# Fileset

[nanomaterials-13-02170.pdf](https://mdr.nims.go.jp/filesets/356fc548-5064-4e57-b628-847d9e64c93b/download)

## Creator

[竹口　雅樹](https://orcid.org/0000-0002-0282-6020)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[The Atomic Observation of the Structural Change Process in Pt Networks in Air Using Environmental Cell Scanning Transmission Electron Microscopy](https://mdr.nims.go.jp/datasets/f95a0d72-8fd7-4b68-ae15-c768cb8013ed)

## Fulltext

The Atomic Observation of the Structural Change Process in Pt Networks in Air Using Environmental Cell Scanning Transmission Electron MicroscopyCitation: Takeguchi, M.; Takei, T.;Mitsuishi, K. The AtomicObservation of the Structural ChangeProcess in Pt Networks in Air UsingEnvironmental Cell ScanningTransmission Electron Microscopy.Nanomaterials 2023, 13, 2170.https://doi.org/10.3390/nano13152170Academic Editors: Elvio Carlinoand Antonietta TaurinoReceived: 3 July 2023Revised: 21 July 2023Accepted: 24 July 2023Published: 26 July 2023Copyright: © 2023 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributed under the terms andconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).nanomaterialsArticleThe Atomic Observation of the Structural Change Process in PtNetworks in Air Using Environmental Cell ScanningTransmission Electron MicroscopyMasaki Takeguchi * , Toshiaki Takei and Kazutaka MitsuishiNational Institute for Materials Science, 1-2-1 Sengen, Tsukuba 305-0047, Japan; takei.toshiaki@nims.go.jp (T.T.);mitsuishi.kazutaka@nims.go.jp (K.M.)* Correspondence: takeguchi.masaki@nims.go.jpAbstract: The structural change in Pt networks composed of multiple chain connections amonggrains was observed in air at 1 atm using atomic-resolution environmental cell scanning transmissionelectron microscopy. An aberration-corrected incident electron probe with a wide convergence anglemade it possible to increase the depth resolution that contributes to enhancing the signal-to-noiseratio of Pt network samples in air in an environmental cell, resulting in the achievement of atomic-resolution imaging. The exposure of the Pt networks to gas molecules under Brownian motion,stimulated by electron beams in the air, increases the collision probability between gas molecules andPt networks, and the Pt networks are more intensely stressed from all directions than in a situationwithout electron irradiation. By increasing the electron beam dose rate, the structural change of the Ptnetworks became significant. Dynamic observation on an atomic scale suggested that the structuralchange of the networks was not attributed to the surface atomic-diffusion-induced step motion butmainly caused by the movement and deformation of unstable grains and grain boundaries. Theoxidized surface layers may be one of the factors hindering the surface atomic step motion, mitigatingthe change in the size of the grains and grain boundaries.Keywords: scanning transmission electron microscopy; environmental cell; Pt network; atomicresolution; catalyst; grain; grain boundary; atomic diffusion1. IntroductionThe search for high-performance and long-lasting catalysts remains a critical issue fordeveloping electrochemical-based devices such as water electrolysis cells, fuel cells, andrechargeable batteries [1–3], which require improvements in power density and stability.The mass activity of catalysts can be enhanced by increasing the specific activity and activesurface area. Using precious metal nanoparticles supported by carbon-based materials isthe most common approach in developing the above catalysts. However, under reactiveenvironments, carbon supports are likely to be damaged, and nanoparticles often cannotmaintain their shape, size, and positions, resulting in the active surface area decreasing inthe short term [4,5]. Furthermore, high costs and supply limitations are associated withprecious metals, especially Pt-group metals. Therefore, precious metal alloys with moreearth-abundant materials, including metal-organic frameworks, core–shell structures, andhigh-entropy alloys, have been investigated to not only reduce the amount of preciousmetal but also to make them resistive to the reactive environment [6–9]. The improvementin stability has also become more crucial, and various strategies have been proposed.Bulk-like nanoarchitectures with high surface area, such as nanopores, nanosheets, andself-supporting networks, are promising in terms of both stability and activity, leading tolong-lasting high-performance catalysts [10–17]. However, catalytic reactions still causestructural changes in their nanostructures, followed by catalytic deactivation, limitingthe catalysis lifetime. For instance, in the case of nanoporous catalysts, coarsening of theNanomaterials 2023, 13, 2170. https://doi.org/10.3390/nano13152170 https://www.mdpi.com/journal/nanomaterialshttps://doi.org/10.3390/nano13152170https://doi.org/10.3390/nano13152170https://creativecommons.org/https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://www.mdpi.com/journal/nanomaterialshttps://www.mdpi.comhttps://orcid.org/0000-0002-0282-6020https://doi.org/10.3390/nano13152170https://www.mdpi.com/journal/nanomaterialshttps://www.mdpi.com/article/10.3390/nano13152170?type=check_update&version=1Nanomaterials 2023, 13, 2170 2 of 10nanoporous structure occurs during the catalytic reaction, and its performance graduallyworsens during the process [17]. The degradation mechanism of network-structure catalystsis considered to be similar.The coarsening process of nanopores in nanoporous gold (NPG) in an environmentof 1 vol% CO/air with a pressure range of 1–30 Pa at room temperature was previouslyinvestigated on an atomic scale using high-resolution environmental transmission elec-tron microscopy (TEM) [17]. The NPG with complex bicontinuous porous structures wasfabricated by dealloying an Ag65Au35 alloy leaf. The average diameter of both the goldligaments and nanopores was approximately 30 nm. It was revealed that the coarsening ofthe nanopores and the thinning of the ligaments were attributed to the rapid diffusion ofgold atoms at surface steps and the surface segregation of residual silver atoms. Becauseno significant structural change of the NPG sample was observed in heating experimentsup to 300 ◦C under a vacuum in TEM, it was concluded that the CO/air catalytic reactionmay