一般財団法人材料科学技術振興財団　MST Product Lineup

In semiconductor device manufacturing, from the perspective of improving yield, it is necessary to enhance the cleanliness of the backside of the wafer and to remove substances that remain on the bevel area of the wafer. In this study, we conducted TOF-SIMS analysis of the bevel inclined surface to evaluate the distribution of contamination (Figure 2). Additionally, by comparing the mass spectra of the adhered substances with those of the normal area and the contamination source, we found that the adhered substances matched the metal (Cr) and organic components of the contamination source. TOF-SIMS can capture the contamination generation process in the bevel area (edge and inclined surface).

We investigated the shape and positional relationship of the poly-Si gate, n-type source layer, and p-type body layer of commercially available power transistors (DMOSFETs). By overlaying AFM images, we can also understand the positional relationship with the gate.

By overlaying the luminescent image and the IR image, it is possible to identify leak locations under microscopic observation. If there are large-scale appearance anomalies such as cracks or electrostatic breakdown, abnormalities can also be confirmed with an IR microscope. Additionally, if luminescence cannot be detected due to light shielding of the emitter electrode, the collector electrode is removed, and near-infrared light is detected from the collector side. An example is presented where a high-voltage power supply capable of applying up to 2000V is used to operate a power device with high voltage resistance and low leakage current, and the failure location is identified using an emission microscope.

By using a high-voltage power supply (capable of applying up to 2000V), it is possible to induce breakdown in diodes with high breakdown voltage. In this case, a SiC Schottky diode with a breakdown voltage of 600V was operated, and by applying high voltage in the reverse direction, breakdown was induced. After removing the cathode electrode through polishing, emission microscopy observations were conducted to identify the location of the breakdown current generation. Commercially available products were used for the measurements.

In SCM, it is possible to identify the p/n polarity of semiconductors and visualize the shape of the diffusion layer. This method has been utilized for Si devices, but it can also be applied to SiC devices in areas where the carrier concentration is sufficiently high. This document presents the results of SCM analysis conducted on a cross-section of a SiC Planar Power MOS.

SiC is used as a power device material due to its physical properties, but unlike Si, the diffusion of dopants after thermal treatment following ion implantation is difficult. Therefore, it is necessary to control the distribution in the depth direction through multi-step ion implantation. SIMS analysis can evaluate impurity concentrations in the depth direction with high sensitivity (below ppm), making it suitable for assessing the distribution of dopant elements in SiC. It is also possible to evaluate the distribution of light elements (such as H, C, O, F, etc.). Please contact us regarding any elements you would like to evaluate.

We disassembled a commercially available SiC power MOSFET and conducted imaging SIMS measurements to evaluate the concentration distribution of the dopant elements Al, N, and P in a 20 µm square area to a depth of 0.5 µm. We will present a case where we extracted the depth profile concentration distribution of Al, N, and P localized within the sample surface from the data processing after the imaging SIMS measurements.

We disassembled a commercially available SiC Schottky diode and conducted imaging SIMS measurements to evaluate the concentration distribution of the dopant element Al in a 40μm square area to a depth of 0.5μm. From the data processing after the imaging SIMS measurements, we extracted the depth-wise concentration distribution of Al localized within the sample surface and present a case study comparing the concentration distributions for each pattern.

SiC has been actively researched and utilized in recent years for applications such as power devices. Due to the various polytypes of SiC, there is a problem where stacking defects, which can lead to disordered stacking arrangements, easily occur. One method for detecting these defects is photoluminescence (PL), which analyzes the light emitted when a sample is stimulated with light. We will introduce a case where mapping measurements were conducted to detect light emission caused by defects.

Imaging SIMS measurements were conducted on a 50μm square area on the emitter side of the NPT-IGBT. Figure 1 shows the ion images of 11B and As obtained from the analysis. It can be seen that 11B and As are injected into the same area. Additionally, while conventional analysis calculates the average concentration of each element over the entire detection area, imaging SIMS measurements allow for the extraction of partial depth profiles, enabling the evaluation of the concentration distribution of dopants localized in the plane (Figure 2).

SiC power devices are expected to reduce power loss and handle large power in a compact form as power conversion elements. We will introduce a case where the thickness and density of the gate oxide film, necessary for improving the characteristics of the device, were evaluated using XRR (X-ray reflectivity) and the bonding state was assessed using XPS (X-ray photoelectron spectroscopy).

In the SiC power MOSFET (Figure 1), the concentration distribution of the dopant element Al in the SiC beneath the gate pad was evaluated using SIMS from both the surface side and the backside (Figure 2). Regardless of the direction of analysis, the distribution beyond a depth of approximately 0.5 μm also matches well, suggesting that the spread of the Al concentration distribution reflects the actual elemental distribution rather than being measurement-induced. SSDP-SIMS analysis is possible even on hard substrates like SiC, which are difficult to process. Please feel free to consult us first.

