Search

Numerical investigations have been performed for modeling the global temperature field of an industrial liquid phase epitaxy (LPE) facility and to estimate the temperature fluctuations in a Te-rich solution during the LPE growth. The numerical results agreed well with experimental data and therefore provide reliable reference points for experimenters for further improvements of the growth conditions. (© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

In this communication, the latest experimental results are presented on a composition of Cd1−xZnxTe1±δ, both x and δ, determined by the high-precision, vapor-pressure scanning method. The space arrangement of the single-phase solidus volume of Cd1−xZnxTe1±δ in the pressure-temperature-composition (P-T-X) phase space has been derived from direct vapor-pressure measurements in the temperature range up to 1,375 K. From the experimental data for x=0.0, 0.05, 0.1, 0.15, 0.25, 0.5, 0.75, 0.8, 0.9, and 1, the complete phase diagram of Cd1−xZnxTe1±δ was constructed. Transformation of the diagram was traced for the whole range of existence of the Cd1−xZnxTe1±δ solid solution as a function of ZnTe concentration. Increase of ZnTe in the solid solution results in the extension of the homogeneity range and a shift of the solidus toward Te, so that already for x≥0.15 the solidus does not contain the stoichiometric, (nCd+nZn):nTe=50 mol.% plane. From detailed experimental studies, maximum non-stoichiometry as a function of the temperature was determined in a wide temperature range for Te and metal solubility. Temperature dependencies of the partial pressures of vapor-phase species for different crystal compositions, XS=const, were deduced from the vapor-pressure experiment for the x=0.05 section of the diagram. From these results, relationships are derived between crystal composition, vapor pressure, and composition of the conjugated vapor.

The sessile drop technique has been used to measure the wetting angle and the surface tension of molten germanium (Ge) on various substrate materials. Sapphire, fused silica, glassy carbon, graphite, SiC, carbon-based aerogel, pyrolytic boron nitride (pBN), AlN, Si3N4, and CVD diamond were used as substrate materials. In addition, the effects of different cleaning procedures and surface treatments on the wetting behavior were investigated. The highest wetting angles with values around 170 deg. were found for pBN substrates under active vacuum or with a slight overpressure of 5N Argon or forming gas (2% Hydrogen in 5N Argon). The measurement of the surface tension and its temperature dependence for Ge under a forming gas atmosphere resulted in gamma(T) = 591 - 0.077 (T-T(sub m).

Transverse rotating magnetic fields (Bmax=7.5 mT, frot=50 Hz) were applied to the floating zone growth of doped silicon. Non-periodic dopant fluctuations caused by time-dependent thermocapillary convection were considerably reduced by the rotating field. The radial segregation profiles (measured by a spreading resistance probe) became more homogeneous and more symmetric. The transition from a regime dominated by time-dependent thermocapillary convection to a flow state characterized by the rotating magnetic field was determined. This threshold depends on the height of the melt as well as the melt diameter (crystals between 8 and 14 mm have been investigated) and the efficiency of the applied field increases with larger melt zones. For a melt of 14 mm in diameter and an aspect ratio of 1 it is in the range of 2.5–3.75 mT/50 Hz (corresponding to a Taylor number of Ta=9.3×103–2.1×104). The change from a time-dependent 3D-flow without field to a quasi-axisymmetric 2D-flow with the magnetic field is corroborated by numerical simulations of the flow field: the thermocapillary driven irregular flow rolls are transformed to a nearly axisymmetric flow with high azimuthal flow velocities but reduced axial and radial components.

Ge 1− x Si x ( x ⩽10 at%) single crystals were grown with the float zone technique using a monoellipsoid mirror furnace. The feed rod consisted of pre-synthesized Ge 0.95 Si 0.05 polycrystalline material with an initial composition of pure germanium. Several boron-doped (1–2×10 17 at cm −3 ) crystals were grown using 〈1 0 0〉 Ge seeds. Taking advantage of the pre-synthesized feed rods, a defined macrosegregation could be achieved in the grown crystals with a linear slope at the beginning (≈0.5 at% mm −1 ) followed by a plateau region with a constant silicon distribution (Si concentration up to 10 at%, fluctuation rate: ⩽±0.3 at%). The etch pit density was in the range of 7×10 3 –2×10 4 cm −2 . Micrographs of the etched crystals show sharp changes in interface curvature at the crystal edges. These distortions of the interface morphology are a direct function of the Si concentration; they are considered to be caused by solutal Marangoni convection. Theoretical considerations show that the flow direction and strength vary significantly from a solutal Marangoni convection regime directly in front of the solid–liquid interface to a thermal Marangoni convection regime within the bulk melt.

