Dominant Impurity Research Articles

Growth of GaN on Si substrates – roles of BP thin layer

Abstract Materials for short-wavelength laser emitting devices grown on silicon substrates for opto-electronic coupled integrated circuits have been investigated. One of the most promising materials for the laser is GaN and its mixed alloys. Potential buffer materials that can be grown on Si substrates and that provide a lattice match with GaN are surveyed. Boron monophosphide (BP) is one of the best materials for such a buffer layer. In this paper, we describe the growth of BP epitaxial layers on Si(1 0 0) substrates with area as large as 10×10 mm 2 . The dominant impurity in the BP layers, measured by SIMS analysis is silicon (10 19 atoms / cm 3 ) . As a preliminary experimental, GaN was deposited on the BP/Si substrates at various growth temperatures. Cubic type GaN has a tendency to grow dominantly below 850 °C. The growth of BP on Si and GaN on BP/Si will be discussed.

Growth of c-GaN on Si(100)

One hundred percent cubic GaN with 2.5 μm thickness was successfully obtained on Si(100) substrates using BP thin buffer layer. BP buffer layer was grown on silicon(100) substrates with a dimension of 10×10 mm 2. The dominant impurity in BP layer incorporated during epitaxial growth was silicon. The conduction type is readily varied depending on gas flow ratio BCl 3/PCl 3. GaN was grown of BP/Si(100) using trimethylgallium and monomethylhydrazine as a starting materials. The growth of BP and GaN and their characterizations will be discussed.

Small-angle X-ray scattering in type Ia diamonds

The results of a small-angle X-ray scattering (SAXS) investigation of natural and synthetic diamonds, with different types of nitrogen-related defects, are presented. They reveal a presence of impurity-rich spherical clusters with a diameter of approximately 90 Å in an IaB diamond. SAXS scattering in an IaA diamond has a much lower intensity, and originates from larger clusters with a considerable variation in size. It is likely that the dominant impurity in these clusters is nitrogen. Present data suggest the possibility that some of the nitrogen-related point defects (A- and B-centers) are not randomly distributed in the diamond lattice, but instead are concentrated in clusters. It is suggested that, although optical manifestations of the A-nitrogen are similar in different diamonds, IaA diamonds can be divided into two sub-groups. The IaA 1 group includes annealed, synthetic, and some natural diamonds. They contain randomly distributed nitrogen pairs, which do not influence mechanical properties and do not produce X-ray scattering. In the IaA 2 group, the diamond aggregation process went further and the formation of nitrogen-containing clusters began. The process of nitrogen aggregation is described in terms of the decay of a supersaturated solid solution of nitrogen in diamond. The driving force for the creation of impurity clusters is the lowering of the free energy of nitrogen dissolution in the diamond lattice, and the decrease of strain.

XPS study of Nb-doped oxygen sensing TiO 2 thin films prepared by sol-gel method

Titanium dioxide (TiO 2) thin films have been prepared using the sol-gel method and subsequently doped with niobium oxide (Nb 2O 5) for use in oxygen sensing applications. The chemical composition of the resultant film sensor surface has been investigated using X-ray photoelectron spectroscopy (XPS). Films are essentially stoichiometric with carbon as the dominant impurity at the surface. The film electrical resistance has been examined for detection of oxygen at concentrations of 1 ppm to 1%. Doping resulted in a 40% increase in the oxygen gas sensitivity of the thin films at an operating temperature, as low as 190°C.

Ore petrology, chemistry, and timing of electrum in the Archean hypozonal Transvaal lode gold deposit, Western Australia