This paper presents examples of evaluating the depth concentration distribution of the dopant elements N, Al, and P in commercially available SiC power MOSFETs using multiple analysis modes. Depending on the analysis objectives, it is possible to measure multiple elements simultaneously without measuring each element separately under optimal conditions. Examples are provided for cases of simultaneous multi-element measurement and measurements focused on each individual element.

We will introduce a case study evaluating the distribution of the diffusion layer in commercially available SiC power MOSFET devices. In the SiC MOSFET manufacturing process, the channel is formed through ion implantation, activation heat treatment, and epitaxial layer formation. During the active layer formation process, we understood the device structure through TEM observation and evaluated the diffusion layer distribution of the p-type/n-type cross-section and the epitaxial layer from SCM measurements, as well as the depth concentration distribution of dopant elements (N, Al, P) from SIMS measurements. Measurement methods: SIMS, SCM, TEM Product field: Power devices Analysis purpose: Trace concentration evaluation, shape evaluation, product investigation For more details, please download the materials or contact us.

We will introduce an analysis case of commercially available SiC power MOSFET devices. In SiC materials, it is essential to control the materials in a system that includes not only Si but also C, which differs from the conventional manufacturing methods of Si semiconductors. In the process of forming ohmic junctions between the contact electrodes and the SiC layer, we evaluated the elemental distribution and crystal phases, including C, using EDX/EELS analysis with TEM and electron diffraction.

In lock-in thermal analysis, it is desirable to increase the frequency to narrow down the hotspots; however, there is a problem that the sensitivity decreases. Therefore, it is important to shift the measurement conditions from the high-frequency side to the low-frequency side and identify the frequency at which the heating signal begins to be obtained. In this case, we will introduce an example where the heating location associated with leakage current was identified non-destructively in a cylindrical package. Thus, it is possible to identify heating locations even in samples with complex three-dimensional structures, which are difficult to analyze using the liquid crystal method.

In power semiconductor devices, crystal defects may be formed within the Si substrate for lifetime control. This presents a case study evaluating the carrier concentration distribution due to differences in thermal treatment conditions of hydrogen ions, one of the elements used to create the lifetime control region.

XPS can evaluate the chemical bonding state of the sample surface, and further detailed assessments can be made through waveform analysis. Additionally, by using assumed parameters in the waveform analysis results, it is also possible to calculate the thickness of surface oxide films, etc. In this document, we will introduce a case where the composition and state evaluation of the SiC surface was performed, along with the calculation of the oxide film thickness from the obtained peak intensity.

We will extract small pieces from the wafer chip without breaking it and thin them down for high-resolution TEM observation and analysis. Furthermore, by cutting and preparing samples while leaving the areas to be analyzed, we will conduct TEM observation and analysis of the target areas from any desired direction and provide the data. Through advanced TEM sample preparation techniques, we will meet various observation, analysis, and evaluation needs.

The detection sensitivity in SIMS analysis depends on the amount of sputtered sample per unit time. Depending on the element, significantly improved sensitivity can be achieved by limiting the impurities to one element, allowing evaluation down to ppt (parts per trillion) levels of less than 5E13 atoms/cm3, which is effective for assessing low-concentration impurities in IGBT devices and high-purity wafers. This document presents examples of ultra-high sensitivity evaluations of low-concentration impurities in silicon.

In SIMS analysis of layered structure samples, there is a concern that in structures where the layer of interest is located at a deep position from the sample surface, the depth resolution may degrade due to the influence of the concentration distribution of the upper layers. In such cases, it is effective to remove the upper layers through pretreatment before analysis. This document presents an example of selectively and progressively removing layers from InP/InGaAs-based SHBT (Single Heterojunction Bipolar Transistor) samples.

In general, the quantification of major elements with concentrations exceeding a certain percentage in SIMS is considered to be low. However, by using the M Cs+ (M: element of interest) detection mode with Cs+ as the primary ion, it is possible to determine the compositional distribution of major element elements in the depth direction. An example of depth compositional evaluation for Al, Ga, and In in GaN-based LED structures is presented.

For extremely thin films with a thickness of a few nanometers or less, such as natural oxide films on silicon wafers and silicon nitride thin films, we will measure the Si2p spectrum of the sample's surface. By performing waveform analysis on the obtained spectrum, we will determine the proportion of each bonding state and estimate the film thickness from this result and the average free path of photoelectrons (Equation 1).

