Stefaan Forment

A comparative study of electrochemically formed and vacuum-deposited n-GaAs/Au Schottky barriers using ballistic electron emission microscopy (BEEM)

A comparative study between n-GaAs/Au contacts, formed by electrochemical deposition or by vacuum evaporation, is presented. The main parameter, the barrier height ?B, was determined using three methods, i.e.?classical current-voltage and capacitance-voltage measurements as well as STM-based ballistic electron emission microscopy (BEEM). The latter method allowed us to determine the distribution of ?B over the contact area on a nanometre scale and showed that the electrochemically made contacts are inhomogeneous. The main result, confirmed by the three methods, was that ?B was higher for the electrochemically deposited contacts than for the evaporated ones. This higher value is attributed to O- groups, present at the interface during the electrochemical metallization, and forming interfacial dipoles with the Au, leading to an increase of the barrier.

Design of an instrument for measuring the spectral bidirectional scatter distribution function

The spectral bidirectional scatter distribution function (BSDF) offers a complete description of the spectral and spatial optical characteristics of a material. Any gloss and color measurement can be related to a particular value of the BSDF, while accurate luminaire design with ray tracing software requires the BSDF of reflectors and filters. Many measuring instruments, each having particular advantages and limitations, have been reported in the literature, and an overview of these instruments is included. A measuring instrument that allows for an absolute determination of the spectral BSDF with a full three dimensional spatial coverage in both reflectance and transmittance mode, a broadband spectral coverage, a large dynamic range, a reasonable acquisition time, and a large sample illumination area is presented. The main instrument characteristics are discussed, and the measurement capabilities are illustrated.

The effect of Schottky metal thickness on barrier height inhomogeneity in identically prepared Au/n-GaAs Schottky diodes

We have studied identically prepared Au(5 nm)/n-GaAs (35 dots) and Au(65 nm)/n-GaAs (38 dots) Schottky barrier diodes (SBDs) on the same n-type GaAs single crystal. A GaAs wafer has been prepared by the usual chemical etching, and evaporation of the metal has been carried out in a conventional vacuum system. The effective Schottky barrier heights (SBHs) and ideality factors obtained from the current–voltage (I–V) characteristics have differed from diode to diode. The SBH for the Au(5 nm)/n-GaAs diodes have ranged from 0.839 to 0.943 eV and the ideality factor n from 1.011 to 1.150. The SBH for the Au(65 nm)/n-GaAs diodes have ranged from 0.828 to 0.848 eV and the ideality factor n from 1.026 to 1.069. Our aim is to find the laterally homogeneous SBH values of the SBDs depending on Schottky metal thickness. The lateral homogeneous SBH values of 0.940 eV for the Au(5 nm)/n-GaAs and 0.866 eV for the Au(65 nm)/n-GaAs diodes have been calculated from a linear relationship between barrier height (BH) and the ideality factor, which can be explained by lateral inhomogeneities of the SBH, respectively.

Influence of hydrogen treatment and annealing processes upon the Schottky barrier height of Au/n-GaAs and Ti/n-GaAs diodes

After annealing in a H2 atmosphere at different temperatures, 100-oriented n-GaAs substrates were metallized with Au and Ti layers of different thicknesses to form Au/n-GaAs and Ti/n-GaAs Schottky diodes. The Schottky barrier height (SBH) variation and its dependence on subsequent N2 annealing for these Schottky diodes have been studied by different measurement techniques (I–V, C–V and BEEM) to obtain reliable values. These methods show that pre-metallization annealing leads to a less homogeneous metal semiconductor (MS) interface. In case of thick Au layers the effective barrier height is reduced as soon as the H2 annealing temperature reaches 300 °C. However, GaAs samples covered with thin Au layers or Ti layers do not exhibit such a barrier height reduction. The lower value in the case of thick Au layers is attributed to H+ groups, present at the interface due to the annealing in H2 atmosphere, forming interfacial dipoles with Au, leading to an inhomogeneous barrier and a decrease of the effective barrier height. It seems that these dipoles disappear again in the case of thin Au layers or are not formed with Ti. Post-metallization N2 annealing at higher temperatures lowers the barrier height for all samples. The resulting barrier inhomogeneities are explained and analysed using the bond polarization theory of Tung (2001 Phys. Rev. B 64 205310, 2001 Mater. Sci. Eng. R 35), a recent Schottky barrier formation model and the pinch-off model (Tung 1992 Phys. Rev. B 45 13509).

