Advance View

Electron Holography Study of Secondary Electron Distribution around ChargedEpoxy Resin

Takafumi Sato1, Naoya Tsukida1,+1, Mitsuaki Higo1,+2, Hideyuki Magara1, Zentaro Akase1,Daisuke Shindo1,2,+3 and Nobuhiko Ohno3,4

1Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan2RIKEN Center for Emergent Matter Science (CEMS), Wako 351-0198, Japan3Jichi Medical University, Shimotsuke 329-0498, Japan4National Institute for Physiological Sciences, Okazaki 444-8787, Japan

The accumulation and distribution of electron-induced secondary electrons around epoxy resin are studied using electron holography. Thedistribution of secondary electrons is determined to be sensitive to the surface conditions of the epoxy resin, particularly to the presence ofconductive materials on the surface that are introduced during specimen preparation processes. These results provide a deeper understanding ofthe charging and discharging mechanisms for epoxy resin, and the behavior of secondary electrons around these materials. This study alsoprovides a new perspective for the visualization of various forms of electron behavior around insulating materials such as epoxy resin throughthe control of their surface characteristics. [doi:10.2320/matertrans.MI201904]

(Received April 23, 2019; Accepted July 23, 2019; Published September 6, 2019)

Keywords: electron holography, electric field, secondary electron, epoxy resin, surface condition, EPMA, FIB

1. Introduction

Transmission electron microscopy (TEM) techniques havebeen widely utilized to study the microstructures of variousmaterials.1­4) Among these techniques, electron holographyhas the unique feature of enabling the visualization ofelectromagnetic fields at the nanoscale.5­8) As such, bothelectric9­13) and magnetic14­23) fields have been extensivelystudied for a variety of materials. However, when insulatingmaterials are examined using TEM, care should be takenwith respect to the charging effect because the specimensbecome positively charged as a result of secondary electronemission.24­26) The additional electric field due to thecharging effect tends to modify the inherent electromagneticfield of the specimen. Severe charging effects also causespecimen drift and anomalous image contrast. Various fixingand coating techniques with conductive materials have beenwidely utilized to suppress specimen charging and realizeimproved observation.27­29) However, the detailed mecha-nisms of the charging and discharging effect, and thebehavior of secondary electrons around the charged speci-mens have yet to be clarified.

We have recently conducted a systematic investigation onthe charging phenomenon for insulating materials usingelectron holography. With biological specimens, e.g., micro-fibrils in sciatic nerve tissues, in addition to the electric field,the distribution of secondary electrons around the chargedspecimens can be observed by the detection of electric fieldvariations due to the motion of secondary electrons. Forexample, the orbits of secondary electrons around micro-fibrils were clarified using reconstructed phase30) andamplitude31) images obtained with electron holography. The

accumulation of secondary electrons in the branched regionsof microfibrils was also studied.32,33) These studies observedthe motion and accumulation of secondary electrons in theupper-edge regions of microfibrils that were far from thesupporting film. The reason for this precaution is thatcharging of insulating materials is strongly affected by theirradiation of secondary electrons from the substrate and/orsupporting film, as reported elsewhere.34,35) In the presentstudy, particular attention was given to the effect ofconductive materials placed on the surfaces of epoxy resin.The electric field fluctuation due to the motion of secondaryelectrons was then studied in relation to the surfaceconditions of charged epoxy resin.

2. Experimental

Epoxy resin (EPO­TEKμ 375, Epoxy Technology Co.,Ltd.) was used as a TEM specimen. A method for thepreparation of a shape-controlled epoxy resin specimen isdescribed in the following. Initially, the resin and the curingagent were mixed in a ratio of 10:1. This mixture was heatedat a temperature of around 80°C for 90min to accelerate thereaction and curing process to prepare a bulk epoxy resinspecimen. Sectioned flakes were obtained by slicing a bulkspecimen with an ultramicrotome (Leihurt Ultracut S, LeicaMicrosystems) and these were fixed on a thin Cu plateusing a conductive epoxy resin. Figure 1 shows an opticalmicroscopy image of the geometrical configuration of asectioned epoxy resin flake and the Cu plate. The shapeand size of the sectioned flake were adjusted using a focusedion beam (FIB) apparatus (JIB-4500, JEOL). Control of thespecimen shape and the thinning processes were performedvery carefully because the thin specimen rod of epoxy resinwas easily deformed and broken during the microfabricationprocesses.

