Detailed microstructural studies were performed on Ni/Si ohmic contacts to silicon carbide in order to investigate the effect of initial Ni:Si ratio in as-deposited structures on the occurrence of characteristic defects in Ni silicide layers, such as voids, layer discontinuities, rough surface or rough interface. The chosen range of investigated Ni:Si ratios corresponded to delta-Ni2Si as a dominant phase after complete annealing sequence. Strong effect of the initial stoichiometry on the ohmic contact's microstructure was observed. The highest Ni concentration significantly lowered the temperature at which roughening of the surface and the interface occurred. The middle value of investigated concentrations resulted in the rough interface after high temperature annealing, while the lowest investigated Ni content preserved smooth interface but introduced large voids and layer discontinuities. After the first annealing step, gamma-Ni31Si12 and/or delta-Ni2Si phases were detected. In the ohmic contacts (after two-step annealing sequence), beside delta-Ni2Si, the metastable, high temperature phase theta-Ni2Si was detected (also referred to as Ni3Si2 center dot h). This phase can exist within a relatively broad range of Ni:Si stoichiometry. The stoichiometry change toward higher Si content, which occurs during high temperature annealing, was realized through this phase. Superstructures were detected in theta-Ni2Si (Ni3Si2 center dot h) and in-gamma-Ni31Si12 grains. The effect of the stoichiometry change on the morphology of the Ni silicide layers is discussed. (C) 2016 Elsevier B.V. All rights reserved.

The new approach to fabrication process of nickel-based ohmic contacts to silicon carbide (SiC) is presented. During the first annealing step (300 degrees C), the amorphous Ni-Zr layer retards diffusion between two nickel suicide layers, thus handling the contradictory requirements for optimal Ni:Si ratio. Different stoichiometry obtained in each silicide layer allows to preserve smooth interface with SiC and simultaneously to avoid relatively easily meltable Si-rich Ni-Si phases during high temperature annealing (1000 degrees C) and therefore prevents morphology degradation. After annealing at 1000 degrees C only one final nickel suicide layer is present and Zr atoms are agglomerated at its surface. Morphology of the final suicide layer is substantially improved when compared to typical Ni-based contacts obtained by similar high-temperature annealings. The improved microstructure of the ohmic contact is a promising advantage in terms of SiC devices reliability. (C) 2015 Elsevier B.V. All rights reserved.

Scanning distortions are a well-known issue in the scanning electron microscopy (SEM) [1, 2], and the problem is related to the focused ion beam (FIB) instrumentation as well [3]. The characteristic jagging of vertical edges in SEM images may disturb observations and impede imaging. The same effect occurring during Focused Ion Beam (FIB) operations causes imperfect patterning of designed shapes and affects the FIB micro/nano machining. In effect, it physically harms or damages patterned structures or devices.

Distribution of chemical composition in nickel-based ohmic contacts to n-type 4H-SiC was investigated with XEDS mapping performed on plan view and cross-sectional TEM samples. Obtained results indicate that local deviations in stoichiometry from that of Ni2Si phase significantly contribute to degradation of surface morphology.

The use of focused ion beam (FIB) for research or processing of nanostructures requires very accurate beam positioning. However, numerous reasons for beam-position fluctuations exist. When FIB is used for specimen imaging, then these beam fluctuations cause the image jitter, blur or specimen-edge deformation. Similarly, beam fluctuations decrease the spatial resolution of FIB-based technological processes of milling or deposition. The sources of fluctuations are electromagnetic interference (EMI), floor vibrations and airborne acoustic noise. Our work concerns acoustic noise impact on focused ion beam fluctuations. The measurements were carried out on Helios NanoLab 600 DualBeam system with ion and electron beam columns. Reference specimens were imaged by electron or ion beam while acoustic waves of different frequency, magnitude and direction were intentionally generated nearby the system. It was found that while EMI-related distortions are caused by a wide and continuous spectrum of frequencies, for acoustic noise the strong deformations of image occur only at several resonant frequencies (mainly in the range 100–400Hz). Comparison of results obtained for either electron or ion beam allowed to attribute different resonant peaks to various FIB-system components (ion column, electron column, specimen stage). Spectral analysis showed that resonant components of the acoustic noise surrounding the system cause beam-position fluctuations in the range of several nanometers, highly unfavourable for nanotechnological works on FIB. The noise is generated mainly by various parts of the system itself. A method was also developed to identify whether the observed beam-position fluctuations originate from acoustic noise or from electromagnetic interference. It was possible because electromagnetic field impacts charged particles along their entire path while the acoustic vibrations act only on the mechanic elements of the system. Therefore the electromagnetic fluctuations are dependent on the particle velocity (i.e. the beam energy) while the acoustic fluctuations are independent of it. It was found that distortions (of FIB image and of patterns performed by FIB technological processing) caused by ion-beam position fluctuations due to acoustic noise can be reduced. The reduction can be achieved by selection of appropriate parameters of FIB process e.g. working distance and scanning parameters (like scan rate and scan direction).