induce gold atom diffusion and silver atom segregation. Moreover, high-resolution insitu tensile TEM observations of the deformation and fracture processes of NPG ligamentsshowed that the combination of dislocation plasticity and stress-driven surface atom diffu-sion can promote the plastic instability of ligaments, finally resulting in the brittle failure ofthe nanoporous structures [18]. Although the results of gas-environmental and tensile insitu TEMs might be locally limited phenomena, they help clarify what might happen innanoporous and network-structure catalyst architectures under practical conditions.In the catalytic nano-architectures, thermal stress causes local mechanical stress onall parts from all directions. NPG has intrinsically robust ligaments because of theirrelatively large diameter, and therefore the degradation mechanism is mainly attributed tonanopore coarsening and ligament thinning, promoted by atomic diffusions, as describedabove. In contrast, network-structure catalysts are predicted to be more fragile becausethey are composed of multiple chain-like connections of grains and grain boundaries (GB)containing many defects, with sizes smaller than those of NPG ligaments. Nevertheless,it has been reported that self-supporting network-structure catalysts can maintain theinitial activity longer than the nanoparticles on supports [9,13]. However, the atomicstructural change mechanism of the network catalysts under a reactive environment hasrarely been investigated.In the present work, the structural change dynamics of Pt networks in an oxidationenvironment of air at 1 atm and room temperature were observed using environmentalcell (EC) scanning TEM (STEM) with an atomic resolution. An aberration-corrected STEMprobe with a convergence semi-angle of 28 mrad provided a small depth of the field,thereby dramatically improving the signal-to-noise ratio (SNR) of the Pt networks in anEC, resulting in the amelioration of the STEM resolution. The Pt networks were composedof chain-connected grains and GBs and were deformed during STEM observation. Underelectron beam irradiation, primary electrons collide with gas molecules in air and transfermomentum to them, stimulating their Brownian motion. Thus, the collision probabilitybetween gas molecules and Pt networks increases, and the Pt networks are more intensivelystressed from all directions than the situation without electron irradiation. The presentSTEM observations showed no remarkable change in size in the grains and GBs, whileunstable grains moved and atoms fluctuated at the GBs and surfaces. Moreover, thesurfaces were covered with oxide layers. Based on these results, the factors that influencethe structural stability of the Pt network in the oxidation environment were considered.2. Materials and Methods2.1. Sample Preparations in Environmental CellsFigure 1a shows a cross-section of our custom EC, which was assembled by sandwich-ing a pair of 6 mm × 6 mm Si chips with an electron-transparent silicon nitride membranewindow of 30 µm × 300 µm in size [19–21]. A spacer of SiO2 with a height of 200 nm wasdeposited on the bottom chip. The thickness of the silicon nitride windows was measuredto be approximately 30 nm with a cross-sectional TEM of a specimen that was producedNanomaterials 2023, 13, 2170 3 of 10from another Si chip using the focused ion beam lift-out process. Amorphous carbon with4 nm thickness was sputter-coated on both sides of the Si chips using a precision etchingcoating system (Model 682, Gatan, Pleasanton, CA, USA). Next, Pt was deposited on theinner side of the top chip using an auto fine coater (JFC-1600, JEOL, Tokyo, Japan), as shownin Figure 1b. Through deposition, Pt nanoparticles formed and then coalesced with eachother, finally resulting in the formation of network structures composed of chain-connectedgrains, as shown in Figure 1d. The image is a high-angle annular dark-field (HAADF) STEMimage, and the image intensity of the Pt networks is almost the same throughout, except forin some regions. This is because the network structure was grown two-dimensionally onthe silicon nitride membrane surface. The brighter areas correspond to three-dimensionalshaped structures, but their numbers are small. After the Pt deposition, the top chip wasplaced on the bottom chip in air so that the rectangular silicon nitride windows of thetop and bottom chips could be orthogonally arranged to obtain a 30 µm × 30 µm squareviewing field. Finally, the EC was attached to a single-tilt TEM holder (EM-21010 (SCSH),JEOL, Japan) to complete an EC-TEM holder, as shown in Figure 1c.Nanomaterials 2023, 13, x FOR PEER REVIEW 3 of 10   membrane window of 30 µm × 300 µm in size [19–21]. A spacer of SiO2 with a height of 200 nm was deposited on the bottom chip. The thickness of the silicon nitride windows was measured to be approximately 30 nm with a cross-sectional TEM of a specimen that was produced from another Si chip using the focused ion beam lift-out process. Amor-phous carbon with 4 nm thickness was sputter-coated on both sides of the Si chips using a precision etching coating system (Model 682, Gatan, Pleasanton, CA, USA). Next, Pt was deposited on the inner side of the top chip using an auto fine coater (JFC-1600, JEOL, To-kyo, Japan), as shown in Figure 1b. Through deposition, Pt nanoparticles formed and then coalesced with each other, finally resulting in the formation of network structures com-posed of chain-connected grains, as shown in Figure 1d. The image is a high-angle annular dark-field (HAADF) STEM image, and the image intensity of the Pt networks is almost the same throughout, except for in some regions. This is because the network structure was grown two-dimensionally on the silicon nitride membrane surface. The brighter areas correspond to three-dimensional shaped structures, but their numbers are small. After the Pt deposition, the top chip was placed on the bottom chip in air so that the rectangular silicon nitride windows of the top and bottom chips could be orthogonally arranged to obtain a 30 µm × 30 µm square viewing field. Finally, the EC was attached to a single-tilt TEM holder (EM-21010 (SCSH), JEOL, Japan) to complete an EC-TEM holder, as shown in Figure 1c.  