- Email: [email protected]
Atomic-scale electron microscopy at ambient pressure
ARTICLE IN PRESS Ultramicroscopy 108 (2008) 993– 998
Contents lists available at ScienceDirect
Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic
Atomic-scale electron microscopy at ambient pressure J.F. Creemer a,, S. Helveg b, G.H. Hoveling c, S. Ullmann b, A.M. Molenbroek b, P.M. Sarro a, H.W. Zandbergen d a
DIMES-ECTM, Delft University of Technology, P.O. Box 5053, 2600 GB Delft, The Netherlands Haldor Topsøe A/S, Nymøllevej 55, DK-2800 Kgs. Lyngby, Denmark DEMO, Delft University of Technology, P.O. Box 5031, 2600 GA Delft, The Netherlands d Kavli Institute of NanoScience, HREM, Delft University of Technology, P.O. Box 5046, 2600 GA Delft, The Netherlands b c
a r t i c l e in fo
abstract
Article history: Received 27 January 2008 Received in revised form 31 March 2008 Accepted 15 April 2008
We demonstrate a novel nanoreactor for performing atomic-resolution environmental transmission electron microscopy (ETEM) of nanostructured materials during exposure to gases at ambient pressures and elevated temperatures. The nanoreactor is a microelectromechanical system (MEMS) and is functionalized with a micrometer-sized gas-flow channel, electron-transparent windows and a heating device. It fits into the tip of a dedicated sample holder that can be used in a normal CM microscope of Philips/FEI Company. The nanoreactor performance was demonstrated by ETEM imaging of a Cu/ZnO catalyst for methanol synthesis during exposure to hydrogen. Specifically, the nanoreactor facilitated the direct observation of Cu nanocrystal growth and mobility on a sub-second time scale during heating to 500 1C and exposure to 1.2 bar of H2. For the same gas reaction environment, ETEM images show atomic lattice fringes in the Cu nanocrystals with spacing of 0.18 nm, attesting the spatial resolution limit of the system. The nanoreactor concept opens up new possibilities for in situ studies of nanomaterials and the ways they interact with their ambient working environment in diverse areas, such as heterogeneous catalysis, electrochemistry, nanofabrication, materials science and biology. & 2008 Elsevier B.V. All rights reserved.
PACS: 68.37.Og 85.85.+j 82.65.+r 78.67.Bf Keywords: Environmental TEM ETEM Microelectromechanical systems MEMS Gas–solid interactions Nanocrystals Methanol synthesis catalyst
1. Introduction It has been a longstanding challenge in transmission electron microscopy (TEM) to study chemical processes at gas–solid interfaces in situ with atomic-scale resolution [1–6]. Highresolution TEM (HRTEM) is a powerful technique for atomic-scale imaging of nanomaterials that are kept under high vacuum conditions (Ca. 10 9 bar). Applying HRTEM for studies of gas–solid interactions is extremely demanding, because the HRTEM image resolution is degraded by scattering of the electron beam on gas atoms and by specimen drift due to heating. A key requirement is therefore to limit the number of gas atoms in the microscope, which can be reached by limiting the pressure or by confining the gas along the path of the beam into a layer that is as thin as possible. Confinement has been attained by two methods: by differentially pumped vacuum systems [1–5] and by windowed
Corresponding author. Tel.: +3115 278 6277; fax: +3115 262 2163.
E-mail address: [email protected] (J.F. Creemer). 0304-3991/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ultramic.2008.04.014
cells [1,6–8]. These capabilities for performing in situ studies of gas–solid interactions are referred to as environmental TEM (ETEM). Recently, both methods have provided ETEM images showing atomic lattice fringes with a spacing smaller than 0.2–0.3 nm at pressures up to about 10 10 3 bar and temperatures of up to several hundred degrees Celsius [3–7,9]. Specifically, in Ref. [5], the gas layer has a thickness of 5 mm, corresponding to a gas density of 3 103 atoms/nm2 along the beam direction. By means of atomic-resolution ETEM, significant new insights have been obtained into the mechanisms of gas–solid reactions in a variety of nanostructured materials, such as heterogeneous catalysts [3–7,9]. In general, however, caution is needed because the information is obtained at gas pressures of a few millibars, and not at ambient pressures in which many nanomaterials find technological application. It is important to bridge this so-called pressure gap and to clarify how the structural response of the nanomaterials at lower pressures relates to their dynamic behavior at the ambient pressure conditions, because the properties of nanomaterials may depend strongly on their detailed structure and morphology [10,11].
ARTICLE IN PRESS 994
J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
Here we report on a nanoreactor that enables high-resolution ETEM of nanomaterials during exposure to heat and gases at ambient pressure. This was achieved by miniaturization of the gas volume and heater into a microelectromechanical system (MEMS). The performance of the nanoreactor is demonstrated by the in situ imaging of Cu nanocrystals in a catalyst for methanol synthesis. The nanoreactor allows the observation of nanocrystal growth and mobility on a sub-second time scale during heating to 500 1C and exposure to 1.2 bar of H2. In the same environment, ETEM images show atomic lattice fringes in the Cu nanocrystals with spacing that demonstrated a spatial resolution of 0.18 nm.
