- Email: [email protected]
PEEM study of work function changes in Cu, Au and Pd metal surfaces with surface acoustic wave propagation
Surface Science 600 (2006) 2644–2649 www.elsevier.com/locate/susc
PEEM study of work function changes in Cu, Au and Pd metal surfaces with surface acoustic wave propagation Hiroshi Nishiyama, Yasunobu Inoue
*
Analysis Center and Department of Chemistry, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka 940-2188, Japan Received 20 December 2005; accepted for publication 29 March 2006 Available online 4 May 2006
Abstract The effects of surface acoustic wave (SAW) on the work function of Cu, Au and Pd metal surfaces with different surface structures were studied by photoelectron emission microscopy (PEEM). SAW propagation produced bright PEEM images for Cu, Au and Pd metal surfaces consisting of high-index planes and step sites, whereas it yielded dark images for the metals exposing low-index planes, indicating that the SAW enhanced photoemission from rough metal surfaces containing coordinatively-unsaturated metal atoms and lowered that from densely packed smooth metal surfaces. Changes in the PEEM images with SAW-on and SAW-off were reversible and were associated with decreases and increases in the work function of the metal surfaces, respectively. The SAW caused periodic and vertical lattice displacement, and it was demonstrated that large lattice displacement was responsible for work function changes from coincidence between the patterns of photoemission and lattice displacement. A mechanism for work function changes is proposed on the basis of effects on the spatial structures and electronic properties of metal surfaces. 2006 Elsevier B.V. All rights reserved. Keywords: Surface acoustic waves; Piezoelectric effect; Photoelectron emission microscopy; Work function; Surface structures; Copper; Silver; Gold; Polycrystalline surfaces; Metallic films
1. Introduction Work function is one of the most important surface properties and determines the chemical and physical properties of metal surfaces. It is intrinsic to the metal–metal atom distance and metal atom arrangements. Hence, it is interesting to investigate how the work function of metal surfaces varies as the metal lattice is dynamically distorted. A surface acoustic wave (SAW) is generated on a poled ferroelectric crystal by imposing rf electric power to interdigital transducer (IDT) electrodes deposited on the crystal, and can cause the lattice distortion of thin metal films deposited on its propagation path. We have previously demonstrated that the SAW is capable of activating thin polycrystalline metal (Ag, Pd, Ni) film surfaces, mark-
*
Corresponding author. Tel.: +81 258 47 9832; fax: +81 258 47 9830. E-mail address: [email protected] (Y. Inoue).
0039-6028/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.susc.2006.03.047
edly enhancing their catalytic activities for ethanol and CO oxidation [1–4]. King et al. observed that the SAW enhanced the catalytic activity of a Pt{1 1 0} single-crystal film for CO oxidation [5–8]. Furthermore, it exhibited interesting effects that were able to accelerate a single path in two parallel reaction paths on catalysts. For example, in Cu metal-catalyzed ethanol decomposition, the SAW dominantly enhanced ethylene production with little change in acetaldehyde production [9]. These results strongly suggest that the SAW affects the electronic structures of metal surfaces, and it is thought that the effects on catalyst activation and reaction selectivity are closely associated with changes in the work function. To confirm this, it is useful to determine photoemission characteristics directly related to the work function of metals during SAW propagation. Photoelectron emission microscopy (PEEM) provides a magnified image of the surface thanks to contrasts in the intensity of photoelectrons emitted from surfaces irradiated by UV light and yields a two-dimensional pattern of
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
differences in the work function of metal surfaces [10]. In a previous preliminary study, changes in the intensity of photoemission for a polycrystalline Cu thin film with SAW propagation were observed. The study showed the importance of conducting research to find out whether or not similar SAW effects were observed in other metal systems. Furthermore, the advantages of PEEM can be exploited to seek a positional correlation between lattice distortion and work function changes. In the present study, a PEEM apparatus was designed to measure the PEEM images of metal surfaces during SAW propagation. Polycrystalline Cu, Au and Pd thin films were deposited on the propagation path of Rayleigh SAW. To investigate the SAW effects on the photoemission behavior of metal surfaces with different surface structures, the metal surfaces were treated in different manners: a smooth surface was produced by annealing, and a rough surface was fabricated by sputtering. To characterize the surface structures of these metal surfaces on an atomic scale, an infrared reflection absorption spectroscopy (IRAS) method was employed using CO as a probe molecule. 2. Experimental section For the generation of a Rayleigh SAW, a 128-rotated y-cut LiNbO3 single crystal with a thickness of 0.5 mm was used. The two IDT electrodes, designed to generate an acoustic wave with a wavelength of 200 lm and a frequency of 20 MHz, were fabricated photolithographically on the crystal plane. Polycrystalline Cu, Au and Pd films (13 · 13 mm) were deposited at a thickness of 80 nm in the middle of the substrate crystal surface between the two IDT electrodes so that the SAW propagated through the metal phase. The metal films were prepared by evaporation of the metals with electron beam heating in vacuum and deposition at a rate of 0.5 nm s 1 at 473 K. After deposition, a sample was quickly transferred to the chamber of a separate PEEM apparatus. The metal films were either annealed at 573 K for 2.5 h under ultra high vacuum (UHV) conditions (<10 7 Pa) (the metal surface is referred to as an annealed surface) or sputtered at 373 K at 1000 V and then subjected to heat treatment at 573 K for 2.5 h in UHV conditions (the surface is referred to as a sputtered surface). Experiments were carried out using annealed and sputtered samples prepared separately and were repeated with fresh samples. A PEEM apparatus was designed to measure the photoemission characteristics during SAW propagation. Fig. 1 schematizes the PEEM system, together with a sample holder. A SAW sample was mounted on a copper plate furnished with a PBN heating plate. The temperature of the sample was changed from room temperature to 573 K. The SAW sample was placed just below the PEEM optics system to be as close as possible (4 mm) to the inlet, and irradiated at an incident angle of 70 using a 200 W deuterium lamp. The ejected electrons introduced into the PEEM lens system were captured by a phosphor screen.