The Transvaal lode gold deposit is hosted in ultramafic schists and graphite-bearing pelites that are metamorphosed to lower and mid-amphibolite facies assemblages. The orebody is structurally controlled by the Transvaal shear zone which overprinted and reactivated the penetrative regional fabric. The main tabular orebody is formed of a massive quartz sulfide vein and related splays and adjacent hydrothermally altered wall rock. Prograde alteration in the ultramafic rocks consists of a proximal diopside-actinolite-quartz zone which grades outward into a distal amphibole (hornblende-actinolite)-biotite-plagioclase-quartz schist. In pelitic metasedimentary rocks, a proximal biotite-muscovite-quartz-graphite zone grades outward into a distal cordierite-quartz-biotite-graphite schist. Retrograde chlorite, talc, and zoisite replaced prograde hydrothermal alteration minerals and exhibit generally postkinematic fabrics. The dominant sulfide assemblages in the hydrothermal alteration zones at Transvaal are pyrrhotite-loellingite, pyrrhotite-arsenopyrite 1, and loellingite-arsenopyrite 1. Other assemblages include chalcopyrite-pyrrhotite, pyrite-loellingite, and pyrrhotite-cubanite. These assemblages are interpreted to consist of a mixture of equilibrium assemblages that relate to a prograde hydrothermal alteration event and disequilibrium assemblages that formed during cooling and reequilibration of the rocks. Arsenopyrite 1 is interpreted to have formed both during the prograde alteration event and during high-temperature retrogressive replacement of loellingite. Average compositions of arsenopyrite 1 in equilibrium with loellingite and pyrrhotite indicate paleotemperatures of about 510 degrees + or - 20 degrees C, about 40 degrees C lower than estimates of the peak metamorphic temperatures. Low and high Fe pyrrhotite and arsenopyrite 2, for instance, would not be in equilibrium with other sulfides and are interpreted to be retrograde.Both electrum and gold occur in the Transvaal gold deposits but electrum is the more abundant gold-bearing phase and contains between 24 and 56 wt percent silver. The compositions of electrum grains correspond to gold finenesses of about 431 to 732 with silver the dominant impurity. Electrum occurs mostly at the interface between arsenopyrite 1-loellingite, or as inclusions in loellingite, pyrrhotite, and arsenopyrite 1. More rarely, it occurs as as fracture fill within arsenopyrite 1. The differing silver contents of electrum can be explained by differing fluid-rock reactions. Low Ag electrum (about 24 wt % Ag) occurs where the lode is entirely hosted by ultramafic schists, whereas high Ag electrum is at the ultramafic schist-graphitic-pelite contact. It is proposed that fluid interaction with reducing pelites produced Ag-rich electrum through destabilization of gold bisulfide and silver chloride complexes, whereas fluid that reacted with more oxidized ultramafic rocks formed Au-rich electrum as silver chloride complexes were not destabilized.The main phase of hydrothermal alteration, which included the crystallization of prograde silicates, pyrrhotite, and loellingite, was syntectonic with respect to the Transvaal shear zone and at approximately peak metamorphic conditions. Electrum in equilibrium with pyrrhotite and/or loellingite was thus deposited at synpeak metamorphic conditions. However, the majority of electrum grains coexist with arsenopyrite 1 and composite loellingite-arsenopyrite 1 grains and are interpreted to have formed from submicroscopic gold released from loellingite during high-temperature retrogressive replacement by arsenopyrite. These are also considered to be part of the main gold-bearing hydrothermal event.Retrograde low-temperature, postmetamorphic silicate and sulfide minerals such as chlorite, muscovite, pyrite, and cubanite are evidence for processes that occurred after the main, gold-depositing hydrothermal activity took place. Renewed flux of hydrothermal fluids during the protracted metamorphic history of the terrane could have produced the low-temperature retrogression; the occurrence of low-temperature sulfides is likely the result of in situ reequilibration during cooling of the rocks.

Effect of Thin-Film Texture and Zirconium Diffusion on Reliability against Electromigration in Chemical-Vapor-Deposited Copper Interconnects

The effects of film texture and zirconium diffusion on the reliability of chemical-vapor-deposited (CVD) copper interconnects were examined. The crystallographic orientation of CVD copper depends strongly on the substrate structure and pre-treatment sequences. Sputter pre-etching of TiN liner greatly enhanced the (111) orientation of sputter deposited copper film. The CVD copper grew epitaxially on the (111) oriented sputtered copper seed layer. Dominant impurities in CVD copper are oxygen and fluorine, which can be reduced by hydrogen annealing at 400°C. By improving the texture, the lifetime against electromigration extends about twice longer than that of as-deposited CVD copper film. Copper zirconium alloy wire with a surface layer of the inter-metallic compound CuZr2 further extended the electromigration lifetime due to thermal diffusion of zirconium into copper. The electromigration lifetime of this alloy wire was significantly improved with only a small increase in electrical resistivity.