2. Nanoreactor design A window cell is the preferred configuration for ambient pressure conditions [1]. It uses two electron-transparent windows allowing a much thinner gas layer than in a differentially pumped system. The windows come at the expense, however, of extra electron scattering. The window cell approach was previously used for electron microscopy at 1 bar with an image resolution of several nanometers [1]. Our nanoreactor is designed as a window cell and can reach atomic-scale resolution by using MEMS technology [12–17]. With MEMS, the thickness of the gas column and the windows can be miniaturized. The MEMS approach further enables the integration of a heater and a thermal sensor for variable-temperature experiments at a sub-second time scale. Fig. 1a illustrates the nanoreactor. It consists of two facing dies made with thin-film technology on a silicon substrate. Each die has a central hole of 1 mm2 that is covered by a 1.2 mm thick membrane of SiNx. The opposing membranes form the top and
bottom die
membrane top die
gas inlet
bottom of a shallow gas-flow channel. The minimum height of the channel is 4 mm, determined by disc-shaped spacers integrated in one of the membranes (Fig. 1c). This height corresponds to an atomic density along the beam direction of only 0.2 103 atoms/nm2 at 1 bar and room temperature. Moreover, the spacers prevent stiction of membranes, which would permanently block the channel. To increase the electron transparency, ultrathin windows were formed in recessions in the central part of the membranes (Fig. 1a, c and d). A window consists of a 10 nm thin film of SiNx [18], adding 1.0 103 atoms/nm2 in the beam direction. The SiNx is amorphous to avoid effects of electron diffraction contrast [19]. The windows must withstand a pressure difference of more than 1 bar towards the vacuum of the TEM. Their lateral dimensions are therefore limited to about 10 mm, as determined from plate theory [20,21]. For the same reason they are ellipsoidal in shape. In the areas between the windows, the heater is embedded in the form of a spiralled thin-film Pt wire (see Fig. 1c) [22–24]. The wire is covered on all sides by SiNx and can be resistively heated to 500 1C over a prolonged period of time [25,26]. The heated gas volume is small and well-insulated from the rest of the system. This limits the power consumption and thereby the thermal expansion of the system components, which is very beneficial for low specimen drift. The temperature of the window area can be derived from the local electrical resistance, which is measured through the four electrical connections and has an estimated uncertainty of 10%. The temperature dependency of the resistance was derived from oven experiments and previous work [22]. For operation in a TEM, the nanoreactor is mounted on a custom-made TEM specimen holder (Fig. 1b). Specifically, the nanoreactor is leak-tight interfaced with two exterior tubes to
heater, windows
gas outlet
1 mm
3 mm
100 µm
10 µm
Fig. 1. Illustration of the nanoreactor device. (a) Schematic cross-section of the nanoreactor. (b) Optical image of the TEM holder with the integrated nanoreactor and the four electrical probe contacts. (c) Optical close-up of the nanoreactor membrane. The bright spiral is the Pt heater. The small ovaloids are the electron-transparent windows. The circles are the SiO2 spacers that define the minimum height of the gas channel. (d) A low-magnification TEM image of a pair of superimposed 10 nm thick windows. Their alignment creates a highly electron-transparent (bright) square through which high-resolution TEM imaging can be performed.
ARTICLE IN PRESS J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
introduce a flowing gas phase. Four tungsten probe needles make electrical connections to the heater and thermal sensor. The tip of the holder has a thickness of 3.56 mm.
3. ETEM experiments using the nanoreactor The performance of the nanoreactor was tested and demonstrated by imaging the formation and the structure of Cu nanocrystals on a ZnO support. This system is commonly used as catalyst for methanol synthesis and for conversion of hydrocarbons in fuel cells. Also, it is a prototype example of the industrially important group of 3d transition metal catalysts. These are particularly challenging to image at the atomic scale, because they require a resolution below 0.2 nm. This resolution limit was previously demonstrated with ETEM at about 1 10 3 bar pressure using the differentially pumped vacuum system [27]. The Cu/ZnO catalyst was prepared by dry mixing ZnO nanopowder (Advanced Nanotechnology Limited, Australia) with copper acetylacetonate. The precursor was decomposed by calcination in air at 200 1C and organic residues were removed by subsequent calcination in air at 400 1C. The preparation resulted in a catalyst precursor consisting of mainly CuO dispersed on ZnO nanocrystallites, as qualitatively indicated by electron energy loss spectra of the near-edge structure at the Cu L2,3 ionization edge. The nanoreactors were loaded with the catalyst precursor after their assembly by flushing with a suspension of catalyst powder in ethanol, followed by ethanol evaporation at room temperature. This resulted in catalyst grains being deposited on the windows as shown in Fig. 2. The ETEM experiments were carried out using a Philips CM300ST FEG equipped with a differentially pumped environmental cell [5,27]. This microscope was used because of its proven gas supply system, which was connected to the sample holder instead of the environmental cell. The environmental cell itself was kept evacuated. Later experiments confirmed that the pressurized nanoreactors could also be operated in a regular Philips/FEI CM300-T. The information limit of the ETEM is below 0.14 nm. A Tietz FastScan-F114 CCD was used for recording TEM images and was operated with exposure times in the interval from 10 to 1 frame/s. A Gatan Image Filter (GIF-2000) was used for acquisition of electron energy loss (EEL) spectra. EEL spectra were recorded in image mode with a dispersion of 0.1 eV/pixel, covering an energy interval of about 100 eV, and with a spectral resolution of 1.5–2.0 eV corresponding to the full-width at half-maximum of the zero-loss peak. All spectra included the zero-loss peak for energy calibration. Prior to use, the nanoreactor and gas tubes connecting the nanoreactor to the feed gas bottle were evacuated by a turbo pump to a base pressure of at least 1–3 10 5 bar as measured by a MicroPirani gauge (MKS 910 DualTrans Transducer). During exposure to hydrogen (nominal purity 99.9999%, Air Liquide Alphagaz 2 H2), the pressure at the nanoreactor in- and outlets was kept at the same level and measured by a Piezo sensor (MKS 910 DualTrans Transducer) mounted on the gas supply tubes at the inlet. The calibration of the Piezo sensor was checked by exposure to ambient air. After insertion and evacuation of a nanoreactor in the transmission electron microscope, the nanoreactor was exposed to H2. Electron energy loss spectroscopy (EELS) was performed at the hydrogen K-ionization edge. For the gas pressure exceeding 0.4 bar, EELS shows a distinct hydrogen K-edge peak that appears above the plasmon ridge of SiNx confirming the presence of H2 (Fig. 3a). The gas pressure also caused the membranes to bulge, thereby increasing the height of the gas channel. The increased height was evaluated by consecutive focussing on catalyst grains on opposite windows and determining the difference between the
995
focal distances (Fig. 2c). At 1.2 bar the channel height was 35 mm, in agreement with analytical estimates [20,21]. The height increase may be reduced by bonding the spacers to the top membrane. However, with a height of 35 mm, the density of scatters is only 2 103 atoms/nm2. Even with this increase, the total density of gas and window atoms was limited to 4 103 atoms/nm2. This is comparable to the density of 3 103 atoms/nm2 that previously allowed atomic-resolution TEM at much lower gas pressures [5]. The nanoreactor should therefore provide similar resolving power, but at ambient pressure.