Traveling heater method (THM) experiments on InP crystal growth from indium solution (SKAT-FD) were performed during the Russian FOTON-11 mission in October 1997. The aim of these experiments was to ascertain the effectiveness of controlling heat and mass transfer in the liquid phase by applying a rotating magnetic field (RMF). As an effort to understand the parameters influencing the crystal growth a mathematical model was developed for numerical simulation of these experiments. The SKAT-FD crystals were grown with RMF at an induction of 1 and 2 mT. As the calculations show, such magnetic fields influence remarkably the hydrodynamics of the melt, and consequently, the growth process. The change of the magnetic induction value (or switching off the RMF) led to an abrupt change of the growth rate, measured from artificial dopant striations produced in the growing crystals by the heat-pulse technique.

S-doped InP crystals were grown from an In-solution in a THM arrangement during the EURECA-1 mission using the automatic mirror furnace (AMF) under microgravity conditions. The InP-crystals showed growth rate oscillations in the range of 0.8–2.0 μm/min with a period of 45 min, which is half the time of an orbit of the EURECA-1 satellite. The fluid dynamic conditions during this experiment are modelled using a 3D model with the commercial software FIDAP. The time-dependent simulation, considering thermal as well as solutal buoyancy effects, indicates a weak laminar flow in this material system. Due to the fixed attitude of the satellite with respect to the sun the residual gravity vector rotates relative to the sample axis. This rotation results in a significant change of the residual convection level. Its maximum is obtained, when the gravity is parallel to the growth interface (two times during each orbit). Due to a high Schmidt number Sc≈25 a time-dependent influence of the convection on the concentration field can also be observed.

USA Several microgravity experiments on sounding rockets and the Space Shuttle have shown that time-dependent thermocapillary (Marangoni) convection is the major cause for the formation of dopant striations in floating-zone grown semiconductor crystals, at least in small-scale systems not employing RF heating. To quantify this correlation, a silicon floating-zone experiment was performed during the TEXUS36 flight (February 7, 1998) in the monoellipsoid mirror furnace TEM02-ELLI. During the experiment, temperature fluctuations in the silicon melt zone and the microscopic growth rate were simultaneously measured. Temperature fluctuations of 0.5 C - 0.7 C with main frequencies between 0.1 Hz and 0.3 Hz were detectable. The microscopic growth rate fluctuated considerably around the average growth rate of 1 mm/min: rates from 4 mm/min to negative values (backmelting) were observed. Dopant striations are clearly visible in the Sb-doped crystal. They were characterized by Spreading Resistance measurements and Differential Interference Contrast microscopy. The frequencies associated with the dopant inhomogeneities correspond quite well with those of the temperature fluctuations and microscopic growth rates. 3D numerical simulations were performed to predict the optimum position of the temperature sensor, to evaluate characteristic temperature amplitudes and frequencies, and to give insight into the instability mechanisms of Marangoni convection in this configuration. The simulations were in good agreement with the experimental values, showing temperature fluctuations with frequencies ? 0.25 Hz and amplitudes up to 1.8 C at a position equivalent to that of the sensor tip in the experiment.

