Quantitative Dopant Profiling Research Articles

The Mechanistic Determination of Doping Contrast from Fermi Level Pinned Surfaces in the Scanning Electron Microscope Using Energy-Filtered Imaging and Calculated Potential Distributions

Abstract Secondary electron (SE) doping contrast in the scanning electron microscope is correlated with Fermi level pinned surfaces of Si samples prepared using HF-based wet-chemical treatment or focused ion beam (FIB) micromachining en route to quantitative dopant profiling. Using energy-resolved SE imaging techniques and finite-element analyses of surface states and surface junction potentials, we clarified the surface band-bending effects post-NH4F-treatment, consistent with brighter p-contrast from degenerately doped (>1019 cm−3) regions. In general, SE spectromicroscopy scan measurements unambiguously indicated heavy suppression of patch fields, while the empirical discovery of scan frequency-modulated contrast inversion due to Chee et al. [Springer Proceedings in Physics, 120, pp. 407–410 (2008)] is ascribable to competing fixed oxide charge and dynamic charge injection phenomena (particularly at dwell times >29 μs). Leveraging numerical simulations of electric potentials and variable-voltage experimental data, the theoretical model based on amorphization damage-mediated Fermi level pinning is elucidated for Ga+ FIB-processed site-specific doping contrast on patch field-free surfaces. This work successfully argues against the notion that doping contrast ultimately or exclusively entails patch fields or adventitious metal–semiconductor contacts.

Unravelling new principles of site-selective doping contrast in the dual-beam focused ion beam/scanning electron microscope

Doping contrast using the secondary electron (SE) signal in the scanning electron microscope (SEM) can satisfy the International Roadmap for Semiconductors (ITRS) requisites for quantitative dopant profiling of next-generation integrated circuits and devices, but only if adopting a site-selective specimen preparation procedure. In this study, site-specific dopant profiling was performed on the trench side-wall cut by a 30-kV Ga+ focused ion beam (FIB) into silicon p–n junction specimens and milled using successively lower voltages in the dual-beam FIB/SEM. Although depositing the protective platinum strap on the surface effectively controls ‘curtaining’ effects at low final milling voltages, significantly reduced doping contrast from the side-wall compared to that from a cleaved surface subjected to the same ion-beam energy is ascribed to the material affected by a previous milling step, as well as the dissimilar geometries of milling and imaging. New principles underpinning the doping contrast mechanism were surveyed taking into account the depth and concentration of ion implantation and amorphization damage as a linear function of the final milling voltage. Patch fields are suppressed, but the bulk doping-dependent surface band-bending fields at the amorphous-crystalline interface is crucial for doping contrast. In general, as the milling voltage decreases the doping contrast increases, which although reaches up to only half that attainable from a freshly-cleaved specimen, is usable, and demonstrates the feasibility of site-specific dopant profiling in situ in the SEM.

Enhancing doping contrast and optimising quantification in the scanning electron microscope by surface treatment and Fermi level pinning

Recent advances in two-dimensional dopant profiling in the scanning electron microscope have enabled a high throughput, non-contact process diagnostics and failure analysis solution for integrated device manufacturing. The routine (electro)chemical etch processes to obtain contamination-free, hydrogen-terminated silicon surfaces is industrially important in ULSI microfabrication, though doping contrast, which is the basis for quantitative dopant profiling, will be strongly altered. We show herein that ammonium-fluoride treatment not only enabled doping contrast to be differentiated mainly by surface band-bending, but it enhanced the quality of linear quantitative calibration through simple univariate analysis for SE energies as low as 1 eV. Energy-filtering measurements reveal that the linear analytical model broached in the literature (c.f. Kazemian et al., 2006 and Kazemian et al., 2007) is likely to be inadequate to determine the surface potential across semiconductor p-n junctions without suitable deconvolution methods. Nevertheless, quantification trends suggest that energy-filtering may not be crucial if patch fields and contamination are absolutely suppressed by the appropriate edge termination and passivation.

Quantitative Dopant Profiling by Energy Filtering in the Scanning Electron Microscope

Two-dimensional dopant mapping using secondary electrons (SEs) in the scanning electron microscope (SEM) is a technique under intense research and development due to improvements in instrumental resolution and its potential to enable rapid low-cost diagnostics in process optimization for microelectronic and optoelectronic fabrication. However, an initial drawback of the technique is the lack of a complete quantitative model to obtain accurate information on dopant profiles. This paper focuses on detailed studies of dopant profiling quantification on a wide range of doped silicon homojunctions in an SEM under various operating conditions for the first time. Doping type and geometry-dependent effects from the specimen are taken into account, and improved sensitivity, resolution, and quantification accuracy due to energy filtering of the SEs are demonstrated. The resulting SE intensities under varying operating conditions are surveyed with an indication of aspects of the mechanism responsible for doping contrast, including the patch fields, surface band bending, and inelastic mean free path effects in the specimen, as well as the angular velocities of the SEs. Although all samples satisfied the classical logarithmic doping dependence of contrast approximation, some of them required specialized energy-filtering techniques. The theoretical model based on energy filtering to determine the built-in voltage across the $p\!-\!n$ junction that is in the literature does not accurately apply to $n^{+}\!-i-n$ homojunctions. The surface band bending and inelastic mean free path dependence of the SE kinetic energy models have been derived in this paper, which accurately describe the quantitative contrast characteristics from donor distributions, thereby producing more accurate quantification of this technique in semiconductor research and industry.