4. Atomic-resolution ETEM of Cu nanocrystals at ambient pressure conditions To transform the CuO precursor into the catalytically active metallic Cu nanocrystallites, the catalyst was heated in the H2 atmosphere to the maximum operation temperature of 500 1C. Specifically, the temperature was raised rapidly from room temperature to 500 1C within 68.5 s in eight quasi-instantaneous steps (see Fig. 3b). Meanwhile, a time-lapsed image series was recorded with a frame rate of 6.9 frames/s (Fig. 2b). As can be seen in Fig. 2, the ZnO crystallites appeared within the precursor with facetted, compact shapes with diameters of 20–100 nm (Fig. 2a). The CuO appeared as smaller patches of more irregular shapes at the edges of the ZnO. As the temperature was increased above to about 260 1C, the CuO patches broke up into several particles with diameters between 5 and 10 nm (Fig. 2b–e). The darker contrast and compact shape of these particles suggest that they consist of metallic Cu. This was confirmed by HRTEM images. Surprisingly, some nanocrystals migrated immediately after their formation, but the nanocrystals remained immobile at later stages (Fig. 2c,d). This could reflect that Cu strongly pins at sites that are nonuniformly distributed over the ZnO surfaces [28]. After growth, inspection of neighboring windows revealed Cu nanocrystals of similar size and shape, indicating that the ensemble of nanocrystals was not noticeably influenced by the electron beam. Nanocrystals were also similar on opposite top and bottom windows, indicating sufficient temperature uniformity. The images in Fig. 2 show that the position of the catalyst grains inside the nanoreactor was remarkably stable. The drift that remained was quantified by measuring the specimen displacements between consecutive images [29]. The analysis shows two distinct patterns (Fig. 3b). At constant temperatures, the drift rate was below one pixel (0.63 nm) per frame (145 ms). Followed over a longer time scale, drift built up a visible displacement in one specific direction. This drift, however, decayed exponentially with a time constant of 137 s and with an initial rate (at 500 1C) of only 3.3 nm/s. This was sufficiently slow to be compensated for by tracking an area of interest. When the temperature was stepwise increased, the drift rate immediately responded with a pulse that was shorter than the exposure time of one frame. This corresponded to a small, quasi-instantaneous displacement of 0.40 nm/1C, which could be compensated for by tracking as well. The low drift rates and the rapid equilibration confirmed that heat was generated very locally and that the surrounding Si dies were well-insulated. This was further supported by the power consumption, which was only 30 mW at 500 1C. The state of the nanocrystals was inferred from atomicresolution ETEM images. Fig. 4 shows a close-up of the catalyst during exposure to 1.2 bar H2 at 500 1C. The image shows clear atomic lattice fringes in both the brighter ZnO support crystallites and the darker Cu nanocrystals (Fig. 4a). A Fourier transform of the image provided an unambiguous identification of the lattice spacings and thus of the materials (Fig. 4b). Specifically, lattice
ARTICLE IN PRESS 996
J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
Fig. 2. Image sequences of the Cu nanocrystal growth and mobility on ZnO. Nanocrystals (darker contrast) form from CuO precursors (blue arrows) during heating from RT to 500 1C in 1.2 bar H2 (see Fig. 3b). After growth, nanocrystals can exhibit transient mobility (white square). Crystallites on the opposite window are seen out of focus (black arrows in (a) and (c)). The frames are recorded at (a) RT, (b) 260 1C, (c) 330 1C, (d) 365 1C, (e) 410 1C, and (f) 500 1C. All frames are averaged over four consecutive images. The exposure time for each image is 0.145 s.
fringes with spacings of 0.21 and 0.18 nm could be recorded in the nanocrystals, corresponding to the (111) and (2 0 0) planes of Cu, respectively. This means that, even with the present reaction
environment, the image resolution was better than 0.18 nm and close to the resolution limit of 0.14 nm of the host TEM instrument [5,27]. The image also shows a faint low-frequency contrast of the
ARTICLE IN PRESS J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
997
Intensity (scaled, arbitrary units)
1.22 bar H2 0.73 bar H2 0.41 bar H2 vacuum
0 0
5
10 15 20 Electron energy loss (eV)
25
30
600 500
1,000
400 100 300 10 200 1
Temperature (°C)
Image displacement (nm/frame)
10,000
100
0.1
0 0
20
40 Time (s)
60
80
Fig. 3. Nanoreactor performances. (a) EEL spectra at different H2 pressures. The evacuated nanoreactor displays a SiNx plasmon peak with maximum at 23.4 eV [30]. The hydrogen K ionization-edge at 12.8 eV [30] appears distinctly above 0.4 bar. All spectra are normalized by scaling the intensity at 23.4 eV to match the evacuated reactor. (b) Thermal drift velocity (blue points) as a function of time during heating from RT to 501 1C (red curve). The drift velocity is determined as the displacement rate between consecutive images using automated crosscorrelation [29]. The manual corrections for accumulated displacement appear as artificial drift (light-blue points).
amorphous windows. Similar lattice-resolved images were recorded of several different nanocrystals imaged in two different nanoreactors.