2645
Cooled CCD Phosphor screen
UHV conditions
PEEM lens optics
Electric field shield plate UV source e -
SAW sample
metal film Terminator
Network tuner
Amplifier Power reflection meter
Network analyzer
Fig. 1. The schema of a PEEM apparatus and an electric circuit for SAW generation.
The brightness changes in PEEM images caused by photoelectrons were monitored with a cooled CCD camera (Hamamatsu Orca-ER). Measurements were performed at room temperature under UHV conditions (<1 · 10 8 Pa). The SAW was generated at a rf power of 0.2– 1.0 W. The PEEM images and intensity changes with SAW-on and SAW-off were taken repeatedly to confirm reproducibility. 3. Results and discussion In a previous study [11,12], we used infrared absorption spectroscopy (IRAS) for CO adsorption in order to determine the atomic-scale structure of polycrystalline metal surfaces prepared by annealing and sputtering since the vibration frequencies of the adsorbed CO used as a probe molecule provide information on the structure of Cu surfaces. Only a symmetric single peak was observed at 2070 cm 1 on an annealed Cu surface, whereas a complex spectrum consisting of four peaks at 2072, 2087, 2097, and 2104 cm 1 was obtained for a sputtered Cu surface. On the basis of the comparison of CO adsorption on Cu single crystals, the single peak at 2070 cm 1 for the
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
100 0
100
2400
200
300 400 Time /s
SAW-on
500
600
SAW-on
SAW-off
2360
SAW-off
2320 0
100
200
(b)
300 400 Time /s
500
600
Fig. 3. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for the high-index Cu metal surfaces.
remained bright during SAW propagation. Turning SAW off caused the images to become dark. The following SAW-on/SAW-off run also exhibited reproducible results with regard to image brightness. The PEEM intensity is given in Fig. 3(b). With SAW-on, the PEEM intensity rapidly increased and reached a constant level. With SAW-off, the intensity decreased to the original level after a transient fluctuation. The intensity changes were nearly the same as those in the second SAW-on/SAW-off run. Fig. 4(a) and (b) show changes in the pixel images and the PEEM intensity for a low-index Au surface, respectively. Although the pixel contrast between dark and bright images was obscure compared to that for the Cu surfaces,
SAW-off
SAW-off
SAW-on SAW-off SAW-on
SAW-off
300 200
200
100 0
100
200
300 400 Time /s
500
600
700
2440 SAW-off
2420
SAW-on
2400 2380
SAW-off
SAW-on
2360 0
100
200
300 400 Time /s
500
600
700
Fig. 2. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for the low-index Cu metal surface.
(a) PEEM intensity /arb. units
PEEM intensity /arb. units
SAW-off
200
(a)
100
(b)
SAW-on
400 SAW-off SAW-on SAW-off SAW-on
300
(a)
SAW-off SAW-on SAW-off
300
Pixel
Pixel
400
400
Pixel
annealed Cu surface was assigned to the stretch frequency of CO adsorbed on Cu(1 1 1) [13,14]. Although a SAW sample was exposed to air in sample transfer, the appearance of the strong CO peaks adsorbed on the annealed Cu surface indicated that annealing at 573 K produced available Cu(1 1 1) plane in line with the previous findings that deoxydation occurred from Cu(1 1 1) surface at a temperature of >570 K [15]. In the case of the sputtered Cu surface, the peaks at 2072, 2087, and 2097 cm 1 were assigned to the CO stretch frequencies adsorbed on Cu (1 1 1), (1 0 0) [16,17], and (1 1 0) [18,19] planes, respectively. The large peak at 2104 cm 1 was associated with the frequency of CO on high-index Cu planes such as (3 1 1) and (7 5 5) planes and step sites [20]. Thus, the annealed and the sputtered Cu surfaces exposed the low-index plane such as Cu(1 1 1) and a high density of high-index planes and step sites, respectively. Fig. 2(a) shows the pixel images of a low index Cu surface with SAW-off and SAW-on under UHV conditions. SAW-on produced dark images that persisted while the SAW was turned on. With SAW-off, the images became bright. Alternating SAW-on and SAW-off repeated the dark and bright images in a reproducible manner. The PEEM intensity was monitored to quantitatively examine the image changes. Fig. 2(b) shows changes in the PEEM intensity in the presence and absence of the SAW. At an initial stage of SAW-on, after a transient fluctuation, a sharp drop in intensity occurred, followed by a slow decrease. Upon turning the SAW off, the intensity returned to a slightly higher level. The second SAW-on caused a similar PEEM intensity decrease. The recovery of the intensity with SAW-off brought the intensity back to nearly the same level as the first run. Fig. 3(a) shows the pixel images for a high-index Cu surface. With SAW-on, the images became bright. The images
PEEM intensity /arb. units
2646
0
100
2960
200
300 400 Time /s
SAW-off
2940
500
600
SAW-off
SAW-on
2920
(b)
SAW-on 0
100
200
300 400 Time /s
500
600
Fig. 4. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for low-index Au metal surfaces.