The electron holography experiment was conducted usingan electron microscope (JEM-3000F, JEOL) equipped with a

+1Graduate Student, Tohoku University. Present address: NACHI-FUJIKOSHI CORP., Toyama 930-8511, Japan

+2Graduate Student, Tohoku University. Present address: Yamaha MotorCo. Ltd., Iwata 438-8501, Japan

+3Corresponding author, E-mail: daisuke.shindo.b1@tohoku.ac.jp

Materials TransactionsSpecial Issue on Development and Application of Advanced Electron Microscopy Techniques for Materials Science©2019 The Japan Institute of Metals and Materials

Advance View

thermal field emission gun having a biprism, and anothermicroscope (HF-3000X, Hitachi High-Technologies)equipped with a cold field emission gun having threebiprisms in the imaging system. The accelerating voltageused for both the microscopes was 300 kV. The exposure timewas set to 6 s and 3 s for observations with the JEM-3000FTEM and HF-3000X TEM, respectively. Holograms wereobtained for the reconstructed phase and amplitude imagesusing Fourier transform operation.

Surface contaminants on the TEM specimens wereanalyzed using an electron probe microanalysis (EPMA;JXA-8530F, JEOL) system with an acceleration voltage of15.0 kV. Elemental mapping images were obtained for C, Gaand Al elements using the EPMA system.

To simulate the electric potential distribution around theepoxy resin, ELFIN and ELF/MAGIC software (ELFCorporation) were used for the electric field analysis. Thesesimulations were performed on the basis of Maxwell’sequations.

Figure 2 shows a schematic illustration of the experimentalconditions to form a hologram using TEM with a biprism.The hologram is obtained by application of a voltage tointerfere an object wave with a reference wave. Phase andamplitude information can be extracted by processing theseholograms through the Fourier transform operation.31,33)

Reconstructed phase images clarify the electric potentialdistribution, while reconstructed amplitude images detect thefluctuation of electric field due to the motion of secondaryelectrons.

3. Results and Discussion

Figure 3 shows a low-magnification TEM image of anepoxy resin specimen shaped by FIB. The Y-shape epoxyresin imitates the shape of the biological specimen observedin the previous study.32,33) Figure 4(a) shows a hologram ofthe specimen, and Fig. 4(b) shows the reconstructed phaseimage, where the purple region corresponds to the epoxyresin. A simulation was conducted based on a simplifiedthree-dimensional model whose thickness is uniformly10 µm, and a shape of top view and a potential distributionon its top surface are shown in Fig. 4(c). The potentials onthe side and bottom surface are set at 0.0V. A three-dimensional electric potential distribution around the modelwas calculated. A region between «10 µm from center of thespecimen along a beam direction was considered in thesimulation. In this simulation, an electric potential caused bysecondary electrons distributed around the specimen was notconsidered. Figure 4(d) shows a phase image simulated withthe above model. A modulation of a phase shift of thereference wave was considered in Fig. 4(d). A width of aninterference fringes region was assumed to be 4.4 µm. Aresult of a calculated phase image without consideration ofthe modulation of the phase shift of the reference wave isshown in Fig. 4(e) for comparison. Contour lines around thespecimen in Fig. 4(e) correspond to electrical equipotentiallines projected along the incident electron beam direction.

Figure 5(a) shows a reconstructed amplitude image, wherethe dark regions correspond to fluctuations in electric fielddue to the motion of secondary electrons.31,33) Color imagesare presented in Fig. 5(b), where the color scale bar in thelower part indicates the relation between the visibility ofthe interference fringes and the contrast of the reconstructedamplitude images. Here, the incident electron intensity ishomogeneous in the observed area outside the chargedspecimen. The decrease of visibility of interference fringesresults from the electric field fluctuation, which wasdiscussed in previous paper.33) On the other hand, this low

Specimen(Epoxy resin)

Cu plate

Conductive epoxy

Fig. 1 Optical microscopy image showing the geometrical configuration ofsectioned epoxy resin flake and a Cu plate.

Object wave

Hologram

Specimen

Reference wave

Biprism

Fig. 2 Experimental conditions to form a hologram using TEM with abiprism.