5. Conclusions The present work demonstrates a MEMS-based nanoreactor for environmental transmission electron microscopy with high spatiotemporal resolution of nanomaterials in ambient environments, representative for their functioning state. The atomic-scale insight can be crucial for understanding how nanomaterial properties are affected by their working conditions. The concept of the nanoreactor should be versatile enough to be extended towards other environments, including higher pressures, heavier constituents, and even liquids. Moreover, the nanoreactor can be beneficially combined with other characterization techniques, including mass spectrometry, optical, and X-ray techniques. The proposed nanoreactor opens up new possibilities for in situ studies of functional nanomaterials and the ways they interact with ambient environments, and it is generally applicable to a
Fig. 4. A representative HRTEM image of the Cu/ZnO catalyst during exposure to 1.2 bar hydrogen at 500 1C. (a) The image displays lattice fringes of a twinned Cu nanocrystal and of the ZnO support. (b) A Fourier transform of a. The bright dots represent sets of lattice fringes. Their lattice spacing corresponds to the distance to the origin and reveals the crystallographic identity. The large, red circle corresponds a spacing of 0.21 nm. The smallest, resolved lattice spacing is 0.18 nm.
variety of fields, such as heterogeneous catalysis, electrochemistry, nanofabrication, materials science and biology.
Acknowledgments This work is supported by STW, Applied Science Foundation of NWO and the Ministry of Economic Affairs with financial
ARTICLE IN PRESS 998
J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
contributions from FEI Company, Haldor Topsøe A/S, and the European ESTEEM project. We thank the DIMES Technology Centre, the Nanofacility, IMT Neuchaˆtel and D. Briand for support in the cleanroom fabrication, F. Tichelaar, U. Ziese, and P. Kooyman for TEM characterization. FEI Company, and especially M. Stekelenburg, we thank for valuable discussions and technical input. Haldor Topsøe A/S acknowledges CTCI Foundation, Taiwan for participation in establishment of its ETEM facility. References [1] E.P. Butler, K.F. Hale, Dynamic Experiments in the Electron Microscope, Practical Methods in Electron Microscopy, vol. 9, North-Holland, Amsterdam, 1981. [2] R.T.K. Baker, P.S. Harris, J. Phys. E 5 (1972) 793. [3] E.D. Boyes, P.L. Gai, Ultramicroscopy 67 (1997) 219. [4] R. Sharma, P.A. Crozier, Environmental Transmission Electron Microscopy in Nanotechnology, Microscopy in Nanotechnology, Kluwer Academic, New York, 2005. [5] P.L. Hansen, S. Helveg, A.K. Datye, Adv. Catal. 50 (2006) 77. [6] G.M. Parkinson, Catal. Lett. 2 (1989) 303. [7] S. Giorgio, S. Sao Joao, S. Nitsche, D. Chaudanson, G. Sitja, C.R. Henry, Ultramicroscopy 106 (2006) 503. [8] T.L. Daulton, B.J. Little, J.W. Kim, S. Newell, K. Lowe, Y. Furukawa, J. JonesMeehan, D.L. Lavoie, JEOL News 37E (2002) 6. [9] S. Helveg, C. Lo´pez-Cartes, J. Sehested, P.L. Hansen, B.S. Clausen, J.R. RostrupNielsen, F. Abild-Pedersen, J.K. Nørskov, Nature 427 (2004) 426. [10] N.I. Jaeger, Science 293 (2001) 1601. [11] H. Topsøe, J. Catal. 216 (2003) 155.
[12] M.J. Williamson, R.M. Tromp, P.M. Vereecken, R. Hull, F.M. Ross, Nat. Mater. 2 (2003) 532. [13] E.D. Green, G.S. Kino, J. Vac. Sci. Technol. B 9 (1991) 1557. [14] M. Zhang, E.A. Olson, R.D. Twesten, J.G. Wen, L.H. Allen, I.M. Robertson, I. Petrov, J. Mater. Res. 20 (2005) 1802. [15] E.A. Olson, M.Y. Efremov, M. Zhang, Z. Zhang, L.H. Allen, J. Microelectromech. Syst. 12 (2003) 355. [16] L.A. Arana, S.B. Schaevitz, A.J. Franz, M.A. Schmidt, K.F. Jensen, J. Microelectromech. Syst. 12 (2003) 600. [17] R.M. Tiggelaar, P. van Male, J.W. Berenschot, J.G.E. Gardeniers, R.E. Oosterbroek, M.H.J.M. de Croon, J.C. Schouten, A. van den Berg, M.C. Elwenspoek, Sensors Actuators A 119 (2005) 196. [18] P.J. French, P.M. Sarro, R. Malle´e, E.J.M. Fakkeldij, R.F. Wolffenbuttel, Sensors Actuators A 58 (1997) 149. [19] T. Doll, M. Hochberg, D. Barsic, A. Scherer, Sensors Actuators A 87 (2000) 52. [20] J. Yang, O. Paul, Sensors Actuators A 97–98 (2002) 520. [21] C. van Rijn, M. van der Wekken, W. Nijdam, M. Elwenspoek, J. Microelectromech. Syst. 6 (1997) 48. [22] D. Briand, A. Krauss, B. van der Schoot, U. Weimar, N. Barsan, W. Go¨pel, N.F. de Rooij, Sensors Actuators B 68 (2000) 223. [23] A. Hierlemann, O. Brand, C. Hagleitner, H. Baltes, Proc. IEEE 91 (2003) 839. [24] F. Solzbacher, C. Imawan, H. Steffels, E. Obermeier, M. Eickhoff, Sensors Actuators B 77 (2001) 111. [25] D. Briand, F. Beaudoin, J. Courbat, N.F. de Rooij, R. Desplats, P. Perdu, Microelectron. Reliab. 45 (2005) 1786. [26] S.L. Firebaugh, K.F. Jensen, M.A. Schmidt, J. Microelectromech. Syst. 7 (1998) 128. [27] P.L. Hansen, J.B. Wagner, S. Helveg, J.R. Rostrup-Nielsen, B.S. Clausen, H. Topsøe, Science 295 (2002) 2053. [28] O. Dulub, L.A. Boatner, U. Diebold, Surf. Sci. 504 (2002) 271. [29] B. Schaffer, W. Grogger, G. Kothleitner, Ultramicroscopy 102 (2004) 27. [30] R.F. Egerton, Electron Energy-Loss Spectroscopy in the Electron Microscope, second ed., Plenum Press, New York, 1996.
