Scanning Capacitance Microscopy Signal Research Articles

Assessing the electrical activity of individual ZnO nanowires thermally annealed in air.

ZnO nanowires (NWs) are very attractive for a wide range of nanotechnological applications owing to their tunable electron concentration via structural and surface defect engineering. A 2D electrical profiling of these defects is necessary to understand their restructuring dynamics during engineering processes. Our work proposes the exploration of individual ZnO NWs, dispersed on a SiO2/p++-Si substrate without any embedding matrix, along their axial direction using scanning capacitance microscopy (SCM), which is a useful tool for 2D carrier profiling. ZnO NWs are hydrothermally grown using 0–20 mM ammonium hydroxide (NH4OH), one of the reactants of the hydrothermal synthesis, and then annealed in a tube oven at 350 °C/1.5–15 h and 450 °C/15 h. While the as-grown ZnO NWs are highly conductive, the annealed ones exhibit significant SCM data with a high signal-to-noise ratio and temperature-dependent uniformity. The SCM signal of ZnO NWs is influenced by both their reduced dimensionality and the electron screening degree inside them. The electrical activity of ZnO NWs is only observed below a critical defect concentration that depends on the annealing temperature. Optimal SCM signals of 200 and 147 mV are obtained for samples with 0 and 20 mM NH4OH, respectively, and annealed at 350 °C/15 h. The corresponding electron concentrations of 3.27 × 1018 and 4.58 × 1018 cm−3 were estimated from the calibration curve, respectively. While thermal treatment in air of ZnO NWs is an effective approach to tune the defect density, 2D electrical mapping enables identifying their optimal electrical characteristics, which could help to boost the performance of final devices exploiting their coupled semiconducting–piezoelectric properties.

Voltage modulation efficiency in scanning capacitance microscopy

This paper reports a novel investigation of the voltage modulation efficiency (VME) in scanning capacitance microscopy (SCM). A signal intensity model was used to define the VME, which is dependent on the impedance components in an SCM setup. In SCM, the VME was found to play a key mediating role in the close relationship between the signal intensity and the modulation voltage, providing an indicator for the surface treatment and the back-contact process of an SCM specimen. We observed that, for silicon-based specimens, ultraviolet-assisted oxidation and microwave annealing improved the specimen surface and the back-contact, respectively, which increased the VME. It was also found that a high modulation voltage and a large back-contact area may induce a significant stray capacitance around the conductive tip and, hence, lower the VME. The VME degradation not only decreased the SCM signal intensity but also reduced the image contrast in the regions with high carrier concentrations. In addition, our experimental results further revealed that the signal intensity model also provided a promising opportunity to establish a precise and quantitative method for measuring the thickness of insulating layers.

Surface electrical characterization of defect related inhomogeneities of AlGaN/GaN/Si heterostructures using scanning capacitance microscopy

Large mismatch of lattice constants and thermal expansion coefficients between materials in AlGaN/GaN/Si heterostructures lead to high density of structural defects and material inhomogeneities in the epitaxial layers. The influence of growth conditions on the local electronic properties of the fabricated AlGaN/GaN/Si heterostructures was investigated using scanning capacitance microscopy (SCM). The significant differences of spatial electronic properties in the nanometer scale for the heterostructures grown with and without application of SiN nanomasking layer were revealed. Analysis of two dimensional images of the SCM signal and dC/dVAmp = f(VDC) spectrum allowed to conclude that negative charge accumulated at dislocations in the epitaxial layers could affect the surface electronic states of the AlGaN barrier layer but could not have major impact on the two dimensional electron gas formation at the AlGaN/GaN interface.

Nanoscale calibration of n-type ZnO staircase structures by scanning capacitance microscopy

Cross-sectional scanning capacitance microscopy (SCM) was performed on n-type ZnO multi-layer structures homoepitaxially grown by molecular beam epitaxy method. Highly contrasted SCM signals were obtained between the ZnO layers with different Ga densities. Through comparison with dopant depth profiles from secondary ion mass spectroscopy measurement, it is demonstrated that SCM is able to distinguish carrier concentrations at all levels of the samples (from 2 × 1017 cm−3 to 3 × 1020 cm−3). The good agreement of the results from the two techniques indicates that SCM can be a useful tool for two dimensional carrier profiling at nanoscale for ZnO nanostructure development. As an example, residual carrier concentration inside the non-intentionally doped buffer layer was estimated to be around 2 × 1016 cm−3 through calibration analysis.

Contrast distortion induced by modulation voltage in scanning capacitance microscopy

With a dark-mode scanning capacitance microscopy (SCM), we directly observed the influence of SCM modulation voltage (MV) on image contrasts. For electrical junctions, an extensive modulated area induced by MV may lead to noticeable changes in the SCM signal phase and intensity, resulting in a narrowed junction image and a broadened carrier concentration profile. This contrast distortion in SCM images may occur even if the peak-to-peak MV is down to 0.3 V. In addition, MV may shift the measured electrical junction depth. The balance of SCM signals components explain these MV-induced contrast distortions.