Deep level transient spectroscopy study of nickel-germanide Schottky barriers on n-type germanium

Nickel-germanide Schottky barriers have been made on n-type germanium and evaluated by deep level transient spectroscopy in order to detect possible metal indiffusion during the 30s rapid thermal annealing (RTA) employed for the germanidation. It is shown that while no electron traps have been found for the 300 and 350°C RTA step, the double acceptor level at EC−0.3eV of substitutional nickel was observed for the 400 and 450°C samples. The corresponding concentration profile increases exponentially towards the surface from which an effective diffusion coefficient of ∼5×10−10cm2∕s at 450°C has been derived.

Electrochemical formation and properties of n-GaAs/Au and n-GaAs/Ag Schottky barriers: Influence of surface composition upon the barrier height

The barrier height ΦB of electrochemically formed n-GaAs/Au Schottky barriers is investigated as a function of deposition potential VD . Two potential regions can be distinguished: one in which ΦB is comparable to the value of the contacts made by evaporation and another one in which ΦB is higher. This is explained on the basis of a different surface composition of GaAs in the respective potential regions. At more positive VD , gold will be deposited upon a surface at which the Ga atoms are –O− terminated. This leads to the formation of a Auδ+–Oδ− dipole layer at the interface and hence to an increase in ΦB . At more negative VD , metallic Ga will be formed during the cathodic decomposition of n-GaAs. As a consequence, surface Ga–O− groups will no longer be present, no surface dipole layers will be formed and ΦB will be lower. This interpretation is supported by results, obtained on n-GaAs/Ag Schottky barriers formed in solutions with different pH.

Stray light performance of a combined monochromator–spectrograph UV irradiance measuring instrument

Critical UV applications such as disinfections and dermatological treatments call for accurate UV spectral irradiation measurements. Single monochromators and spectrographs are not able to provide satisfying accuracy due to poor stray light rejection. The main drawbacks of a double monochromator scanning system are the need for motion synchronization, the inconvenience in performing field measurements and the time-consuming scanning approach. In this paper, a combination of a monochromator with a large bandpass and a spectrograph is proposed, which allows for accurate UV spectral irradiance measurements with short measurement times of less than 1 s, and yet with acceptable stray light rejection. Stray light levels are determined for both monochromatic and heterochromatic primary radiation using slit functions and cut-off filters. These experiments revealed a considerable UV stray light response reduction for the monochromator–spectrograph combination compared to single spectrograph instruments. The stray light performance in the 250–400 nm range matches with results obtained with double monochromator instruments. With this setup, the quality of a double dispersing instrument is combined with the quick data acquisition of a spectrograph.

Point-Defect Generation in Ni-, Pd-, and Pt-Germanided Schottky Barriers on N-Type Germanium Substrates

In this paper, the creation of point defects in n-type germanium by Ni-, Pd or Pt-germanidation is investigated by means of Deep Level Transient Spectroscopy (DLTS). The germanidation is achieved by a Rapid Thermal Annealing (RTA) step between 300 and 500oC of a 30 nm sputtered metal layer. Contrasting behaviour is found between Ni, on the one hand, and Pd and Pt, on the other: in-diffusion of nickel is found starting from 400oC, while in the second case, vacancy-related deep levels have been observed at the lower RTA temperatures. Evidence will be provided that these defects are most likely formed during the sputtering of the heavy metal atoms, introducing radiation damage in the Ge substrate. Finally, the impact of these deep levels on the current-voltage (I-V) characteristics of the obtained metal-germanide Schottky barriers will be discussed.

Near-field and far-field goniophotometry of narrow-beam LED arrays

In lighting calculations and simulations, the emission of a light source is conventionally modeled using the far-field luminous intensity distribution. However, the advent of luminaires including large arrays of LEDs with focusing optics creating narrow beams has made the traditional limiting photometric distance to reach far-field conditions less easy to determine. Furthermore, even correct far-field data can lead to erroneous predictions when illuminances are determined on a task surface which is positioned within the near-field region. A near-field representation could overcome these problems, but experimental validation for such LED arrays is lacking. This paper reports on near-field and far-field laboratory experiments using an array of two and five narrow-beam LEDs. A near-field approach makes discussions to determine the far-field photometric distance superfluous and leads to correct illuminances at any location with respect to the array, irrespective of the dimensions of the array and the beam angle of the individual components. Introducing the near-field representation of light sources in lighting design offers more accurate predictions when luminaires based on LED arrays with focusing optics are involved.