Epoxy specimen

Fig. 3 TEM image of epoxy resin at low magnification.

T. Sato et al.2

Advance Viewvisibility of interference fringes results in lower amplitudein the reconstructed amplitude images. We measured thevisibility of interference fringes and the contrast of thereconstructed amplitude image in the same points of variousregions. The relationship was obtained in the color scale baras shown in Fig. 5(b). As indicated by the white arrows, faintred color regions are observed in the lower right and leftparts, which correspond to the region of relatively highelectric potential in Fig. 4(c). This feature is significantlydifferent from biological specimens, i.e., microfibrils ofsciatic nerve tissue, in which secondary electrons tend toaccumulate at the top region between two branches.Figures 6(a) and 6(b) show the reconstructed amplitudeimages around microfibrils of sciatic nerve tissue observedin the previous paper.32) The bright red color regions areconsidered to correspond to the part where the fluctuation ofelectric potential due to the motion of the secondary electronsis prominent. This point will be discussed later. It is also

noted that the red color regions shift gradually with thepassage of time along the inside surfaces of the two branchesof microfibrils.32)

To confirm the distribution of secondary electrons aroundthe branched region of the epoxy resin specimen, hologramswere observed under different experimental conditions, asshown in Fig. 7. In Fig. 7(a), interference fringes areobserved in one part of the two branches. Note that Fresnelfringe contrasts were significantly suppressed in the holo-gram because the hologram was acquired with the double-biprism interferometry.36) Figure 7(b) shows the recon-structed phase image of the region. Figure 7(g) is a resultof a simulation of a phase image using a three-dimensionalpotential model shown in Fig. 7(f ). The width of theinterference region was assumed to be 2.8 µm. Figure 7(g)basically represents the asymmetry of the phase distributionalong the branch of the epoxy resin in the experimental phase

1μm

Fig. 4 (a) Hologram of Y-shaped epoxy resin. (b) Reconstructed phase image of epoxy resin treated with FIB. (c) Estimated electricpotential distribution on the specimen surface. (d) Simulated phase image of epoxy resin for the image shown in (b). (e) Simulated phaseimage without consideration of the modulation of the phase shift of the reference wave.

1μmVisibility

Amplitude

5% 30%

Fig. 5 (a) Reconstructed amplitude image of epoxy resin of which theshape was controlled by the FIB method and (b) the corresponding colorimage.

500 nmVisibility

Amplitude

5% 30%

Fig. 6 (a), (b) Reconstructed amplitude images around microfibrils ofsciatic nerve tissue. The contrast change was caused by the passage oftime for observation.

Electron Holography Study of Secondary Electron Distribution around Charged Epoxy Resin 3

Advance View

image of Fig. 7(b). Note that the electric potentialdistribution is slightly different from that of Fig. 4(c) dueto the specimen conditions, such as illumination intensityand area of the incident electron beam. Figure 7(c) shows thereconstructed amplitude image, and the corresponding colorimage is presented in Fig. 7(d). It is found that the lower partof the branch indicated by the arrow “L” (8.0V) shows a faintred color region being different from the upper part indicatedby the arrow “U” (4.0V). Figure 7(e) shows the averagedintensity distribution of the reconstructed amplitude imagealong the direction of the black arrow in the rectangularregions indicated in Fig. 7(c). The amplitude of the lowerpart is lower than that of the upper part of the branch. Thesefeatures are consistent with the results in Fig. 5(b).

The difference between the present results and thoseobtained in previous observations with Y-shaped biologicalspecimens is considered to result from the presence ofmetallic elements on the surface of the epoxy resin due tothe FIB thinning process. To confirm this, EPMA elementalmapping analysis was performed. Figure 8(a) shows ascanning electron microscopy (SEM) image, and Figs. 8(b)to 8(d) show elemental mapping images that for C, Ga andAl. The epoxy resin contains carbon; therefore, the red color,which indicates carbon-rich areas, is reasonable. On the otherhand, Figs. 8(c) and 8(d) show that Ga and Al are alsopresent. The Ga is caused by the Ga ion beam used in the FIBthinning process. The presence of Al is considered to be dueto redeposition from the Al specimen supporting plate of theFIB system. The presence of these metallic elements on thesurface of the Y-shaped epoxy resin is attributed to the smallaccumulation of secondary electrons at the top edge region.

