Skip to main content Accessibility help
×
Home

Information:

  • Access

Figures:

Actions:

      • Send article to Kindle

        To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Novel Method for Preparing Transmission Electron Microscopy Samples of Micrometer-Sized Powder Particles by Using Focused Ion Beam
        Available formats
        ×

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Novel Method for Preparing Transmission Electron Microscopy Samples of Micrometer-Sized Powder Particles by Using Focused Ion Beam
        Available formats
        ×

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Novel Method for Preparing Transmission Electron Microscopy Samples of Micrometer-Sized Powder Particles by Using Focused Ion Beam
        Available formats
        ×
Export citation

Abstract

The preparation of transmission electron microscopy (TEM) samples from powders is quite difficult and challenging. For powders with particles in the 1–5 μm size range, it is especially difficult to select an adequate sample preparation technique. Epoxy is commonly used to bind powder, but drawbacks, such as differential milling originating from unequal milling rates between the epoxy and powder, remain. We propose a new, simple method for preparing TEM samples. This method is especially useful for powders with particles in the 1–5 μm size range that are vulnerable to oxidation. The method uses solder as an embedding agent together with focused ion beam (FIB) milling. The powder was embedded in low-temperature solder using a conventional hot-mounting instrument. Subsequently, FIB was used to fabricate thin TEM samples via the lift-out technique. The solder proved to be more effective than epoxy in producing thin TEM samples with large areas. The problem of differential milling was mitigated, and the solder binder was more stable than epoxy under an electron beam. This methodology can be applied for preparing TEM samples from various powders that are either vulnerable to oxidation or composed of high atomic number elements.

Introduction

Transmission electron microscopy (TEM) is used in various fields for the analysis of the microstructure, crystal structure, and chemical composition of materials. A TEM specimen is typically thinner than 100 nm because TEM uses the electrons transmitted through a specimen (Williams & Carter, 2009). Specimen preparation is therefore the most crucial step as it determines the researcher’s ability to obtain accurate information from the analyzed materials. A number of studies on TEM specimen preparation methodology have been published, and several ways to prepare TEM specimens, such as the dispersion method, ultramicrotomy, mechanical polishing, ion milling, and focused ion beam (FIB) milling have been recognized. However, despite the many methods and techniques available, the preparation of TEM samples from powders with particles a few micrometers in size remains a challenge due to the difficulty in selecting an adequate sample preparation method. When the size of the powder is smaller than 100 nm, it can be observed directly because the electrons are able to penetrate through the powder. A TEM sample can be prepared simply by dispersing a drop of the ultrasonicated powder suspension onto a copper grid coated with an amorphous carbon film (Ayache et al., 2010). If the particle size of the powder is larger than 100 nm, it cannot be observed directly because electrons cannot be transmitted through the particles (Wen et al., 2007; Danaie & Mitlin, 2009; Benslim et al., 2010; Hashemi-Sadraei et al., 2011). Thus, appropriate methods are needed to prepare TEM samples from powders with a variety of sizes.

A simple way to prepare a TEM sample from powders is to slice a thin specimen directly using ultramicrotomy. Ultramicrotomy is mainly used on biological samples, polymers, and powders. In the case of a powder, the sample is embedded in an epoxy then sliced by a glass or diamond knife. The sliced flakes are floated on water and collected onto a grid supported by a film grid (Wei & Li, 1997; Williams & Carter, 2009; Litynska-Dobrzynska et al., 2010; Hoffmann et al., 2012). Unfortunately, ultramicrotomy has several disadvantages and limitations. It can cause artificial defects in hard materials due to the physical stress induced when slicing using a diamond knife. The diamond knife can also fracture the sample, and many defects can be introduced during cutting, which changes the microstructure. In addition, in the case of samples vulnerable to oxidation, floating the slices on a suitable inert medium other than water is required, and it is difficult to find liquids with the appropriate surface tension. Thus, it is difficult to use ultramicrotomy to prepare TEM samples from metal powders that have high hardness.