Figure 1. (a) Schematic illustrations showing the cross-section of the EC. (b) The geometry of Pt network samples, a silicon nitride membrane, and carbon-coated layers. (c) HAADF-STEM image of Pt networks in air. (d) Photo of the EC attached to a TEM holder. 2.2. Environmental Cell Scanning Transmission Electron Microscopy The EC-TEM holder was introduced into an aberration-corrected transmission elec-tron microscope (JEM-ARM200F, JEOL, Japan) equipped with a Schottky field emission gun. First, the XYZ position of the viewing field of a 30 µm × 30 µm window was memo-rized, and then the EC-TEM holder was removed. Next, another TEM holder in which a conventional carbon film-supported Cu microgrid was mounted was introduced into the microscope, and the electron-optics system and aberration correctors were finely tuned. Then, the EC-TEM holder was again introduced into the microscope, the viewing field position was recalled, and the STEM imaging was performed. An aberration-corrected beam probe with an energy of 200 keV and a current of 52.5 pA passing through a con-denser aperture (28 mrad convergence semi-angle) was used. The STEM resolution and dwell time were 512 × 512 pixels and 5.0 µs/pixel, respectively. STEM movies were rec-orded using a high-definition video function in the Gatan Microscopy Suite® (Gatan, USA) with a rate of 10 frames/s.   Figure 1. (a) Schematic illustrations showing the cross-section of the EC. (b) The geometry of Ptnetwork samples, a silicon nitride membrane, and carbon-coated layers. (c) HAADF-STEM image ofPt networks in air. (d) Photo of the EC attached to a TEM holder.2.2. Environmental Cell Scanning Transmission Electron MicroscopyThe EC-TEM holder was introduced into an aberration-corrected transmission electronmicroscope (JEM-ARM200F, JEOL, Japan) equipped with a Schottky field emission gun.First, the XYZ position of the viewing field of a 30 µm × 30 µm window was memo-rized, and then the EC-TEM holder was removed. Next, another TEM holder in which aconventional carbon film-supported Cu microgrid was mounted was introduced into themicroscope, and the electron-optics system and aberration correctors were finely tuned.Then, the EC-TEM holder was again introduced into the microscope, the viewing fieldposition was recalled, and the STEM imaging was performed. An aberration-correctedbeam probe with an energy of 200 keV and a current of 52.5 pA passing through a condenseraperture (28 mrad convergence semi-angle) was used. The STEM resolution and dwell timewere 512 × 512 pixels and 5.0 µs/pixel, respectively. STEM movies were recorded using ahigh-definition video function in the Gatan Microscopy Suite® (Gatan, USA) with a rate of10 frames/s.Nanomaterials 2023, 13, 2170 4 of 102.3. Achievement of Atomic Resolution by Improving the Depth Resolution of a STEM ProbeDuring the STEM observation of samples in an EC, the materials of the two membranewindows and the media packed in the cell deteriorate the incident electron beam probeshape. The lateral spatial resolution of EC-STEM is given as follows [22–24]:∆d =√do2 + dSNR2 + dblur2 + dcc2 (1)where do is the resolution determined by the geometrical optics of an instrument (i.e., it isdefined by spherical aberration and a condenser aperture function), dSNR is the resolutionlimited by the SNR of an object image, dblur is the resolution originating from the beamblurring caused by elastic scattering, and dcc is the resolution limited by chromatic aber-ration [22–24]. Although the geometrical optics and chromatic aberration are fixed by amicroscope imaging condition, the influence of the beam blurring of the STEM probe canbe minimized when the sample is placed on the inner side of the top membrane. However,the SNR depends on the thicknesses of the membranes and media, which predominantlylimit the resolution of EC-STEM [22]. Reducing the membrane thickness to increase SNR israrely recommended because the robustness of the membrane also decreases.The improvement of the depth resolution in STEM (i.e., shortening the elongation of theSTEM probe shape along the vertical direction), is the most effective way to enhance the contrastof the sample against the background intensity, i.e., directly relating to the SNR, effectivelyincreasing the resolution in EC-STEM. Since the depth resolution is inversely proportionalto a square of the convergence semi-angle in an aberration-corrected probe-forming lens, thewider the convergence semi-angle, the smaller the depth resolution [25–31]. This technique isespecially practical for atomic-sized objects and has been exploited to three-dimensionallyvisualize atoms on surfaces and inside samples [30,31]. In the present work, an aberration-corrected STEM probe formed with a condenser aperture with a 28 mrad convergencesemi-angle was used, for which the depth resolution was theoretically determined tobe approximately 3 nm, assuming that the chromatic aberration was negligible. Usingthis probe, atomic-resolution STEM observations can be performed in air for samplessandwiched between the silicon nitride membranes of an EC.2.4. The Dose Rate of STEM ObservationThe dose rate is a significant factor in environmental electron microscopy because theinteraction between incident electrons and materials in the cell (i.e., the sample and itssurrounding media) causes damage to the sample and radiolysis of media molecules. Thelatter generates chemically reactive species such as radicals. This issue requires considera-tion, in particular, concerning how much of the natural phenomena or structures can beunderstood from the observed results and how the results should be compensated to clarifythe real underlying mechanism or systems. The direct interaction (elastic scattering) be-tween primary electrons and the sample causes knock-on damage when the electron energyis higher than the threshold value of the displacement energy of atoms in the sample. Theelastic scattering of electrons by molecules in media under Brownian motion also occurs,and momentum is transferred from electrons to the molecules, accelerating the Brownianmotion speed, as previously mentioned. The inelastic scattering of primary electrons resultsin the emission of secondary electrons from the sample, media molecules, and cell windowmembranes. Secondary electrons are the main factor in radiolysis, especially in the case ofthe liquid cell [32]. The radiolysis products are chemically