High resolution Auger electron spectroscopy studies on (100) and (111) facets of chemical vapor deposited diamond

Diamond thin films were grown by microwave plasma and hot filament chemical vapor deposition (MPCVD and HFCVD, respectively) techniques. Films were systematically characterized by x-ray diffraction, micro-Raman spectroscopy, scanning electron microscopy (SEM), and Auger electron spectroscopy (AES). Although the results obtained using various characterization techniques are broadly similar, there are however subtle differences. For instance, Raman spectra show a sharp peak at ≃ 1332 cm−1 corresponding to natural diamond in both types of films. The intensity and the position of the non-diamond band in the two sets of films differ. While the maxima of the non-diamond band in HFCVD film lies at 1450 cm−1, in MPCVD film it occurs at 1525 cm−1. Also the values of FWHM in HFCVD film (≃ 7.5 cm−1) are smaller than the MPCVD films (≃ 9.5 cm−1). This may indicate that the concentration of non-diamond carbon impurities on the grain boundaries of HFCVD films are really small. SEM results on the other hand indicate that the grain size of the MPCVD films is larger than HFCVD films. AES was performed in a survey scan (beam size ∼10 μm × 8 μm) and high resolution (beam size ≃ 0.2 μm) mode with an initial aim to investigate the surface characteristics and environment of carbon atoms of the diamond films. In the survey scan, the spectra show a line shape typical of CVD diamond films. Anomalous results were obtained when the AES was performed on (100) and (111) facets in high resolution mode. This may be explained in terms of the surface reconstruction taking place due to hydrogen desorption via core-hole Auger decay process. Auger depth profiles were also obtained on the facets which reveal that Si, O, and N are the dominant impurities. The impurity content of HFCVD films is observed to be lower by a factor of 2 as compared to MPCVD films.

Ab initiostudy of point defects inCdF2

The plane-wave pseudopotential method is used to study point defects in ${\mathrm{CdF}}_{2}$. We present comprehensive results for the native defects as well as for dominant impurities. In addition to ${\mathrm{F}}_{i},$ ${\mathrm{V}}_{\mathrm{Cd}}$ and ${\mathrm{O}}_{\mathrm{F}}$ were found to be easily formed compensating acceptors. For In and Ga impurities the experimentally observed large Stokes shift could not be established, and the results rule out symmetric atomic relaxation as the mechanism leading to the bistable behavior. The limitations of the present approach utilizing density-functional theory and the local-density approximation in the case of ionic materials are addressed.

Grain boundary self-diffusion in Cu polycrystals of different purity

The temperature dependence of grain boundary (GB) self-diffusion in Cu polycrystals was systematically investigated using the 64Cu radiotracer and the serial sectioning technique. Two different high-purity Cu materials were used: 99.9998% Cu and 99.999% Cu. The difference in impurity content was checked by time-of-flight secondary-ion mass spectrometry (SIMS). The two materials were studied in the temperature ranges 1066-720 K (99.9998% Cu) and 973-784 K (99.999% Cu). A significant dependence of the GB self-diffusion on the material purity was observed, resulting in a different Arrhenius behaviour of the diffusivity δD GB where δ is the GB width and D GB is the GB diffusion coefficient. The Arrhenius parameters are 3.89 × 10 −16 m 3/s and 72.47 kJ/mol for the highest purity material, and 1.16 × 10 −15 m 3/s and 84.75 kJ/mol for the high purity material. The differences in GB diffusion coefficients and activation enthalpies are explained in terms of strong impurity-Cu atom bonds in the boundary. In the present investigation sulphur is the dominant impurity. Remarks are made relating the impurity segregation and the reduction of the GB diffusivity to the lifetimes of thin film metallic interconnects determined by electromigration effects in GBs. Using a semi-empirical relation of Borisov et al. the GB energy, γ GB, of large angle boundaries and its temperature dependence were determined from the self-diffusion data of Cu. The resulting values of γ GB (e.g. 650 mJ/m 2 at 1000 K) are in good agreement with GB energy measurements and model calculations for Cu.