Characterization of Ge Implanted with Ni and Hf Ions

Since the early days of Ge research, the impact of transition metals on the electrical characteristics (resistivity, carrier lifetime, ...) has been well documented [1]. The development of advanced CMOS on Ge substrates has renewed the interest of metal contamination during processing on the properties of the material. Besides the traditional fast diffusing transition metals (Ni, Cu and Fe) other less common metals are nowadays also of interest: the implementation of so-called high-k gate dielectrics introduces metals like Hf, La,... in the process flow [2]. However, not much is known about the properties (diffusivity, deep levels) of these metals in Ge. In an attempt to fill the gap in our knowledge, two doses of Ni and Hf have been introduced in 100mm diameter nand p-type Ge substrates by ion implantation, placing the peak concentration at 45 and 43nm deep, respectively. Annealing has been performed at 350C for 1min (Ni) and 500°C for 5min (Hf) and at 700C for 1hr. The latter anneal serves to activate the metals and remove the implantation damage. A combination of chemical (SIMS), structural (TEM) and electrical techniques (Microwave Absorption lifetime measurements, DLTS) has been applied. As will be shown, both Ni and Hf reduce significantly the carrier lifetime, which can be ascribed to the deep levels found by DLTS and the extended defects observed in TEM. Both SIMS (Secondary Ion Mass Spectroscopy) and x-TEM (cross section Transmission Electron Microscopy) will probe the upper layer of the substrate. Hf was detected by TofSIMS. The as-implanted profile corresponds to a typical implantation profile. Ni was detected by SIMS and as expected the peak position was ~50nm for the as-implanted sample. When annealing at 350°C, a segregation towards the substrate takes place as well as a loss of the original dose. The sample annealed at 700°C shows an unexpected profile: the dose that can be extracted from the profile is higher than the dose for the as-implanted sample. By further analysis using other techniques (TEM and optical microscopy), we could detect some extended defects at the surface of this sample. These defects can play a role in the SIMS measurement but do not clarify the origin of the profile. Further investigation is needed. TEM images were obtained for the Ni implanted samples on n-type Ge. A peak concentration of 5.10 at/cm Ni atoms at a position of 45nm, results in a 200nm amorphous Ge layer. When annealing these samples at 350°C (1min), a 30 nm amorphous layer remains and a high density of defects is still present at the amorphous/crystalline interface. When annealing at 700°C, the Ge crystal is fully regrown but large defects are observed at the surface. The formation of NiGe precipitates could lay at the origin of these defects. However, further investigation is needed to support this assumption (XRD measurements). TEM images were also made for the Hf implanted in n-type Ge. Here, 2 series of samples were studied: samples with a peak concentration at 5.10at/cm and at 5.10at/cm. Some differences are observed. The lower concentration gives rise to a partially amorphous layer of 50nm. While the higher concentration shows a fully amorphous layer of 85nm. When annealing the low-dose material, the Ge lattice is recrystallized but small defects are still present in the surface layer. The higher dose, gives rise to fully recrystallized Ge. Both series are fully recrystallized at 700°C. These physical observations can now be linked to some electrical measurements. MCLT (Minority Carrier Lifetime) was measured by microwave photoconductivity. A passivation layer was applied to reduce surface recombination effects. Ge was implanted with Ge to study the influence of the damage on the LT. A Ge sample that was not implanted nor annealed acts as a reference. After implantation, the LT is determined mainly by the amorphous top layer. The LT is significantly reduced due to this damage. When annealing these samples at intermediate temperature, the damage is significantly reduced and this explains the increase in LT. Annealing at 700°C restores the LT back to the original level. The same trends are observed for the Ni and Hf implanted samples. However, annealing at 700°C restores the lattice but is also activating the metal impurities and the life time drops significantly. This indicates that both Hf and Ni impurities introduce trap levels in Ge and are able to influence the MCLT when activated. A more direct way to study these levels introduced by metals in Ge, is by DLTS (Deep Level Transient Spectroscopy). Measurements are planned in the future and will contribute to a deeper insight in the position of these trap levels. To conclude a study was performed on the influence of metals on the physical characteristics of Ge. It can be seen that damage introduced by implantation is determining the lifetime of the carriers. The damage can be restored by annealing but the temperature on which full recrystallization is seen, depends on the metal implanted and on the implantation dose. When restoring the Ge lattice, MCLT values are increasing. For all implantation conditions, full recrystallization is achieved by annealing at 700°C (1h). Only for the Ni implantation, extended defects were seen at the surface. The origin of these defects is yet unclear but the formation of NiGe precipitates could explain this observation. Both Ni and Hf are introducing trap levels in Ge. This can be seen from the fact that after full recrystallization, traps are activated and MCLT values drop significantly for both metals. A further study by DLTS on the position of these trap levels is currently been undertaken.