Contents lists available at ScienceDirect
Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic
Atomic-scale electron microscopy at ambient pressure J.F. Creemer a,, S. Helveg b, G.H. Hoveling c, S. Ullmann b, A.M. Molenbroek b, P.M. Sarro a, H.W. Zandbergen d a
DIMES-ECTM, Delft University of Technology, P.O. Box 5053, 2600 GB Delft, The Netherlands Haldor Topsøe A/S, Nymøllevej 55, DK-2800 Kgs. Lyngby, Denmark DEMO, Delft University of Technology, P.O. Box 5031, 2600 GA Delft, The Netherlands d Kavli Institute of NanoScience, HREM, Delft University of Technology, P.O. Box 5046, 2600 GA Delft, The Netherlands b c
a r t i c l e in fo
abstract
Article history: Received 27 January 2008 Received in revised form 31 March 2008 Accepted 15 April 2008
We demonstrate a novel nanoreactor for performing atomic-resolution environmental transmission electron microscopy (ETEM) of nanostructured materials during exposure to gases at ambient pressures and elevated temperatures. The nanoreactor is a microelectromechanical system (MEMS) and is functionalized with a micrometer-sized gas-flow channel, electron-transparent windows and a heating device. It fits into the tip of a dedicated sample holder that can be used in a normal CM microscope of Philips/FEI Company. The nanoreactor performance was demonstrated by ETEM imaging of a Cu/ZnO catalyst for methanol synthesis during exposure to hydrogen. Specifically, the nanoreactor facilitated the direct observation of Cu nanocrystal growth and mobility on a sub-second time scale during heating to 500 1C and exposure to 1.2 bar of H2. For the same gas reaction environment, ETEM images show atomic lattice fringes in the Cu nanocrystals with spacing of 0.18 nm, attesting the spatial resolution limit of the system. The nanoreactor concept opens up new possibilities for in situ studies of nanomaterials and the ways they interact with their ambient working environment in diverse areas, such as heterogeneous catalysis, electrochemistry, nanofabrication, materials science and biology. & 2008 Elsevier B.V. All rights reserved.
PACS: 68.37.Og 85.85.+j 82.65.+r 78.67.Bf Keywords: Environmental TEM ETEM Microelectromechanical systems MEMS Gas–solid interactions Nanocrystals Methanol synthesis catalyst
1. Introduction It has been a longstanding challenge in transmission electron microscopy (TEM) to study chemical processes at gas–solid interfaces in situ with atomic-scale resolution [1–6]. Highresolution TEM (HRTEM) is a powerful technique for atomic-scale imaging of nanomaterials that are kept under high vacuum conditions (Ca. 10 9 bar). Applying HRTEM for studies of gas–solid interactions is extremely demanding, because the HRTEM image resolution is degraded by scattering of the electron beam on gas atoms and by specimen drift due to heating. A key requirement is therefore to limit the number of gas atoms in the microscope, which can be reached by limiting the pressure or by confining the gas along the path of the beam into a layer that is as thin as possible. Confinement has been attained by two methods: by differentially pumped vacuum systems [1–5] and by windowed
Corresponding author. Tel.: +3115 278 6277; fax: +3115 262 2163.
E-mail address: [email protected] (J.F. Creemer). 0304-3991/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ultramic.2008.04.014
cells [1,6–8]. These capabilities for performing in situ studies of gas–solid interactions are referred to as environmental TEM (ETEM). Recently, both methods have provided ETEM images showing atomic lattice fringes with a spacing smaller than 0.2–0.3 nm at pressures up to about 10 10 3 bar and temperatures of up to several hundred degrees Celsius [3–7,9]. Specifically, in Ref. [5], the gas layer has a thickness of 5 mm, corresponding to a gas density of 3 103 atoms/nm2 along the beam direction. By means of atomic-resolution ETEM, significant new insights have been obtained into the mechanisms of gas–solid reactions in a variety of nanostructured materials, such as heterogeneous catalysts [3–7,9]. In general, however, caution is needed because the information is obtained at gas pressures of a few millibars, and not at ambient pressures in which many nanomaterials find technological application. It is important to bridge this so-called pressure gap and to clarify how the structural response of the nanomaterials at lower pressures relates to their dynamic behavior at the ambient pressure conditions, because the properties of nanomaterials may depend strongly on their detailed structure and morphology [10,11].
ARTICLE IN PRESS 994
J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
Here we report on a nanoreactor that enables high-resolution ETEM of nanomaterials during exposure to heat and gases at ambient pressure. This was achieved by miniaturization of the gas volume and heater into a microelectromechanical system (MEMS). The performance of the nanoreactor is demonstrated by the in situ imaging of Cu nanocrystals in a catalyst for methanol synthesis. The nanoreactor allows the observation of nanocrystal growth and mobility on a sub-second time scale during heating to 500 1C and exposure to 1.2 bar of H2. In the same environment, ETEM images show atomic lattice fringes in the Cu nanocrystals with spacing that demonstrated a spatial resolution of 0.18 nm.