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
400
SAW-off
SAW-on SAW-off
SAW-on SAW-off
Pixel
300 200 100 0
100
200
PEEM intensity /arb. units
(a)
2880 2870 2860 2850 2840 2830
300 400 Time /s
500
600
SAW-on SAW-on SAW-off SAW-off 0
100
200
(b)
300 400 Time /s
500
600
Fig. 5. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for high-index Au metal surfaces.
dark images were obtained with SAW-on, which became bright with SAW-off. The PEEM intensity decreased with SAW-on and was recovered with SAW-off. Fig. 5(a) and (b) show the results for high-index Au surfaces. Brightdark images with SAW-on and SAW-off appeared, whose changes were exactly the opposite as those observed for the low-index Au surface. Fig. 6 shows changes in the pixel image and intensity of a sputtered Pd surface exposing high-index planes. SAW-on produced bright images and a strong PEEM intensity. The changes were reversible with SAW-on and SAW-off. The annealed Pd surface showed poor reproducibility of the PEEM images with SAW-on and SAW-off. This is possibly due to the difficulty of pro-
Pixel
400
SAW-off
SAW-on SAW-off
2647
duction of low index surfaces by annealing. It should to be noted that Cu, Au (and Pd) caused analogous PEEM intensity changes with SAW propagation that were different between low- and high-index metal surfaces. The PEEM image and intensity is associated with the work function of metal surfaces in the observed area. A PEEM intensity enhancement corresponds to a decrease in the work function and vice versa. The PEEM intensity behavior clearly demonstrated that the SAW increased the work function of the low-index Cu and Au surfaces, whereas it decreased those of the high-index planes and step sites. Interestingly, the changes occurred reproducibly, and the decreases in the work function for the smooth metal surfaces with SAW-on were exactly the opposite of the enhancement of the work function for the rough metal surfaces. A laser Doppler method was used to measure surface distortion due to SAW propagation, and it was shown that Rayleigh SAW caused large lattice displacement (standing wave) vertical to the surface [21]. The top-to-top distance of standing waves was 100 lm, equivalent to half the wavelength (200 lm) obtained from the finger space of the IDT. In order to find out which part of the lattice displacement was responsible for changes in the PEEM intensity, the position of the laser spot for the Doppler displacement measurements was scanned in the direction of SAW propagation, as shown in Fig. 7, and the PEEM images were obtained as a two-dimensional pattern. Fig. 8 shows the PEEM images for a high-index Au surface over a 200-lm region. White color corresponded to the emitted current before turning on the SAW. Red represents a large amount of photoelectron emission, indicative of a work function decrease. The red regime was concentrated at the two ends of the pattern, and faint in the central part. The comparison of the red color and lattice displacement pattern indicated that the red color zone corresponded to a large lattice displacement. Fig. 9 shows the results for a
SAW-on SAW-off
300
SAW pro
pagatio
100 0
100
200
PEEM intensity /arb. units
(a)
500
600
on 6 Z-axis /nm 5 Height 4 3 (Lattice displacement) 2 1
2760
750
SAW-on 2720
500
SAW-on SAW-off SAW-off
2680 0
(b)
300 400 Time /s
n directi
Widt h
200
100
200
0
300 400 Time /s
500
600
Fig. 6. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for high-index Pd metal surfaces.
200 μm 250
500
Y-axis /μm
250 750
0
X-axis /μm : Analyzed area
Fig. 7. A laser Doppler pattern of Rayleigh SAW and analyzed area for photoemission measurements.
2648
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
Fig. 8. Two-dimensional lattice displacement (a) and PEEM image (b) for high-index Au surfaces.
According to a Jellium model [22,23] for the work function of a metal surface, free electrons at metal surfaces spill out toward the vacuum (just outside) from the topmost surface. The spill-out electrons and positive charges remaining inside the metal produce an electric double layer. The potential barrier of the layer is related to the work function of the metal. The dependence of work function on the crystallographic orientations is due to the differences in the electric layer arising from differences in spill-out electron density among the crystal faces. Generally, the work function of a close-packed smooth metal surface is larger than that of a surface that is rougher on an atomic scale (a less close-packed surface). For example, the work function of Cu metal is 4.94, 4.59 and 4.48 eV, [24] and that of Ag is 4.74, 4.64, and 4.52 eV [25] for the (1 1 1), (1 0 0) and (1 1 0) planes, respectively. The opposite changes in the work function with SAW-on between the high- and lowindex metal surfaces are due to atomic-scale structural differences between the surfaces. The high-index planes and step sites consist largely of coordinatively-unsaturated surface atoms. The surface atoms are loosely bound, and lattice distortion is known to be significantly concentrated on unsaturated and irregular sites such as dislocations and vacancies [26]. Thus, it is possible that the vertical lattice displacement induced by SAW influences the position of the metal atoms and permits the unsaturated metal ion cores to stick out into the electric double layer. Protruding surface metal atoms make the electric double layer thin, leading to decreases in the work function. Fig. 10 shows a model. The model suggests that the work function decreases in the high-index planes and step sites with SAWon propagation are associated with spatial changes in coordinatively-unsaturated surface atoms. In contrast, the low-index metal surfaces are closely packed, and it is unlikely that the lattice displacement by the SAW directly influences the atom positions and atom-atom distance. In our previous studies on resonance
Fig. 9. Two-dimensional lattice displacement (a) and PEEM image (b) for low-index Au surfaces.
low-index Au surface. Blue was indicative of decreases in photoelectrons, i.e., increases in the work function. Blue appeared strongly at both end regions, where lattice displacement was the largest. Thus, there was a spatial correlation between the sites of work function enhancement and large lattice displacement. These findings indicate that a large vertical lattice displacement was responsible for changes in the work function.