Calibrated nanoscale dopant profiling using a scanning microwave microscope

The scanning microwave microscope is used for calibrated capacitance spectroscopy and spatially resolved dopant profiling measurements. It consists of an atomic force microscope combined with a vector network analyzer operating between 1–20 GHz. On silicon semiconductor calibration samples with doping concentrations ranging from 1015 to 1020 atoms/cm3, calibrated capacitance-voltage curves as well as derivative dC/dV curves were acquired. The change of the capacitance and the dC/dV signal is directly related to the dopant concentration allowing for quantitative dopant profiling. The method was tested on various samples with known dopant concentration and the resolution of dopant profiling determined to 20% while the absolute accuracy is within an order of magnitude. Using a modeling approach the dopant profiling calibration curves were analyzed with respect to varying tip diameter and oxide thickness allowing for improvements of the calibration accuracy. Bipolar samples were investigated and nano-scale defect structures and p-n junction interfaces imaged showing potential applications for the study of semiconductor device performance and failure analysis.

Kelvin probe force microscopy in the presence of intrinsic local electric fields

AbstractKelvin probe force microscopy (KPFM) is used to investigate the electrostatic force between a conductive probe and doped semiconductors. The observed frequency dependence of the probed KPFM bias is strongly related to sample‐specific intrinsic local electric fields. Equilibrium drift and diffusion of excess charge carriers at low operation frequencies influence the characteristics of the asymmetric electric dipole in the surface region of the investigated semiconductors during the KPFM measurement. The sample‐specific KPFM background signal does not influence the frequency‐dependent lateral variation of the electrical signal. The KPFM bias probed on doped semiconductor nanostructures with high or small enough operation frequencies allows for quantitative dopant profiling or investigation of diffusion processes in internal electric fields, respectively.

A quantitative model for doping contrast in the scanning electron microscope using calculated potential distributions and Monte Carlo simulations

This paper describes the use of a Monte Carlo model incorporating a finite-element method computing the electrostatic fields inside and outside a semiconductor, plus a ray-tracing algorithm for determining the doping contrast observed in a scanning electron microscope (SEM). This combined numerical method also enables the effects on the doping contrast of surface band-bending to be distinguished from those of external patch fields outside the specimen, as well as any applied macroscopic external fields from the detection system in the SEM. Good agreement of our new theory with experiment is obtained. The contrast characteristics in energy-filtered secondary electron images are also explained. The results of this work lead to a more advanced understanding of the doping contrast mechanisms, thereby enabling quantitative dopant profiling using the SEM.

Progress towards quantitative dopant profiling in the SEM

Dopant contrast from secondary electron (SE) imaging is essentially determined by the local electronic band structure at the surface of the solid specimen being analysed in the scanning electron microscope (SEM). To study dopant contrast mechanisms on inhomogeneously doped semiconductors, we have developed a quantitative model by using a numerical method based on a combination of finite-element carrier and potential modelling and Monte Carlo simulation of SE emission and contrast detection. This combined model enables distinction between the effects of specimen internal fields (surface band-bending), microscopic external fields (patch fields) and any macroscopic external fields applied in the SEM.

Quantitative dopant profiling in semiconductors: A Kelvin probe force microscopy model

Kelvin probe force microscopy (KPFM) is used to investigate the electrostatic force between a conductive probe and nanostructured Si with shallow or buried selectively doped regions under ambient conditions. A unique KPFM model correlates the measured Kelvin bias with the calculated Fermi energy, and thus allows quantitative dopant profiling. We show that due to an asymmetric electric-dipole formation at the semiconductor surface the measured Kelvin bias is related with the difference between Fermi energy and respective band edge, and independent of the probe potential.

Quantitative dopant profiling of laser annealed focused ion beam-prepared silicon p-n junctions with nanometer-scale resolution

Silicon p-n junction specimens have been prepared by focused ion beam milling for examination using off-axis electron holography. By using 28nS pulsed XeCl excimer laser irradiation we have reduced the total electrically “inactive” thickness in these specimens from 142to5nm. A platinum layer has been sputtered onto the specimen surfaces to remove the build up of charge from the region of interest during examination. Subsequently, a value of the built in potential has been determined directly from a phase image which is consistent with theory.