reactive with the sample andoften cause unexpected results. Hence, minimizing the dose rate is the general way tomitigate unwanted physical and chemical damage to the samples.Whereas electrons continuously illuminate a whole viewing area in TEM, in STEM,a finely focused probe is raster-scanned, and each point in the observation area is onlyexposed for the length of the dwell time during every pass. The dose rate of STEM isdescribed as:Dose rate = 6.24 × 106 JD2[e/nm2s]=JD2[pA/nm2](2)Nanomaterials 2023, 13, 2170 5 of 10where J is the beam current, and D is the length of one side of the viewing area, assumingthat the viewing area is a rectangle. In this study, the STEM beam of 52.5 pA gives adose rate of the order of 10−1 pA/nm2, even when magnified 10 million times (enough toachieve atomic-resolution imaging), which is almost the same order as that in conventionalhigh-resolution liquid TEM [33,34].3. Results and DiscussionThe morphology change of the Pt networks was small during HAADF-STEM imagingwhen searching the field of view and focusing at a magnification of less than 1 million times.However, when the magnification was increased to 5 million times, the network structuresstarted transforming, as shown in Figure 2, in which 2a to 2d were taken every 1.5 s underthe dose rate of 3.3 × 10−2 pA/nm2. This gradual transformation was potentially causedby the random attacks of gas molecules and atoms accelerated by electron collisions, ratherthan the direct electron sputtering, because the transformation speed of Pt networks wasdramatically increased when the Pt networks were covered with water in the EC andnegligible when they were in a vacuum. It should be noted that the connected parts weremore stable than the isolated islands. The islands indicated by white arrows 1 and 2 movedand finally attached to the neighboring parts. The Video S1 shows the gradual change ofthe whole morphology of the network structures.Nanomaterials 2023, 13, x FOR PEER REVIEW 5 of 10   𝐷𝑜𝑠𝑒 𝑟𝑎𝑡𝑒 = 6.24 × 10  [e nm s⁄ ] = [pA/nm ]  (2)where J is the beam current, and D is the length of one side of the viewing area, assuming that the viewing area is a rectangle. In this study, the STEM beam of 52.5 pA gives a dose rate of the order of 10−1 pA/nm2, even when magnified 10 million times (enough to achieve atomic-resolution imaging), which is almost the same order as that in conventional high-resolution liquid TEM [33,34].  3. Results and Discussion The morphology change of the Pt networks was small during HAADF-STEM imag-ing when searching the field of view and focusing at a magnification of less than 1 million times. However, when the magnification was increased to 5 million times, the network structures started transforming, as shown in Figure 2, in which 2a to 2d were taken every 1.5 s under the dose rate of 3.3 × 10−2 pA/nm2. This gradual transformation was potentially caused by the random attacks of gas molecules and atoms accelerated by electron colli-sions, rather than the direct electron sputtering, because the transformation speed of Pt networks was dramatically increased when the Pt networks were covered with water in the EC and negligible when they were in a vacuum. It should be noted that the connected parts were more stable than the isolated islands. The islands indicated by white arrows 1 and 2 moved and finally attached to the neighboring parts. The Video S1 shows the grad-ual change of the whole morphology of the network structures.  Figure 2. HAADF-STEM images showing the morphology change of Pt networks in air under the dose rate of 3.3 × 10−2 pA/nm2. (a–d) were taken every 1.5 s. The islands indicated by white arrows 1 and 2 moved up and finally attached to the neighboring parts. The morphology change became more remarkable as the magnification increased. Figure 3a–c shows a series of atomic-resolution HAADF-STEM images, revealing multiple connected grains, captured and cropped from a STEM movie (Video S2), of which the STEM magnification and dose rate were 10 million times and 1.3 × 10−1 pA/nm2, respec-tively. The elapsed time is displayed at the bottom right of each image. The Pt networks in this viewing area change slowly for 46 s. In Figure 3a–c, each bright dot corresponds to an atomic column in the grains and is marked by red, blue, and green circles in Figure 3d–f, respectively. Figure 4a is a schematic drawing of an atomic model showing the orienta-tional relationship of the face-centered-cubic (FCC) structure viewed from [110]. In Fig-ure 3a–c, the three grains show a clear atomic structure with <011> incidence, and they are connected with their {111} planes. Figure 4b shows an overlayed image of only the colored circles from Figure 3d–f. The change of atomic structure in the central part of each grain is smaller than that in other parts, such as the surfaces and GBs. In Figure 4b, atomic col-umns with less movement in the grains are filled with the same colors as the circles. The group of color-filled circles can be regarded as stable grain cores. In Figure 3d–f, the grain cores are labeled as GC1, GC2, and GC3. The crystallographic orientation relations be-tween GC1 and GC2 are well-matched, whereas GC3 had a mirror image relation with them (i.e., a twin relation). GC1 and GC2 have three GBs, whereas GC3 has only two GBs. In Figure 3, all the images are displayed by aligning them so that the outlines of GC1 and GC2 match. The positions of not only the outlines but also internal atomic columns are Figure 2. HAADF-STEM images showing the morphology change of Pt networks in air under thedose rate of 3.3 × 10−2 pA/nm2. (a–d) were taken every 1.5 s. The islands indicated by white arrows1 and 2 moved up and finally attached to the neighboring parts.The morphology change became more remarkable as the magnification increased.Figure 3a–c shows a series of atomic-resolution HAADF-STEM images, revealing multipleconnected grains, captured and cropped from a STEM movie (Video S2), of which the STEMmagnification and dose rate were 10 million times and 1.3 × 10−1 pA/nm2, respectively.The elapsed time is displayed at the bottom right of each image. The Pt networks in thisviewing area change slowly for 46 s. In Figure 3a–c, each bright dot corresponds to anatomic column in the grains and is marked by red, blue, and green circles in Figure 3d–f,respectively. Figure 