Heat and mass transfer in float-zone processing are strongly influenced by convective flows in the zone. They are caused by buoyancy convection, thermocapillary (Marangoni) convection, or artificial sources such as rotation and radio-frequency heating. Flows in conducting melts can be controlled by the use of magnetic fields, either by damping fluid motion with static fields or by generating a defined flow with rotating fields. The possibilities of using static and rotating magnetic fields in silicon floating-zone growth have been investigated by experiments in axial static fields up to 5 T and in transverse rotating magnetic fields up to 7.5 mT. Static fields of a few 100 mT already suppress most striations but are detrimental to the radial segregation by introducing a coring effect. A complete suppression of dopant striations caused by time-dependent thermocapillary convection and a reduction of the coring to insignificant values, combined with a shift of the axial segregation profile toward a more diffusion-limited case, is possible with static fields greater than or equal to 1 T. However, under certain conditions the use of high axial magnetic fields can lead to the appearance of a new type of pronounced dopant striations, caused by thermoelectromagnetic convection. The use of a transverse rotating magnetic field influences the microscopic segregation at quite low inductions, of the order of a few millitesla. The field shifts time- dependent flows and the resulting striation patterns from a broad range of low frequencies at high amplitudes to a few high frequencies at low amplitudes.

Rotating or alternating magnetic fields are widely used in the industrial steel casting process or in metallurgical manufacturing. For the growth of single crystals, these techniques attracted a rapidly increasing attention within the last years: a well defined melt flow leads to a more homogeneous temperature and concentration distribution in the melt and consequently improves the growth process. Rotating magnetic fields (RMF) might be used instead of crucible and/or crystal rotation avoiding mechanically induced disturbances or might be added to conventional rotation mechanisms to gain a further flow control parameter. Compared to static magnetic fields, rotating ones are distinguished by a much lower energy consumption and technical effort. Furthermore, there are no reports on detrimental effects such as the generation of thermoelectromagnetic convection or coring effects in the grown crystals. One advantage of rotating magnetic fields is the possibility to apply them even to melts with a rather low electrical conductivity like e.g. aqueous solutions. High flow velocities are already generated with moderate fields. Therefore the field strength has to be adjusted with care because otherwise undesirable Taylor vortices might be induced. In the last years, the potential of rotating magnetic fields for crystal growth processes was demonstrated for model arrangements using e.g. gallium or mercury as a test liquid as well as for a variety of growth techniques like Float Zone, Czochralski, Bridgman, or Travelling Heater Methods: Fluctuations of the heat transport due to time-dependent natural convection have could be reduced by more than an order of magnitude or the mass transport could be improved with respect to the a better radial symmetry and/or a more homogeneous microscopic segregation.

Heat and mass transfer in semiconductor float-zone processing are strongly influenced by convective flows in the zone, originating from sources such as buoyancy convection, thermocapillary (Marangoni) convection, differential rotation, or radio frequency heating. Because semiconductor melts are conducting, flows can be damped by the use of static magnetic fields to influence the interface shape and the segregation of dopants and impurities. An important objective is often the suppression of time-dependent flows and the ensuing dopant striations. In RF-heated Si-FZ - crystals, fields up to O.STesla show some flattening of the interface curvature and a reduction of striation amplitudes. In radiation-heated (small-scale) SI-FZ crystals, fields of 0.2 - 0.5 Tesla already suppress the majority of the dopant striations. The uniformity of the radial segregation is often compromised by using a magnetic field, due to the directional nature of the damping. Transverse fields lead to an asymmetric interface shape and thus require crystal rotation (resulting in rotational dopant striations) to achieve a radially symmetric interface, whereas axial fields introduce a coring effect. A complete suppression of dopant striations and a reduction of the coring to insignificant values, combined with a shift of the axial segregation profile towards a more diffusion-limited case, are possible with axial static fields in excess of 1 Tesla. Strong static magnetic fields, however, can also lead to the appearance of thermoelectromagnetic convection, caused by the interaction of thermoelectric currents with the magnetic field.

Semiconductor crystal growth experiments performed in Germany often used the Floating-zone technique (FZ). Recent Results of the FZ growth of Silicon, Si and Galliumantimonide, GaSb in connection with fluid dynamics will be reported.

GeSi single crystals in the range Si⩽10 at% have been grown on Ge 〈1 1 1〉-seed crystals by the vertical Bridgman method in a radiation heated mirror furnace. The diameter of the crystals was 9 mm, their length about 30–40 mm. Axial and radial macrosegregation have been measured by energy dispersive analysis by X-ray (EDAX), microsegregation was investigated using interference contrast microscopy. The average EPD is in the range 6×10 4 –1×10 5 cm −2 and is enhanced at the seed/crystal interface and in the part grown in the vicinity of the container wall. Measurements with a Bartels five crystal diffractometer of the 〈1 1 1〉 peak resulted in FWHM (full-width at half-maximum) values of 20 to 40 arcsec, compared to the theoretical value of 17 arcsec. The FWHM values of the bound excitons, determined by photoluminescence, are in the range 2.5–4.5 meV, which are the best values reported hitherto, and show a linear enlargement with increasing silicon concentration. The segregation coefficients of Ga, In, P, As, and Sb fit well to a linear interpolation between the Ge and Si values.