The miniaturization of devices has increased the need for evaluating the depth distribution of impurities in extremely shallow regions. To conduct an accurate assessment, SIMS analysis using a primary ion beam with lower energy (below 1 keV) is required. Figure 1 shows examples of measurements of Si wafers implanted with BF2+ 1 keV, P+ 1 keV, and As+ 1 keV, using a primary ion beam energy of 250 eV to 300 eV.

In semiconductor device manufacturing, it is necessary to remove metals remaining on the backside of the wafer from the perspective of improving yield, and it is important to quantitatively understand the amount of metal components remaining. To investigate the concentration distribution of metals remaining on the backside within a range of 500 µm from the bevel area, we conducted evaluations using TOF-SIMS. TOF-SIMS has the spatial resolution to detect metal components only near the bevel area, and by using standard samples with known concentrations, it is possible to quantitatively calculate the concentrations.

We will introduce a case where the thickness of extremely thin films, such as natural oxide films on silicon wafers and silicon nitride thin films, which are less than a few nanometers thick, was calculated using XPS analysis. By measuring the Si2p spectrum of the surface of the Si wafer and performing waveform analysis on the obtained spectrum, we can determine the ratio of the presence of each bonding state, and from this result, it is possible to estimate the film thickness based on the average free path of photoelectrons (Equation 1). XPS allows for non-destructive and simple calculation of thin film thickness on substrates as broad average information.

Typically, SIMS measurements are conducted using chips with a flat surface of a few millimeters in size, but analysis can also be performed on small chips or samples with special shapes, typically less than 1 mm in size, by applying a fixed pre-treatment. Some samples that require investigation of trace components may have shapes that are not suitable for analysis under normal conditions, such as tiny chips or wire-like samples (Figure 1). In such cases, analysis is performed after fixation (Figure 2). Additionally, analysis may be possible for cross-sections, side surfaces, or samples with special shapes by applying pre-treatment.

From the perspective of cost reduction, the use of high-resistance Si substrates for power devices made of GaN is expected. However, it is said that if Al and Ga diffuse to the surface of the Si substrate during high-temperature film formation, a low-resistance layer is formed, leading to leakage. Therefore, we will introduce a case where SIMS analysis was conducted to evaluate the presence or absence of Al and Ga diffusion into the Si substrate. To accurately assess trace diffusion, measurements were conducted from the Si substrate side towards the GaN layer.

GaN-based LEDs have become widely used for lighting applications. To enhance the light extraction efficiency, textured surfaces may be created; however, these textures can lead to a degradation in depth resolution during depth analysis. We will present cases where flattening processing was applied to textured surfaces to mitigate the degradation of depth resolution and evaluate the depth concentration distribution, as well as cases where analysis was conducted from the backside (substrate side).

We will introduce a case where the depth concentration distribution of dopant elements such as Mg and Si in GaN-based LED structures was evaluated using multiple analysis modes. By selecting the optimal analysis mode according to the purpose in SIMS analysis, more precise evaluations can be achieved, so please feel free to consult with us.

SiC power devices are expected to reduce power loss and handle high power in a compact form as power conversion elements. The quality evaluation of SiC substrates, which is necessary for manufacturing these devices, has become a challenge. We propose a method to evaluate and quantify the crystal orientation, in-plane defect distribution, surface roughness, and impurities of SiC substrates, as well as to visualize these factors. Measurement methods: XRD, AFM, PL, SIMS Product fields: Power devices, lighting Analysis purposes: Trace concentration evaluation, structural evaluation, shape evaluation, failure analysis, defect analysis For more details, please download the materials or contact us.

SRA (Spreading Resistance Analysis) is a method that involves polishing a sample at an angle, making contact with two probes on the polished surface, and measuring the spreading resistance (Figure 1). By evaluating the carrier concentration distribution, it is possible to gain insights into the activation status of dopants. As an example, we present a case of SRA conducted on the central and outer regions of a diode chip surface after package opening (Figure 2).

By comparing the impurity concentration distribution obtained from SIMS with the carrier concentration distribution from SRA, not only can we understand the activation state of the dopants, but the presence of unknown impurities and inherent structural issues are also suggested in areas that do not align. Using Profile Viewer, it is possible to overlay and analyze SIMS data and SRA data on your own. As an example, we will introduce a case where SIMS and SRA were performed on a commercially available diode chip, specifically on the central part of the surface and the outer edge (Figure 1) as well as the backside after the package was opened.

In GaN-based LEDs, it is said that the diffusion of the dopant element Mg into the active layer leads to a decrease in luminous efficiency. This document presents a case study where SIMS analysis was conducted on GaN-based LED structural samples from both the surface side and the sapphire substrate side (back side) to evaluate the depth profile of Mg concentration.