Impurities and impurity transport in the spheromak SPHEX

Absolute spectroscopic measurements are made of all significant ionization states of the dominant impurities in SPHEX during sustained operation, and these are compared with a diffusive transport model to derive average plasma parameters. Statistical tests are used to establish the reliability of these results. A line ratio measure is also made and compares well with the ionization balance. The electron power balance in these discharges is estimated assuming that only Spitzer dissipation (a small fraction of the total source power) heats the electrons. Titanium gettering is used to reduce recycling, and thus to measure the fraction of impurity due to recycling relative to that directly injected from the Marshall gun. It is established that during sustainment the electrons are radiation dominated everywhere, so that the rise in when Ti is used is quite moderate, from 30 to 50 eV. However, the impurity transport rate is substantially reduced by gettering, implying that high transport rates are not intrinsic to the spheromak configuration. Some more qualitative data pertaining to the decaying spheromak is also presented.

The measurement of the intrinsic impurities of molybdenum and carbon in the Alcator C-Mod tokamak plasma using low resolution spectroscopy

The intrinsic impurity content of molybdenum and carbon was measured in the Alcator C-Mod tokamak using low resolution, multilayer mirror (MLM) spectroscopy ( Delta lambda ~1-10 AA). Molybdenum was the dominant high-Z impurity and originated from the molybdenum armour tiles covering all of the plasma facing surfaces (including the inner column, the poloidal divertor plates and the ion cyclotron resonant frequency (ICRF) limiter) at Alcator C-Mod. Despite the all metal first wall, a carbon concentration of 1 to 2% existed in the plasma and was the major low-Z impurity in Alcator C-Mod. Thus, the behaviour of intrinsic molybdenum and carbon penetrating into the main plasma and the effect on the plasma must be measured and characterized during various modes of Alcator C-Mod operation. To this end, soft X-ray extreme ultraviolet (XUV) emission lines of charge states, ranging from hydrogen-like to helium-like lines of carbon (radius/minor radius, r/a~1) at the plasma edge to potassium to chlorine-like (0.4<r/a<0.6) and magnesium- to sodium-like (r/a<0.4) lines of molybdenum in the main plasma, were measured using a novel, low resolution, photometrically calibrated polychromator with MLMs as dispersive elements. The MLM spectra were investigated in detail, and comparisons with high resolution spectroscopy were made. The utility of low resolution spectroscopy to diagnose tokamak plasmas is presented, and meaningful information about impurity behaviour was obtainable owing to the specific choice of the observed spectral regions. Ab initio physics rates from the HULLAC atomic physics package were input into the collisional radiative (CR) model and the multiple ionization state transport (MIST) code, and both MIST and the CR model were used in the interpretation of the molybdenum spectrum. The carbon spectrum was interpreted using the MIST code and the direct impact rates of Itikawa (Itakawa, Y., et al., At Data Nucl. Data Tables 33 (1985) 149), which were incorporated into the collisional radiative model. The intrinsic i

Experiments in FTU with different limiter materials

Over the last few years, a great deal of effort has been devoted to solving the problem of power and particle handling in divertors, which has been recognized as a critical issue for the operation of a magnetic fusion reactor. In particular, the choice of materials for plasma facing components has been examined with a view to developing heat and erosion resistant materials for divertor target plates. A large database on the behaviour of low-Z (carbon or beryllium) materials in tokamaks is available, while for high-Z materials there is little experience in the present generation of magnetic fusion devices. Frascati Tokamak Upgrade (FTU), a high field compact tokamak, has devoted part of its experimental campaign to studying the plasma characteristics when its limiter material is changed from the usual Inconel (nickel) to molybdenum and tungsten. Siliconization of the machine has also allowed the comparison of plasma performance when a relatively low-Z (silicon) ion is the dominant impurity. In this article, results are reported concerning the plasma operation, the differences in plasma characteristics and radiation losses, the impurity generation mechanisms and the relative impurity concentrations in the core plasma. A simulation of the experimental results, made with a self-consistent edge-core coupled model is presented, in order to provide evidence of the main physics mechanisms responsible for the observed behaviour