Fig. 10. A model for work function decreases in a high-index metal surface with SAW. Coordinatively-unsaturated surface atoms stick out as a result of the lattice displacement by SAW, and the protruding, positively charged metal cores produce a thin electric double layer.
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
oscillation (RO) of the bulk acoustic wave generated on ferroelectric substrates by imposing rf electric power, it has been demonstrated that the work function of annealed Pd and Ag metal films deposited on the ferroelectric substrates increased when lattice displacement caused by RO was vertical to the surface, but little varied when it was parallel to the surface [27]. Based on the fact that the work function increased in the case that the direction of motion of the vertical lattice displacement was the same as the direction of electron spill-out, a model was proposed that the density of spill-out electrons increased by the ROinduced vertical lattice displacement, which made the electric double layer thicker and hence enhanced the work function of the metals [27]. The present SAW caused lattice displacement vertical to the metal surface to the same extent as the RO did. Thus, the similarity of lattice displacement between RO and SAW explains increases in the work function of low-index metal planes. The present PEEM study clearly showed that the SAW affected the photoemission behavior differently depending on the roughness of the metal surface. Interestingly, the periodic and vertical lattice displacement caused by SAW has marked influences on the spatial structure and electronic properties of metal surfaces. The present findings are encouraging for the development of a physical method to artificially change the work function of metal surfaces, and hence their chemical properties, for example, as catalysts. Acknowledgements We thank Dr. N. Saito for his assistance with Doppler shift measurements. This work was supported by a Grant-in-Aid for Encouragement of Young Scientists B (16750010) and for Scientific Research A (15206088) from The Ministry of Education, Science, Sports and Culture of Japan.
2649
References [1] Y. Inoue, M. Matsukawa, K. Sato, J. Am. Chem. Soc. 111 (1989) 8965. [2] Y. Inoue, M. Matsukawa, K. Sato, J. Phys. Chem. 96 (1992) 2222. [3] Y. Inoue, M. Matsukawa, H. Kawaguchi, J. Chem. Soc., Faraday Trans. 88 (1992) 2923. [4] H. Nishiyama, N. Saito, M. Yashima, Y. Watanabe, Y. Inoue, Faraday Discuss. 107 (1997) 425. [5] S. Kelling, T. Mitrelias, Y. Matusmoto, V.P. Ostanin, D.A. King, Faraday Discuss. 107 (1997) 435. [6] S. Kelling, T. Mitrelias, Y. Matusmoto, V.P. Ostanin, D.A. King, J. Chem. Phys. 107 (1997) 5609. [7] T. Mitrelias, T.S. Kelling, R.I. Kvon, V.P. Ostanin, D.A. King, Surf. Sci. 417 (1998) 97. [8] S. Kelling, S. Cerasari, H.H. Rotermund, G. Ertl, D.A. King, Chem. Phys. Lett. 293 (1998) 325. [9] H. Nishiyama, N. Saito, T. Yashima, K. Sato, Y. Inoue, Surf. Sci. 427/428 (1999) 152. [10] W. Engel, M.E. Kordesch, H.H. Rotermund, S. Kubala, A. von Oertzen, Ultramicroscopy 36 (1991) 148. [11] H. Nishiyama, Y. Inoue, J. Phys. Chem., B 107 (2003) 8738. [12] H. Nishiyama, Y. Inoue, Sur. Sci. 594 (2005) 156. [13] M.A. Chesters, S.F. Parker, R. Raval, Surf. Sci. 165 (1986) 179. [14] B.E. Hayden, K. Kretzschmar, A. Bradshaw, Surf. Sci. 155 (1985) 553. [15] J. Bloch, D.J. Bottomley, S. Janz, H.M. van Driel, J. Chem. Phys. 98 (1993) 9167. [16] R. Ryberg, Surf. Sci. 114 (1982) 627. [17] K. Horn, J. Pritchard, Surf. Sci. 55 (1976) 701. [18] P. Hollins, K.J. Davies, J. Pritchard, Surf. Sci. 138 (1984) 75. [19] K. Horn, M. Hussain, J. Pritchard, Surf. Sci. 63 (1977) 244. [20] J. Pritchard, T. Catterick, R.K. Gupta, Surf. Sci. 76 (1975) 1. [21] N. Saito, H. Nishiyama, K. Sato, Y. Inoue, Chem. Phys. Lett. 297 (1998) 72. [22] N.D. Lang, W. Kohn, Phys. Rev. B 3 (1971) 1215; N.D. Lang, W. Kohn, Phys. Rev. B 1 (1970) 4555. [23] H.L. Skriver, N.M. Rosengaard, Phys. Rev. B 46 (1992) 7157. [24] P.O. Gartland, S. Berge, B. Slagsvold, J. Phy. Rev. Lett. l (1972) 738. [25] A.W. Dweydari, C.H.B. Mee, Phys. Status Solidi. A 17 (1973) 247; A.W. Dweydari, C.H.B. Mee, Phys. Status Solidi. A 27 (1975) 223. [26] C. Kittel, Introduction to Solid State Physics, seventh ed., John Wiley & Sons, Inc., 1996, pp. 590. [27] N. Saito, Y. Inoue, J. Phys. Chem. 106 (2002) 5011.
