs and AlGaAs leads -to, very significant reductions in unintentional carbon incorporation compared to corresponding layers grown using trimethylgallium (TMGa). However, in a continuing effort to generate even further reductions in impurity levels, the present paper provides the first reported comparison of the tri-isopropylgallium (TIPGa) and tri-tertiarybutylgallium (TTBGa) precursors for CBE growth applications. The use of TTBGa is found to lead to unacceptably low GaAs growth rates, an effect which is attributed to a steric influence on the chemisorption process. In contrast, the TIPGa-grown GaAs layers exhibit very important improvements in electrical and optical properties compared to corresponding TEGa-grown material. Results of initial AlGaAs growth experiments performed using TIPGa are also presented.
The effect of AsH3 pressure on Si planar doping on GaAs in atmospheric pressure metalorganic vapor phase epitaxy was investigated by changing the gas supply sequence: supplying arsine (AsH3) (1) before, (2) with and (3) after disilane (Si2H6) supply for planar doping. Two types of competitive behavior of AsH3 were found. One is promotion of incorporation of Si into the GaAs surface and the other is desorption of adsorbed Si from the surface. On the basis of understanding the Si planar doping behavior, splendid wafers of Si planar-doped AlGaAs/InGaAs pseudomorphic HEMT structures were successfully grown using Si2H6 as a dopant source. The uniformity of sheet carrier concentration n(s) (2.44 X 10(12) cm-2 at 300 K) is +/-2% on a 3-inch wafer and the variation in averages of wafers is within +/-2.9%. The uniformity of electron mobility mu (4700 cm2/V.s at 300 K) is +/-3.3% on a 3-inch wafer and the variation in averages of wafers is within +/-4.3%.
We have investigated the effects of methoxide (-OMe), one of the main oxygen impurities in trimethylaluminum (TMA), on AlGaAs layers grown by metalorganic vapor phase epitaxy (MOVPE). -OMe enriched TMA was prepared by partial oxidation to yield 5300 wt.ppm of -OMe in the liquid phase TMA. The vaporization behavior of -OMe in TMA was studied. There was a strong correlation found between the concentration of -OMe in the TMA and that of oxygen atoms in the AlGaAs layers. The electrical characteristics of the layers co-doped with disilane and -OMe showed that oxygen-related centers compensated the Si donor. The effects of Al composition, growth temperature and As/(Al + Ga) ratio on the behavior of oxygen incorporation were also investigated. It was found that the incorporation behavior of the oxygen was similar to that of other group VI (S, Se, Te) impurities.
C-doped GaAs layers were grown by the atmospheric pressure metalorganic vapor phase epitaxy (AP-MOVPE) method using TMGa and TMAs at low temperature and low V/III ratio. Carbon concentrations up to 7 × 1019 cm-3 were obtained, and their mobility was higher than for Zn-doped GaAs. Using this C-doped GaAs as the base layer, AlGaAs/GaAs HBTs were fabricated, and current gain up to 50 was demonstrated for the sample with a base layer of 700 Å in thickness and 4 × 1019 cm-3 in concentration. The uniformity across the wafer and the reproducibility of the current gains were considered to be good enough for the use of this epitaxial layers as HBT ICs.
We report on a study of buffer layer leakage currents in AlGaN/GaN HEMT structures with AlN buffer layers grown on 6H-SiC substrates by RF-plasma molecular beam epitaxy. We find that the use of a thin Be:GaN layer dramatically reduces the leakage currents in AlGaN/GaN HEMTs. Four structures were grown under the same conditions, except for the Be concentration, using quarters of the same n(+) SiC substrate. Device isolation is improved by a factor of 10(3) and the transistor pinch-off characteristics are improved. Details of the growth conditions and measurements are presented. Published by Elsevier Science B.V.
Semi-insulating Fe-doped InGaAsP alloys of λ = 1.0, 1.15, 1.35 and 1.55 μm wavelength compositions lattice matched to InP have been grown and characterized for the first time. The resistivity is maximum for λ = 1.0 μm InGaAsP at a value of 1×108 ω cm, and decreases exponentially with the arsenic mole fraction for the longer wavelength alloys. The solubility of Fe in the InGaAsP is fixed at the InP solubility value of 5×1016 cm-3 at 630°C until phosphorus is eliminated from the alloy lattice. Diffusion of Fe in InGaAsP is noted, where the diffusion tendency increases with the arsenic mole fraction. Limiting the total Fe concentration to the solubility limit of Fe in InGaAsP greatly reduces diffusion of Fe.
Summary form only given. The InAsSb/AlSbAs heterojunction system is promising for mid-infrared (2-5 μm) lasers, since it combines the narrow band gap of an InAsSb alloy (< 0.4 eV at 77K) with a large conduction band offset at the hetero-boundary. The use of a strained quantum well (QW) active layer instead of thick InAsSb one is believed to increase the differential gain and reduce considerably the Auger recombination which limits the maximum operating temperature of such lasers. We report on MBE growth and photoluminescence (PL) properties of compressively strained InAsSb/AlSbAs single QW heterostructures. A set of norminally-undoped structures with well widths varying from 4 to 20 nm was grown pseudomorphically on GaSb [001] substrates using a two-stage growth regime. PL measurements of the structures were performed at 80K. Temperature and pump-power dependences of the PL intensities versus various structure parameters will be discussed.