4a is a schematic drawing of an atomic model showing the orientationalrelationship of the face-centered-cubic (FCC) structure viewed from[110]. In Figure 3a–c,the three grains show a clear atomic structure with <011> incidence, and they are connectedwith their {111} planes. Figure 4b shows an overlayed image of only the colored circlesfrom Figure 3d–f. The change of atomic structure in the central part of each grain is smallerthan that in other parts, such as the surfaces and GBs. In Figure 4b, atomic columns withless movement in the grains are filled with the same colors as the circles. The group ofcolor-filled circles can be regarded as stable grain cores. In Figure 3d–f, the grain coresare labeled as GC1, GC2, and GC3. The crystallographic orientation relations betweenGC1 and GC2 are well-matched, whereas GC3 had a mirror image relation with them(i.e., a twin relation). GC1 and GC2 have three GBs, whereas GC3 has only two GBs. InFigure 3, all the images are displayed by aligning them so that the outlines of GC1 andGC2 match. The positions of not only the outlines but also internal atomic columns areconsistent. Notably, they are stable, despite atoms in their GBs continuing to fluctuate.As for two GBs supporting GC3, the GB between GC2 and GC3 is a twin boundary, andanother, the GB between GC3 and the grain on the left side of GC3, appears to contain aNanomaterials 2023, 13, 2170 6 of 10crystallographic mismatch and be strained because the grain is largely misaligned fromthe <011> incidence. Hence, GC3 was unstable, unlike GC1 and GC2, and its internal andsurrounding atoms were more likely to fluctuate.Nanomaterials 2023, 13, x FOR PEER REVIEW 6 of 10   consistent. Notably, they are stable, despite atoms in their GBs continuing to fluctuate. As for two GBs supporting GC3, the GB between GC2 and GC3 is a twin boundary, and an-other, the GB between GC3 and the grain on the left side of GC3, appears to contain a crystallographic mismatch and be strained because the grain is largely misaligned from the <011> incidence. Hence, GC3 was unstable, unlike GC1 and GC2, and its internal and surrounding atoms were more likely to fluctuate.  Figure 3. (a–c) Atomic-resolution HAADF-STEM images showing the structural change of Pt net-works in air under the dose rate of 1.3 × 10−1 pA/nm2. The elapsed time is displayed at the right of each image. (d–f) Atomic columns in Pt networks in (a–c) are marked by red, blue, and green circles.  Figure 4. (a) Schematic drawing of an atomic model and the orientational relationship of the FCC structure viewed from [110] . (b) Overlayed image of only the colored circles from Figure 3d–f. Atomic columns with less movement in the grains are filled with the same colors as the circles. The group of color-filled circles can be regarded as stable grain cores, labeled as GC1, GC2, and GC3. In addition to the fluctuation of atoms at the GBs (i.e., unstable GBs) among GC1, GC2, and GC3, the atoms continuously rearranged on their surfaces. Figure 5a shows a HAADF-STEM image captured from a STEM movie (Video S2), and Figure 5b shows a fast Fourier transformation (FFT) pattern calculated from a region inside a yellow rectan-gle in 5a, in which diffraction spots corresponding to the Pt {111} and Pt {200} planes are Figure 3. (a–c) Atomic-resolution HAADF-STEM images showing the structural change of Pt net-works in air under the dose rate of 1.3 × 10−1 pA/nm2. The elapsed time is displayed at the right ofeach image. (d–f) Atomic columns in Pt networks in (a–c) are marked by red, blue, and green circles.Nanomaterials 2023, 13, x FOR PEER REVIEW 6 of 10   consistent. Notably, they are stable, despite atoms in their GBs continuing to fluctuate. As for two GBs supporting GC3, the GB between GC2 and GC3 is a twin boundary, and an-other, the GB between GC3 and the grain on the left side of GC3, appears to contain a crystallographic mismatch and be strained because the grain is largely misaligned from the <011> incidence. Hence, GC3 was unstable, unlike GC1 and GC2, and its internal and surrounding atoms were more likely to fluctuate.  Figure 3. (a–c) Atomic-resolution HAADF-STEM images showing the structural change of Pt net-works in air under the dose rate of 1.3 × 10−1 pA/nm2. The elapsed time is displayed at the right of each image. (d–f) Atomic columns in Pt networks in (a–c) are marked by red, blue, and green circles.  Figure 4. (a) Schematic drawing of an atomic model and the orientational relationship of the FCC structure viewed from [110] . (b) Overlayed image of only the colored circles from Figure 3d–f. Atomic columns with less movement in the grains are filled with the same colors as the circles. The group of color-filled circles can be regarded as stable grain cores, labeled as GC1, GC2, and GC3. In addition to the fluctuation of atoms at the GBs (i.e., unstable GBs) among GC1, GC2, and GC3, the atoms continuously rearranged on their surfaces. Figure 5a shows a HAADF-STEM image captured from a STEM movie (Video S2), and Figure 5b shows a fast Fourier transformation (FFT) pattern calculated from a region inside a yellow rectan-gle in 5a, in which diffraction spots corresponding to the Pt {111} and Pt {200} planes are Figure 4. (a) Schematic drawing of an atomic model and the orientational relationship of the FCCstructure viewed from[110]. (b) Overlayed image of only the colored circles from Figure 3d–f.Atomic columns with less movement in the grains are filled with the same colors as the circles. Thegroup of color-filled circles can be regarded as stable grain cores, labeled as GC1, GC2, and GC3.In addition to the fluctuation of atoms at the GBs (i.e., unstable GBs) among GC1,GC2, and GC3, the atoms continuously rearranged on their surfaces. Figure 5a shows aHAADF-STEM image captured from a STEM movie (Video S2), and Figure 5b shows a fastFourier transformation (FFT) pattern calculated from a region inside a yellow rectangle in5a, in which diffraction spots corresponding to the Pt {111} and Pt {200} planes are observed.It should be noted that there are extra spots inside the Pt {111} spots. The radius of theextra spots from the center spot is approximately 10% smaller than that of the Pt {111} spots.Figure 5c shows an intensity line profile along a blue line in 5a, illustrating that the atomicNanomaterials 2023, 