Introduction

Over the years, the focused ion beam (FIB) technique has evolved as a versatile tool in materials characterization around the world. Despite being its major focus on applications in semiconductor industry, the FIB has been instrumental also in other fields such as metallurgy, geology, forensics, and pharmaceuticals, to name a few [1]. One of the chief applications of FIB technique has been its application in site-specific specimen preparation for transmission electron microscopy (TEM) characterization, for which it is known to be highly skillful [2, 3]. Also, the past few years have seen a tremendous progress in applications of FIB in specimen preparation for micromechanical characterization within the scanning electron microscope (SEM) which is proving to be handy with microstructurally challenging samples such as irradiated specimens and in situations, where sample quantity is limited to carry out bulk mechanical characterizations [4, 5]. In addition to these, FIB is also useful in failure analysis [6, 7], cross-section imaging of complex microstructures, 3D materials characterization [8], energy storage materials research [9], and even been used along with other analytical tools such as Auger electron spectroscopy and secondary ion mass spectroscopy (SIMS) as an analytical tool [10].

One of the most undervalued applications of FIB must certainly be its ability to mill and allow cross-sectional imaging of complex microstructures. In situations, where, conventional imaging does not yield substantial information about the microstructure, FIB can be a handy tool and make a great value addition to the amount of information, we can gather from SEM analysis in a dual-beam FIB-SEM instrument. This, in combination with the analytical tools within SEM such as energy-dispersive spectroscopy (EDS) and electron backscatter diffraction (EBSD), FIB can be very effective and has immense applications in the domains of metallurgy and materials science. With the rapid progress in additive manufacturing, this could also be very useful in powder metallurgy domain with its ability to investigate the internal microstructures of powder samples. It is evident that subsurface microstructure analysis through FIB itself is a productive characterization and warrants full attention. In addition to this, a more exciting feature for a material scientist here is that this analysis can be carried out as part of the TEM sample preparation protocol itself. Instead of thinning the sample down to electron transparency, the thinning process may be paused and incorporate few additional steps to prepare the sample for a qualitative and quantitative SEM investigation and gain useful additional information about the microstructure of the sample.

Despite its usefulness, however, it has not been used extensively in materials research, and it may be due to the rather sophisticated applications of FIB, as mentioned above, which takes precedence in many cases. In this study, an attempt is made to evaluate the subsurface microstructural features of Fe-based composite systems (Fe–15 wt.% Y2O3 and Fe–15 wt.% ZrO2) by making use of dual-beam FIB-SEM system, and this will be demonstrated as a part of TEM specimen preparation protocol itself. The composite systems studied here are model systems for understanding the dispersoid behavior in an oxide dispersion strengthened (ODS) steel, which is being developed as a candidate nuclear core structural material for next-generation fast breeder reactors (FBR’s) [11,12,13,14]. ODS steels consist of a fine (~ 0.3–0.5 wt.%) distribution of ceramic dispersoids in their matrix which are known to arrest the dislocation motion thereby extending the lifetime of structural alloys inside a nuclear reactor. However, the systems used in this study consist of 15 wt.% of ceramic dispersoids along with Fe and can be considered as model systems. Employing model systems with concentrated amount of dispersoids (~5–15 wt.%) has been a widely adopted approach [15,16,17,18], and accordingly, the same is used here as well. These model systems will allow us to evaluate the behavior of dispersoids with processing and thermal treatments conveniently than the actual ODS steel, and detailed characterizations of their initial microstructures are reported earlier as a function of processing variables [19, 20]. Among the two systems under investigation, Y2O3 is the widely used ceramic dispersoid in ODS steels, and ZrO2 is the candidate dispersoid which is being evaluated and compared with Y2O3 for its dispersion behavior. The interest in ZrO2 added steel emerge from the recent studies in both ZrO2 and Zr added ODS alloys, which have shown promising mechanical, thermal properties along with control over the dispersoid size during the consolidation process [21,22,23]. In view of this, the objective of this research work is to understand the behavior of these two model ODS alloy systems at high temperature and a comparative analysis of ZrO2 and Y2O3 dispersoids. The competence of FIB_SEM technique and its complimentary nature with TEM and EDS techniques for subsurface microstructural analysis of model ODS alloy systems is also demonstrated.