The solubility of AlPO4 was measured by a weight loss method using small Pt capsules over the range 170–525°C, 1–30 kpsi (69–2067 bar), 0–14.6 m (molar) H3PO4. The temperature coefficient of solubility, at constant pressure, was found to be negative and procedures for accurate solubility measurements under these conditions were developed. It is shown that a similar “retrograde” dependence of solubility on temperature at constant pressure in the system SiO2–H2O can be explained on the basis of the pressure-volume-temperature behavior of the solvent resulting in a positive temperature coefficient of solubility at constant solvent density or percent fill. It is suggested that the same effect might obtain for AlPO4–H3PO4–H2O at some p-v-T conditions but from an analysis of growth experiments at constant percent fill the temperature coefficient of solubility at constant fill must be negative over most of the range of growth conditions. The temperature dependence of solubility of AlPO4 obeys the Van 't Hoff equation with a change in the heat of solution at about 300°C suggesting a change in the dissolving mechanism. The solubility dependence on H3PO4 concentration suggests a dissolving reaction involving species such as Al(H2PO4)3 or Al(H2PO4)-36, a result consistent with the observed pressure dependence. Imperfections in AlPO4 were studied by macroscopic and microscopic examination in an index matching medium, etching and X-ray linewidth and topographic measurements. Growth on non-equilibrium faces is shown to produce the hydrothermal analogue of dendritic growth (“crevice flaws”) leading to solvent bubble entrapment (“veils”), strains and cracks. The use of etching to select seeds free of misoriented regions, etching to remove seed preparation damage and growth on equilibrium faces were all shown to improve the quality of grown material.
We developed a method for simulating birefringence of an annealed ingot of calcium fluoride single crystal caused by the residual stress after annealing process. The method comprises the heat conduction analysis that provides the temperature distribution during the ingot annealing, the elastic thermal stress analysis using the assumption of the stress-free temperature that provides the residual stress after annealing, and the birefringence analysis of an annealed ingot induced by the residual stress. The finite element method was applied to the heat conduction analysis and the elastic thermal stress analysis. In these analyses, the temperature dependence of material properties and the crystal anisotropy were taken into account. In the birefringence analysis, the photoelastic effect gives the change of refractive indices, from which the optical path difference in the annealed ingot is calculated by Jones calculus. An approximate method for calculating the optical path difference using the average stress along the wave normal is also proposed and the relation between the Jones calculus and the approximate method is discussed. It is found that the result of the approximate method agrees very well with that of the Jones calculus. The distribution of the optical path difference in the annealed ingot obtained from the present calculation agrees reasonably well with that of the experiment. Its calculated value also agrees reasonably well with that of the experiment, when a stress-free temperature is adequately selected.
We report on the chemical synthesis, thermal decomposition studies and the first low pressure MOVPE growth experiments for a new class of metalorganic As compounds, designed as substitutes for the highly toxic AsH3. The key feature of these molecules is the in-situ formation of As-H-functions, formed by thermal decomposition under beta-elimination only in the hot temperature zone of the MOVPE reactor. As a model precursor, the As-source diethyl-tertiarybutyl-arsine (DETBAs) has been synthesized. The expected thermal decomposition mechanism of this molecule under beta-elimination, studied in an ersatz reactor system at low pressure, is proved by the detection of the products diethylarsine (DEAsH) and the stable isobutene (C4H8) in the mass spectra as cracking pattern. In first LP-MOVPE growth experiments, DETBAs has been used in combination with the standard group (III) sources TMGa and TMAl to realize GaAs and (AlGa)As bulk layers as well as AlAs/GaAs superlattices. The influence of growth parameters on the structural and electrical quality of the epilayers is presented and discussed by means of high-resolution X-ray diffraction, interference microscopy and Hall investigations. Smooth surface morphologies and narrow X-ray diffraction linewidths are achieved for GaAs (12" FWHM) and (AlGa)As (13" FWHM), as well as for AlAs/GaAs superlattices (SL main peak: 11.5" FWHM; SL sattelite peak: 15.3" FWHM). All layers exhibit n-type carrier concentrations, caused by residual impurities in the used batches of DETBAs. Best results show impurity levels of N(D) - N(A) = 3 x 10(16) cm-3 with electron mobilities of mu = 4800 cm2/V.s at 77 K. These first results render this class of molecules as very interesting and promising substitutes for the group V hydrides.
GaN grown by molecular beam epitaxy (MBE) has been studied extensively during the last decade. Because molecular nitrogen is too inert to be used in MBE growth, two different approaches have been used to provide active nitrogen, plasma-assisted MBE (PA-MBE) and ammonia-MBE. We have shown that PA-MBE using an RF plasma source can produce high quality AlN, GaN, InN and mixed group III-nitrides for a variety of applications. Nitrogen incorporation in GaAs is now extensively used for long wavelength laser applications. However, the incorporation of As in GaN during MBE has not been studied as intensively so far.