PEEM study of work function changes in Cu, Au and Pd metal surfaces with surface acoustic wave propagation Hiroshi Nishiyama, Yasunobu Inoue
*
Analysis Center and Department of Chemistry, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka 940-2188, Japan Received 20 December 2005; accepted for publication 29 March 2006 Available online 4 May 2006
Abstract The effects of surface acoustic wave (SAW) on the work function of Cu, Au and Pd metal surfaces with different surface structures were studied by photoelectron emission microscopy (PEEM). SAW propagation produced bright PEEM images for Cu, Au and Pd metal surfaces consisting of high-index planes and step sites, whereas it yielded dark images for the metals exposing low-index planes, indicating that the SAW enhanced photoemission from rough metal surfaces containing coordinatively-unsaturated metal atoms and lowered that from densely packed smooth metal surfaces. Changes in the PEEM images with SAW-on and SAW-off were reversible and were associated with decreases and increases in the work function of the metal surfaces, respectively. The SAW caused periodic and vertical lattice displacement, and it was demonstrated that large lattice displacement was responsible for work function changes from coincidence between the patterns of photoemission and lattice displacement. A mechanism for work function changes is proposed on the basis of effects on the spatial structures and electronic properties of metal surfaces. 2006 Elsevier B.V. All rights reserved. Keywords: Surface acoustic waves; Piezoelectric effect; Photoelectron emission microscopy; Work function; Surface structures; Copper; Silver; Gold; Polycrystalline surfaces; Metallic films
1. Introduction Work function is one of the most important surface properties and determines the chemical and physical properties of metal surfaces. It is intrinsic to the metal–metal atom distance and metal atom arrangements. Hence, it is interesting to investigate how the work function of metal surfaces varies as the metal lattice is dynamically distorted. A surface acoustic wave (SAW) is generated on a poled ferroelectric crystal by imposing rf electric power to interdigital transducer (IDT) electrodes deposited on the crystal, and can cause the lattice distortion of thin metal films deposited on its propagation path. We have previously demonstrated that the SAW is capable of activating thin polycrystalline metal (Ag, Pd, Ni) film surfaces, mark-
*
Corresponding author. Tel.: +81 258 47 9832; fax: +81 258 47 9830. E-mail address: [email protected] (Y. Inoue).
0039-6028/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.susc.2006.03.047
edly enhancing their catalytic activities for ethanol and CO oxidation [1–4]. King et al. observed that the SAW enhanced the catalytic activity of a Pt{1 1 0} single-crystal film for CO oxidation [5–8]. Furthermore, it exhibited interesting effects that were able to accelerate a single path in two parallel reaction paths on catalysts. For example, in Cu metal-catalyzed ethanol decomposition, the SAW dominantly enhanced ethylene production with little change in acetaldehyde production [9]. These results strongly suggest that the SAW affects the electronic structures of metal surfaces, and it is thought that the effects on catalyst activation and reaction selectivity are closely associated with changes in the work function. To confirm this, it is useful to determine photoemission characteristics directly related to the work function of metals during SAW propagation. Photoelectron emission microscopy (PEEM) provides a magnified image of the surface thanks to contrasts in the intensity of photoelectrons emitted from surfaces irradiated by UV light and yields a two-dimensional pattern of
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
differences in the work function of metal surfaces [10]. In a previous preliminary study, changes in the intensity of photoemission for a polycrystalline Cu thin film with SAW propagation were observed. The study showed the importance of conducting research to find out whether or not similar SAW effects were observed in other metal systems. Furthermore, the advantages of PEEM can be exploited to seek a positional correlation between lattice distortion and work function changes. In the present study, a PEEM apparatus was designed to measure the PEEM images of metal surfaces during SAW propagation. Polycrystalline Cu, Au and Pd thin films were deposited on the propagation path of Rayleigh SAW. To investigate the SAW effects on the photoemission behavior of metal surfaces with different surface structures, the metal surfaces were treated in different manners: a smooth surface was produced by annealing, and a rough surface was fabricated by sputtering. To characterize the surface structures of these metal surfaces on an atomic scale, an infrared reflection absorption spectroscopy (IRAS) method was employed using CO as a probe molecule. 2. Experimental section For the generation of a Rayleigh SAW, a 128-rotated y-cut LiNbO3 single crystal with a thickness of 0.5 mm was used. The two IDT electrodes, designed to generate an acoustic wave with a wavelength of 200 lm and a frequency of 20 MHz, were fabricated photolithographically on the crystal plane. Polycrystalline Cu, Au and Pd films (13 · 13 mm) were deposited at a thickness of 80 nm in the middle of the substrate crystal surface between the two IDT electrodes so that the SAW propagated through the metal phase. The metal films were prepared by evaporation of the metals with electron beam heating in vacuum and deposition at a rate of 0.5 nm s 1 at 473 K. After deposition, a sample was quickly transferred to the chamber of a separate PEEM apparatus. The metal films were either annealed at 573 K for 2.5 h under ultra high vacuum (UHV) conditions (<10 7 Pa) (the metal surface is referred to as an annealed surface) or sputtered at 373 K at 1000 V and then subjected to heat treatment at 573 K for 2.5 h in UHV conditions (the surface is referred to as a sputtered surface). Experiments were carried out using annealed and sputtered samples prepared separately and were repeated with fresh samples. A PEEM apparatus was designed to measure the photoemission characteristics during SAW propagation. Fig. 1 schematizes the PEEM system, together with a sample holder. A SAW sample was mounted on a copper plate furnished with a PBN heating plate. The temperature of the sample was changed from room temperature to 573 K. The SAW sample was placed just below the PEEM optics system to be as close as possible (4 mm) to the inlet, and irradiated at an incident angle of 70 using a 200 W deuterium lamp. The ejected electrons introduced into the PEEM lens system were captured by a phosphor screen.