Materials and Methods

The elemental Fe, ZrO2, and Y2O3 powder were commercially procured. Two composite systems with compositions Fe–15 wt.% Y2O3 and Fe–15 wt.% ZrO2 are synthesized through ball milling under argon atmosphere. A rotation speed of 300 rpm is maintained for both the systems, and milling was carried out for 60 h. A detailed powder characterization is previously reported for both the systems [19, 20].

Further, the milled powders were annealed at 1000 °C for 1 h in an evacuated quartz tube, to evaluate the thermal behavior of composite powder. The temperature chosen is typical of ODS steel powder consolidation through hot isostatic press or spark plasma sintering, to simulate the actual processing conditions. The FEI make dual-beam (e-beam and Ga+ beam) FIB-SEM instrument, Helios Nanolab 600i attached with an Apollo X Silicon Drift Detector is used for SEM imaging, FIB sample preparation, and subsequent EDS analysis.

In all the three samples, the powder particles are prepared for SEM analysis by spreading a miniscule amount on a carbon tape, followed by a thin gold coating. For the FIB analysis, no special sample preparation was made in addition to the regular SEM analysis. However, while choosing a powder particle for FIB analysis, a reasonably large (~ 15–20-μm diameter) and relatively flat particle was chosen so as to have sufficient sample region for FIB analysis.

TEM analysis is carried out with a Philips make CM-200 Analytical Transmission Electron Microscope, operated at 200 kV. The instrument is equipped with a TVIPS make 2K×2K CCD camera and Oxford make EDS for chemical analysis. The ImageJ software is used for the image analysis and processing.

Results and Discussion

Microstructural Investigations Through SEM

The results of preliminary SEM analysis of Fe–15 wt.% ZrO2 are shown in Fig. 1a–d. Figure 1a shows the as milled powder sample after 60 h of ball milling, where the individual powder particles are clearly observed. The micrograph distinctly allows us to investigate the shape and size of the powder particles, and statistical analysis showed the particle size to be in the range of ~ 5–15 μm, as reported previously [19]. Figure 1b shows the micrograph of the same sample annealed for 1 h at 1000 °C. A drastic change in the microstructure can be observed here with agglomerated powder particles measuring up to 150–200 μm. This is not surprising, as the exposure to high-temperature results in neighboring powder particles to come together and coarsen. The agglomeration of individual powder particles with compositional inhomogeneity is further evident as shown by the inset of Fig. 1b. The inset micrograph, obtained at a higher magnification, shows the clearly grown crystallites of powder particles, measuring up to 0.5–3 μm. It is evident here, that individual composite particle as shown in Fig. 1a is no longer visible, but rather agglomeration and compositional non-uniformity is apparent. To further confirm the chemical inhomogeneity, EDS analysis was carried out on the same sample on two distinct crystallites marked 1 and 2, and results are shown in Fig. 1c and d, respectively. Distinct and dominant peaks for Fe are observed in Fig. 1c; whereas, Fig. 1d shows the dominant peak for Zr, further strengthening the microstructural observations.

Figure 1
figure 1

(a) Secondary electron (SE) micrograph of as milled Fe–15 wt.% ZrO2 powder. (b) SE micrograph of annealed powder and inset showing the same sample at a higher magnification. (c) and (d) Energy-dispersive spectra from the annealed sample which shows Fe-rich and Zr-rich regions, respectively. Figure (e) and (f) SE micrographs of annealed Fe–15 wt.% Y2O3 samples at lower and higher magnifications along with (g) and (h) showing the energy-dispersive spectra from Fe-rich and Y-rich regions, respectively.