Note that in the elemental mapping analysis by EPMAsystems, the signals obtained strongly depend on the shape ofspecimens and also geometrical configuration of the speci-men and the detector. Thus, from these analyses it is difficultto quantitatively evaluate the distribution of metallicelements. Nevertheless, it is considered that the final thinningprocess was mainly done at the top edge region between thebranches and relatively high density of metallic elements is

VVisibility

Amplitude

5% 30%

Distance / μm

Am

plitu

de in

tens

ity /

arb.

uni

t

1 μm

Fig. 7 (a) Interference fringes observed in one part of the two branches. An enlarged hologram of the red square region is presented in theupper right. (b) Reconstructed phase image. (c) Reconstructed amplitude image and (d) corresponding color image. (e) Intensity ofreconstructed amplitude image along the black arrow for the rectangular regions indicated in (c). (f ) Estimated electric potentialdistribution of the specimen surface. (g) Simulated phase image of epoxy resin for the image shown in (b).

1 μm

Fig. 8 (a) SEM image, and (b)­(d) elemental mapping images correspond-ing to C, Ga and Al.

T. Sato et al.4

Advance View

expected to form at the top edge region resulting in lowelectric potential of the surface and thus the low density ofsecondary electrons around the surface.

To examine the effect of these metallic elements on thesurface, the secondary electron distributions were comparedfor different surface conditions of the epoxy resin specimens,i.e., one specimen prepared by ultramicrotomy and the otherby the FIB system. Unlike when the FIB system was used,a clear specimen surface without metallic elements wasobtained by ultramicrotomy. Figure 9(a) shows a recon-structed phase image of a thin film of epoxy resin (dark brownregion) prepared by ultramicrotomy. In order to estimate theelectric potential of a top surface of the specimen, wesimulated the phase images by the same method shown inFig. 4. The thickness of the model was 2.0 µm. The electricpotentials on side and bottom surfaces of the model were set at0.0V. The width of interference fringes region was assumed tobe 2.8 µm. Compared with the simulation results shown inFig. 9(b), the electric potential of the epoxy resin is estimatedto be 1.2V. This electric potential is relatively low comparingwith the side surface of the Y-shaped epoxy specimen(Fig. 4(c)). Although there are no metallic elements on thesurface of the specimen prepared by ultramicrotomy, thecharging effect is suppressed due to the irradiation ofsecondary electrons from the specimen support plate nearthe observed area. The reconstructed amplitude image inFig. 9(c) shows red color regions around the surface of theepoxy resin, which are considered to correspond to a highdensity of secondary electrons strongly interacting with thesurface of the positively charged specimen. Particularly in theconcave region, as indicated by the arrow in Fig. 9(c), a brightred color region is evident, which corresponds to a large

fluctuation of the electric field due to the interaction ofaccumulated secondary electrons with the surface. Afterobserving the hologram of the epoxy resin, both sides ofthe thin specimen were irradiated with a weak Ga-ion beam.The beam intensity was 0.85 © 10 3mC·m¹2, which is 200times smaller than that typically used for polishing specimens.The reconstructed phase image in Fig. 9(d) shows that theelectric potential of the specimen becomes 1.0V by computersimulation (Fig. 9(e)). The electric potential of the specimenbefore and after was not significantly different due to the weakGa-ion beam. It should also be noted that the shape of thespecimen in Fig. 9(d) is almost the same as in Fig. 9(a), whichindicates that the irradiation damage by Ga ions is small. Inthe reconstructed amplitude image in Fig. 9(f ), the coloredregions observed in Fig. 9(c) are no longer visible. Thus, thedistribution of secondary electrons that interact strongly withthe surface of the positively charged epoxy resin is sensitiveto the presence of metallic elements on the surface.