2645
Cooled CCD Phosphor screen
UHV conditions
PEEM lens optics
Electric field shield plate UV source e -
SAW sample
metal film Terminator
Network tuner
Amplifier Power reflection meter
Network analyzer
Fig. 1. The schema of a PEEM apparatus and an electric circuit for SAW generation.
The brightness changes in PEEM images caused by photoelectrons were monitored with a cooled CCD camera (Hamamatsu Orca-ER). Measurements were performed at room temperature under UHV conditions (<1 · 10 8 Pa). The SAW was generated at a rf power of 0.2– 1.0 W. The PEEM images and intensity changes with SAW-on and SAW-off were taken repeatedly to confirm reproducibility. 3. Results and discussion In a previous study [11,12], we used infrared absorption spectroscopy (IRAS) for CO adsorption in order to determine the atomic-scale structure of polycrystalline metal surfaces prepared by annealing and sputtering since the vibration frequencies of the adsorbed CO used as a probe molecule provide information on the structure of Cu surfaces. Only a symmetric single peak was observed at 2070 cm 1 on an annealed Cu surface, whereas a complex spectrum consisting of four peaks at 2072, 2087, 2097, and 2104 cm 1 was obtained for a sputtered Cu surface. On the basis of the comparison of CO adsorption on Cu single crystals, the single peak at 2070 cm 1 for the
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
100 0
100
2400
200
300 400 Time /s
SAW-on
500
600
SAW-on
SAW-off
2360
SAW-off
2320 0
100
200
(b)
300 400 Time /s
500
600
Fig. 3. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for the high-index Cu metal surfaces.
remained bright during SAW propagation. Turning SAW off caused the images to become dark. The following SAW-on/SAW-off run also exhibited reproducible results with regard to image brightness. The PEEM intensity is given in Fig. 3(b). With SAW-on, the PEEM intensity rapidly increased and reached a constant level. With SAW-off, the intensity decreased to the original level after a transient fluctuation. The intensity changes were nearly the same as those in the second SAW-on/SAW-off run. Fig. 4(a) and (b) show changes in the pixel images and the PEEM intensity for a low-index Au surface, respectively. Although the pixel contrast between dark and bright images was obscure compared to that for the Cu surfaces,
SAW-off
SAW-off
SAW-on SAW-off SAW-on
SAW-off
300 200
200
100 0
100
200
300 400 Time /s
500
600
700
2440 SAW-off
2420
SAW-on
2400 2380
SAW-off
SAW-on
2360 0
100
200
300 400 Time /s
500
600
700
Fig. 2. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for the low-index Cu metal surface.
(a) PEEM intensity /arb. units
PEEM intensity /arb. units
SAW-off
200
(a)
100
(b)
SAW-on
400 SAW-off SAW-on SAW-off SAW-on
300
(a)
SAW-off SAW-on SAW-off
300
Pixel
Pixel
400
400
Pixel
annealed Cu surface was assigned to the stretch frequency of CO adsorbed on Cu(1 1 1) [13,14]. Although a SAW sample was exposed to air in sample transfer, the appearance of the strong CO peaks adsorbed on the annealed Cu surface indicated that annealing at 573 K produced available Cu(1 1 1) plane in line with the previous findings that deoxydation occurred from Cu(1 1 1) surface at a temperature of >570 K [15]. In the case of the sputtered Cu surface, the peaks at 2072, 2087, and 2097 cm 1 were assigned to the CO stretch frequencies adsorbed on Cu (1 1 1), (1 0 0) [16,17], and (1 1 0) [18,19] planes, respectively. The large peak at 2104 cm 1 was associated with the frequency of CO on high-index Cu planes such as (3 1 1) and (7 5 5) planes and step sites [20]. Thus, the annealed and the sputtered Cu surfaces exposed the low-index plane such as Cu(1 1 1) and a high density of high-index planes and step sites, respectively. Fig. 2(a) shows the pixel images of a low index Cu surface with SAW-off and SAW-on under UHV conditions. SAW-on produced dark images that persisted while the SAW was turned on. With SAW-off, the images became bright. Alternating SAW-on and SAW-off repeated the dark and bright images in a reproducible manner. The PEEM intensity was monitored to quantitatively examine the image changes. Fig. 2(b) shows changes in the PEEM intensity in the presence and absence of the SAW. At an initial stage of SAW-on, after a transient fluctuation, a sharp drop in intensity occurred, followed by a slow decrease. Upon turning the SAW off, the intensity returned to a slightly higher level. The second SAW-on caused a similar PEEM intensity decrease. The recovery of the intensity with SAW-off brought the intensity back to nearly the same level as the first run. Fig. 3(a) shows the pixel images for a high-index Cu surface. With SAW-on, the images became bright. The images
PEEM intensity /arb. units
2646
0
100
2960
200
300 400 Time /s
SAW-off
2940
500
600
SAW-off
SAW-on
2920
(b)
SAW-on 0
100
200
300 400 Time /s
500
600
Fig. 4. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for low-index Au metal surfaces.