Full size image

Further, it is observed that the results of SEM analysis of Fe–15 wt.% Y2O3 are no different from above. Figure 1e and f shows the SEM micrograph of the sample at lower and higher magnifications after annealing for 1 h at 1000 °C. Similar to our previous observations, an agglomerated microstructure is evident with particle sizes about ~ 100–200 μm which is observed. The presence of chemical inhomogeneity with Fe- and Y-rich regions (marked as 1 and 2 in Fig. 1f) is prevalent here as well, as evident from the EDS analysis in Fig. 1g and h, respectively.

It may be concluded here that; SEM analysis of the annealed powder sample does give preliminary information about the powder morphology and microstructure along with chemical information through EDS. However, more details about the dispersoid and matrix microstructure and quantitative measurements are not possible here owing to the agglomerated microstructure and further analysis through advanced techniques such as TEM becomes essential. However, the specimen preparation for TEM analysis of this agglomerated microstructure is also challenging owing to the difficulty in sample preparation of micron-sized particles. The conventional argon ion milling is challenging in terms of time required and difficulty but still it does not ensure a good sample despite the best efforts. Thus, alternate methods have to be incorporated for specimen preparation and inclusion of FIB technique thus becomes inevitable here. In the next section, application of FIB in subsurface microstructure analysis of the two composite model ODS systems is discussed, which has been carried out as part of the TEM specimen preparation protocol itself. This analysis is believed to provide valuable insights on the grain growth, matrix–dispersoid microstructure, and chemistry in these model ODS systems.

Preparation of Specimen for Subsurface Microstructure Analysis Using FIB_SEM

As discussed previously, the FIB has been employed here as part of TEM specimen preparation protocol itself, but with special emphasis on subsurface microstructure analysis. At the outset, the selection of region of interest is carried out by carefully choosing an area which is nearly flat, at least to an extent of about 15–20 μm. Although a flat region is not always guaranteed, an attempt to look for it is important, as the uneven surface add more challenges to the TEM specimen preparation. The platinum coating to mark and protect the selected region will be carried out next, with the assistance of electron beam (e−-beam) and Ga+ beam (I-beam), in two layers. The adhesion of the coating is known to be better with initial e−-beam deposition, which is then followed by a relatively thicker and faster deposition through I-beam for superior protection of region of interest. Such a region of interest, coated with platinum is shown in Fig. 2a, where a rectangular-shaped coated region, measuring 15 μm × 1.5 μm, and a thickness of 2 μm is shown. Since the e− beam and I-beam are at 52° to each other, aligning the instrument becomes important, such that region of interest can be seen by both beams simultaneously, and accordingly, the alignment procedures are performed at this stage. Further, to separate the rectangular region of interest from rest of the sample, milling is carried out using I-beam operated at high beam current and voltage of 21 nA and 30 kV, respectively. This is known as trench cutting, and this step is followed by further milling the rectangular region at moderate currents and voltages of I-beam to reduce the thickness down to about 1.5 μm, such that only the platinum-coated region of interest is remained. Following this, an undercut is performed on the milled rectangular region, detaching the region of interest from the bulk of the sample, such that it is connected to the sample only by a small rectangular region, as shown in Fig. 2b.

Figure 2
figure 2

(a–e) SE micrographs depicting the steps involved in specimen preparation through FIB

Full size image

The next step in the process is separating the region of interest from the bulk of the sample and welding it to an external copper grid, which is compatible with TEM specimen holders. For this, we first attach the region of interest to a motorized Tungsten needle inside the instrument and drive it closer to the sample and then weld the region of interest to the needle with platinum. Driving the needle close enough to the sample without disturbing the region of interest in very intricate and needs careful execution. The welded specimen is then removed completely from the sample and slowly driven away. The specimen free from the bulk sample and welded to the needle is shown in Fig. 2c. The specimen is then driven slowly toward the copper grid carefully and once again, by making use of platinum, welded onto one of the grid posts, as shown in Fig. 2d, and the sample is pointed with an arrow mark. The needle is then detached from the grid and driven away to its park position. At this stage, the first part of preparation protocol is completed. A closer look up of the prepared specimen at this stage is shown in Fig. 2e, and the specimen typically has a dimension of 15 × 5 × 1.5 μm, at this stage.