Cathodoluminescence and photoluminescence measurements are commonly accepted as revealing local properties of a specimen region excited by a beam of electrons or photons. However, in the presence of a strong electric eld (e.g. a junction) an electron (or light/laser) beam-induced current is generated, which spreads over the structure. A secondary non-local electroluminescence, generated by this current and detected together with the expected luminescence signal, may strongly distort measurement results. This was con rmed by cathodoluminescence measurements on test structures prepared by focused ion beam on AlGaAs/GaAs/InGaAs laser heterostructures. Methods for minimizing the distortion of measured luminescence signals are presented.

The capabilities and applications of the focused ion beam (FIB) technology for detection of an electrochemical signal in nanoscale area are shown. The FIB system, enabling continuous micro- and nanofabrication within only one equipment unit, was used to produce a prototype of electrochemical nanometer-sized electrode for sensor application. Voltammetric study of electrochemically active compound (ferrocenemethanol) revealed the diffusion limiting current (12 pA), corresponding to a disc (planar) nanoelectrode with about 70 nm diameter of contact area. This size is in a good accordance with the designed contact-area (50 nm × 100 nm for width × thickness) of the FIB-produced nanoelectrode. It confirms that produced nanoelectrode is working properly in liquid solution and may enable correct measurements in nanometer-sized regions.

Ni/Si multilayer contact structures to 4H-SiC after subsequent annealing steps are investigated with electron microscopy methods. After high temperature annealing step, specific defects in the contact structures are observed. The influence of phase transformations during annealings on the morphology on the contacts is discussed and the explanation of formation mechanism of voids within contact layer is proposed.

Cathodoluminescence (CL) in SEM and electroluminescence (EL) techniques are widely used for investigation of optical properties of electronic structures. It is assumed that the CL signal represents the local properties of the region irradiated by the electron beam. However, this assumption is true if there is no electric field in the excitation region. In the opposite case the electron-hole pairs generated by the electron beam are separated by the electric field and the local voltage source is generated. The voltage is distributed all over the structure through resistance paths and it causes a current flow, which presence affects the registered CL signal. A range of this effect depends on the resistance distribution within the structure and on the value of electron beam current. The range may be much longer than the diffusion length of minority carriers. When EL signal is measured in SEM, it represents the mean properties of the whole structure. The comparison of the CL results with the EL ones detected in SEM gives valuable information about the examined structures, as will be shown in the present investigations. The AlGaAs/GaAs heterostructures with 8 nm InGaAs quantum well have been examined. For that aim also special test structures for CL and EL measurements have been formed on standard epitaxial structures.

Transmission electron microscopy (TEM) techniques were used for characterization of annealing (400, 600 and 800 °C) influence on the structural properties of the HfO2 film (45 nm thick) deposited on Si substrate. Such structures are considered as high-k dielectric materials for application in novel semiconductor devices. The studies showed that independently of the annealing temperature a very thin and flat amorphous layer is formed between HfO2 layer and Si substrate. This result was also found in the non-annealed sample. EDXS examination confirmed that the stoichiometry for the hafnium oxide layer in each sample corresponds to 1:2 for Hf:O (i.e. to HfO2). TEM images revealed differences in the microstructure of HfO2 layers in annealed samples, however the layers have similar thickness and interface roughness in all studied samples.

Nickel silicide ohmic contacts to 4H n-SiC were investigated using electron microscopy. Ni/Si multilayer structures were fabricated using magnetron sputtering technique. The Ni to Si layer thickness ratio was chosen to achieve the stoichiometry of Ni,Si phase. The deposited structure was subjected to a two step annealing procedure. First annealing step was performed at 600 degrees C, the second at 1050 degrees C or 1100 degrees C. Microstructure and morphology after each annealing step were characterized using scanning and transmission electron microscopy. The specific voids and discontinuities of the layer were observed after annealing at high temperature. Phase compositions were investigated with electron diffraction technique. After annealing at 600 degrees C the phases Ni2Si, Ni3Si2 and Ni31Si12 were detected. High temperature annealing resulted in the presence of only Ni2Si phase. The influence of phase transformations on the morphology of the contacts is discussed. Explanation of the origin of layer discontinuities is proposed. [doi:10.2320/matertrans.MB201014]

Cathodoluminescence (CL) in scanning electron microscopy (SEM) is commonly accepted as revealing local properties of a specimen region illuminated by an electron beam. CL is widely used to visualize defects in semiconductor structures. However, the presence of a strong electric field in, for example, heterojunctions or p–n junctions causes a separation of generated electron–hole (e–h) pairs and suppresses recombination in the specimen region excited by the beam. As a result CL – a radiative recombination – becomes quenched. At the same time, electron beam-induced current (EBIC) flows throughout the structure, which may produce secondary electroluminescence that is registered by the CL detector. Consequently, the CL measurement is distorted and if there are defects in the structure, they remain unrevealed. The current study shows that registration of the CL signal for different values of electron beam current (including high ones) enables true defect detection in semiconductor layers with built-in electric field. Results for a special test structure prepared with focused ion beam on AlGaAs/GaAs laser heterostructures with an 8nm InGaAs quantum well are presented.