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
400
SAW-off
SAW-on SAW-off
SAW-on SAW-off
Pixel
300 200 100 0
100
200
PEEM intensity /arb. units
(a)
2880 2870 2860 2850 2840 2830
300 400 Time /s
500
600
SAW-on SAW-on SAW-off SAW-off 0
100
200
(b)
300 400 Time /s
500
600
Fig. 5. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for high-index Au metal surfaces.
dark images were obtained with SAW-on, which became bright with SAW-off. The PEEM intensity decreased with SAW-on and was recovered with SAW-off. Fig. 5(a) and (b) show the results for high-index Au surfaces. Brightdark images with SAW-on and SAW-off appeared, whose changes were exactly the opposite as those observed for the low-index Au surface. Fig. 6 shows changes in the pixel image and intensity of a sputtered Pd surface exposing high-index planes. SAW-on produced bright images and a strong PEEM intensity. The changes were reversible with SAW-on and SAW-off. The annealed Pd surface showed poor reproducibility of the PEEM images with SAW-on and SAW-off. This is possibly due to the difficulty of pro-
Pixel
400
SAW-off
SAW-on SAW-off
2647
duction of low index surfaces by annealing. It should to be noted that Cu, Au (and Pd) caused analogous PEEM intensity changes with SAW propagation that were different between low- and high-index metal surfaces. The PEEM image and intensity is associated with the work function of metal surfaces in the observed area. A PEEM intensity enhancement corresponds to a decrease in the work function and vice versa. The PEEM intensity behavior clearly demonstrated that the SAW increased the work function of the low-index Cu and Au surfaces, whereas it decreased those of the high-index planes and step sites. Interestingly, the changes occurred reproducibly, and the decreases in the work function for the smooth metal surfaces with SAW-on were exactly the opposite of the enhancement of the work function for the rough metal surfaces. A laser Doppler method was used to measure surface distortion due to SAW propagation, and it was shown that Rayleigh SAW caused large lattice displacement (standing wave) vertical to the surface [21]. The top-to-top distance of standing waves was 100 lm, equivalent to half the wavelength (200 lm) obtained from the finger space of the IDT. In order to find out which part of the lattice displacement was responsible for changes in the PEEM intensity, the position of the laser spot for the Doppler displacement measurements was scanned in the direction of SAW propagation, as shown in Fig. 7, and the PEEM images were obtained as a two-dimensional pattern. Fig. 8 shows the PEEM images for a high-index Au surface over a 200-lm region. White color corresponded to the emitted current before turning on the SAW. Red represents a large amount of photoelectron emission, indicative of a work function decrease. The red regime was concentrated at the two ends of the pattern, and faint in the central part. The comparison of the red color and lattice displacement pattern indicated that the red color zone corresponded to a large lattice displacement. Fig. 9 shows the results for a
SAW-on SAW-off
300
SAW pro
pagatio
100 0
100
200
PEEM intensity /arb. units
(a)
500
600
on 6 Z-axis /nm 5 Height 4 3 (Lattice displacement) 2 1
2760
750
SAW-on 2720
500
SAW-on SAW-off SAW-off
2680 0
(b)
300 400 Time /s
n directi
Widt h
200
100
200
0
300 400 Time /s
500
600
Fig. 6. Changes in (a) PEEM images and (b) intensity with SAW-on and SAW-off for high-index Pd metal surfaces.
200 μm 250
500
Y-axis /μm
250 750
0
X-axis /μm : Analyzed area
Fig. 7. A laser Doppler pattern of Rayleigh SAW and analyzed area for photoemission measurements.
2648
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
Fig. 8. Two-dimensional lattice displacement (a) and PEEM image (b) for high-index Au surfaces.
According to a Jellium model [22,23] for the work function of a metal surface, free electrons at metal surfaces spill out toward the vacuum (just outside) from the topmost surface. The spill-out electrons and positive charges remaining inside the metal produce an electric double layer. The potential barrier of the layer is related to the work function of the metal. The dependence of work function on the crystallographic orientations is due to the differences in the electric layer arising from differences in spill-out electron density among the crystal faces. Generally, the work function of a close-packed smooth metal surface is larger than that of a surface that is rougher on an atomic scale (a less close-packed surface). For example, the work function of Cu metal is 4.94, 4.59 and 4.48 eV, [24] and that of Ag is 4.74, 4.64, and 4.52 eV [25] for the (1 1 1), (1 0 0) and (1 1 0) planes, respectively. The opposite changes in the work function with SAW-on between the high- and lowindex metal surfaces are due to atomic-scale structural differences between the surfaces. The high-index planes and step sites consist largely of coordinatively-unsaturated surface atoms. The surface atoms are loosely bound, and lattice distortion is known to be significantly concentrated on unsaturated and irregular sites such as dislocations and vacancies [26]. Thus, it is possible that the vertical lattice displacement induced by SAW influences the position of the metal atoms and permits the unsaturated metal ion cores to stick out into the electric double layer. Protruding surface metal atoms make the electric double layer thin, leading to decreases in the work function. Fig. 10 shows a model. The model suggests that the work function decreases in the high-index planes and step sites with SAWon propagation are associated with spatial changes in coordinatively-unsaturated surface atoms. In contrast, the low-index metal surfaces are closely packed, and it is unlikely that the lattice displacement by the SAW directly influences the atom positions and atom-atom distance. In our previous studies on resonance
Fig. 9. Two-dimensional lattice displacement (a) and PEEM image (b) for low-index Au surfaces.