In XPS analysis, in addition to the photoelectron peaks used for evaluation, photoelectron peaks from other orbitals and Auger peaks from X-ray excitation are also detected. Depending on the combination of elements, these sub-peaks can overlap with the target peak and interfere with the evaluation. *Typically, a strong photoelectron peak emitted from an inner shell level close to the outer shell is used. In the quantitative calculations of XPS analysis, the removal of such interfering peaks is primarily performed using the following two methods: 1. Removal using sensitivity coefficient ratios 2. Removal using waveform separation

SR measurement involves diagonally polishing the sample and moving two probes in contact with the sample surface while measuring the electrical resistance directly beneath. The depth from the sample surface at a certain measurement point is determined by the product of the value of Sinθ (bevel angle) when the angle of diagonal polishing of the sample is θ and the distance from the bevel edge to that measurement point. The distance from the bevel edge is calculated as [X-step] × [number of measurement points].

1. Measure the spreading resistance (Ω) directly beneath the surface of the sample that has been polished at an angle by contacting two probes. 2. Create a calibration curve from the measurements of the standard sample and use it to convert the resistance to resistivity (Ω·cm). Additionally, perform distribution corrections as needed through volume correction. *) The calibration curve varies depending on the conductivity type (p-type/n-type) and the surface orientation. 3. Use the relationship between resistivity and carrier concentration*) to calculate the carrier concentration (/cm³).

Imaging SIMS is effective for obtaining spatial distribution information of hydrogen and impurities at ppm levels. By using a RAE (Resistive Anode Encoder) detector for projection-type imaging SIMS, it is possible to obtain distribution images with a larger diameter and deeper regions compared to the commonly used scanning methods in imaging analysis.

- Photoluminescence measurement (PL measurement) can be conducted not only at room temperature but also by placing the sample in a cryostat for low-temperature measurements. Low-temperature measurements tend to show an increase in peak intensity and a decrease in peak full width at half maximum compared to room temperature measurements, which may provide insights into various energy levels. - Below, we will explain the precautions for low-temperature measurements and the differences in spectral shapes between room temperature and low-temperature measurements.

This text introduces an example of evaluating the composition and bonding states of GaN films used in LEDs and power devices using XPS. Understanding how the composition and bonding states change due to film formation conditions and surface treatments is effective for process management. It is important to appropriately select the X-rays used during the evaluation according to the purpose. Measurement conditions for each focus will also be presented.

XPS is a surface-sensitive technique, so carbon derived from organic contaminants due to the atmosphere is detected at a significant level. Reducing the influence of this carbon from organic contaminants is important for evaluating the original composition of the film. Typically, Ar ion sputtering is used to remove organic contaminants, but damage caused by sputtering may prevent the evaluation of the film's original composition and bonding states. We present an example where the influence of carbon from organic contaminants was reduced by removing the surface oxide layer using wet etching instead of Ar ion sputtering.

In XPS analysis, the binding state evaluation of the material surface is conducted by observing the energy of photoelectrons obtained through X-ray irradiation. It allows for the assessment of whether metal elements are in an oxidized state, and for elements with significant energy shifts (chemical shifts) due to oxidation, it also enables the evaluation of the presence and proportion of multiple valences. Below are the main metal elements and oxides for which multiple valence evaluations are possible.

The electron diffraction method using an electron microscope is classified into three types based on the way the electron beam is incident on the sample. The characteristics of each type and examples of data are presented. It is necessary to choose the appropriate method according to the size of the evaluation object and the analysis purpose.

GaN, which is being utilized as a power device and optical device, has a hexagonal wurtzite structure and exhibits crystallographic asymmetry (Ga polarity and N polarity) in the c-axis direction. The growth processes of epitaxial films differ between Ga polarity and N polarity, and the surface physical properties and chemical reactivity of the crystal also vary. In this document, the polarity of GaN was evaluated through annular bright field (ABF)-STEM observation. As a result, the positions of the Ga sites and N sites were identified, allowing for a visual clarification of the characteristics of Ga polarity and N polarity. Measurement method: TEM Product fields: Power devices, optical devices Analysis objectives: Shape evaluation, structural evaluation, thickness evaluation For more details, please download the document or contact us.

GaN-based high electron mobility transistors, known as GaN HEMTs (High Electron Mobility Transistors), achieve a two-dimensional electron gas layer through an AlGaN/GaN heterostructure, resulting in high electron mobility. This document presents a case study of the disassembly and evaluation of commercially available GaN HEMT devices. Using HAADF-STEM, we confirmed the crystallographic integrity near the AlGaN/GaN interface. Additionally, we assessed the compositional distribution using EELS. Measurement methods: TEM, EELS. Product field: Power devices. Analysis purpose: Structural evaluation, film thickness evaluation. For more details, please download the document or contact us.