2. Nanoreactor design A window cell is the preferred configuration for ambient pressure conditions [1]. It uses two electron-transparent windows allowing a much thinner gas layer than in a differentially pumped system. The windows come at the expense, however, of extra electron scattering. The window cell approach was previously used for electron microscopy at 1 bar with an image resolution of several nanometers [1]. Our nanoreactor is designed as a window cell and can reach atomic-scale resolution by using MEMS technology [12–17]. With MEMS, the thickness of the gas column and the windows can be miniaturized. The MEMS approach further enables the integration of a heater and a thermal sensor for variable-temperature experiments at a sub-second time scale. Fig. 1a illustrates the nanoreactor. It consists of two facing dies made with thin-film technology on a silicon substrate. Each die has a central hole of 1 mm2 that is covered by a 1.2 mm thick membrane of SiNx. The opposing membranes form the top and
bottom die
membrane top die
gas inlet
bottom of a shallow gas-flow channel. The minimum height of the channel is 4 mm, determined by disc-shaped spacers integrated in one of the membranes (Fig. 1c). This height corresponds to an atomic density along the beam direction of only 0.2 103 atoms/nm2 at 1 bar and room temperature. Moreover, the spacers prevent stiction of membranes, which would permanently block the channel. To increase the electron transparency, ultrathin windows were formed in recessions in the central part of the membranes (Fig. 1a, c and d). A window consists of a 10 nm thin film of SiNx [18], adding 1.0 103 atoms/nm2 in the beam direction. The SiNx is amorphous to avoid effects of electron diffraction contrast [19]. The windows must withstand a pressure difference of more than 1 bar towards the vacuum of the TEM. Their lateral dimensions are therefore limited to about 10 mm, as determined from plate theory [20,21]. For the same reason they are ellipsoidal in shape. In the areas between the windows, the heater is embedded in the form of a spiralled thin-film Pt wire (see Fig. 1c) [22–24]. The wire is covered on all sides by SiNx and can be resistively heated to 500 1C over a prolonged period of time [25,26]. The heated gas volume is small and well-insulated from the rest of the system. This limits the power consumption and thereby the thermal expansion of the system components, which is very beneficial for low specimen drift. The temperature of the window area can be derived from the local electrical resistance, which is measured through the four electrical connections and has an estimated uncertainty of 10%. The temperature dependency of the resistance was derived from oven experiments and previous work [22]. For operation in a TEM, the nanoreactor is mounted on a custom-made TEM specimen holder (Fig. 1b). Specifically, the nanoreactor is leak-tight interfaced with two exterior tubes to
heater, windows
gas outlet
1 mm
3 mm
100 µm
10 µm
Fig. 1. Illustration of the nanoreactor device. (a) Schematic cross-section of the nanoreactor. (b) Optical image of the TEM holder with the integrated nanoreactor and the four electrical probe contacts. (c) Optical close-up of the nanoreactor membrane. The bright spiral is the Pt heater. The small ovaloids are the electron-transparent windows. The circles are the SiO2 spacers that define the minimum height of the gas channel. (d) A low-magnification TEM image of a pair of superimposed 10 nm thick windows. Their alignment creates a highly electron-transparent (bright) square through which high-resolution TEM imaging can be performed.
ARTICLE IN PRESS J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
introduce a flowing gas phase. Four tungsten probe needles make electrical connections to the heater and thermal sensor. The tip of the holder has a thickness of 3.56 mm.
3. ETEM experiments using the nanoreactor The performance of the nanoreactor was tested and demonstrated by imaging the formation and the structure of Cu nanocrystals on a ZnO support. This system is commonly used as catalyst for methanol synthesis and for conversion of hydrocarbons in fuel cells. Also, it is a prototype example of the industrially important group of 3d transition metal catalysts. These are particularly challenging to image at the atomic scale, because they require a resolution below 0.2 nm. This resolution limit was previously demonstrated with ETEM at about 1 10 3 bar pressure using the differentially pumped vacuum system [27]. The Cu/ZnO catalyst was prepared by dry mixing ZnO nanopowder (Advanced Nanotechnology Limited, Australia) with copper acetylacetonate. The precursor was decomposed by calcination in air at 200 1C and organic residues were removed by subsequent calcination in air at 400 1C. The preparation resulted in a catalyst precursor consisting of mainly CuO dispersed on ZnO nanocrystallites, as qualitatively indicated by electron energy loss spectra of the near-edge structure at the Cu L2,3 ionization edge. The nanoreactors were loaded with the catalyst precursor after their assembly by flushing with a suspension of catalyst powder in ethanol, followed by ethanol evaporation at room temperature. This resulted in catalyst grains being deposited on the windows as shown in Fig. 2. The ETEM experiments were carried out using a Philips CM300ST FEG equipped with a differentially pumped environmental cell [5,27]. This microscope was used because of its proven gas supply system, which was connected to the sample holder instead of the environmental cell. The environmental cell itself was kept evacuated. Later experiments confirmed that the pressurized nanoreactors could also be operated in a regular Philips/FEI CM300-T. The information limit of the ETEM is below 0.14 nm. A Tietz FastScan-F114 CCD was used for recording TEM images and was operated with exposure times in the interval from 10 to 1 frame/s. A Gatan Image Filter (GIF-2000) was used for acquisition of electron energy loss (EEL) spectra. EEL spectra were recorded in image mode with a dispersion of 0.1 eV/pixel, covering an energy interval of about 100 eV, and with a spectral resolution of 1.5–2.0 eV corresponding to the full-width at half-maximum of the zero-loss peak. All spectra included the zero-loss peak for energy calibration. Prior to use, the nanoreactor and gas tubes connecting the nanoreactor to the feed gas bottle were evacuated by a turbo pump to a base pressure of at least 1–3 10 5 bar as measured by a MicroPirani gauge (MKS 910 DualTrans Transducer). During exposure to hydrogen (nominal purity 99.9999%, Air Liquide Alphagaz 2 H2), the pressure at the nanoreactor in- and outlets was kept at the same level and measured by a Piezo sensor (MKS 910 DualTrans Transducer) mounted on the gas supply tubes at the inlet. The calibration of the Piezo sensor was checked by exposure to ambient air. After insertion and evacuation of a nanoreactor in the transmission electron microscope, the nanoreactor was exposed to H2. Electron energy loss spectroscopy (EELS) was performed at the hydrogen K-ionization edge. For the gas pressure exceeding 0.4 bar, EELS shows a distinct hydrogen K-edge peak that appears above the plasmon ridge of SiNx confirming the presence of H2 (Fig. 3a). The gas pressure also caused the membranes to bulge, thereby increasing the height of the gas channel. The increased height was evaluated by consecutive focussing on catalyst grains on opposite windows and determining the difference between the
995
focal distances (Fig. 2c). At 1.2 bar the channel height was 35 mm, in agreement with analytical estimates [20,21]. The height increase may be reduced by bonding the spacers to the top membrane. However, with a height of 35 mm, the density of scatters is only 2 103 atoms/nm2. Even with this increase, the total density of gas and window atoms was limited to 4 103 atoms/nm2. This is comparable to the density of 3 103 atoms/nm2 that previously allowed atomic-resolution TEM at much lower gas pressures [5]. The nanoreactor should therefore provide similar resolving power, but at ambient pressure.