(Invited) Scanning Probe Analysis of Dielectrics on High Mobility Substrates

In this work we employ electrical scanning probe techniques to evaluate on a nanoscale the influence of semiconductor surface morphology on the electrical properties of dielectrics. Gate leakage, reliability and oxide thickness uniformity are assessed by conductive atomic force microscopy (C-AFM). Interface traps are analyzed by scanning capacitance microscopy (SCM). A correlation between nanoscale and macroscale trends is established following reverse processing of the devices. The impact of surface morphology on localized dielectric properties is evaluated by studying MOSFETs having the same silicon channel strain but different dielectric/substrate interface roughness. The correlation between surface morphology, leakage and SCM signal suggests that for dielectrics on high mobility substrates prone to surface roughening, the morphology as well as the dielectric itself play a role in the final dielectric quality. Optimized epitaxial growth will therefore enable improved device performance, reliability and variability. The results have implications for all technologies which employ relaxed SiGe templates.

Statistical Method for the Estimation of Carrier Concentration in Semiconductor Devices Using Scanning Capacitance Microscopy

Scanning capacitance microscopy (SCM) is a variation of atomic force microscopy (AFM) based techniques. SCM has been used to characterize two-dimensional electric properties in semiconductor devices, such as the free carrier distribution. We have demonstrated quantitative analysis of the estimated carrier concentration using statistically derived correlation functions with SCM signals. It was also applied to silicon carbide (SiC) materials as well as the conventional measurements for silicon devices. The statistical method does not require transforming the signals into physical quantities and is applicable to other next-generation AFM based techniques.

Characterization of thickness variations of thin dielectric layers at the nanoscale using scanning capacitance microscopy

In this work, the applicability of scanning capacitance microscopy (SCM) for film thickness characterization and its sensitivity to the surface roughness on nanoscale were examined experimentally. SiO2 layers with different film thicknesses (between 5 and 19 nm) were analyzed by conventional capacitance-voltage (C-V) measurements and using SCM in the scanning capacitance spectroscopy (SCS) mode. The influence of the film thickness on the SCM signal was studied in detail by comparison of modeled data with experimental data. The dC/dV-V characteristics measured by SCS at the nanoscale could be correlated with derivatives of conventionally measured C-V curves as well as simulated C-V characteristics for the different film thicknesses. Quantitatively comparing their peak areas, it was found that the dC/dV signal of SCS correlates with the change in the insulator thickness. The sensitivity of SCM for the detection of local variations of dielectric-layer thicknesses at the nanoscale was demonstrated by SCM mapping of crystalline high-k layers, where spatial differences of the SCM signal could be directly correlated with changes in the topography caused by film thickness variations.

Visualization of MeV ion impacts in Si using scanning capacitance microscopy

Scanning capacitance microscopy (SCM) of $3\phantom{\rule{0.3em}{0ex}}\mathrm{MeV}$ ${\mathrm{Au}}^{2+}$ ion implanted Si have been performed for doses between $2\ifmmode\times\else\texttimes\fi{}{10}^{8}$ and $5\ifmmode\times\else\texttimes\fi{}{10}^{9}\phantom{\rule{0.3em}{0ex}}{\mathrm{cm}}^{\ensuremath{-}2}$. The measurements show a random pattern of reduced SCM signal (charge trapping) correlated with the ion impacts. These features have a lateral dimension of $150--600\phantom{\rule{0.3em}{0ex}}\mathrm{nm}$ and reveal a pronounced dose dependence. It is argued that the Fermi level near the impacts and along the ion tracks is modified (pinned) due to deep acceptor states formed by the penetrating ions. Substantial evidence for this argument is provided by SCM images obtained at different temperatures, where a strong correlation is revealed between the probing frequency and the emission rate of the single negative acceptor level of divacancy. To the best of our knowledge, this is a direct observation of signatures for individual ion impacts in Si by an electrical scanning technique.

Scanning probe microscopy of single Au ion implants in Si

We have studied 5 MeV Au 2+ ion implantation with fluences between 7 × 10 7 and 2 × 10 8 cm − 2 in Si by deep level transient spectroscopy (DLTS) and scanning capacitance microscopy (SCM). The DLTS measurements show formation of electrically active defects such as the two negative charge states of the divacancy ( V 2( /–) and V 2(–/0)) and the vacancy–oxygen (VO) center. It is observed that the intensity of the V 2( /–) peak is lower compared to that of V 2(–/0) by a factor of 5. This has been attributed to a highly localized distribution of the defects along the ion tracks, which results in trapping of the carriers at V 2(–/0) and incomplete occupancy of V 2( /–). The SCM measurements obtained in a plan view show a random pattern of regions with a reduced SCM signal for the samples implanted with fluence above 2 × 10 8 cm − 2 . The reduced SCM signal is attributed to extra charges associated with acceptor states, such as V 2(–/0), formed along the ion tracks in the bulk Si. Indeed, the electron emission rate from the V 2(–/0) state is in the range of 10 kHz at room temperature, which is well below the probing frequency of the SCM measurements, resulting in “freezing” of electrons at V 2(–/0).

