Phase Mapping Using a Combination of Multi-Functional Scanning Electron Microscopy Detectors and Imaging Modes

Abstract

Microstructure degradation and phase transformations are critical concerns in nickel-based superalloys during thermal exposure. Understanding the phase transformation mechanism requires the detailed mapping of the distribution of each phase at different degradation stages and in various precipitation sizes. However, differentiating between phases in large areas, typically on the scale of millimeters and often relying on scanning electron microscopy (SEM) techniques, has traditionally been a challenging task. In this study, we present a novel and efficient phase mapping method that leverages multiple imaging detectors and modes in SEM. This approach allows for the relatively rapid and explicit differentiation and mapping of the distribution of various phases, including MC, M₂₃C₆, γ′, and η phases, as demonstrated in a typical superalloy subjected to aging experiments at 800 °C.

1. Introduction

Nickel-based superalloys are extensively utilized in extreme service conditions owing to their remarkable properties, including high strength and high corrosion resistance. These exceptional attributes primarily arise from a substantial volume fraction of alloying elements such as Cr, Mo, Ti, and C. However, the incorporation of these elements inevitably leads to the formation of complex phases, including carbides and intermetallic compounds, alongside the major strengthening phase γ′-Ni₃Al [1]. During service and thermal exposure processes, these phases undergo degradation and may transform into other phases. A specific phase of concern is the η-Ni₃Ti phase, which is believed to transform from γ′-Ni₃Al [2,3]. A fully grown η phase is generally considered unfavorable for high-temperature properties as it depletes γ′ and increases the local strain concentration [4,5]. Therefore, these phase transformations are important as both the original degraded phases and newly transformed phases will greatly affect the properties of alloys. Understanding the η phase nucleation and growth mechanism and controlling the precipitation of the η phase are of both scientific and industrial significance. Recently, it was found that the degradation of MC carbides is closely related to the precipitation of the η phase, following the transformation reaction MC + γ → M₂₃C₆ + η [6,7]. However, despite long-standing documentation of the degradation of MC carbides, the detailed transformation mechanisms associated with η formation remain unclear. For instance, the precipitation sequences of the transformed phases and the specific relationships between η, MC, and γ′ have rarely been reported [8,9]. Understanding these relationships and mechanisms is critical for controlling η precipitation in superalloys.

One of the challenges in studying these transformations is the difficulty in observing initial precipitates of small size (~nm) and distinguishing each phase in large areas. While traditional transmission electron microscopy (TEM) can explicitly identify every phase, its application is limited to small areas. Although modern SEM techniques offer high resolution in the nanometer range, they face challenges in providing sufficient phase contrast to accurately map different phase distributions. This is especially true when dealing with very similar compositions, as exemplified by the difficulty in distinguishing γ′-Ni₃(Al,Ti) and η-Ni₃Ti in normal SEM images [9,10]. The compositions of these phases are closely related, making their distinction challenging even with the successful implementation of backscatter electron detectors. Addressing these challenges is crucial for advancing our understanding of the intricate phase transformations in nickel-based superalloys during thermal exposure, such as the η formation process, and ultimately for controlling detrimental phase precipitations. Using the advanced detection system in newly developed SEM shows a potential way to improve the contrast of the second phase in metals [11].

In the current study, we developed a relatively rapid and user-friendly approach to identify and map different phases in SEM by combining various images captured from different detectors and imaging modes, demonstrating the feasibility of achieving excellent phase contrast in advanced SEM. The advantage of this method is its capability to distinguish different phases with high resolution over large areas without requiring difficult sample preparation, compared to other characterization methods. In the current research, using this method, the η phase was distinguished from MC and γ′ phases, providing a precise configuration of the phase distribution in a quasi-3D manner. This will facilitate our understanding of the degradation process of MC carbides and the associated η phase formation mechanism. This method and technique are also applicable to other phase systems and are likely to be useful in resolving questions such as phase transformation mechanisms on multiple scales.