Nondestructive Quantitative Dopant Profiling Technique by Contact Radiography

We have developed the only non-destructive technique to profile graded dopants in ICF shells to the precision required by the NIF specifications (Doping level must be accurate to 0.03 at. % and its radial distribution accurate to submicron precision). This quantitative contact radiography method was based on precision film digitization and a dopant simulation model. The measurements on Cu/Be and Ge/CH shells agree with those from electron microprobe and X-ray fluorescence.

Effect of experimental parameters on doping contrast of Si p?n junctions in a FEG-SEM

A study of the effect of experimental conditions such as magnification and scan time on the doping contrast values from Si p–n junctions in the scanning electron microscope has been conducted in order to help enable quantitative dopant profiling. It is found that the contrast is substantially affected by sample charging or contamination as well as by electron–hole pair generation depending on the size of the scanned field area and scan time. At high magnifications and long scan times, a large number of electron–hole pairs are generated, which reduce the built-in voltage and, therefore, make quantification difficult. However, at lower magnifications and long scan times, or at all magnifications and TV scan times, electron–hole pair generation has a negligible effect on the doping contrast, and hence quantification is possible: in particular, the built-in voltage can be measured. One way to achieve this is using a reference sample. Recommendations for optimum experimental conditions for achieving a quantitative procedure are given.

Effect of experimental parameters on doping contrast of Si p–n junctions in a FEG-SEM

A study of the effect of experimental conditions such as magnification and scan time on the doping contrast values from Si p–n junctions in the scanning electron microscope has been conducted in order to help enable quantitative dopant profiling. It is found that the contrast is substantially affected by sample charging or contamination as well as by electron–hole pair generation depending on the size of the scanned field area and scan time. At high magnifications and long scan times, a large number of electron–hole pairs are generated, which reduce the built-in voltage and, therefore, make quantification difficult. However, at lower magnifications and long scan times, or at all magnifications and TV scan times, electron–hole pair generation has a negligible effect on the doping contrast, and hence quantification is possible: in particular, the built-in voltage can be measured. One way to achieve this is using a reference sample. Recommendations for optimum experimental conditions for achieving a quantitative procedure are given.

Sub-10 nm lateral spatial resolution in scanning capacitance microscopy achieved with solid platinum probes

Sub-10 nm resolution can be obtained in scanning capacitance microscopy (SCM) if the probe tip is approximately of the same size. Such resolution is observed, although infrequently, with present commercially available probes. To acquire routine sub-10 nm resolution, a solid Pt metal probe has been developed with a sub-10 nm tip radius. The probe is demonstrated by SCM imaging on a cross-sectioned 70 nm gatelength field-effect transistor (FET), a shallow implant (n+/p, 24 nm junction depth), and an epitaxial staircase (p, ∼75 nm steps). Sub-10 nm resolution is demonstrated on the FET device over the abrupt meeting between a silicon-on-insulator oxide layer and a neighboring Si region. Comparable resolution is observed on the implant structure, and quantitative SCM dopant profiling is performed on it with sub-10 nm accuracy. Finally, the epitaxial staircase structure is quantitatively profiled demonstrating the accuracy obtained in quantitative profiling with the tips.

Quantitative electrostatic force measurement in AFM

We describe a method for measuring forces in the atomic force microscope (AFM), in which a small amplitude oscillation (∼1 Å p–p) is applied to a stiff (∼40 N/m) cantilever below its first resonant frequency, and the force gradient is measured directly as a function of separation. We have used this instrument to measure electrostatic forces by applying an ac voltage between the tip and the sample, and observed a variation in contact potential difference with separation. We also show how the benefits of this instrument may be exploited to make meaningful capacitance measurements, especially at small tip–surface separations, and demonstrate the potential of this technique for quantitative dopant profiling in semiconductors.

Advances in experimental technique for quantitative two-dimensional dopant profiling by scanning capacitance microscopy

Several advances have been made toward the achievement of quantitative two-dimensional dopant and carrier profiling. To improve the dielectric and charge properties of the oxide–silicon interface, a method of low temperature heat treatment has been developed which produces an insulating layer with consistent quality and reproducibility. After a standard polishing procedure is applied to cross-sectional samples, the samples are heated to 300 °C for 30 min under ultraviolet illumination. This additional surface treatment dramatically improves dielectric layer uniformity, scanning capacitance microscopy (SCM) signal to noise ratio, and C–V curve flat band offset. Examples of the improvement in the surface quality and comparisons of converted SCM data with secondary ion mass spectrometry (SIMS) data are shown. A SCM tip study has also been performed that indicates significant tip depletion problems can occur. It is shown that doped silicon tips are often depleted by the applied SCM bias voltage causing errors in the SCM measured profile. Worn metal coated and silicided silicon tips also can cause similar problems. When these effects are tested for and eliminated, excellent agreement can be achieved between quantitative SCM profiles and SIMS data over a five-decade range of dopant density using a proper physical model. The impact of the tip size and shape on SCM spatial accuracy is simulated. A flat tip model gives a good agreement with experimental data. It is found that the dc offset used to compensate the C–V curve flat band shift has a consistently opposite sign on p- and n-type substrates. This corresponds to a positive surface on p-type silicon and to a negative surface on n-type silicon. Rectification of the large capacitance probing voltage is considered as a mechanism responsible for the apparent flat band shift of (0.4–1) V measured on the samples after heating under UV irradiation. To explain the larger flat band shift of (1–5) V, tip induced charging of water-related traps is proposed and discussed.