In a conventional TEM specimen preparation protocol, the sample is further thinned down up to electron transparency, with not much focus on the cross-sectional microstructure of the specimen. However, during the current study, keen observations are made using imaging and chemical analysis, and results are discussed in the next section.

Subsurface Microstructural Analysis of FIB Milled Specimen Through SEM and EDS

To investigate the subsurface, the grid welded specimen was thinned down to about 1 μm and cleaned using low current and low-voltage I-beam. Since the focus here is subsurface analysis and not thinning down the sample for TEM analysis, it would be ideal to keep a reasonable thickness of the specimen such that structural instability does not hinder the analysis. However, care must be taken to clean the surface from any undue gallium traces on the surface due to high-power milling process, and thus, low-power cleaning is an essential step here. The results of subsurface microstructural analysis of Fe–15 wt.% ZrO2 composite material are shown in Fig. 3. Figure 3a shows the cross-sectional secondary electron micrograph of the specimen taken at a specific tilt angle, such that electron beam encounters the specimen to raster scan the entire region of interest. Region of interest is found to be sharp and distinct, with crystallites of different morphologies observed with much clearer view of their spatial distribution. EDS analysis is employed on the same specimen to map the region as per its chemistry, and results are collectively showed in Fig. 3b. The topmost region in Fig. 3b shows the mapped region of interest and elemental map** corresponding to Fe, Zr, and O; is shown below it, sequentially. Complimenting the microstructural observations, the EDS analysis suggests that there is a clear separation of Fe- and Zr-rich regions with majority of Zr and O which are pushed toward the grain boundary between two Fe crystallites. This is validated by the co-existing Zr and O signals at the sandwiched region and their clear absence from the Fe-rich region. This result is convincingly bringing out the agglomeration behavior of ZrO2 crystallites under the influence of high temperature and in addition also providing insights about the spatial distribution of these crystallites.

Figure 3
figure 3

(a) SE micrograph showing subsurface microstructure of annealed Fe–15 wt.% ZrO2 sample. (b) EDS elemental map** of the same sample; depicting the elemental distribution in the same region.

Full size image

A similar analysis of subsurface microstructure is also carried out for the Fe–15 wt.% Y2O3 system to understand the behavior of individual crystallites and its microstructural comparison with Fe–15 wt.% ZrO2 system. The results of this analysis along with chemical analysis through EDS elemental map** are shown in Fig. 4. The sample preparation for a damage free surface which is ready for the analysis was similar to the previous sample as explained previously. Figure 4a shows the secondary electron micrograph of the sample, where a completely different perspective of the microstructure in comparison with Fig. 1e and f can be seen. The FIB milled cross-sectional micrograph clearly brings about the growth of individual crystallites and their spatial distribution, similar to the previous case. In addition to the large crystallites, the presence of fine crystallites within the large Fe grain is also observed, some of which are shown with the circles drawn in Fig. 4a. Further, the results in Fig. 4a are complimented with the EDS analysis and depicted in Fig. 4b, where the elemental maps corresponding to Fe, Y, O and overlay of all the maps are shown. A clear enhancement in Y and O signals is observed in crystallites which were observed to be bright through SEM micrograph in Fig. 4a. Correspondingly, the region which is devoid of Y signal is dominated with Fe signals, showing the large growth of Fe grains. The co-existing Y and O region is most likely to be Y2O3, which indicates a large growth of Y2O3 crystallites. The crystals are observed to possess perfect edges with hexagonal shape in some cases, in contrast with ZrO2 crystallites, which showed no such perfect crystalline shape. In addition, there seems to be regions (marked with a circle and shown with an arrow mark) where Fe and O are enriched together with clear lack of Y signal, possibly hinting us at a potential new phase formation of Fe and O chemistry.

Figure 4
figure 4

(a) SE micrograph showing subsurface microstructure of annealed Fe–15 wt.% Y2O3 sample. (b) EDS elemental map** of the same sample; depicting the elemental distribution in the same region.