2. Materials and Experimental Methods

The nickel-based superalloy Waspaloy was chosen for this study. The alloy has a measured chemical composition of 3.1% Ti, 1.6% Al, 19.5% Cr, 13.5% Co, 4.3% Mo, 0.07% C, 0.05% Zr, and the balance Ni and was aged at 800 °C for 10,000 h. The specimens were gradually ground, polished, and culminated with 0.05 µm colloidal silica. Electrolytic etching (10% phosphoric acid at 3–5 V for 3–5 s to remove the γ matrix) was used for microstructure observation by using a field emission gun scanning electron microscope (SEM) equipped with a trinity detection system (FEI Apreo 2). The specific usage of different detectors and imaging modes in SEM is the core issue of the current research, which will be explained in detail in the Section 3. An electron probe X-ray micro-analyzer (SHIMADZU EPMA-1720, Shimadzu, Kyoto, Japan) was used for elemental composition detection for phases. The FIB (Focused Ion Beam, FEI Helios NanoLab 600i, Thermo Fisher Scientific, Waltham, MA, USA) lift-out technique was used to extract specific sites of aged samples for further TEM analyses (FEI-Talos F200X, Thermo Fisher Scientific, Waltham, MA, USA).

3. Results and Discussion

Figure 1 presents a typical grain boundary structure containing primary MC carbides after heat treatment, before the aging experiment. In the heat-treated sample, grain boundaries exhibit continuous M₂₃C₆ particles and occasional blocky MC carbides, with other MC carbides residing within the grains.

In Figure 2 and Figure 3, we illustrate two instances of phase mapping through image combinations. Specifically, these examples showcase the decomposition of MC carbides on grain boundaries with the subsequent formation of newly transformed γ′, η, and M₂₃C₆ carbides after 10,000 h of aging at 800 °C. Long-term thermal exposure induces major changes in the microstructure, notably the degradation of MC carbides and the formation of new phases such as η. Figure 2a displays a typical secondary electron image taken using the Everhart–Thornley Detector (ETD) in “Standard” mode (“ETD + Standard”). This image exhibits poor contrast, with discrete black dots in the central area representing mainly MC carbides (confirmation of the MC phase will be provided in Figure 4). Characterization via the Angular Backscattered (ABS) Detector in “Standard” mode (“ABS + Standard”) improves the contrast of MC, as shown in Figure 2b. Additionally, the η phase with a whitish contrast can be roughly distinguished, although with lower resolution compared to Figure 2c. While the backscatter imaging technique shown in Figure 2b has been successful in phase contrast imaging, it usually requires a high-energy electron beam (high voltage and high beam current, in this case, 20 kV/3.2 nA), leading to lower resolution. Figure 2c presents a similar backscatter image but utilizes the in-lens detector (T1) in “Optiplan” mode (“T1 + Optiplan”) with low voltage and current (2 kV/50 pA). This detector and mode offer decent phase contrast with high resolution. The fine structure of the η phase, indicated by the red arrow in Figure 2c, is now visible compared to the same area in Figure 2b. It is important to note that the phase contrast between η and γ′ can vary between the “T1 + Optiplan” image (the top part in Figure 2c) and the “ABS + Standard” image (the bottom part in Figure 2b). Figure 2d displays another “Optiplan” image using the ETD detector (“ETD + Optiplan”), providing a much better contrast for M₂₃C₆ due to the subtle difference in height between these phases. Combining images from different detectors and modes, and leveraging the distinct contrast of each image, allows for the exact phase distributions of all phases in the reaction area to be obtained and schematically shown in Figure 2e.

Figure 3 presents another example closely resembling the one in Figure 2. Additionally, SEM-EDS mapping for the area was conducted and is illustrated in Figure 3e. EDS point or mapping is frequently employed to ascertain composition and serves as a supplementary technique for phase identification. In this case, particles rich in Cr and Mo can be confidently identified as M₂₃C₆-type carbides and the surrounding γ′ particles rich in Ti and Al. However, due to the resolution limitation, distinguishing between the distribution of η and γ′ is challenging since they share similar compositions. It is also difficult to identify the MC phase via EDS when the size of MC is small (in nm), as seen in Figure 2 and Figure 3a–d. For the purpose of comparison, EPMA mapping, which is more sensitive to chemical composition than EDS, was conducted for the same area and is displayed in Figure 3f. It exhibits a similar composition distribution to EDS mapping and demonstrates higher sensitivity, such as clearly higher Ti content in the MC particles area. Nevertheless, similar limitations observed in EDS, such as the difficulty of mapping the η phase distribution, also apply to EPMA.