Gallium Arsenide layers grown using low pressure Vapour-Phase Epitaxy (LP-VPE) were studied with CV, Hall-measurements and Photoluminescence. The results have been analysed for the general investigation of the influence of material properties on particle detector performance for X-ray applications. This p-type material exhibits a free carrier concentration of 1.3 × 10 11 cm −3 at room temperature and was compensated by the presence of shallow donors and deep acceptors. Because of this, the detector performance was restricted by the space-charge density of the order of 10 14 cm −3 .

Numerical stability analysis and general flow characterization were performed for the flow induced by transverse rotating magnetic fields in cylindrical containers. Starting from Richardson’s theoretical critical value for the generation of Taylor vortices in infinite cylinders, overstability in finite cylinders with different aspect ratios was investigated. Decreasing aspect ratio leads to a more stable configuration. For small cylinders and reasonable frequencies of the magnetic field, instabilities arise at low magnetic inductions, e.g., Bc≈0.5 mT [gallium as model liquid, with radius of the cylinder R=12.5 mm, aspect ratio (h/d)=2 and frequency f=50 Hz]. As a source of instabilities Taylor vortices have been identified. They are transported by the secondary flow, emerging in finite cylinders due to the imbalance of pressure gradient and centrifugal force close to the top and bottom, thus generating time-dependent flow. Their generation is statistical and manifests itself in, e.g., nonperiodic temperat...

A numerical modelling of the heat transport in a three-zone resistance furnace has been performed.The results of the calculations agree very well with experimentally obtained temperature profiles. A strong dependence of the calculated temperature field on the thermal conductivity of the furnace material is observed and can be explained by nonisotropic effects like inhomogeneities in the insulation.The simulations can also predict the influence of modifications in the furnace set-up like an additional electromagnet instead of an outer insulation layer. Further, the computer model is used to reveal the influence of different positions of a growth ampoule. The temperature field in the ampoule shows a decrease of the temperature gradients at the phase boundary at the end of the growth process and a change of the interface shape from concave to convex.

Undoped and chlorine-doped CdTe single crystals of 2.4 cm diameter were grown from the vapour phase by a modified Markov method. Boundary conditions for the control of partial pressures were realized applying a heat sink at different temperatures. The influence of the sink on dopant incorporation was investigated by photoluminescence. Transitions of free and bound excitons dominate the PL spectra in cases of undoped CdTe. The characteristic luminescence of the A-centre was identified in CdTe : Cl. Alpha measurements yielded a product of carrier mobility and lifetime of 2 × 10−4 cm2/V.

Silicon floating-zone experiments in axial magnetic fields up to 5 T have been performed and the macroscopic and microscopic segregation have been determined. Although the axial segregation is shifted towards the diffusive case with the application of higher fields, a pure diffusion-controlled regime cannot be attained even with fields of 5 T. The deterioration of the radial profiles experienced with smaller fields, due to the separation of the flow field in a quiescent center and a periphery mixed by thermocapillary (Marangoni) convection, can be reduced considerably with higher fields. A complete suppression of dopant striations caused by time-dependent thermocapillary convection is possible with fields of several Tesla. However, the use of high axial magnetic fields can sometimes lead to the appearance of a new type of pronounced dopant striations of oscillatory nature. These striations can be attributed to thermoelectromagnetic convection, caused by the interaction of thermoelectric currents with the magnetic field, and often suggest an annular flow pattern. The fact, that the direction and magnitude of thermocurrents depend on the interface shape, temperature field, and composition, explains the variety of striation patterns obtained, as well as the possibility to grow striation-free crystals.