In order to apply the mechanical polishing method, which is the most basic method for preparing TEM specimens from metals and ceramics, powders should be formed into a bulk sample by embedding the powders in epoxy (Zhou et al., 2003 a ; Wen et al., 2015). After making the bulk material composed of the powder and binder, mechanical polishing and ion milling are used to make the final samples (Zhou et al., 2001, 2003 b ; Chung et al., 2002; He et al., 2003; Thornton et al., 2007). Although many studies have sought to optimize this approach, several issues have yet to be resolved. For example, differential milling between the epoxy and the material of interest cannot be solved completely without changing the embedding agent. Epoxy resin is based on carbon, which is a low atomic number element, whereas metal powders are often composed of relatively high atomic number elements. Thus, the ratio between the epoxy and powder during fabrication of the epoxy–powder mixture is an important factor that must be decided based on the size of the powder particles and the constituent elements. Furthermore, even when the powders and epoxy are mixed well, it is difficult to achieve uniform distribution of the powders in epoxy. Areas that are not well mixed and have a lower fraction of powder are weak, and thus they are milled faster than the areas rich in powder during ion milling. As a result, the formation of large, uniformly thin areas is hindered. In addition, the epoxy has to permeate into all the spaces between the particles using the minimum possible quantity of epoxy. If this process is not well controlled, particles may not be bound to the epoxy and can drop out. In the case of materials that are easily oxidized, the mixing of the powder and epoxy and the subsequent curing of the mixture must be conducted under an inert atmosphere with exacting requirements. Finally, the mechanical polishing that thins the sample, as well as ion milling of the mixtures, are difficult and time-consuming processes, and success is highly dependent on the skill of the researcher.

FIB milling is a method for directly producing TEM specimens by milling them using accelerated Ga ions. The lift-out technique of the FIB method is widely used (Kitano et al., 1995; Prenitzer et al., 1998; Rea et al., 2005). As in ultramicrotomy and mechanical polishing, a binder is needed in the case of 1–5 µm sized powders because they cannot be otherwise secured during ion milling. Although epoxy embedding raises the same difficulties in differential milling that it did in mechanical polishing, we can monitor the milling procedure using electron imaging. As such, this method is more controllable than that using mechanical polishing followed by ion milling.

It is evident that the differential milling of the binder and powders, arising due to differences in milling rates, always occurs during ion milling, and its elimination requires a fundamental solution. Therefore, we used solder as an embedding agent for 1–5 µm powders, which is the size range for which a definite sample preparation technique does not exist. After embedding the powder in the solder, the FIB lift-out technique was used to prepare TEM specimens, which were then compared with specimens made via the conventional epoxy mounting method. Our newly proposed method was effective in preparing thin TEM samples with large areas, and the problem of differential milling was mitigated. Furthermore, the oxidation of the powder was easily prevented, and the solder binder was more stable under an electron beam than epoxy.

Materials and Methods

A Nd–Fe–B powder, prepared by using both spray drying and reduction/diffusion processes, was used to compare the sample preparation methods (Lin et al., 1997). The particle size of the Nd2Fe14B powder varied from 1 to 5 μm. This powder was appropriate for investigating the different sample preparation methods because the size makes it difficult to unambiguously select a proper method. In addition, Nd2Fe14B compounds are very vulnerable to oxidation. Gatan G-1 epoxy (Gatan Inc., Pleasanton, CA, USA) and conventional low-temperature Sn-based solder (245 Flux-Cored Wire; Kester, Itasca, IL, USA) were used as embedding agents.