Transmission electron microscopy methods were used to determine the impact of two different implantation processes on the morphology of platinum silicide layers constituting low Schottky barrier contacts intended as the source/drain in MOS transistors. These processes are very promising candidates for the reduction of the Schottky barrier height (SBH) of contacts and are realized by (i) implantation-through-metal (ITM) followed by dopant-segregation induced by silicidation annealing and (ii) implantation-through-silicide (ITS) followed by dopant-segregation due to the post-silicidation annealing. The studies showed that depending on the type and conditions of the process (ITM or ITS with various post-silicidation annealing temperatures) different morphologies of PtSi layers and PtSi/Si interfaces roughnesses are observed. Better quality silicide layers and silicide/silicon interfaces were found for samples after the ITS process with post-silicidation annealing at 500°C than for samples after the ITM process or the ITS process with post-silicidation annealing at temperatures not exceeding 400°C. The observed microstructure of grains and interfaces in these samples, along with the impact of the dopant-segregation, may significantly influence the SBH value. The diffraction patterns and EDXS measurements revealed that regardless of the process type, the formed silicide layer is always PtSi.

Spatially and spectrally resolved cathodoluminescence (CL) studies performed in a scanning electron microscope (SEM) or a scanning transmission electron microscope (STEM) are widely applied to determine the luminescence spectrum, map the optical activity and reveal defects in semiconductor structures. It is commonly recognized that the CL signal represents the local properties of the structure region excited by the electron beam. The present investigations show that if the structure under study contains regions with a strong electric field (e.g. p-n junction), the CL signal may much depend on phenomena far from the excitation region. The range of CL-results distortion extends from negligible changes to completely wrong output. It depends on various parameters described in the paper. The CL results obtained for AlGaAs/GaAs laser heterostructures with InGaAs quantum well are presented.

The formation of ytterbium silicide fabricated by annealing at 480 °C for one hour has been studied by means of high resolution transmission electron microscopy (HRTEM) and energy dispersive X-ray spectroscopy (EDS). The annealing process has been performed under ultra high vacuum (UHV) conditions. The formation of an amorphous silicide layer was observed between the Yb-layer and the silicon substrate in the as-deposited sample. Ytterbium silicide observed after annealing consists of two different layers: crystalline and amorphous ones. The studies confirmed that the formed crystalline layer is of the YbSi2−x phase, however, the structure is different from the hexagonal AlB2 type.

In this paper, we present results of transmission electron microscopy studies on erbium silicide structures fabricated under various thermal conditions. A titanium cap has been used as a protective layer against oxidation during rapid thermal annealing of an erbium layer in a temperature range of 300-700 degrees C. Both layers (200 nm Ti and 25 nm Er) were deposited by electron-beam sputtering. The investigations have shown that the transformation of the 25-nm-thick erbium into erbium silicide is completed after annealing at 500 degrees C. At higher temperatures, the formation of a titanium silicide layer above erbium silicide is observed. The lowest Schottky barrier has been measured in the sample annealed at 700 degrees C.

The dependence of spatially and spectrally resolved cathodoluminescence in a scanning electron microscope on resistances in semiconductor structures, especially on the layer resistance, is reported. This previously unstudied dependence is utilized for thin-layer sheet-resistance measurement. The method is illustrated by an assessment of lateral confinements in semiconductor-laser heterostructures.

Different types of distortions in scanning electron microscopy require different methods of their elimination, and therefore influence of these types on particular elements of the SEM system should be known. The proposed method allows for separation of the direct influence of the magnetic field on the electron beam in the SEM chamber from its influence in the SEM column and from the distortions generated in the SEM scanning block. For this purpose, a series of distorted images is registered for several working distances (between the final aperture of the electron column and the specimen) and for several energies of the electrons. Magnitudes of the distortions are measured on these images. For each applied electron energy, the dependence of the results versus the working distance is approximated with the second-order polynomial function. The analysis of the polynomial coefficients allows for the separation of the above-mentioned kinds of distortions. The presented method enables a selection of the most efficient solution to the distortions reduction. It utilizes the SEM itself and does not need any additional equipment.

We report on transmission electron microscopy studies of Au/Pt/Ti/Pt(10-30 nm) contact structures for high power GaAs/InGaAs semiconductor lasers. The studies showed that annealing at 450 degrees C of contact structures causes the reaction of whole Pt with substrate components (Ga and As) and the formation of Pt-GaAs interlayers with smooth interfaces as required for such structures. Annealing of the structures at 470 and 490 degrees C unfavourably affects the contact structure. At this condition, the strong downward diffusion of Au and Pt from the top layers causes a formation of Au-Pt pits, which break the Ti barrier. Transmission electron microscopy observation revealed that Au/Pt/Ti/Pt(10-30 nm) system annealed at 450 degrees C is appropriate for practical applications. The EDS technique used to identify the phase composition in the Pt(30nm)/GaAs structure (specially produced for the EDS analysis) annealed at 450 degrees C showed that two layers were formed as a result of the reaction of the whole Pt layer with GaAs, and they consist of Ga, Pt and As. The top layer has the highest concentration of Ga. However, the bottom layer, which is close to the substrate, has the highest concentration of As.