Growth of ZnTe by metalorganic vapor phase epitaxy: Surface adsorption reactions, precursor stoichiometry effects, and optical studies

The metalorganic vapor phase epitaxy growth of ZnTe by di-isopropyl-telluride and di-methyl-zinc (Me2Zn) precursors is investigated by studying the epilayer growth rate as a function of both growth temperature and precursor transport rates. The ZnTe growth is a thermally activated process involving the heterogeneous pyrolysis of both Zn and Te alkyls onto the ZnTe surface. The growth rate dependence on growth conditions is explained in terms of surface adsorption-desorption reactions, assuming that the incorporation of Zn and Te atoms into ZnTe takes place through their selective adsorption on different surface lattice sites. There is also evidence that the occurrence of a competitive species for the surface adsorption of Zn atoms, which is identified as the CH3⋅ (methyl) radical, is produced by the pyrolysis of Me2Zn. Photoluminescence (PL) and absorption measurements performed on ZnTe allow to identify two new donor-acceptor pair (DAP) bands, originated from the recombination of a Ga donor with two acceptor centers, whose ionization energies are 56 meV for the higher energy band and around 140–150 meV for the lower energy one. Hall measurements show that the 56 meV acceptor is responsible of the p-type conductivity of the layers. The nature of the impurities originating such PL features is discussed with the support of secondary ion mass spectrometry. It is shown that Ga, Si, and C are dominant impurities in the layers, whereas Cu does not occur in our ZnTe. Unintentional C doping occurs in ZnTe as a consequence of the strong methyl and iso-propyl radical surface adsorption. We show that C is incorporated as an acceptor in ZnTe, originating the DAP bands observed in the PL spectra. Within this view, the 56 meV ionization energy acceptor is tentatively assigned to substitutional C atoms on Te lattice sites.

Soft X-ray induced darkening of glass surfaces

A study of the coloration of glass surfaces induced by soft X-rays was made using X-ray photoelectron spectroscopy and optical absorption measurements. Sodium was found to be the dominant metal impurity in the glass that exhibits the surface coloration. An analysis of the relative atomic concentration of the impurity shows an increase of sodium with increasing exposure of the glass surface to the soft X-rays. The results suggest that the color-center sites involve trapped photoelectrons near the surface region, which can cause mobile sodium ions to move toward the exposed glass surface.

Effect of the oxide-desorption temperature on the substrate–epilayer interface charge in organometallic vapor-phase epitaxy of GaAs

We found that the use of relatively low temperatures during oxide desorption in organometallic vapor-phase epitaxy of GaAs reduced the epilayer–substrate interface charge. No differences in Hall mobility were found between the samples grown with 550 and 650 °C desorption. O, C, Si, and S were the dominant impurities at the interface. Si and S, which are shallow donors, induce the interface-charge accumulation. Low-temperature desorption leaves oxygen at the surface, suppressing the activation of shallow dopants. A high level (≳1×1012/cm2) of Cl was detected on every GaAs substrate we measured, but its effect on epitaxy is not clear.

New Developments of Less Toxic Group-V Precursors for the Metalorganic Vapour Phase Epitaxy of III–V-Semiconductors: In -Situ-Formation of As–H Functions by Thermal β-Elimination of Specific As-Trialkyl Compounds