Figure 1 shows the sample-embedding procedure using the G-1 epoxy. The powder was mixed with the G-1 epoxy and poured into a 3-mm-diameter Cu tube (Figs. 1a–1c). Next, the mixture was densified by inserting a rod to apply pressure such that the volume fraction of the epoxy was minimized (Fig. 1d). The mixture was cured under pressure and the epoxy-embedded sample was prepared by cutting after curing (Figs. 1e, 1f). Figure 2 shows a schematic and photographs of the proposed sample-embedding procedure using solder and hot mount equipment (Labopress 3; Struers, Cleveland, OH, USA). The powder was placed on the movable lower ram and then a plate-shaped solder piece prepared beforehand was placed above the powder. Next, an electrically conductive polymer (black glass-filled epoxy powder; Allied High Tech Products, Inc., Rancho Dominguez, CA, USA) was poured to encapsulate the solder and powder samples. After applying heat (180°C) and pressure (30 kN) for 6 min, the resultant solder-embedded specimen was used to prepare a TEM sample. After completing the mounting process, both samples were mechanically polished with SiC paper, diamond suspension (3 and 1 μm), and colloidal silica (0.05 μm). Polishing was performed under water-free conditions in order to avoid oxidation of the powder. In addition, the load was reduced to minimize deformation during mechanical polishing. The polished samples were observed using scanning electron microscopy (SEM, JSM-7000F; JEOL Ltd., Akishima, Tokyo, Japan). The TEM specimens were produced by FIB (JIB-4601F; JEOL Ltd., Akishima, Tokyo, Japan) using the lift-out technique. After lift-out, the specimens were subjected to rough thinning using 30 kV and 40 pA conditions down to a 250-nm thickness. After that, conditions were tuned to 10 kV and 10 pA to obtain a thickness below 100 nm. To reduce surface damage sustained during Ga ion milling, the final thinning step was conducted using 3 kV and 30 pA for 5 min. The specimens were then analyzed using TEM (JEM-ARM200F; JEOL Ltd., Akishima, Tokyo, Japan).

Figure 1 Schematic of the sample-embedding procedure using the G-1 epoxy. a–c: The powders were mixed with the G-1 epoxy and poured into a 3-mm diameter Cu tube. d: The powder–epoxy mixture was densified by inserting a rod and applying pressure to minimize the fraction of epoxy. e,f: The mixture was cured under pressure, and the epoxy-embedded sample was prepared by cutting after curing.

Figure 2 Schematic and photographs of the sample-embedding procedure using conventional solder. The powder was put on the movable lower ram and then a plate-shaped solder prepared beforehand was put above the powder. a: Next, electrically conductive polymer was poured in so as to encapsulate the solder and powder sample. b: After applying heat (180°C) and pressure (30 kN) for 6 min, the solder-embedded sample was complete.

Results and Discussion

Figure 3a shows an SEM image around the FIB lift-out region of the G-1 epoxy-embedded sample. Figure 3b shows an SEM image of the TEM specimen obtained by lift-out from the trench after the final thinning. Yellow arrows (medium dark areas) and red arrows (dark areas) indicate the Nd2Fe14B powder and holes, respectively. The residual bright areas are the G-1 epoxy. As shown in Figures 3a and 3b, the fraction of powder to epoxy is not high. Although different fractions of epoxy and powder were investigated in the mixing process, it was difficult to enhance the fraction of powder further to densely fill the Cu tube. It is assumed that there is a limit on how densely the powder can be filled because a certain minimum amount of epoxy is needed to bind the powder well. In addition, the Cu tube was sealed during the pressing. Thus, it was difficult to pack the powder much more densely than that shown in Figure 3b. The G-1 epoxy was more vulnerable to the Ga ion beam than the Nd2Fe14B powder and, as a result, it was etched far faster. Therefore, the more the TEM specimen was thinned, the more easily it was bent and perforated. Figure 3c is a tilted view of Figure 3b and it shows bending and large holes. The red arrows in Figure 3c indicate the same positions as in Figure 3b. The bending and large holes disrupted the fabrication of a uniformly thin TEM specimen. As such, in the case of the G-1 epoxy-embedded sample, bending and perforation of the thin specimen occurred due to the difference in the milling rates between epoxy and powder, and consequently, it was difficult to make the areas of interest thin enough for TEM analysis. Rugged steps in the G-1 epoxy-embedded TEM specimen are shown in Figure 3c. Furthermore, Figure 3d shows a TEM image of the fabricated TEM specimen. The Nd2Fe14B powder was considerably oxidized and covered with an ~100-nm-thick oxide layer. It seems that the particles were oxidized when mixing them with epoxy. No matter how hard we tried, a certain amount of time was required for mixing, pouring the mixture into the Cu tube, and curing. This means that the sample was exposed to air for some time. In conclusion, the conventional embedding method using G-1 epoxy is not appropriate for preparing TEM specimens from powders such as Nd2Fe14B, which have a large difference in atomic number with the carbon-based epoxy. Furthermore, the method is not suitable for easily oxidized samples.