Intermittent contact scanning capacitance microscopy-An improved method for 2D doping profiling

In the present study an improved method for 2D doping profiling of semiconductor device structures is presented. The method combines the capabilities of scanning capacitance microscopy (SCM) with the advantages of intermittent contact atomic force microscopy (IC-AFM) and is called intermittent contact scanning capacitance microscopy (IC-SCM). Compared with standard SCM, IC-SCM provides mechanically stable measurement conditions because tip wear is nearly eliminated. Furthermore, background signals without local information are suppressed by demodulating the SCM signal at higher harmonics of the tapping tip frequency. Both, reduced tip wear and higher harmonics demodulation yield improved spatial image resolution at less tip degradation compared with standard SCM.

Experimental investigation of interface states and photovoltaic effects on the scanning capacitance microscopy measurement for p-n junction dopant profiling

Controlled polishing procedures were used to produce both uniformly doped and p-n junction silicon samples with different interface state densities but identical oxide thicknesses. Using these samples, the effects of interface states on scanning capacitance microscopy (SCM) measurements could be singled out. SCM measurements on the junction samples were performed with and without illumination from the atomic force microscopy laser. Both the interface charges and the illumination were seen to affect the SCM signal near p-n junctions significantly. SCM p-n junction dopant profiling can be achieved by avoiding or correctly modeling these two factors in the experiment and in the simulation.

Experimental investigation of the dielectric-semiconductor interface with scanning capacitance microscopy

An experimental investigation of how interface states effect scanning capacitance microscopy (SCM) measurements is presented. Different sample polishing procedures were used to make SCM samples that would have different interface state densities, but identical oxide thicknesses. By comparing SCM signals of these samples, the effect of interface states could be singled out. The interface states of these SCM samples were found to have an amphoteric energy distribution. The magnitude of the maximum SCM signals (maximum d C/d V in d C/d V versus dc bias, V dc, plots) is independent of the interface-trapped charges, while the full width at half maximum (FWHM) of the d C/d V– V dc curves is broadened with the interface states. The physics of SCM interface states effect is also discussed.

Experimental aspects and modeling for quantitative measurements in scanning capacitance microscopy

In this article we discuss the reliability of a quantification method for scanning capacitance microscopy (SCM) measurements based on the calculation of a calibration curve. We demonstrate that an accurate control of the conductive tip coating stability, low temperature oxidation and Si–SiO2 interface microroughness allows one to fabricate a nanometric metal-oxide-semiconductor device (nanoMOS), whose dC∕dV-V characteristics measured on a set of different concentration levels can be reproduced by simulation of an ideal nanoMOS with the realistic three-dimensional geometry. We also studied the impact of tip coating (metal and conductive diamond coated tips) and oxidation method (wet and UV∕ozone oxides) on the reproducibility of the measured SCM signal for different concentration levels both on p- and n-type Si staircase calibration samples, and we demonstrated that the UV∕ozone oxidation associated with the use of a diamond tip is the best solution. The experimental calibration curve obtained by this choice is well fitted by the calculated calibration curve. The maximum experimental errors affecting the measured SCM signal depending on doping concentrations have been determined and these errors have been used to estimate the maximum errors on the concentration values calculated by applying the theoretical calibration curve.

On calculating scanning capacitance microscopy data for a dopant profile in semiconductors

One scanning probe method that holds great promise for two-dimensional dopant profiling in semiconductors is scanning capacitance microscopy (SCM). The usual way to calculate the SCM signal (the derivative capacitance) for a dopant profile is to use the finite-element method to solve Poisson’s equation. The domain region includes the doped semiconductor substrate region, the oxide layer region, and the air region bordering the metallic probe-tip of the SCM. The bias between the probe and sample is determined by using Dirichlet boundary conditions along the boundary of the probe-tip and the grounded plane deep within the doped semiconductor substrate. Managing this mesh-object (domain geometry and boundary conditions) requires sophisticated software. To help simplify these calculations, it is proposed that the SCM signal be found by using the areal capacitance distribution of the oxide and the air near the probe-tip to specify a natural boundary condition along the probed surface of the doped semiconductor sample, so that only the doped semiconductor substrate region needs to be meshed. For this configuration, it is found that linear finite-elements can give inaccurate results, and that cubic finite-elements can give accurate results. To help speed the calculations, it is proposed that a reasonably coarse mesh may be used, as well. The method is applied to a model one-dimensional ion-implanted dopant profile in a two-dimensional geometry, and the results of calculation for the forward problem are compared with that of using Dirichlet boundary conditions along the probe-tip.