Finally, we discuss the fluctuation of electric potentialdistribution around the charged specimen observed in thereconstructed amplitude images taking account of thesystematic studies on the charging effect of insulatingmaterials performed so far. First we note that there existpossible contributions to the cause of the fluctuations ofelectric potential around the charged specimens, i.e.,secondary electron motions and electric potential change ofthe specimens. The charging effect of insulating materialsresults from the unbalance of secondary electron emissionand supply of electrons, such as irradiation of secondaryelectrons from the substrate. When the experimentalconditions such as irradiation intensity and illumination areaof the incident electron beam are constant, the charging state

VVisibility

Amplitude

5% 30%

Epoxy

Epoxy

1μm

Epoxy

1μm

Epoxy

Fig. 9 (a) Reconstructed phase image of ultramicrotomed epoxy resin. (b) Simulated phase image with surface electric potential of 1.2V.(c) Reconstructed amplitude image of ultramicrotomed epoxy resin. (d) Reconstructed phase image of epoxy resin treated with FIB.(e) Simulated phase image with electric potential of 1.0V. (f ) Reconstructed amplitude image of epoxy resin treated with FIB.

Electron Holography Study of Secondary Electron Distribution around Charged Epoxy Resin 5

Advance View

of the insulating specimen becomes stationary. In that case, itis considered that drastic changes of the electric potential ofthe charged insulating specimens do not occur.35) In the caseof microfibrils of sciatic nerve tissue as shown in thereconstructed amplitude images presented in Fig. 6, if thereexist considerable fluctuations of electric potential of thespecimen, the red color regions are expected to be observedaround the specimen monotonously as discussed in theprevious paper.33) In the observed images, however, the redcolor regions are observed mainly around the inside surfacesbetween the two branches of microfibrils. Furthermore, thecolor regions are localized and shifted gradually along theinside surfaces of the two branches. These features are wellexplained through the interaction between the secondaryelectrons and the surfaces of charged specimens as discussedin the previous papers.32,33) In the epoxy specimen shown inFig. 9(c), red color regions are observed around the chargedspecimen. As noted above, the bright red colored region isprominent in the concave region indicated by an arrow inFig. 9(c). This tendency is also explained by the interactionbetween the secondary electrons and the surface of thecharged epoxy specimen. Thus, the red color regions aroundthe charged insulating materials observed in the reconstructedamplitude images can be interpreted as the fluctuations ofelectric potential mainly caused by the motions of secondaryelectrons around the charged specimen. Accordingly, thisinterpretation may be applicable to the contrast of amplitudeimages of Figs. 5(b) and 7(d). On the other hand, as clarifiedin this study, the existence of metallic elements on the surfaceof charged specimens drastically reduces the fluctuation ofelectric potential due to the secondary electron motions. Inaddition, we should note that the reconstructed amplitudeimages do not visualize the whole information about thedistribution of secondary electrons whose density is relativelylow at the region far from the charged specimen and evennear the charged specimen with the existence of metallicelements on the surface.

4. Conclusions

In summary, the electric potential of charged epoxy resinwas quantitatively evaluated by simulation. The distributionof secondary electrons around the charged epoxy resin wasvisualized by the reconstructed amplitude imaging process.The distribution of secondary electrons was determined to besensitive to the presence of metallic elements on the surface.Therefore, the distribution of secondary electrons can becontrolled by adjustment of the surface conditions of theepoxy resin specimen. This study is expected to open a newperspective for electron holographic visualization of variousforms of electron behavior around insulating materials bycontrol of the surface conditions.

Acknowledgments

The authors wish to thank Dr. K. Niitsu (Kyoto University)for the collaboration and useful discussions.

REFERENCES

1) K.W. Urban: Science 321 (2008) 506­510.2) P.A. Midgley and R.E. Dunin-Borkowski: Nat. Mater. 8 (2009) 271­

280.3) C. Xiao, N. Fujita, K. Miyasaka, Y. Sakamoto and O. Terasaki: Nature

487 (2012) 349­353.4) D.D. Bock, W.-C.A. Lee, A.M. Kerlin, M.L. Andermann, G. Hood,

A.W. Wetzel, S. Yurgenson, E.R. Soucy, H.S. Kim and C. Reid: Nature471 (2011) 177­182.

5) A. Tonomura: Electron Holography, 9 chapters, (Springer, Berlin,1999).