Figure 3 The results of the focused ion beam (FIB) milling of the sample prepared using the G-1 epoxy embedding. a: Scanning electron microscopy (SEM) image around the FIB lift-out region of the G-1 epoxy-embedded sample. b: SEM image of the thin-foil transmission electron microscopy (TEM) sample prepared by FIB lift-out after the final thinning. Yellow arrows (medium dark areas) and red arrows (dark areas) indicate the Nd2Fe14B particles and perforations, respectively. The bright areas are the G-1 epoxy. c: Tilted view of (b) that shows bending and large holes. The red arrows indicate the same positions as in (b). d: TEM image of the fabricated sample. The Nd2Fe14B powder was considerably oxidized and covered with an ~100-nm-thick oxide layer.

Figure 4a shows an SEM image of the TEM specimen fabricated using FIB from the powder sample embedded in solder. The Nd2Fe14B powder is much more densely packed here than it was in the TEM specimen from the epoxy-embedded case (Fig. 3b). Yellow arrows and blue arrows indicate the Nd2Fe14B particles and Sn–Pb solder, respectively. The Sn–Pb solder, which is a low-temperature solder, melted slightly during hot mounting due to the heat applied, and the solder therefore penetrated easily into the gaps in the powder. The method proposed here, which uses solder as an embedding agent, has definite advantages in controlling the density of the powder and wetting between the embedding agent and powder in a very short time, and it can be performed with a reproducibility that is user-independent. Furthermore, the differences in the atomic numbers between the Sn–Pb solder and the Nd–Fe–B powder are smaller than in the case of the carbon-based epoxy embedding agent. As a result, bending and perforation problems decreased, and consequently, the TEM specimen of the Nd2Fe14B powder had a relatively uniform thickness without large holes or steps. In fact, it was difficult to completely avoid the curtaining artifact and the development of minor perforations when thinning the sample down to a few tens of nanometers, as shown in Figure 4a, because the sample was a multi-phase composite. In addition, this method has an advantage with respect to avoiding oxidation. It was difficult to control and optimize the fraction of epoxy and powder when the embedded mould was prepared using the G-1 epoxy. The epoxy also made the powder vulnerable to oxidation during the preparation step. On the other hand, when preparing the solder-embedded mould, oxidation could be easily controlled because hot mounting was conducted immediately after covering the solder and the powder with conductive polymer. It was not necessary to make a compact of the powder. Merely pouring the powder onto the lower ramp of the hot mount equipment was sufficient. Figures 4b and 4c show TEM and scanning transmission electron microscopy (STEM) images from the TEM specimen (Fig. 4a) fabricated using solder. This method provided a TEM specimen that was thin enough to easily obtain high-resolution images and STEM high-angle annular dark-field images. The oxidized layer on the surface of the powder was ~20 nm thick, which is much thinner than in the epoxy-embedded case, where the thickness was ~100 nm. Despite the above advantages of the new method, it is necessary to be aware of several factors. As shown in Figure 4a, the analyst should be careful in interpreting images from the TEM specimen fabricated from the new method because it could be difficult to distinguish between the powders (yellow arrows) and the solder agent (blue arrows). Furthermore, solder cannot be used as an embedding agent for certain materials that react with the Sn–Pb alloy. Although precautions must be taken to avoid such issues, the proposed method for TEM sample preparation using FIB and solder as an embedding agent can be effective in preparing thin samples with large areas and uniform thicknesses of powders with 1–5 μm particle sizes, especially those vulnerable to oxidation.