13, 2170 7 of 10spacing of surface layers is approximately 10% larger than that of the bulk ones. FromFigure 5b,c, it is suggested that the lattice spacing of a few atomic layers of Pt {111} surfacesexpanded by approximately 10% along the surface direction. Figure 4b (and Video S2)shows surface atoms repeatedly appearing and disappearing randomly, but no atomicstep motion was recognized. This differs from the surface step growth reported in thecase of catalytically reacting surfaces of NPG in a CO/air environment [17]. Yoshida et al.reported that the environmental TEM of Pt nanoparticles supported on cerium oxides underoxidation conditions showed the transformation of the nanoparticle surfaces to oxideslayers, for which the lattice constant was larger than that of Pt, and the oxide layers grewfurther by increasing the oxygen partial pressure and electron beam intensity [35]. Sincethe formation of the surface layers with an expanded lattice spacing in our observation isconsistent with their results, the surface layers covering the network structures may be anoxide phase.Nanomaterials 2023, 13, x FOR PEER REVIEW 7 of 10   observed. It should be noted that there are extra spots inside the Pt {111} spots. The radius of the extra spots from the center spot is approximately 10% smaller than that of the Pt {111} spots. Figure 5c shows an intensity line profile along a blue line in 5a, illustrating that the atomic spacing of surface layers is approximately 10% larger than that of the bulk ones. From Figure 5b,c, it is suggested that the lattice spacing of a few atomic layers of Pt {111} surfaces expanded by approximately 10% along the surface direction. Figure 4b (and Video S2) shows surface atoms repeatedly appearing and disappearing randomly, but no atomic step motion was recognized. This differs from the surface step growth reported in the case of catalytically reacting surfaces of NPG in a CO/air environment [17]. Yoshida et al. reported that the environmental TEM of Pt nanoparticles supported on cerium oxides under oxidation conditions showed the transformation of the nanoparticle surfaces to ox-ides layers, for which the lattice constant was larger than that of Pt, and the oxide layers grew further by increasing the oxygen partial pressure and electron beam intensity [35]. Since the formation of the surface layers with an expanded lattice spacing in our observa-tion is consistent with their results, the surface layers covering the network structures may be an oxide phase.  Figure 5. (a) HAADF-STEM image of the Pt networks captured from a STEM movie (Video S2). (b) FFT pattern calculated from a region inside a yellow rectangle in (a), in which diffraction spots cor-responding to Pt {111} and Pt {200} planes are observed. (c) Intensity line profile along the blue line in (a). The present observations showed that the morphology changes were mainly due to the movement of unstable grains, as shown in Video S1. Furthermore, no atomic step mo-tion was noted, and neither surface growth nor etching was observed, although the atoms at the surfaces and GBs fluctuated, as shown in Video S2. Unstable grains and chained grains may deform or move to attach to neighboring parts in the network structures, and a GB-containing strain may attempt to relax its strain energy via atomic reconstruction or fracturing themselves. The connections of the grains through their {111} planes with small mismatches may be stable. Considering that the surface self-diffusion coefficients of a Pt atom on Pt {111} and an Au atom on Au {111} at 27 °C are calculated to be 5.3 × 10−8 and 8.3 × 10−7 cm2/s, respectively, using parameters in Ref. [36], diffusion-induced step motion of Pt atoms is intrinsically not likely to occur compared with that of Au atoms. In addition, oxygen molecules can be spontaneously dissociated and chemically adsorbed on Pt sur-faces to be a surface oxide layer, while the chemisorption of oxygen on Au rarely occurs. Thus, differing from the case of NPG, the passive oxide surface on Pt in the present case Figure 5. (a) HAADF-STEM image of the Pt networks captured from a STEM movie (Video S2).(b) FFT pattern calculated from a region inside a yellow rectangle in (a), in which diffraction spotscorresponding to Pt {111} and Pt {200} planes are observed. (c) Intensity line profile along the blueline in (a).The present observations showed that the morphology changes were mainly due to themovement of unstable grains, as shown in Video S1. Furthermore, no atomic step motionwas noted, and neither surface growth nor etching was observed, although the atomsat the surfaces and GBs fluctuated, as shown in Video S2. Unstable grains and chainedgrains may deform or move to attach to neighboring parts in the network structures, anda GB-containing strain may attempt to relax its strain energy via atomic reconstructionor fracturing themselves. The connections of the grains through their {111} planes withsmall mismatches may be stable. Considering that the surface self-diffusion coefficients ofa Pt atom on Pt {111} and an Au atom on Au {111} at 27 ◦C are calculated to be 5.3 × 10−8and 8.3 × 10−7 cm2/s, respectively, using parameters in Ref. [36], diffusion-induced stepmotion of Pt atoms is intrinsically not likely to occur compared with that of Au atoms. Inaddition, oxygen molecules can be spontaneously dissociated and chemically adsorbedon Pt surfaces to be a surface oxide layer, while the chemisorption of oxygen on Au rarelyoccurs. Thus, differing from the case of NPG, the passive oxide surface on Pt in the presentcase hindered surface atomic diffusion, which explains why the coarsening and thinningof grains and GBs followed by the rupturing of connected chains barely occurred in thePt networks.Nanomaterials 2023, 13, 2170 8 of 104. ConclusionsPt networks, which were composed of connected polycrystalline chains, were fab-ricated on the inner surface of a silicon nitride window in an EC, and their dynamicstructural change was observed via atomic-resolution EC-STEM in air at 1 atm and roomtemperature. Owing to the shrinkage of the elongation along the beam direction of anaberration-corrected STEM probe with a wide convergence semi-angle, the SNR of thePt network was enhanced, which is a dominant factor in improving the spatial lateralresolution, and atomic-resolution