As MOSFETs scale to deep-submicron dimensions, there has been an increasing demand for silicon epitaxial layers with abrupt doping profiles. For nanoscale devices, arsenic is an attractive n-type dopant because of its high solubility and low diffusivity, but suffers from surface segregation during epitaxy, making high- concentration incorporation with abrupt profiles difficult. In this paper we report results of arsenic incorporation in Si molecular beam epitaxy (MBE) using a unique combination of solid (As, Si) and gas (disilane) sources to achieve these goals.
Quality GaAs buffer layers have been grown in a high throughput MOVPE reactor using a novel electrochemical arsine generator which reduces the required quantity of arsine gas on site to less than 0.1 lbs. Hall, secondary ion mass spectrometry (SIMS), photoluminescence (PL), and deep level transient spectroscopy (DLTS) measurements reveal material properties comparable to those obtained with high purity tank arsine. Background carrier concentrations in the low 10(14) cm-3 range were observed. AlxGa1-xAs layers grown using generator arsine were shown by SIMS to have the same oxygen content as layers grown previously in this reactor using tank arsine but are higher than reported elsewhere. The oxygen contamination of AlxGa1-xAs was observed to decrease with the addition of an arsine purifier. Preliminary observations on the reproducibility of generator arsine are presented.
MgS has a very large bandgap of ∼ 5eV and can form an excellent barrier material for wide-gap II-VI quantum structures. Although its stable crystal structure is rocksalt, our group has recently established a novel molecular beam epitaxy (MBE) technique that allows us to grow zinc-blende MgS lattice matched to GaAs substrates to thicknesses greater than 130 nm [1]. The lattice parameter of zinc-blende MgS is almost the same as that of ZnSe. Accordingly, strain between MgS and CdSe is almost identical to that between ZnSe and CdSe and a transition from 2D to 3D growth is expected with increasing CdSe coverage. Using the barrier material MgS with CdSe dots instead of ZnSe has two advantages. Firstly, large band discontinuities, which are estimated to be 2.1 and 0.9eV for the conduction and valence bands, respectively, will provide strong carrier confinement. Secondly, interdiffusion of MgS and CdSe is inhibited due to the immiscibility of these two materials. The clear material boundary between the dots and the barrier will enhance the confinement even further.
GaInNAs and GaNAs are the representative III-VN semiconductors, which have a near-linear dependence of the bandgap between GaAs and GaN with a much larger alloy bandgap bowing parameter than for any other ternary semiconductors. Recently, GaInNAs and GaNAs have been intensively investigated as new material systems for long wavelength communication devices fabricated on Si and GaAs substrates at the optical-fiber communication wavelength window (1.3∼1.55μm) with superior temperature stability, as well as for multi-junction high efficiency solar cells. However, the crystal quality of GaInNAs, in general, becomes degraded with increasing N composition. This problem must be overcome in order to realize more reliable device-quality GaInNAs. Meanwhile, we have previously shown that the irradiation of atomic H during MBE growth of GaAs, and GaN significantly improves the crystal quality of these films. We investigate the effect of atomic H irradiation in the growth of GaInNAs by RF-MBE in view of the growth dynamics and crystal quality including surface morphology and optical characteristics.
Quaternary InxAlyGa1-x-yN alloys with In mole fraction x ranging from 0 to 0.10 and Al mole fraction y in the range of 0.30-0.40 were grown by RF plasma-assisted molecular beam epitaxy (RF-MBE) by varying the growth temperature and the In flux. The surface morphology and structural quality of InAlGaN layers were optimized using metal-rich growth conditions and a relatively low growth temperature. The layers exhibited strong band-edge photoluminescence (PL) up to room temperature and a minimum line width of 61 meV for 20 K PL emission at 3.0 eV. Abrupt GaN/InAlGaN multiple quantum well structures were grown, with high-resolution X-ray diffraction curves exhibiting satellite peaks and interference fringes. Unwanted In incorporation in the GaN layers by In segregation effects was eliminated. Laser structures with active region consisting of InAlGaN wells and GaN barriers exhibited lasing under optical pumping at room temperature.
A novel structure of the active region of multiple quantum well (MQW) InGaAsP lasers at 1.55 μm is proposed. Both barriers and wells have the same group V (As/P) composition. This minimizes the thermal intermixing problems encountered with more conventional structures. The specific band structure associated with this design (loose hole confinement, strong electron confinement) may also lead to specific, favorable device characteristics such as, for instance, high-speed intensity modulation capability.
Top-cited authors
Ramasamy Perumalsamy
  • Sri Sivasubramaniya Nadar College of Engineering
Gerald B Stringfellow
  • University of Utah
H. Amano
  • Nagoya University
Akira Yoshikawa
  • Tohoku University
Georg Müller
  • Friedrich-Alexander-University of Erlangen-Nürnberg