Two-step dopant diffusion study performed in two dimensions by scanning capacitance microscopy and TSUPREM IV

We report the results of a two-step two-dimensional (2D) diffusion study by scanning capacitance microscopy (SCM) and 2D SUPREM IV process simulation. A quantitative 2D dopant profile of a gate-like structure is measured with the SCM on a cross-sectioned polished silicon wafer. The gate-like structures consist of heavily implanted n+ regions separated by a lighter doped n-type region underneath 0.56 μm gates. The SCM is operated in the constant-change-in-capacitance mode. The 2D SCM data are converted to dopant density through a physical model of the SCM/silicon interaction. This profile has been directly compared with 2D SUPREM IV process simulation and used to calibrate the simulation parameters. The sample is then further subjected to an additional diffusion in a furnace for 80 min at 1000 °C. The SCM measurement is repeated on the diffused sample. This final 2D dopant profile is compared with a SUPREM IV process simulation tuned to fit the earlier profile with no change in the parameters except the temperature and time for the additional diffusion. Our results indicate that there is still a significant disagreement between the two profiles in the lateral direction. SUPREM IV simulation considerably underestimates the diffusion under the gate region.

Quantitative measurement of two-dimensional dopant profile by cross-sectional scanning capacitance microscopy

Quantitative two-dimensional dopant profiling of a gatelike structure is achieved by scanning capacitance microscopy (SCM). A processed silicon wafer is sectioned then polished such that a cross section is made through a gatelike structure. This structure consists of two n+ doped regions separated by a lighter doped n region. The two-dimensional SCM data are converted to dopant density through a physical model of the SCM–silicon interaction. Improvements to the physical model and SCM data to dopant profile algorithm are discussed. Advances in sample preparation are described. Adjustment of the conversion model parameters allows for fitting of the SCM dopant profile to a vertical secondary-ion-mass-spectroscopy profile. The accuracy of this fit is better than 20%. This fit also gives strong evidence that the full two-dimensional profile is truly quantitative.

Quantitative two-dimensional dopant profiling of abrupt dopant profiles by cross-sectional scanning capacitance microscopy

The scanning capacitance microscope (SCM) is capable of quantitative two-dimensional carrier and dopant density mapping with nanometer scale spatial resolution. The method can be applied to either the top or the cross-sectional surface of a silicon sample. The quantitative inversion of SCM data to carrier or dopant density is achieved using a quasi-one-dimensional model. Cross-sectional SCM measurements have been performed on samples that have abrupt dopant density steps. The dopant density in these samples systematically varies from 1017 to 1020 cm−3. The cross-sectional measurements provide a means to directly compare the SCM results with vertical secondary ion mass spectroscopy (SIMS) profiles. A spatial resolution of approximately 25 nm is achieved. A first order quantitative agreement between the SCM and the SIMS profiles is found.

Direct comparison of cross-sectional scanning capacitance microscope dopant profile and vertical secondary ion-mass spectroscopy profile

The scanning capacitance microscope (SCM) has been shown to be useful for quantitative 2D dopant profiling near the surface of silicon. An atomic force microscope is used to position a nanometer scale tip at a silicon surface, and local capacitance change is measured as a function of sample bias. A new feedback method has been recently demonstrated in which the magnitude of the ac bias voltage applied to the sample is adjusted to maintain a constant capacitance change as the tip is scanned across the sample surface. The applied ac bias voltage as a function of position is then input into an inversion algorithm to extract the dopant density profile. The new feedback approach allows for the use of a quasi-1D model in the inversion algorithm. Since there are no alternative 2D dopant profiling techniques which are well established at present, evaluation of the quantitative character of 2D SCM measured profiles has been a challenge. To avoid this obstacle, we have developed sample preparation methods which allow direct comparison of lateral SCM measured profiles on cleaved wafers (cross-sectional plane) with vertical secondary ion-mass spectroscopy (SIMS) profiles. The direct comparison of inverted SCM data and SIMS profiles indicates that quantitative 2D dopant profiling can be achieved by the SCM on a nanometer scale.