imaging was achieved. The exposure of Pt networksto the gas molecules stimulated by electron beams was considered to increase the colli-sion probability between gas molecules and Pt networks, and the Pt networks are moreintensively stressed from all directions than the situation without electron irradiation. Theobservations suggested that the morphology change of the Pt network in an oxidationenvironment was mainly attributed to the deformation and movement of chain-connectedgrains and GBs. Although atoms at the surface and GBs continued fluctuating, the sur-face atomic-diffusion-induced step motion was difficult to observe. Considering that oursample has two-dimensional network features, three-dimensional network catalysts thatcan be fabricated practically using a sophisticated design may show high stability becauseof the increased number of connection points among grains. By modifying our EC toenable the introduction of various gases, the technology can be expanded to more practicalcatalytic reaction experiments in the future. Aberration-corrected EC-STEM using a wideconvergence angle is capable of improving the signal-to-background ratio of the samples inmedia sandwiched within the silicon nitride windows of an EC, opening the way to the insitu nano- and atomic-scale characterizations of catalysts in not only gases but also variouskinds of liquids including electrolytes.Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nano13152170/s1, Video S1: Takeguchi_S1.mp4; Video S2: Takeguchi_S2.mp4.Author Contributions: Conceptualization, M.T.; methodology, M.T.; investigation, M.T.; liquid celland sample preparation, T.T.; discussion, M.T. and K.M.; writing, M.T.; review, M.T. and K.M. Allauthors have read and agreed to the published version of the manuscript.Funding: A part of this work was supported by the Innovative Science and Technology Initiative forSecurity, ATLA, Japan (Grant Number JPJ004596).Data Availability Statement: Not applicable.Acknowledgments: We thank Xiaoguang Li for helping fabricate our ECs and for the fruitful discus-sion. We thank Robert Ireland for editing a draft of this manuscript.Conflicts of Interest: The authors declare no conflict of interest.References1. Ma, T.Y.; Dai, S.; Qiao, S.Z. Self-supported electrocatalysts for advanced energy conversion processes. Mater. Today 2016, 19,265–273. [CrossRef]2. Ehelebe, K.; Schmitt, N.; Sievers, G.; Jensen, A.W.; Hrnjić, A.; Jiménez, P.C.; Kaiser, P.; Geuß, M.; Ku, Y.; Jovanovič, P.; et al.Benchmarking fuel cell electrocatalysts using gas diffusion electrodes: Inter-lab comparison and best practices. ACS Energy Lett.2022, 7, 816–826. [CrossRef]3. Tian, H.; Song, A.; Tian, H.; Liu, J.; Shao, G.; Liu, H.; Wang, G. Single-atom catalysts for high-energy rechargeable batteries. Chem.Sci. 2021, 12, 7656–7676. [CrossRef]4. Xu, K.; Zhao, X.; Hu, X.; Guo, Z.; Ye, Q.; Li, L.; Song, J.; Song, P. The review of the degradation mechanism of the catalyst layer ofmembrane electrode assembly in the proton exchange membrane fuel cell. IOP Conf. Ser. Earth Environ. Sci. 2020, 558, 052041.[CrossRef]5. Debe, M.K. Electrocatalyst approaches and challenges for automotive fuel cells. Nature 2012, 486, 43–51. [CrossRef]6. Jiao, L.; Wang, Y.; Jiang, H.; Xu, Q. Metal–organic frameworks as platforms for catalytic applications. Adv. Mater. 2018, 30, 1703663.[CrossRef] [PubMed]7. Das, S.; Pérez-Ramírez, J.; Gong, J.; Dewangan, N.; Hidajat, K.; Gates, B.C.; Kawi, S. Core-shell structured catalysts forthermocatalytic, photocatalytic, and electrocatalytic conversion of CO2. Chem. Soc. Rev. 2020, 49, 2937–3004. [CrossRef]https://www.mdpi.com/article/10.3390/nano13152170/s1https://www.mdpi.com/article/10.3390/nano13152170/s1https://doi.org/10.1016/j.mattod.2015.10.012https://doi.org/10.1021/acsenergylett.1c02659https://doi.org/10.1039/D1SC00716Ehttps://doi.org/10.1088/1755-1315/558/5/052041https://doi.org/10.1038/nature11115https://doi.org/10.1002/adma.201703663https://www.ncbi.nlm.nih.gov/pubmed/29178384https://doi.org/10.1039/C9CS00713JNanomaterials 2023, 13, 2170 9 of 108. Tajuddin, A.A.H.; Wakisaka, M.; Ohto, T.; Yu, Y.; Fukushima, H.; Tanimoto, H.; Li, X.; Misu, Y.; Jeong, S.; Fujita, J.; et al. Corrosion-resistant and high-entropic non-noble-metal electrodes for oxygen evolution in acidic media. Adv. Mater. 2023, 35, 2207466.[CrossRef]9. Jensen, W.; Sievers, G.W.; Jensen, K.D.; Quinson, J.; Arminio-Ravelo, J.A.; Brüser, V.; Arenz, M.; Escudero-Escribano, M. Self-supported nanostructured iridium-based networks as highly active electrocatalysts for oxygen evolution in acidic media. J. Mater.Chem. A 2020, 8, 1066–1071. [CrossRef]10. Chen, C.; Kang, Y.; Huo, Z.; Zhu, Z.; Huang, W.; Xin, H.L.; Snyder, J.D.; Li, D.; Herron, J.A.; Mavrikakis, M.; et al. Highlycrystalline multimetallic nanoframes with three-dimensional electrocatalytic surfaces. Science 2014, 343, 1339–1343. [CrossRef]11. Sakai, G.; Yoshimura, T.; Isohata, S.; Uota, M.; Kawasaki, H.; Kuwahara, T.; Fujikawa, D.; Kijima, T. Synthesis of Nanogroove-network-structured platinum nanosheets and their carbon-supported forms using a mixed-surfactant templating approach. Adv.Mater. 2007, 9, 237–241. [CrossRef]12. Funatsu, A.; Tateishi, T.; Hatakeyama, K.; Fukunaga, Y.; Taniguchi, T.; Koinuma, M.; Matsuura, H.; Matsumoto, Y. Synthesis ofmonolayer platinum nanosheets. Chem. Commun. 2014, 50, 8503–8506. [CrossRef] [PubMed]13. Sievers, G.W.; Jensen, A.W.; Quinson, J.; Zana, A.; Bizzotto, F.; Oezaslan, M.; Dworzak, A.; Kirkensgaard, J.J.K.; Smitshuysen,T.E.L.; Kadkhodazadeh, S.; et al. Self-supported Pt–CoO networks combining high specific activity with high surface area foroxygen reduction. Nat. Mater. 2021, 20, 208–213. [CrossRef] [PubMed]14. Fujita, T.; Guan, P.; McKenna, K.; Lang, X.; Hirata, A.; Zhang, L.; Tokunaga, T.; Arai, S.; Yamamoto, Y.; Tanaka, N.; et al. Atomicorigins of the high catalytic activity of nanoporous gold. Nat. Mater. 