The bulk damage (namely the introduction rate of the arsenic antisite AsGa and its ionisation ratio, the Fermi-level position and resistivity) was determined as a function of the non-ionising energy loss (NIEL) of hadrons in semi-insulating (SI) GaAs. The study was performed using near-infrared (NIR) absorption and time dependent charge measurements (TDCM) on 23 GeV proton and 192 MeV pion irradiated Liquid Encapsulated Czochralski (LEC) grown GaAs. We have shown that the bulk damage scales linearly with the total NIEL up to a pion fluence of 5 × 1014 particles/cm2. The EL2 introduction rate for the high energetic pions and protons were determined to be 92.7 ± 2.6 cm−1 and 68.1 ± 3.7 cm−1, respectively. A comparison of these values gives a hardness factor for the 192 MeV pions of 9.5 ± 1.4. In addition it was found that the variation of the Fermi-level and resistivity with radiation damage is a function of the initial resistivity of the SI GaAs before irradiation.

V and Ga doped CdTe single crystals with one inch diameter were grown without wall contact in a semi-closed vapour phase system (modified Markov method). By vapour transport modelling, we demonstrate that volatile impurities and excess species are enabled to condense in a heat sink connected to the growth chamber. Vapour composition and component fluxes are controlled by the temperature profile, in particular by the sink temperature. The grown crystals exhibit pronounced {{1 1 1}}, {{1 1 0}} and {{1 0 0}} facetting. The influence of the deep heat sink on interface stability is discussed in terms of growth morphology and formation of inclusions. Piezobirefringence measurements indicate nearly stress-free growth. The resistivities of the grown crystals are up to 3 × 10 9 Ω · cm .

3d-transition metals are known to produce deep levels in III-V semiconductors like gallium arsenide or indium phosphide. In this work, gallium antimonide was doped with the most promising candidates vanadium, titanium and iron according to the predictions of the energetic positions of the 3+/2+-charge transfer levels in the bandgap. For optimum incorporation of dopants several growth methods (travelling heater method THM, Czochralski CZ and liquid phase electroepitaxy LPEE) were used. The grown crystals were characterized by photoluminescence and electrical measurements. No deep levels in the forbidden band were found in the case of vanadium and titanium. Vanadium shows a scavenging effect on unintentional impurities. Titanium tends to form inclusions. In gallium antimonide doped with iron an acceptor was found with an activation energy of 26 meV. The findings support the predictions of Tersoff concerning charge transfer levels.

Vanadium and gallium doped CdTe crystals were grown from the vapour phase (modified Markov method) and (Cd 0.9 Zn 0.1 )Te:V from the melt (vertical Bridgman). The crystals were characterized by photoinduced current transient spectroscopy (PICTS), photoluminescence (PL) and time dependent charge measurements (TDCM). Transitions from different charge states (V 2+ /V 3+ ) of the vanadium donor have been observed in the V-doped crystals by PICTS. A shallow donor level (dE = 0.068 eV) and the Ga A-center have been identified by PICTS and PL measurements in CdTe:Ga. In case of V-doping high resistivity is achieved all over the crystal while Ga-doping results in a high resistivity region only in the middle of the crystal. Calculation of the resistivity by means of a compensation model shows that for both dopants an additional not observed deep donor has to be assumed in order to describe the resistivity distributions.

CdTe 0.9 Se 0.1 :Cl is a detector grade material for gamma and X-rays. Its high resistivity and the high mobility lifetime product yield a high charge collection efficiency of 90 percent. CdTe 0.9 Se 0.1 :Cl was used for the first time to built up a strip detector. The detector performance was investigated by a 57 Co source. The signal behaviour, charge collection efficiency and coupling effects were analyzed for different strips. The comparison between the signal amplitude of all strips showed a good homogeneous response for the device. For a single strip a charge collection efficiency of more than 40 percent was obtained.

Radiation detectors work as Schottky devices. The Schottky contact of the investigated detectors consists of a Ti/Pt/Au layer on the front-side of a 500?m thick semi-insulating GaAs substrate. The back-side is an alloyed Ge/Au ohmic contact.The structure of these contacts are squares with sides of between 1,5mm and 5mm. In order to seperate the behaviour of the contacts, an independent determination of those parameters of the bulk material which are responsible for a satisfactory functioning of the detector are necessary. We present a theoretical model of the detector which simulates detector properties according to relevant parameters such as the resistivity p, the mobility ? of the charge carriers, the energy level EDD and concentration NDD of the bulk-defects and the leakage-current jL of the device. The aim of this study is to determine these parameters experimentally and compare materials from different suppliers. The relevant parameters are then used for the detector simulation.