Figure 4 The results of focused ion beam (FIB) milling of the sample prepared using solder embedding. a: Scanning electron microscopy image of the transmission electron microscopy (TEM) specimen fabricated by FIB. Yellow and blue arrows indicate Nd2Fe14B particles and Sn–Pb solder, respectively. Red arrows indicate perforations. b: TEM and (c) scanning TEM (STEM) images demonstrate that the proposed method could provide TEM specimens that are thin enough to easily obtain high-resolution TEM images and STEM high-angle annular dark-field images.

The proposed method has another advantage in terms of SEM observations and high-current SEM imaging. When compared with the conventional hot-mounting method for powder samples using only a carbon-based conductive polymer (Fig. 5a), solder can easily penetrate into the gaps of the powder and bind the particles more tightly. Therefore, our method has advantages in terms of avoiding the separation of the powder from the binder and edge retention during mechanical polishing (Figs. 5a, 5b). Another significant advantage is that the solder will mitigate electron charging during high-current imaging in an electron microscope, for example, in high-current energy-dispersive spectroscopy and electron backscatter diffraction (EBSD). Although the conductive polymer has electrical conductivity, it is not as good a conductor as solder. Figures 5c and 5d compare SEM images of Nd–Fe–B powder samples prepared using only the conductive polymer with those prepared using our method observed under high current. Clearly, solder can solve the problem of electron charging, whereas the conductive polymer beside the powder shows electron charging. Our novel method was adopted in our previous research on EBSD analysis of Nd–Fe–B nano-grained powders and has been shown to produce good results (Kim et al., 2016).

Figure 5 Comparison between (a,c) conductive polymer embedding and (b,d) solder embedding using (a,b) conventional scanning electron microscopy (SEM) observation and (c,d) high-current SEM imaging.

Conclusion

Powders with particles in the 1–5 μm size range do not have an accepted TEM sample preparation method. Thus, obtaining uniformly thin TEM specimens from such materials is challenging. We propose a novel and simple method for the TEM sample preparation of powders by using solder as an embedding agent in combination with FIB. In this study, a Sn–Pb low-temperature solder was used, and powders were embedded in the solder using a conventional hot-mounting instrument. TEM specimens were fabricated using FIB by the lift-out technique. Using this method, the powders could be packed more densely, and the problem of differential milling was improved considerably. In addition, the proposed method was effective in preventing oxidation of the powder, and the solder binder was more stable under an electron beam than epoxy. Our methodology can be widely applied for preparing TEM samples from various powders that are especially vulnerable to oxidation or are composed of high atomic number elements.

Acknowledgments

This research was supported in part by the Global PhD Fellowship Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (2012H1A2A1009583), in part by the National Research Council of Science and Technology (NST) (CRC-15-06-KIGAM) and NRF grants (NRF-2011-0030058 and NRF-2015R1D1A1A01059653) funded by the Korean government (MSIP). The authors are also grateful for the technical support provided by the Cooperative Center for Research Facilities (CCRF) at Sungkyunkwan University.