Full size image

It is quite evident that both model systems show a substantial grain growth when exposed to high temperatures. However, it may be noted that there are definite differences between the two systems in terms of their grain growth behavior. In comparison with the ZrO2 system, the Y2O3 crystallites seem to have undergone an ordered growth and randomly distributed throughout the specimen region of interest. Contrary to this, the Fe–15 wt.% ZrO2 system does not seem to be growing with clear facets and sharp edges, and the dispersoid crystallite sizes also seem to be finer. Most interestingly the presence of fine Y2O3 inside Fe grains is observed in Fe–15 wt.% Y2O3 system as detailed earlier but no such evidence for ZrO2 is found in Fe–15 wt.% ZrO2 system, which is intriguing and needs a thorough understanding. The possible reason for the difference in grain growth behavior might lie in the microstructures of the ceramic species prior to annealing. It was reported by us previously [20] that, after 30 h of ball milling Y2O3 begins to lose its crystallinity and by 60 h of ball milling, it completely amorphized, as evidenced through XRD and TEM analysis. Incidentally, the Fe–15 wt.% ZrO2 system has shown to be retaining its crystallinity even after 100 h of ball milling for which enough evidence is gathered from XRD and TEM analysis [19]. Thus, it is logical to correlate this dissimilar grain growth behavior between the two systems to the initial microstructures. It is understood that amorphous nature of oxide species in Fe–15 wt.% Y2O3 system and nano-crystallinity in Fe–15 wt.% ZrO2 system has resulted in a rapid grain growth in both the systems. This is also fueled by the fact that ball milling induces a substantial microstrain in the material due to deformation of microstructure, which also drives the grain growth phenomenon.

To seek a clearer understanding, an attempt is made here to evaluate the crystallite sizes of both Y2O3 and ZrO2 from the cross-sectional micrographs shown in Figs. 3a and 4a, and results are summarized in Fig. 5. The measured ferret diameters of the oxide crystallites are reported here as a histogram for both the systems. It is observed that the Y2O3 has a broader size distribution of ~ 50–1600 nm whereas ZrO2 has a relatively narrow distribution of ~ 120–980 nm. Another noteworthy insight here is that the histogram for Y2O3 has a peak between 50 and 200 nm, whereas for ZrO2, crystallite size peaks around 300–500 nm. This shift in the peak toward lower crystallite size for Y2O3, however, is understood as because of the retention and observation of fine crystallites in cross-sectional micrograph, as discussed previously. The absence of fine crystallites of ZrO2 within the Fe grains as observed through cross-sectional SEM micrograph does not necessarily mean their complete absence in the material but rather it could be because of their fine structure and limited analysis area through FIB. It is believed that there is a certain possibility of retention of fine ZrO2 inside the Fe grains which can be evaluated through suitable techniques. An attempt is made in the next section to evaluate the microstructure through TEM for investigating the presence of any retained ZrO2 within the Fe grains, and detailed discussion is presented.

Figure 5
figure 5

Histogram of dispersoid crystallite size for the Fe–15 wt.% Y2O3 and Fe–15 wt.% ZrO2 alloys

Full size image

Subsurface Analysis of Fe–15 wt.% ZrO2 System Through Transmission Electron Microscopy