Te-doped and undoped GaAs single crystals of 6–7 mm diameter and up to 20 mm length have been grown by the floating-zone method. In addition, a Te-doped GaAs single crystal of 20 mm diameter and 20 mm length was successfully grown during the 2nd German Spacelab mission D2. It could be shown that growth rates up to 240 mm/h are possible without losing single crystallinity. Dopant striations due to Marangoni convection have been found, as well as a reduction of the etch pit density compared with that of the starting material.

The influence of growth conditions on radiation detector performance was studied by examining Cd(Te,Se):Cl crystals. They were grown by the travelling heater method under microgravity conditions during the PHOTON 8 mission. An additional forced convection was produced by a rotating magnetic field (400 Hz) of 2 mT. The influence on crystal quality was analyzed by several characterization methods such as time dependent charge measurement, photo-induced current transient spectroscopy and admittance spectroscopy. The charge collection efficiency of detectors was increased from 10% without magnetic field up to 66% by the influence of the rotating magnetic field. The improved detector performance is explained by the increased resistivity and a reduced number of deep levels.

The analysis of striae patterns is commonly restricted to their lateral frequency distribution. Using surface mappings made by shearing interference microscopy (SIM), however, the effects of dopant inhomogeneities on etch roughnesses can be quantitatively detected. The SIM technique allows nanoscale measurements in the vertical direction. Thus, the determination of the height of striations up from ∼ 0.5 nm is possible. The lateral pixel resolution reaches 0.14 μm. Contrarily, wide image fields (maximum 2.2 × 2.2 mm 2 ) permit the recording of macrosegregations. The application on the axial segregation in Si floating-zone crystals (FZ) grown under 1 g and μ g conditions shows a good correlation of spreading resistance (SR) and roughness profiles.

CdTe:Cl crystals were grown from the liquid and from the vapour phase under microgravity (μg) conditions on board the unmanned EURECA I mission. The resistivity distribution of the grown crystals was measured by time dependent charge measurement (TDCM). Photo induced current transient spectroscopy (PICTS) was used to investigate the deep level properties. The axial resistivity distributions of the crystals grown in space differ significantly from 1 g reference crystals. In the case of vapour growth, these differences can be explained by an additional laminar flow under 1 g conditions. Supercooling has to be considered in growth from a Te zone under μg.

GaSb bulk single crystals with low acceptor concentration were grown from a bismuth solution by the traveling heater method. The result is isoelectronic doping by Bi which gives a variation of the opto-electronic properties as a function of grown length as well as a pronounced microscopic segregation. Photoluminescence spectra at 4K show a decrease of the natural acceptor during growth, which is confirmed by Hall measurements. The electrical properties of this isoelectronic doped GaSb are hole concentrations and mobilities of NA − ND = 1.7 × 1016 cm−3 and μ = 870 cm2Vs at room temperature and NA­ND = 1 × 1016 cm−3 and μ = 4900 cm2/Vs at 77K, respectively. The lowest p-type carrier concentration measured at 300K is NA − ND = 3.3 × 1015 cm−3

Time dependent charge measurements (TDCM) are used for noncontact characterization of CdTe crystals doped with vanadium or titanium. Several extensions of the basic technique are presented, which allow for the investigation of the thermal activation energy of the charge carriers, the photosensitivity and the surface photovoltage (SPV). Results of noncontact DLTS measurements show that the formation of defects can change suddenly during the crystal growth process.

The measurement of piezobirefringence in CdTe is a useful tool for the detection of long-range and local stress fields. By means of infrared polariscopy, we obtained both the orientation of principal stresses, and the size mapping of their differences. Wall contact of the crystal, self-seeding and formation of facets at the phase boundary are origins of highly strained regions. Local stress fields can be assigned to microscopic defects such as dislocations and inclusions. We will contribute to the classification of inclusions concerning their size and shape versus local stress generation.

