R.D. Verda

Orientation dependence of blistering in H-implanted Si

The orientation effect on blistering phenomenon in H implanted Si was studied for (100), (111), and (110) Si wafers. It was found that substrate orientation has no observable effects on the underlying blistering mechanisms. Furthermore, the implantation damage, Si–H complex formation in as-implanted samples and surface roughness of the transferred layer appeared to be unaffected by the orientation. However, the blistering kinetics are orientation dependent, with (100) Si having the fastest blistering rate, and (110) Si the slowest. This dependence was attributed to the different density of ruptured Si–Si bonds of different orientations. The magnitude of the observed in-plane compressive stress in the H-implanted Si wafers is rationalized in terms of the formation of platelets in the samples.

Heat resistance of fluorinated diamond-like carbon films

The heat resistance of fluorinated diamond-like carbon (F-DLC) films produced by Plasma Immersion Ion Processing (PIIP) technique was investigated by annealing F-DLC coatings in a vacuum furnace. The growth rate for the F-DLC films was approximately 0.6 μm/h. In order to see the possible change in the composition and properties of the F-DLC films, Rutherford Backscattering Spectrometry (RBS), nanoindentation and contact angle measurements were performed before and after the heat treatments. The results show that the composition and properties of the F-DLC films were unchanged up to heat treatment at 300°C for up to 30 min. Blistering and film delamination occurred for samples treated at 400°C.

The influence of ion-implantation damage on hydrogen-induced ion-cut

Abstract Hydrogen ion-implantation in Si has been shown to be an effective means of cleaving thin layer of Si from its parent wafer. This process has been called smart-cut™, ion-cut, and hydrogen-slicing by various authors. Qualitatively it is known that implanted hydrogen in Si evolves under heating to form bubbles with high internal pressure, which then drives the cleavage process. The general belief has been that the bubbles, which induced the cleavage, occur at the peak in the H-implantation concentration profile. However, our recent experiments have shown that H bubble nucleation and the ultimate cleavage location in Si is controlled by the lattice damage that is generated by the H-implantation process. The stress and strain field in the proton-implantation-induced damage region of the silicon crystal is proposed to explain the observed results.

Hydrogen-implantation induced silicon surface layer exfoliation

Abstract The physical mechanisms of hydrogen-induced silicon surface cleavage were investigated using the combination of cross section transmission electron microscopy (XTEM) and Rutherford back-scattering spectrometry (RBS) channelling analysis. A ⟨100⟩-oriented silicon wafer was implanted with 175 keV protons to a dose of 5 × 1016 cm−2. The implanted wafer was bonded to a SiO2-capped ⟨100⟩-oriented silicon wafer and then heated to an elevated temperature of 600°C to produce exfoliation. The damage region of the implanted silicon was examined by XTEM, which revealed the presence of hydrogen-filled platelets. The depth distribution of the implantation damage was also monitored by RBS in the channelling condition in the as-implanted state as well as after the cleavage of the silicon wafer. A comparison of the RBS and XTEM indicates that the nucleation of hydrogen-filled microcavities and the cleavage of the silicon wafer take place above the hydrogen concentration peak near the implantation damage peak, revealing the crucial role of the implantation damage in the crystal in terms of hydrogen-induced silicon surface layer exfoliation.

Geometric considerations relevant to hydrogen depth profiling by reflection elastic recoil detection analysis

Abstract This work addresses geometric considerations relevant to the accuracy of depth profiling by reflection elastic recoil detection analysis, which becomes an issue when many samples are compared over time or a single sample is repeatedly analyzed following a sequence of processing steps. In such cases, accurate and reproducible geometric alignment and incident beam energy calibration must be performed over time, and are addressed here. Our analysis and experiments show that geometric deviations in the sample tilt angle and incident particle beam steering, as well as deviations from the sample eucentric position, can result in significant errors in depth profiling. As a result, a recommended degree of accuracy is stated for each of these geometric components; techniques are presented to attain this accuracy, or better. The most notable of these techniques is a laser alignment of the sample tilt angle, which has a reproducible accuracy of 0.04°.