6) H. Lichte and M. Lehmann: Rep. Prog. Phys. 71 (2008) 016102.7) E. Völkl, L.F. Allard and D.C. Joy: Introduction to Electron

Holography, (Plenum Press, New York, 1999).8) M. McCartney and D.J. Smith: Annu. Rev. Mater. Res. 37 (2007) 729­

767.9) W.D. Rau, P. Schwander, F.H. Baumann, W. Höppner and A. Ourmazd:

Phys. Rev. Lett. 82 (1999) 2614­2617.10) J. Cumings, A. Zettl, M.R. Cartney and J.C. Spence: Phys. Rev. Lett.

88 (2002) 056804.11) J.J. Kim, D. Shindo, Y. Murakami, W. Xia, J.L. Chou and Y.L. Chueh:

Nano Lett. 7 (2007) 2243­2247.12) M. Nakamura, F. Kagawa, T. Tanigaki, H.S. Park, T. Matsuda, D.

Shindo, Y. Tokura and M. Kawasaki: Phys. Rev. Lett. 116 (2016)156801.

13) R.E. Dunin-Borkowski, M.R. McCartney, R.B. Frankel, D.A.Bazylinski, M. Po’sfai and P.R. Buseck: Science 282 (1998) 1868­1870.

14) K.H. Kim, J.J. Kim, X. Weixing and D. Shindo: Mater. Trans. 48(2007) 2616­2620.

15) M.R. McCartney and Y. Zhu: Appl. Phys. Lett. 72 (1998) 1380­1382.16) Y. Murakami, K. Niitsu, T. Tanigaki, R. Kainuma, H.S. Park and D.

Shindo: Nat. Commun. 5 (2014) 4133­4140.17) H.S. Park, X. Yu, S. Aizawa, T. Tanigaki, T. Akashi, Y. Takahashi, T.

Matsuda, N. Kanazawa, Y. Onose, D. Shindo, A. Tonomura and Y.Tokura: Nat. Nanotechnol. 9 (2014) 337­342.

18) D. Shindo: Mater. Trans. 44 (2003) 2025­2034.19) Z. Akase, P. Young-Gil, D. Shindo, T. Tomida, H. Yashiki and S.

Hinotani: Mater. Trans. 46 (2005) 974­977.20) T. Aiso, D. Shindo and T. Sato: Mater. Trans. 48 (2007) 2621­2625.21) K. Sato, Y. Murakami, D. Shindo, S. Hirosawa and A. Yasuhara: Mater.

Trans. 51 (2010) 333­340.22) J.J. Kim, H.S. Park, D. Shindo, T. Iseki, N. Oshimura, T. Ishikawa and

K. Ohmori: Mater. Trans. 48 (2007) 2606­2611.23) Z. Akase, D. Shindo, M. Inoue and A. Taniyama: Mater. Trans. 48

(2007) 2626­2630.24) K.H. Downing, M.R. McCartney and R.M. Glaeser: Microsc.

Microanal. 10 (2004) 783­789.25) R.M. Glaeser and K.H. Dowing: Microsc. Microanal. 10 (2004) 790­

796.26) M.R. McCartney: J. Electron Microsc. 54 (2005) 239­242.27) P. Echlin: Scan. Electron Microsc. I (1975) 217­224.28) L.B. Munger: Scan. Electron Microsc. I (1977) 481­490.29) S. Ohno, K. Hora, T. Fukuhara and H. Oguchi: Virchows Arch. B 61

(1992) 351­358.30) D. Shindo, J.J. Kim, W.X. Xia, K.H. Kim, N. Ohno, Y. Fujii, N. Terada

and S. Ohno: J. Electron Microsc. 56 (2007) 1­5.31) D. Shindo, J.J. Kim, H.K. Kim, W.X. Xia, N. Ohno, Y. Fujii, N. Terada

and S. Ohno: J. Phys. Soc. Jpn. 78 (2009) 104802.32) D. Shindo, S. Aizawa, Z. Akase, T. Tanigaki, Y. Murakami and H.S.

Park: Microsc. Microanal. 20 (2014) 1015­1021.33) D. Shindo, T. Tanigaki and H.S. Park: Adv. Mater. 29 (2017) 1602216.34) B.G. Frost: Ultramicroscopy 75 (1998) 105­113.35) H. Suzuki, Z. Akase, K. Niitsu, T. Tanigaki and D. Shindo: Microscopy

66 (2017) 167­171.36) K. Harada and A. Tonomura: Appl. Phys. Lett. 84 (2004) 3229­3231.

T. Sato et al.6