Vanadium doped CdTe is studied by electron paramagnetic resonance (EPR), photo-EPR, and optical absorption. The EPR measurements identify vanadium substitutional for Cd in the oxidation states V 2+ and V 3+ . Whereas V 3+ shows the well-known isotropic EPR spectrum, the V 2+ spectrum features triclinic symmetry due to a strong Jahn-Teller coupling. By photo-EPR, the V 2+ /V 3+ -donor level was determined at 0.67 eV below the conduction band. In absorption, additionally to internal crystal field transitions of vanadium, we also observe the recharging process V 3+ to V 2+ .

Low temperature (5 K) spectra of liquid phase electroepitaxially grown sulphur-doped GaSb have been examined. Native acceptor-related transition appears at 777 meV in undoped GaSb and it is shifted by 4–6 meV to lower energy in sulphur-doped layers. The dominant sulphur-related transition S 1 lies at 731–733 meV in different samples. The intensity of this transition is higher than the intensity of the native acceptor-related transition and their ratio increases with the amount of Sb 2 S 3 in the growth melt. Both photoluminescence and photoreflectance measurements have shown that the energy band gap of these sulphur-doped samples ( E G = 732 meV at 295 K) is a little higher in comparison with other published values.

Long-term crystal growth experiments were successfully performed under microgravity conditions during the first flight of the unmanned EURECA-1 mission in the automatic mirror furnace (AMF). Two crystals of sulphur-doped InP with [001] and [111] orientation respectively were grown from indium solution by the travelling heater method (THM). The absence of time dependent buoyancy-driven convection is documented by the lack of type I striations in the space-grown crystals. The sulphur concentration is measured by spatially resolved photoluminescence. As expected, the macrosegregation can be described by a pure diffusion-controlled model which is in good agreement with the findings from the first German spacelab mission D1. Compared to the earth-grown reference samples, both of the space-grown InP crystals show strong disturbances such as inclusions and type II striations. The morphological instabilities are similar to growth disturbances already known from the space-grown MD-ELI-01 from the D1 mission.

Long-term crystal growth experiments were successfully performed under microgravity conditions during the first flight of the unmanned mission EURECA-1 in the automatic mirror furnace (AMF). Two crystals of Al x Ga 1-x Sb with [001] and [111] orientation respectively were grown from gallium solution by the travelling heater method. The grown length of the single crystals is 4.0 mm and 4.2 mm respectively. The space-grown samples show a high microscopic homogeneity which indicates the absence of time dependent convection. From pulse markers a constant growth rate of 0.6 ± 0.1 μm/min is measured which is lower than 0.8 ± 0.1 μm/min obtained in earth grown reference samples. Details about the experiment performance and the growth results are given.

Crystals of the semiconductor Cu0.85In1.05Se2 were grown from the melt. The crystals were characterized by scanning electron microscopy as well as transmission electron microscopy. A coexistence of two phases was found in the entire boules which was detected by energy dispersive spectroscopy and by electron diffraction. High resolution electron microscopy showed a direct intergrowth of both phases.

Surfaces of large single crystals of CdTe grown from the vapour having no contact with the ampoule wall have been investigated by scanning electron microscopy. The morphology of the growth interface has been studied as a function of several growth parameters, such as temperature field, supersaturation and crystal orientation. The formation of flat facets, depressions and pipes parallel to the growth direction is discussed in relation to the growth conditions.

We investigated the optical properties of defects in CdTe and ${\mathrm{Cd}}_{1\mathrm{\ensuremath{-}}\mathit{x}}$${\mathrm{Zn}}_{\mathit{x}}$Te (0x1). Residual impurities give rise to specific far infrared absorptions, while metal vacancy-donor complexes (A centers), identified by optically detected magnetic resonance, are characterized by their near infrared (1.4 eV) photoluminescence (PL) properties. The specific zero-phonon-line positions and phonon couplings are worked out for these complexes involving different group-VII (F, Cl, Br, In) or group-III (In, Al) donors. In addition to the A center PL two emission bands are found at 1.135 and at 1.145 eV. The temperature dependences of the PL show that the 1.145 eV luminescence follows the temperature dependence of the band gap, while the energy position of the 1.135 eV band emission shifts to higher energies with increasing temperature. The A-center PL and the luminescence bands at 1.1 eV are investigated throughout the complete alloy composition range form x=0 to 1. The A center and the 1.135 eV band were found to follow the band-gap shift from CdTe to ZnTe, whereas the 1.145 eV luminescence keeps its emission energy constant.