References

Ayache, J., Beaunier, L., Boumendil, J., Ehret, G. & Laub, D. (2010). Sample Preparation Handbook for Transmission Electron Microscopy: Techniques. New York, NY: Springer.
Benslim, N., Mehdaoui, S., Aissaoui, O., Benabdeslem, M., Bouasla, A., Bechiri, L., Otmani, A. & Portier, X. (2010). XRD and TEM characterizations of the mechanically alloyed CuIn0.5Ga0.5Se2 powders. J Alloys Compd 489, 437440.
Chung, K.H., Lee, J., Rodriguez, R. & Lavernia, E.J. (2002). Grain growth behavior of cryomilled INCONEL 625 powder during isothermal heat treatment. Metall Mater Trans A 33, 125134.
Danaie, M. & Mitlin, D. (2009). TEM analysis and sorption properties of high-energy milled MgH2 powders. J Alloys Compd 476, 590598.
Hashemi-Sadraei, L., Mousavi, S.E., Vogt, R., Li, Y., Zhang, Z., Lavernia, E.J. & Schoenung, J.M. (2011). Influence of nitrogen content on thermal stability and grain growth kinetics of cryomilled Al nanocomposites. Metall Mater Trans A 43, 747756.
He, J.H., Chung, K.H., Liao, X.Z., Zhu, Y.T. & Laveria, E.J. (2003). Mechanical milling-induced deformation twinning in fcc materials with high stacking fault energy. Metall Mater Trans A 34, 707712.
Hoffmann, J., Klimenkov, M., Lindau, R. & Rieth, M. (2012). TEM study of mechanically alloyed ODS steel powder. J Nucl Mater 428, 165169.
Kim, T.H., Oh, J.S., Cha, H.R., Lee, J.G., Kwon, H.W. & Yang, C.W. (2016). Direct observation of texture memory in hydrogenation–disproportionation–desorption–recombination processed Nd-Fe-B magnets using electron backscatter diffraction. Scr Mater 115, 69.
Kitano, Y., Fujikawa, Y., Kamino, T., Yaguchi, T. & Saka, H. (1995). TEM observation of micrometer-sized Ni powder particles thinned by FIB cutting technique. J Electron Microsc 44, 410413.
Lin, J.H., Liu, S.F., Cheng, Q.M., Qian, X.L., Yang, L.Q. & Su, M.Z. (1997). Preparation of Nd-Fe-B based magnetic materials by soft chemistry and reduction-diffusion process. J Alloys Compd 249, 237241.
Litynska-Dobrzynska, L., Dutkiewicz, J., Maziarz, W. & Rogal, L. (2010). TEM and HRTEM studies of ball milled 6061 aluminium alloy powder with Zr addition. J Microsc 237, 506510.
Prenitzer, B.I., Giannuzzi, L.A., Newman, K., Brown, S.R., Irwin, R.B., Shofner, T.L. & Stevie, F.A. (1998). Transmission electron microscope specimen preparation of Zn powders using the focused ion beam lift-out technique. Metall Mater Trans A 29, 23992406.
Rea, K.E., Agarwal, A., Mckechnie, T. & Seal, S. (2005). FIB cross-sectioning of a single rapidly solidified hypereutectic Al-Si powder particle for HRTEM. Microsc Res Tech 66, 1016.
Thornton, J.J., Han, B.Q. & Lavernia, E.J. (2007). Grain growth in cryomilled Ni powder during degassing. Metall Mater Trans A 38, 13431350.
Wei, L.Y. & Li, T. (1997). Ultramicrotomy of powder material for TEM/STEM study. Microsc Res Tech 36, 380381.
Wen, H., Dong, S., He, P., Wang, Z., Zhou, H. & Zhang, X. (2007). Sol-gel synthesis and characterization of ytterbium silicate powders. J Am Ceram Soc 90, 40434046.
Wen, H., Lin, Y., Seidman, D.N., Schoenung, J.M., van Rooyen, I.J. & Lavernia, E.J. (2015). An efficient and cost-effective method for preparing transmission electron microscopy samples from powders. Microsc Microanal 21, 11841194.
Williams, D.B. & Carter, C.B. (2009). Transmission Electron Microscopy—A Textbook for Materials Science. New York, NY: Springer.
Zhou, F., Lee, J., Dallek, S. & Lavernia, E.J. (2001). High grain size stability of nanocrystalline Al prepared by mechanical attrition. J Mater Res 16, 34513458.
Zhou, F., Liao, X.Z., Zhu, Y.T., Dallek, S. & Lavernia, E.J. (2003 a). Microstructural evolution during recovery and recrystallization of a nanocrystalline Al-Mg alloy prepared by cryogenic ball milling. Acta Mater 51, 27772791.
Zhou, F., Nutt, S.R., Bampton, C.C. & Lavernia, E.J. (2003 b). Nanostructure in an Al-Mg-Sc alloy processed by low-energy ball milling at cryogenic temperature. Metall Mater Trans A 34, 19851992.