Depth profiling of hydrogen in crystalline silicon using elastic recoil detection analysis

Abstract Accurate depth profiling of hydrogen in crystalline silicon (c-Si) from reflection elastic recoil detection analysis (ERDA) can be performed using a method that converts the channel difference between surface and bulk signals directly to depth. The method relies on Rutherford backscattering spectrometry (RBS) to unambiguously determine the depth of a buried marker coincident with a hydrogen distribution in each of several silicon calibration standards. A conversion from ERDA channels to depth is extracted from the relationship between the depth to the marker and the channel difference between surface and bulk hydrogen centroids in forward recoil spectra acquired simultaneously. Applying this conversion to ERDA of hydrogen-implanted c-Si taken under the same conditions as those of the silicon calibration standards results in accurate, quantitative depth profiling. A comparison of techniques shows that this method offers distinct advantages over depth profiling by computer simulation of ERDA spectra.

The use of ion channeling and elastic recoil detection in determining the mechanism of cleavage in the ion-cut process

Hydrogen ion implantation in Si has been shown to be an effective means of inducing cleavage in Si and facilitating the transfer of thin slices to other substrates, a process known as ion-cut. In our experiments silicon wafers were implanted with 40 keV protons to a variety of ion doses ranging from 5 � 10 16 to 1 � 10 17 cm � 2 and subsequently annealed at 600 C. Under all of these conditions ion-cutting in the form of ‘‘popping off’’ discrete blisters was obtained. The cleavage mechanisms in these samples were studied through the combined use of Rutherford backscattering spectroscopy in channeling mode, elastic recoil detection analysis, and scanning electron microscopy. Our analyses had shown that the cleavage location in Si is largely controlled by the lattice damage that is generated by the H-implantation process. At lower H doses, the cut location is well correlated with the damage peak and can be explained by damageinduced in-plane stress and the corresponding out-of-plane strain. However, at higher implantation doses the ion-cut location shifts to a portion of the crystal which contains lower damage and sufficient concentration of H. This effect can be explained by the changing fracture mechanics at high H concentrations in heavily damaged Si. 2002 Elsevier Science B.V. All rights reserved.

An energy spread correction for depth profiling by elastic recoil detection analysis

Abstract A technique for hydrogen depth profiling by reflection elastic recoil detection analysis (ERDA) called the channel-depth conversion was introduced in [Verda et al., Nucl. Instr. and Meth. B 183 (2001) 401]. The conversion is determined from ion beam analysis of specially prepared standards and transforms the channel axis of ERDA spectra directly to units of depth. However, the channel-depth conversion does not address energy spread inherent in forward recoil spectra. Energy spread causes a broadening in the energy range of the spectrum, which can lead to errors in depth profiling. This work introduces a technique that addresses energy spread, called the energy spread correction. Together, the energy spread correction and the channel-depth conversion techniques, applied to ERDA spectra in that order, comprise a depth profiling method presented in this work. In the case where forward recoil spectra must be obtained using an absorber foil, we show that the depth profiling method is independent of foil thickness. The method was developed for hydrogen depth profiling by ERDA performed with a reflection geometry setup that requires an absorber foil. But, in principle, this method could be used to profile other elements, and that with transmission geometry setups and/or detection systems that do not require an absorber foil.

ACCURATE HYDROGEN DEPTH PROFILING BY REFLECTION ELASTIC RECOIL DETECTION ANALYSIS

Abstract A technique to convert reflection elastic recoil detection analysis spectra to depth profiles, the channel-depth conversion, was introduced by Verda et al. [Nucl. Instr. and Meth. B 183 (2001) 401]. But the channel-depth conversion does not correct for energy spread, the unwanted broadening in the energy of the spectra, which can lead to errors in depth profiling. A work in progress introduces a technique that corrects for energy spread in elastic recoil detection analysis spectra, the energy spread correction [Nucl. Instr. and Meth. B 187 (2002) 383]. Together, the energy spread correction and the channel-depth conversion comprise an accurate and convenient hydrogen depth profiling method.

Boron-induced redistribution of hydrogen implanted at elevated temperature into crystalline silicon

The redistribution of hydrogen during elevated-temperature implantation of boron-pre-implanted silicon was investigated. By redistribution we mean that the final hydrogen distribution differs from the distribution of a control sample, and is attributed to sample preparation. Samples were prepared with a single boron pre-implantation, with projected range either shallower or deeper with respect to the projected range of a subsequent elevated-temperature hydrogen implantation. For shallower boron, the hydrogen redistribution was a wholesale shift in the entire distribution toward the surface, whereas for deeper boron, partial redistribution of end-of-range hydrogen toward the bulk was observed. Self-implantation experiments show that the wholesale hydrogen redistribution is not due solely to boron pre-implantation damage to the silicon lattice, but is driven by chemical effects attributed to the presence of boron.