To validate the proposed phases shown in Figure 3 (and also indirectly in Figure 2), particularly the η phase and small MC dots, we conducted a site-targeted FIB lift-out on the region in Figure 3a. This was followed by detailed characterization using TEM and STEM (EDS), as depicted in Figure 4. The precise location of the lift-out site is indicated in Figure 4a, which is extracted from Figure 3a. The overview image of the lifted FIB specimen is shown in Figure 4b, and all of the phases in the core area of Figure 4b are schematically presented in Figure 4c, combining information such as chemical composition obtained through STEM-EDS mapping in Figure 4d and the diffraction patterns of each phase (Figure 4i–h). Special attention was given to the small MC dots, as displayed in Figure 4e–h, which correspond to the boxed area in Figure 4b. The EDS mapping in Figure 4g, the selected area diffraction pattern (SADP) in Figure 4i, and the corresponding dark field image in Figure 4h confirm that the small dot-like particles are M(Ti)C-type carbides. These MC carbides align well with the black dot particles observed in Figure 4a (and Figure 3a). Similarly, the η-Ni₃Ti phase and M(Cr)₂₃C₆ carbides are confirmed by the SADP (Figure 4j and Figure 4h, respectively) and the EDS mapping (Figure 4d). It is worth noting that there are two sets of η phases with a rotation angle of 70° around the zone axis of [11$\overline{2}$0]. Overall, the phase distributions in the cross-section plane (schematically shown in Figure 4c) correspond well to the phase distributions in the horizontal plane (schematically shown in Figure 3d). This validates the accurate phase identification and distribution based on the contrast in images as shown in Figure 2e and Figure 3d.

There are several techniques available for phase identification, such as EBSD (electron backscatter diffraction) [12,13], SEM-EDS [14], STEM-EDS [15], and TEM-PED (Precession Electron Diffraction) [16]. Traditionally and theoretically, phase mapping or identification can be achieved using EBSD for bulk samples. However, even with advanced EBSD, it would still be challenging to distinguish the small particles in the size of nanometers (MC particles herein). Additionally, the EBSD scanning process usually takes a long time, especially with smaller step sizes for small particles [17]. Furthermore, preparing samples for EBSD requires higher quality and more time compared to normal sample preparation for SEM observation. Also, it might be possible to identify phases through EDS point analyses with quantitative data in SEM. However, as evidenced in Figure 2e,f, obtaining a clear view of the phase distributions through element mapping is extremely challenging due to resolution limitations. Moreover, element mapping is typically a time-consuming process, even with modern EDS detectors in SEM. TEM (PED) [9] and STEM (EDS) have been proven to be powerful tools for phase identification and mapping with high resolution. However, sample preparation for TEM, either using conventional methods (such as twin-jet polishing) or FIB lift-out, is time consuming. Additionally, the observation area is limited in the TEM specimen. Compared to these aforementioned techniques, the method demonstrated herein is relatively fast and reliable for phase mapping. It involves merely combining a few SEM images with different contrasts for the same area. The advantage of this method lies in the balance of probing small phases (in nanometers) and large searchable areas (in millimeters) in SEM, with relatively easy sample preparation and image acquisition. A brief summary of the comparisons between the method proposed herein and other phase mapping methods is listed in

Table 1. Comparison between the method proposed herein and other phase mapping methods.

Phase Mapping Methods	Advantages	Disadvantages	Note
Imaging with different detectors and modes	Relatively quick and accurate		Proposed herein
EBSD	High accuracy for large-size phases	High demand of sample surface; time consuming for scanning; limited resolution	Widely used for accurate phase mapping for bulk alloys [8]
SEM-EDS		Time consuming for scanning; limited accuracy	Common method for auxiliary phase identification and mapping for bulk alloys
STEM-EDS	High accuracy	Difficulties in specimen preparation; time consuming for scanning; limited research areas	Common method for auxiliary phase identification and mapping for TEM foils
TEM-PED	High accuracy	Difficulties in specimen preparation; time consuming for scanning, limited research areas	Widely used for accurate phase for TEM foils with nanometer resolution [9]

.

4. Conclusions

A method that combines the varying contrasts from different imaging detectors and modes was developed, enabling relatively rapid and precise identification and mapping of phase distributions, particularly for phases such as MC, M23C6, γ, and η. The method reaches a balance of probing small phases (in nanometers) and exploring large areas (in millimeters) in SEM, with relatively easy sample preparation and image acquisition. The method is applicable to other phase and alloy systems and is significant in revealing phase transformation processes and mechanisms.