Thermal decomposition studies and low pressure metalorganic vapour phase epitaxy (MOVPE) growth experiments have been performed using novel, less toxic arsinetrialkyl sources, which decompose by the β-hydride elimination process. Therefore, As–H functions are automatically formed in the hot temperature zone of the MOVPE reactor to enable high quality layer deposition. Decomposition studies have been carried out under growth conditions in the MOVPE reactor by quadrupole mass spectroscopy (QMS). The β-elimination decomposition of the novel As-sources diethyltertiarybutyl-arsine (DE t BAs) and diethylisopropyl-arsine (DE i PrAs) is proven by the detection of the cracking products diethylarsine (DEAsH) and the corresponding alkene fragments isobutene ( C4H8) and propene ( C3H6), respectively. The alkene-to-radical formation for different As- and P-precursors indicate a β-elimination efficiency of about 60% for tertiarybutyl- and 40% for isopropyl-groups, whereas for ethyl groups no significant indication of a β-elimination has been observed. GaAs bulk layers were grown by using trimethylgallium (TMGa) in combination with the β-eliminating As-trialkyl sources DE t BAs and DE i PrAs, respectively, as well as by using TEAs or DEAsH, both decomposing under ethyl-radical formation. In addition, the partially fluorinated source pentafluorethyl-diisopropyl-arsine (PFED i PrAs) has been investigated, however, only an etching of the GaAs substrate surface could be detected independently of the growth conditions. The use of β-eliminating As-trialkyl compounds leads to significantly improved layer properties as compared to TEAs-grown structures. GaAs layers, grown with DE t BAs or DE i PrAs at substrate temperatures of 600–650° C and V/III ratios between 8 and 80 exhibit p-type behaviour with net carrier concentrations of 3–5×1015 cm-3 and hole mobilities of 350–400 cm2/ V·s at 300 K and 3500–4100 cm2/ V·s at 77 K. Carbon is the dominant acceptor impurity in GaAs layers at present, as determined by PL spectroscopy. These results indicate a promising potential of the class of β-eliminating As-compounds as replacements for arsine in the MOVPE.

Preparation and photoluminescence study of high purity CdTe single crystals

CdTe single crystals were grown by vertical Bridgman (VB) and physical vapor transport (PVT) methods. The photoluminescence was measured at 4.2 K to evaluate the crystals and to examine the behavior of residual impurities and the effect of Cd purity. The segregation of dominant residual impurities was confirmed in VB as-grown crystals. The purification effect of PVT was also confirmed. The dominant emission line (A0,X) at 1.5896 eV in the PL spectrum, related to CuCd or VCd, has been found to have a doublet structure shared by two origins, though a complete understanding is still lacking.

Growth and Characterization of GaAs Epitaxial Layers by MOCVD

Abstracttechnique, on both semi-insulating and semi-conducting CraAs substrates with (100) orientation, offset by 2° towards (110) direction. Systematic variation of As/Ga was performed to gain an understanding of growth process, type of formation and other related physical properties. The films were characterized by using the variety of techniques, such as SEM, EDAX, HRTEM, XRD, and PL. Optical and electrical properties of undoped CyaAs epilayers are presented with reference to the growth conditions and AsH3/TMGa ratio. Photoluminescence measurements of GaAs epilayers were recorded at 4.2K and shows the emission of free exciton and confirmed their high purity. The dominant residual impurities in GaAs are presented by using PL. Finally, electrochemical depth profiling exhibited almost homogeneous background carrier distribution and excellent abruptness between the thin GaAs epilayer and substrate.

SIMSによる高純度銅中の微量金属不純物の定量分析

Quantitative analysis of trace amounts of dominant impurities in copper, Ag, Fe, Ni and Pb, were studied using secondary ion mass spectrometry (SIMS). Based on the survey of the secondary ions species, which disturbed the detection of atomic ions of impurities, the removal of these disturbance is proposed as the two methods; applying the high resolution of mass spectrometry, and calculating the isotopic abundance. The precision, detection limit and relative sensitivity factor were experimentally discussed to establish the reliable quantitative SIMS analysis at the local area of high purity copper.

Influence of carbon and oxygen on phosphorus and aluminium co-gettering in silicon solar cells

The impact of carbon and oxygen on the performance of high efficiency photovoltaic devices is still poorly understood. As oxygen is one of the dominant impurities present in silicon, various applications require different levels of oxygen to improve the device performance. In this work, we have studied the effects of the oxygen concentration in silicon on the co-diffusion of phosphorus and aluminium by rapid and conventional processes. We will compare the effectiveness of the co-diffusion of PAl in a variety of silicon materials, as a function of temperature and duration of the thermal process in a classical and a rapid thermal furnace and examine the influence of carbon and oxygen on the gettering efficiency. In particular, we will show that the large enhancement of the minority carrier diffusion length (LD) due to this process can be related to the presence of oxygen and carbon, the influence of which during the thermal cycle is of importance.