Electrical characterization of InGaAs/InP quantum wells by scanning capacitance microscopy

In this work, cross-sectional scanning capacitance microscopy (SCM) is used to investigate InGaAs/InP quantum wells. It is demonstrated that SCM is indeed capable of detecting the electrons in the quantum wells and that the SCM signal shows a systematic trend for the different wells. Clear dips in the dC/dV signal are observed at the InGaAs quantum wells and imply carrier densities higher than the surrounding barriers. It is also shown that the depletion regions in the barriers adjacent to the wells can be resolved. The results show that geometric tip-averaging effects significantly influence the imaging of electrons in quantum wells and limit the lateral resolution.

Ultrathin oxides for the SCM analysis of sub-micron doping profiles

Attenuated total reflection (ATR) spectroscopy and spectroscopic ellipsometry (SE) have been used to characterize oxides used for the scanning capacitance microscopy (SCM) technique. SCM has been used to study boron and phosphorous doped Si test structures epitaxially grown on (100) Si substrates. SCM samples have one-dimensional (1D) doping profiles with sub-micron features, with staircase-like steps in the unipolar sample and a smoother profile in the bipolar sample, as obtained by secondary ion mass spectrometry (SIMS) profiling. Cross-sectional SCM results obtained on samples oxidized by the standard low-temperature UV-ozone method are presented, discussed and compared to results obtained on cleaved samples oxidized by a simple exposure to air. The results show that the native oxide covering a (110) cleaved section may yield SCM images of sufficient quality, with no contrast reversal on a wide range of doping levels, as well as observed on sections prepared with the UV-ozone technique. However, the long-term stability of the SCM signal on native oxides is poor, and UV-ozone oxidation can be used to recover a valid SCM signal. Realistic ultrathin oxide thickness data obtained by SE on (110) substrates are presented together with ATR results, which confirm the superior quality of UV-ozone oxides with respect to other kinds of oxides.

Comparison of experimental and theoretical scanning capacitance microscope signals and their impact on the accuracy of determined two-dimensional carrier profiles

The accuracy with which two-dimensional carrier profiles can be extracted from scanning capacitance microscopy (SCM) images of doped structures in silicon depends on the model used to interpret the SCM differential capacitance data. This work validates models of the SCM by comparing the calculated SCM signal to the measured signal for a variety of sample parameters, measurement conditions, and capacitance sensors from different manufacturers. The magnitude of the capacitance sensor high-frequency voltage is measured and its effect on ΔC–V curves and extracted carrier profiles is quantified. Two nonidealities commonly observed in SCM signals, U-shaped C–V curves and double zero crossing in the SCM signal at the p–n junctions, are related to the measurement parameters and explained.

Characterization of GaAs/AlGaAs laser mesas regrown with semi-insulating GaInP by scanning capacitance microscopy

In this work, cross-sectional scanning capacitance microscopy (SCM) is used to investigate GaAs/AlGaAs buried heterostructure lasers. Laser mesas are buried with GaInP:Fe selectively regrown by hydride vapor phase epitaxy. It is shown that a complete 2D map of the electrical properties of device structure including, delineation of regrown interfaces and electrical nature of the regrown GaInP layer can be obtained. The behavior of the SCM signal with ac bias is used to verify the semi-insulating nature of the regrown layer at different locations of the sample. The measured SCM signal for the regrown GaInP:Fe layer is uniformly zero, indicating very low free carrier densities and confirming semi-insulating properties. This observation strongly suggests, in addition, uniform Fe incorporation in the regrown layers, close to and far away from the mesa. Finally, a nanoscale feature in the SCM contrast appearing as a bright (dark) spot in dC/dV mode (feedback bias mode) is observed at the mesa sidewall close to the interface between the regrown GaInP:Fe and the p-cladding layer. The origin of this contrast is discussed in terms of local band-bending effects and supported by 2D Poisson simulations of the device structure.

Detailed study of scanning capacitance microscopy on cross-sectional and beveled junctions

In this work we have done a systematic study with scanning capacitance microscopy (SCM) on cross-sectional and beveled structures. A study was made on the practical problem of contrast reversal as well as on the effect of carrier spilling related to bevel angle, steepness and substrate concentration of the doping profile. A comparison has been made with the results achieved with spreading resistance profiling and also with theoretical predictions. Finally, the junction displacement for cross-sectional and beveled junctions is studied as a function of the applied bias. It is shown that the junction displacement is much smaller on the beveled surface after demagnification. Furthermore, the large extension of the profile along the beveled surface allows us to study the bias induced variation of the SCM signal within the depletion layer in great detail.