4. Atomic-resolution ETEM of Cu nanocrystals at ambient pressure conditions To transform the CuO precursor into the catalytically active metallic Cu nanocrystallites, the catalyst was heated in the H2 atmosphere to the maximum operation temperature of 500 1C. Specifically, the temperature was raised rapidly from room temperature to 500 1C within 68.5 s in eight quasi-instantaneous steps (see Fig. 3b). Meanwhile, a time-lapsed image series was recorded with a frame rate of 6.9 frames/s (Fig. 2b). As can be seen in Fig. 2, the ZnO crystallites appeared within the precursor with facetted, compact shapes with diameters of 20–100 nm (Fig. 2a). The CuO appeared as smaller patches of more irregular shapes at the edges of the ZnO. As the temperature was increased above to about 260 1C, the CuO patches broke up into several particles with diameters between 5 and 10 nm (Fig. 2b–e). The darker contrast and compact shape of these particles suggest that they consist of metallic Cu. This was confirmed by HRTEM images. Surprisingly, some nanocrystals migrated immediately after their formation, but the nanocrystals remained immobile at later stages (Fig. 2c,d). This could reflect that Cu strongly pins at sites that are nonuniformly distributed over the ZnO surfaces [28]. After growth, inspection of neighboring windows revealed Cu nanocrystals of similar size and shape, indicating that the ensemble of nanocrystals was not noticeably influenced by the electron beam. Nanocrystals were also similar on opposite top and bottom windows, indicating sufficient temperature uniformity. The images in Fig. 2 show that the position of the catalyst grains inside the nanoreactor was remarkably stable. The drift that remained was quantified by measuring the specimen displacements between consecutive images [29]. The analysis shows two distinct patterns (Fig. 3b). At constant temperatures, the drift rate was below one pixel (0.63 nm) per frame (145 ms). Followed over a longer time scale, drift built up a visible displacement in one specific direction. This drift, however, decayed exponentially with a time constant of 137 s and with an initial rate (at 500 1C) of only 3.3 nm/s. This was sufficiently slow to be compensated for by tracking an area of interest. When the temperature was stepwise increased, the drift rate immediately responded with a pulse that was shorter than the exposure time of one frame. This corresponded to a small, quasi-instantaneous displacement of 0.40 nm/1C, which could be compensated for by tracking as well. The low drift rates and the rapid equilibration confirmed that heat was generated very locally and that the surrounding Si dies were well-insulated. This was further supported by the power consumption, which was only 30 mW at 500 1C. The state of the nanocrystals was inferred from atomicresolution ETEM images. Fig. 4 shows a close-up of the catalyst during exposure to 1.2 bar H2 at 500 1C. The image shows clear atomic lattice fringes in both the brighter ZnO support crystallites and the darker Cu nanocrystals (Fig. 4a). A Fourier transform of the image provided an unambiguous identification of the lattice spacings and thus of the materials (Fig. 4b). Specifically, lattice
ARTICLE IN PRESS 996
J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
Fig. 2. Image sequences of the Cu nanocrystal growth and mobility on ZnO. Nanocrystals (darker contrast) form from CuO precursors (blue arrows) during heating from RT to 500 1C in 1.2 bar H2 (see Fig. 3b). After growth, nanocrystals can exhibit transient mobility (white square). Crystallites on the opposite window are seen out of focus (black arrows in (a) and (c)). The frames are recorded at (a) RT, (b) 260 1C, (c) 330 1C, (d) 365 1C, (e) 410 1C, and (f) 500 1C. All frames are averaged over four consecutive images. The exposure time for each image is 0.145 s.
fringes with spacings of 0.21 and 0.18 nm could be recorded in the nanocrystals, corresponding to the (111) and (2 0 0) planes of Cu, respectively. This means that, even with the present reaction
environment, the image resolution was better than 0.18 nm and close to the resolution limit of 0.14 nm of the host TEM instrument [5,27]. The image also shows a faint low-frequency contrast of the
ARTICLE IN PRESS J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
997
Intensity (scaled, arbitrary units)
1.22 bar H2 0.73 bar H2 0.41 bar H2 vacuum
0 0
5
10 15 20 Electron energy loss (eV)
25
30
600 500
1,000
400 100 300 10 200 1
Temperature (°C)
Image displacement (nm/frame)
10,000
100
0.1
0 0
20
40 Time (s)
60
80
Fig. 3. Nanoreactor performances. (a) EEL spectra at different H2 pressures. The evacuated nanoreactor displays a SiNx plasmon peak with maximum at 23.4 eV [30]. The hydrogen K ionization-edge at 12.8 eV [30] appears distinctly above 0.4 bar. All spectra are normalized by scaling the intensity at 23.4 eV to match the evacuated reactor. (b) Thermal drift velocity (blue points) as a function of time during heating from RT to 501 1C (red curve). The drift velocity is determined as the displacement rate between consecutive images using automated crosscorrelation [29]. The manual corrections for accumulated displacement appear as artificial drift (light-blue points).
amorphous windows. Similar lattice-resolved images were recorded of several different nanocrystals imaged in two different nanoreactors.