The TEM studies on the Fe–15 wt.% ZrO2 system have been carried out to demonstrate the efficacy of FIB technique in preparing a high-quality TEM specimen from an agglomerated complex powder system and also to understand the evolution of microstructure of this system with high-temperature annealing. The summarized results from the TEM analysis are shown in Fig. 6. A bright-field TEM micrograph is listed as Fig. 6a which shows a varying mass thickness contrast. Both the dispersoids and Fe grains are visible in the region of interest, although it is tough to deduce the chemical nature of two different grains based on bright-field micrograph alone. The selected area electron diffraction (SAED) analysis is carried out on the sample to compliment the chemical analysis, and the resultant SAED pattern is shown in Fig. 6b. A spot pattern is clearly visible confirming the crystalline nature of the sample and growth of the individual crystallites from its initial nanocrystalline state. Figure 6c shows the dark-field micrograph of the same region, recorded with the Fe (110) SAED spots, wherein the inset shows the indexed SAED pattern, taken along the [001] zone axis. The dark field from the same region for the SAED spot for ZrO2 is shown in Fig. 6d, where a large crystallite with an orientation (111) has appeared bright and distinct. It is interesting to note here that, along with this large crystallite of ZrO2, there are finer regions in this sample, which are appearing brighter, some of which are marked with arrows, which corresponds to the fine ZrO2 in the sample. This presence of fine ZrO2 as observed to TEM dark field clearly indicates that the crystallite size distribution observed previously is incomplete, and the fine ZrO2 crystallite sizes from TEM analysis must be included in Fig. 5, for a complete size distribution of Fe–15 wt.% ZrO2 system. Accordingly, Fig. 7 shows the modified crystallite size distribution histogram for this system, which includes the sizes from both SEM/FIB and TEM dark-field analysis. For the comparison, the histogram for Fe–15 wt.% Y2O3 system is also included. Figure 7 indicates a strong correlation between the size distribution of the two systems with oxides in both systems showing a peak between 10 and 200 nm as highlighted in the figure. The Fe–15 wt.% Y2O3 system shows a relatively broader distribution in comparison with the Fe–15 wt.% ZrO2 system, with former system having many crystallites which have grown into micron dimensions (as marked with a red ellipse), despite being subjected to the same initial processing conditions. This relatively rapid growth of Y2O3 in comparison with ZrO2 is an important observation from the application of ZrO2 as a dispersion strengthener, as we expect the dispersoid to have a lower tendency to grow at elevated temperatures. This observation is also in line with the grain growth activation energy study in this system wherein it has been reported [24] that ZrO2 has an activation energy of ~ 379 kJ/mol which is higher than nanocrystalline Y2O3 [25] and validates that ZrO2 is more resistant to coarsening than Y2O3.

Figure 6
figure 6

(a) Bright-field transmission electron micrograph of annealed Fe–15 wt.% ZrO2 sample. (b) SAED pattern for the same sample; displaying the spotty pattern. (c) and (d) The dark-field micrograph corresponds to Fe (110) and ZrO2 (111) crystallites.

Full size image
Figure 7
figure 7

Modified histogram of dispersoid crystallite size for the Fe–15 wt.% Y2O3 and Fe–15 wt.% ZrO2 model alloy system.

Full size image

Conclusions

Microstructural investigations in two model ODS alloy systems are carried out with the help of FIB-SEM and TEM techniques, and results are analyzed to understand the elevated temperature behavior of dispersoids. The resourcefulness of FIB-SEM technique and its complimenting nature with TEM and EDS in understanding the subsurface microstructure and chemistry of model ODS systems is explored, and crucial insights toward ODS alloy development are gathered.

  1. 1.

    The challenges with agglomerated microstructure analysis through conventional SEM have been circumvented by employing cross-sectional FIB-SEM imaging and chemical analysis.

  2. 2.

    FIB-SEM inspected subsurface microstructure revealed the spatial distribution of dispersoids and matrix grains in both the systems. The ZrO2 dispersoids seem to be pushed toward the boundaries between two Fe grains and show a relatively narrow size distribution without any fine dispersoid retention within the Fe grains. In contrast, Y2O3 crystallites shown a broader size distribution with retention fine dispersoids within the Fe grains.

  3. 3.

    TEM analysis on Fe–15 wt.% ZrO2 system compliment the FIB-SEM results by providing the evidence for fine ZrO2 retention within the Fe grains. Dark-field TEM analysis provided the size distribution for the fine ZrO2 and together, with FIB-SEM size data, provided a complete size distribution histogram, for the Fe–15 wt.% ZrO2 system.

  4. 4.

    The narrow crystallite size distribution for ZrO2 in comparison with Y2O3 is in line with the reported high activation energy for grain growth in this system and further emphasizes the superiority of ZrO2 over Y2O3 as a dispersion strengthener.