2012, 11, 75–780. [CrossRef] [PubMed]15. Lee, J.D.; Qi, Z.; Foucher, A.C.; Ngan, H.T.; Dennis, K.; Cui, J.; Sadykov, I.I.; Crumlin, E.J.; Sautet, P.; Stach, E.A.; et al. Facilitatinghydrogen dissociation over dilute nanoporous Ti–Cu catalysts. J. Am. Chem. Soc. 2022, 144, 16778–16791. [CrossRef]16. Chen, T.; Pang, D.; Kang, J.; Zhang, D.; Guo, L. Network-like platinum nanosheets enabled by a calorific-effect-induced-fusionstrategy for enhanced catalytic hydrogenation performance. Front. Chem. 2022, 9, 818900. [CrossRef]17. Fujita, T.; Tokunaga, T.; Zhang, L.; Li, D.; Chen, L.; Arai, S.; Yamamoto, Y.; Hirata, A.; Tanaka, N.; Ding, Y.; et al. Atomicobservation of catalysis-induced nanopore coarsening of nanoporous gold. Nano Lett. 2014, 14, 172–1177. [CrossRef]18. Liu, P.; Wei, X.; Song, S.; Wang, L.; Hirata, A.; Fujita, T.; Han, X.; Zhang, Z.; Chen, M. Time-resolved atomic-scale observations ofdeformation and fracture of nanoporous gold under tension. Acta Mater. 2019, 165, 99–108. [CrossRef]19. Li, X.; Mitsuishi, K.; Takeguchi, M. Fabrication of a liquid cell for in situ transmission electron microscopy. Microscopy 2021, 70,327–332. [CrossRef]20. Takeguchi, M.; Li, X.; Mitsuishi, K. High-resolution STEM observation of the dynamics of Pt nanoparticles in a liquid. Jpn. J. Appl.Phys. 2022, 61, SD1021. [CrossRef]21. Li, X.; Mitsuishi, K.; Takeguchi, M. Effect of amorphous carbon coating on the performance of liquid phase transmission electronmicroscopy (LP-TEM) and the dynamics of enclosed Pt nano-colloids. Microscopy 2022, 71, 181–186. [CrossRef]22. de Jonge, N.; Ross, F.M. Electron microscopy of specimens in liquid. Nat. Nanotechnol. 2011, 6, 695–704. [CrossRef]23. de Jonge, N. Theory of the spatial resolution of (scanning) transmission electron microscopy in liquid water or ice layers.Ultramicroscopy 2018, 187, 113–125. [CrossRef] [PubMed]24. Jongbaek, S.; Bae, Y.; Park, H.; Kang, S.; Choi, B.K.; Kim, J.; Park, J. Liquid-phase transmission electron microscopy for reliable insituimaging of nanomaterials. Annu. Rev. Chem. Biomol. Eng. 2022, 10, 167–191. [CrossRef]25. van Benthem, K.; Lupini, A.R.; Kim, M.; Baik, H.S.; Doh, S.; Lee, J.; Oxley, M.P.; Findlay, S.D.; Allen, L.J.; Luck, J.T.; et al.Three-dimensional imaging of individual hafnium atoms inside a semiconductor device. Appl. Phys. Lett. 2005, 87, 034104.[CrossRef]26. van Benthem, K.; Lupini, A.R.; Oxley, M.O.; Findlay, S.D.; Allen, L.J.; Pennycook, S.J. Three-dimensional ADF imaging ofindividual atoms by through-focal series scanning transmission electron microscopy. Ultramicroscopy 2006, 106, 1062–1068.[CrossRef] [PubMed]27. Xin, H.L.; Intaraprasonk, V.; Muller, D.A. Depth Sectioning of Individual Dopant Atoms with Aberration-Corrected ScanningTransmission Electron Microscopy. Microsc. Microanal. 2007, 13, 884–885. [CrossRef]28. Hamaoka, T.; Jao, C.; Zhang, X.; Oshima, Y.; Takeguchi, M. Three-dimensional characterization of Guinier–Preston zones in anAl–Cu alloy using depth-sectioning technique. Microscopy 2017, 66, 78–88. [CrossRef]29. Borisevich, A.Y.; Lupini, A.R.; Pennycook, S.J. Depth sectioning with the aberration-corrected scanning transmission electronmicroscope. Proc. Natl. Acad. Sci. USA 2006, 103, 3044–3048. [CrossRef]30. Ishikawa, R.; Shibata, N.; Taniguchi, T.; Ikuhara, Y. Three-dimensional imaging of a single dopant in a crystal. Phys. Rev. Appl.2020, 13, 034064. [CrossRef]31. Ishikawa, R.; Tanaka, R.; Kawahara, K.; Shibata, N.; Ikuhara, Y. Atomic-resolution topographic imaging of crystal surfaces. ACSNano 2021, 15, 9186–9193. [CrossRef] [PubMed]32. Bultema, L.A.; Bücker, R.; Schulz, E.C.; Tellkamp, F.; Gonschior, J.; Miller, R.J.D.; Kassier, G.H. The effect of secondary electronson radiolysis as observed by in liquid TEM: The role of window material and electrical bias. Ultramicroscopy 2022, 240, 113579.[CrossRef] [PubMed]33. Zhu, C.; Liang, S.; Song, E.; Zhou, Y.; Wang, W.; Shan, F.; Shi, Y.; Hao, C.; Yin, K.; Zhang, T.; et al. In-situ liquid cell transmissionelectron microscopy investigation on oriented attachment of gold nanoparticles. Nat. Commun. 2018, 9, 421. [CrossRef]https://doi.org/10.1002/adma.202207466https://doi.org/10.1039/C9TA12796Hhttps://doi.org/10.1126/science.1249061https://doi.org/10.1002/adma.200601127https://doi.org/10.1039/C4CC02527Jhttps://www.ncbi.nlm.nih.gov/pubmed/24947470https://doi.org/10.1038/s41563-020-0775-8https://www.ncbi.nlm.nih.gov/pubmed/32839587https://doi.org/10.1038/nmat3391https://www.ncbi.nlm.nih.gov/pubmed/22886067https://doi.org/10.1021/jacs.2c00830https://doi.org/10.3389/fchem.2021.818900https://doi.org/10.1021/nl403895shttps://doi.org/10.1016/j.actamat.2018.11.022https://doi.org/10.1093/jmicro/dfaa076https://doi.org/10.35848/1347-4065/ac54f0https://doi.org/10.1093/jmicro/dfac012https://doi.org/10.1038/nnano.2011.161https://doi.org/10.1016/j.ultramic.2018.01.007https://www.ncbi.nlm.nih.gov/pubmed/29428430https://doi.org/10.1146/annurev-chembioeng-092120-034534https://doi.org/10.1063/1.1991989https://doi.org/10.1016/j.ultramic.2006.04.020https://www.ncbi.nlm.nih.gov/pubmed/16875782https://doi.org/10.1017/S1431927607073631https://doi.org/10.1093/jmicro/dfw104https://doi.org/10.1073/pnas.0507105103https://doi.org/10.1103/PhysRevApplied.13.034064https://doi.org/10.1021/acsnano.1c02907https://www.ncbi.nlm.nih.gov/pubmed/33983030https://doi.org/10.1016/j.ultramic.2022.113579https://www.ncbi.nlm.nih.gov/pubmed/35780682https://doi.org/10.1038/s41467-018-02925-6Nanomaterials 2023, 13, 2170 10 of 1034. Hong, J.; Bae, J.; Jo, H.; Park, H.; Lee, S.; Hong, S.J.; Chun, H.; Cho, M.K.; Kim, J.; Kim, J.; et al. Metastable hexagonal close-packedpalladium hydride in liquid cell TEM. Nature 2022, 603, 631–636. [CrossRef]35. Yoshida, H.; Omote, H.; Takeda, S. Oxidation and reduction processes of platinum nanoparticles observed at the atomic scale byenvironmental transmission electron microscopy. Nanoscale 2014, 6, 13113–13118. [CrossRef]36. Agrawal, P.M.; Rice, B.M.; Thompson, D.L. Predicting trends in rate parameters for self-diffusion on FCC metal surfaces. Surf. Sci.2002, 515, 21–35. [CrossRef]Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individualauthor(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury topeople or property resulting from any ideas, methods, instructions or products referred to in the content.https://doi.org/10.1038/s41586-021-04391-5https://doi.org/10.1039/C4NR04352Ahttps://doi.org/10.1016/S0039-6028(02)01916-7 Introduction  Materials and Methods  Sample Preparations in Environmental Cells  Environmental Cell Scanning Transmission Electron Microscopy  Achievement of Atomic Resolution by Improving the Depth Resolution of a STEM Probe  The Dose Rate of STEM Observation  Results and Discussion  Conclusions  References