5. Conclusions The present work demonstrates a MEMS-based nanoreactor for environmental transmission electron microscopy with high spatiotemporal resolution of nanomaterials in ambient environments, representative for their functioning state. The atomic-scale insight can be crucial for understanding how nanomaterial properties are affected by their working conditions. The concept of the nanoreactor should be versatile enough to be extended towards other environments, including higher pressures, heavier constituents, and even liquids. Moreover, the nanoreactor can be beneficially combined with other characterization techniques, including mass spectrometry, optical, and X-ray techniques. The proposed nanoreactor opens up new possibilities for in situ studies of functional nanomaterials and the ways they interact with ambient environments, and it is generally applicable to a
Fig. 4. A representative HRTEM image of the Cu/ZnO catalyst during exposure to 1.2 bar hydrogen at 500 1C. (a) The image displays lattice fringes of a twinned Cu nanocrystal and of the ZnO support. (b) A Fourier transform of a. The bright dots represent sets of lattice fringes. Their lattice spacing corresponds to the distance to the origin and reveals the crystallographic identity. The large, red circle corresponds a spacing of 0.21 nm. The smallest, resolved lattice spacing is 0.18 nm.
variety of fields, such as heterogeneous catalysis, electrochemistry, nanofabrication, materials science and biology.
Acknowledgments This work is supported by STW, Applied Science Foundation of NWO and the Ministry of Economic Affairs with financial
ARTICLE IN PRESS 998
J.F. Creemer et al. / Ultramicroscopy 108 (2008) 993–998
contributions from FEI Company, Haldor Topsøe A/S, and the European ESTEEM project. We thank the DIMES Technology Centre, the Nanofacility, IMT Neuchaˆtel and D. Briand for support in the cleanroom fabrication, F. Tichelaar, U. Ziese, and P. Kooyman for TEM characterization. FEI Company, and especially M. Stekelenburg, we thank for valuable discussions and technical input. Haldor Topsøe A/S acknowledges CTCI Foundation, Taiwan for participation in establishment of its ETEM facility. References [1] E.P. Butler, K.F. Hale, Dynamic Experiments in the Electron Microscope, Practical Methods in Electron Microscopy, vol. 9, North-Holland, Amsterdam, 1981. [2] R.T.K. Baker, P.S. Harris, J. Phys. E 5 (1972) 793. [3] E.D. Boyes, P.L. Gai, Ultramicroscopy 67 (1997) 219. [4] R. Sharma, P.A. Crozier, Environmental Transmission Electron Microscopy in Nanotechnology, Microscopy in Nanotechnology, Kluwer Academic, New York, 2005. [5] P.L. Hansen, S. Helveg, A.K. Datye, Adv. Catal. 50 (2006) 77. [6] G.M. Parkinson, Catal. Lett. 2 (1989) 303. [7] S. Giorgio, S. Sao Joao, S. Nitsche, D. Chaudanson, G. Sitja, C.R. Henry, Ultramicroscopy 106 (2006) 503. [8] T.L. Daulton, B.J. Little, J.W. Kim, S. Newell, K. Lowe, Y. Furukawa, J. JonesMeehan, D.L. Lavoie, JEOL News 37E (2002) 6. [9] S. Helveg, C. Lo´pez-Cartes, J. Sehested, P.L. Hansen, B.S. Clausen, J.R. RostrupNielsen, F. Abild-Pedersen, J.K. Nørskov, Nature 427 (2004) 426. [10] N.I. Jaeger, Science 293 (2001) 1601. [11] H. Topsøe, J. Catal. 216 (2003) 155.
[12] M.J. Williamson, R.M. Tromp, P.M. Vereecken, R. Hull, F.M. Ross, Nat. Mater. 2 (2003) 532. [13] E.D. Green, G.S. Kino, J. Vac. Sci. Technol. B 9 (1991) 1557. [14] M. Zhang, E.A. Olson, R.D. Twesten, J.G. Wen, L.H. Allen, I.M. Robertson, I. Petrov, J. Mater. Res. 20 (2005) 1802. [15] E.A. Olson, M.Y. Efremov, M. Zhang, Z. Zhang, L.H. Allen, J. Microelectromech. Syst. 12 (2003) 355. [16] L.A. Arana, S.B. Schaevitz, A.J. Franz, M.A. Schmidt, K.F. Jensen, J. Microelectromech. Syst. 12 (2003) 600. [17] R.M. Tiggelaar, P. van Male, J.W. Berenschot, J.G.E. Gardeniers, R.E. Oosterbroek, M.H.J.M. de Croon, J.C. Schouten, A. van den Berg, M.C. Elwenspoek, Sensors Actuators A 119 (2005) 196. [18] P.J. French, P.M. Sarro, R. Malle´e, E.J.M. Fakkeldij, R.F. Wolffenbuttel, Sensors Actuators A 58 (1997) 149. [19] T. Doll, M. Hochberg, D. Barsic, A. Scherer, Sensors Actuators A 87 (2000) 52. [20] J. Yang, O. Paul, Sensors Actuators A 97–98 (2002) 520. [21] C. van Rijn, M. van der Wekken, W. Nijdam, M. Elwenspoek, J. Microelectromech. Syst. 6 (1997) 48. [22] D. Briand, A. Krauss, B. van der Schoot, U. Weimar, N. Barsan, W. Go¨pel, N.F. de Rooij, Sensors Actuators B 68 (2000) 223. [23] A. Hierlemann, O. Brand, C. Hagleitner, H. Baltes, Proc. IEEE 91 (2003) 839. [24] F. Solzbacher, C. Imawan, H. Steffels, E. Obermeier, M. Eickhoff, Sensors Actuators B 77 (2001) 111. [25] D. Briand, F. Beaudoin, J. Courbat, N.F. de Rooij, R. Desplats, P. Perdu, Microelectron. Reliab. 45 (2005) 1786. [26] S.L. Firebaugh, K.F. Jensen, M.A. Schmidt, J. Microelectromech. Syst. 7 (1998) 128. [27] P.L. Hansen, J.B. Wagner, S. Helveg, J.R. Rostrup-Nielsen, B.S. Clausen, H. Topsøe, Science 295 (2002) 2053. [28] O. Dulub, L.A. Boatner, U. Diebold, Surf. Sci. 504 (2002) 271. [29] B. Schaffer, W. Grogger, G. Kothleitner, Ultramicroscopy 102 (2004) 27. [30] R.F. Egerton, Electron Energy-Loss Spectroscopy in the Electron Microscope, second ed., Plenum Press, New York, 1996.