Semi-insulting CdTe bulk crystals were grown from the vapour phase in both closed and semi-open arrangements. The results of the growth experiments are discussed in terms of various electrical and optical characterization methods. Van der Pauw measurements and time dependent charge measurements (TDCM) were used to determine the resistivity. Deep level defects were investigated by means of photoinduced current transient spectroscopy (PICTS). For one of the most important fields of application, detector spectra of the vapour phase material are measured and discussed.

The transition metal ions titanium and vanadium in CdTe are studied by optical absorption, electron paramagnetic resonance (EPR), magnetic circular dichroism (MCD) and optically detected magnetic resonance (ODMR) spectroscopy. Titanium is observed in two different charge states. Ti2+ (3d2) and Ti3+ (3d1) by EPR. Absorption measurements however show only the 3A2 → 3T1(3P) and 3A2 → 1E(1D) internal transitions of Ti2+. This assignment is confirmed by MCD and ODMR measurements. For vanadium-doped CdTe only the V3+ charge state is observed in EPR. In absorption two pronounced absorption bands are seen which are attributed to internal transitions of V3+.

CdTe: Cl crystals were grown from the vapour phase by the sublimation travelling heater method (STHM) in space on board of the unmanned EURECA I mission. They were characterized by time dependent charge measurement (TDCM) and photoinduced current transient spectroscopy (PICTS). Axial cuts of the space grown crystals investigated by TDCM reveal an axial symmetric resistivity distribution along the growth direction, which can be explained by segregation effects of chlorine.

The stability of an isolated macrostep under a temperature gradient is discussed theoretically for the growth of III–V-semiconductors from a metallic solution, such as liquid phase epitaxy (LPE) and the travelling heater method (THM). To find the distribution of solute at the surface of one macrostep, the two-dimensional diffusion equation is solved, taking into account anisotropic interface kinetics as an important boundary condition. By analyzing the influence of growth rates and additional temperature gradients at the growth face it is found for the behavior of a macrostep, that for decreasing growth rates and for increasing temperature gradients the macrostep gets unstable and disappears. Details about the calculation and a comparison with experimental results from space grown crystals are presented.

The vapour growth of CdTe:V and CdTe:Ti by the sublimation travelling heater method (STHM) is reported for the first time. Characterization of the crystals has been performed by time dependent charge measurement (TDCM) and photoinduced current transient spectroscopy (PICTS). Axial cuts of CdTe:V crystals showed a very homogeneous and high resistivity distribution in the range of 1010 ω · cm. The PICTS spectra depict three levels which can be attributed to transition metal doping. The expected deep donor level can be identified clearly to be at Ev + 0.95 and 0.94 eV in CdTe:V and CdTe:Ti, respectively.

High quality CdTe bulk crystals are difficult to obtain due to various types of defects such as twins, dislocations, inclusions, and sub-grain boundaries. In this paper, the influence of different dopants and growth conditions on formation, frequency, and stereometry of those defects are discussed. The crystals have been grown by means of a sublimation travelling heater method (STHM) using Bridgman seeds. As dopants, the halogens chlorine, bromine, and iodine as well as vanadium and titanium have been used. By means of quantitative near-IR polarizing microscopy, the mapping of long-range stress fields can be related to irregularities in crystal growth. Additionally, local stress birefringence induced by dislocations and inclusions has been observed. The occurrence of crystal perturbations in STHM and Bridgman material is compared.

Several CdTe crystals doped with vanadium (V) or titanium (Ti) were grown by a vertical Bridgman technique and the Sublimation Travelling Heater Method (STHM). Spatially resolved resistivity measurements were performed using the Time Dependent Charge Measurement (TDCM) method. It is shown that V doping leads to a more homogeneous resistivity distribution in contrast to Ti. Photoinduced Current Transient Spectroscopy (PICTS) has been used to characterize the deep levels. Up to three additional levels are introduced through the doping with transition metals.