low-index Au surface. Blue was indicative of decreases in photoelectrons, i.e., increases in the work function. Blue appeared strongly at both end regions, where lattice displacement was the largest. Thus, there was a spatial correlation between the sites of work function enhancement and large lattice displacement. These findings indicate that a large vertical lattice displacement was responsible for changes in the work function.
Fig. 10. A model for work function decreases in a high-index metal surface with SAW. Coordinatively-unsaturated surface atoms stick out as a result of the lattice displacement by SAW, and the protruding, positively charged metal cores produce a thin electric double layer.
H. Nishiyama, Y. Inoue / Surface Science 600 (2006) 2644–2649
oscillation (RO) of the bulk acoustic wave generated on ferroelectric substrates by imposing rf electric power, it has been demonstrated that the work function of annealed Pd and Ag metal films deposited on the ferroelectric substrates increased when lattice displacement caused by RO was vertical to the surface, but little varied when it was parallel to the surface [27]. Based on the fact that the work function increased in the case that the direction of motion of the vertical lattice displacement was the same as the direction of electron spill-out, a model was proposed that the density of spill-out electrons increased by the ROinduced vertical lattice displacement, which made the electric double layer thicker and hence enhanced the work function of the metals [27]. The present SAW caused lattice displacement vertical to the metal surface to the same extent as the RO did. Thus, the similarity of lattice displacement between RO and SAW explains increases in the work function of low-index metal planes. The present PEEM study clearly showed that the SAW affected the photoemission behavior differently depending on the roughness of the metal surface. Interestingly, the periodic and vertical lattice displacement caused by SAW has marked influences on the spatial structure and electronic properties of metal surfaces. The present findings are encouraging for the development of a physical method to artificially change the work function of metal surfaces, and hence their chemical properties, for example, as catalysts. Acknowledgements We thank Dr. N. Saito for his assistance with Doppler shift measurements. This work was supported by a Grant-in-Aid for Encouragement of Young Scientists B (16750010) and for Scientific Research A (15206088) from The Ministry of Education, Science, Sports and Culture of Japan.
2649
References [1] Y. Inoue, M. Matsukawa, K. Sato, J. Am. Chem. Soc. 111 (1989) 8965. [2] Y. Inoue, M. Matsukawa, K. Sato, J. Phys. Chem. 96 (1992) 2222. [3] Y. Inoue, M. Matsukawa, H. Kawaguchi, J. Chem. Soc., Faraday Trans. 88 (1992) 2923. [4] H. Nishiyama, N. Saito, M. Yashima, Y. Watanabe, Y. Inoue, Faraday Discuss. 107 (1997) 425. [5] S. Kelling, T. Mitrelias, Y. Matusmoto, V.P. Ostanin, D.A. King, Faraday Discuss. 107 (1997) 435. [6] S. Kelling, T. Mitrelias, Y. Matusmoto, V.P. Ostanin, D.A. King, J. Chem. Phys. 107 (1997) 5609. [7] T. Mitrelias, T.S. Kelling, R.I. Kvon, V.P. Ostanin, D.A. King, Surf. Sci. 417 (1998) 97. [8] S. Kelling, S. Cerasari, H.H. Rotermund, G. Ertl, D.A. King, Chem. Phys. Lett. 293 (1998) 325. [9] H. Nishiyama, N. Saito, T. Yashima, K. Sato, Y. Inoue, Surf. Sci. 427/428 (1999) 152. [10] W. Engel, M.E. Kordesch, H.H. Rotermund, S. Kubala, A. von Oertzen, Ultramicroscopy 36 (1991) 148. [11] H. Nishiyama, Y. Inoue, J. Phys. Chem., B 107 (2003) 8738. [12] H. Nishiyama, Y. Inoue, Sur. Sci. 594 (2005) 156. [13] M.A. Chesters, S.F. Parker, R. Raval, Surf. Sci. 165 (1986) 179. [14] B.E. Hayden, K. Kretzschmar, A. Bradshaw, Surf. Sci. 155 (1985) 553. [15] J. Bloch, D.J. Bottomley, S. Janz, H.M. van Driel, J. Chem. Phys. 98 (1993) 9167. [16] R. Ryberg, Surf. Sci. 114 (1982) 627. [17] K. Horn, J. Pritchard, Surf. Sci. 55 (1976) 701. [18] P. Hollins, K.J. Davies, J. Pritchard, Surf. Sci. 138 (1984) 75. [19] K. Horn, M. Hussain, J. Pritchard, Surf. Sci. 63 (1977) 244. [20] J. Pritchard, T. Catterick, R.K. Gupta, Surf. Sci. 76 (1975) 1. [21] N. Saito, H. Nishiyama, K. Sato, Y. Inoue, Chem. Phys. Lett. 297 (1998) 72. [22] N.D. Lang, W. Kohn, Phys. Rev. B 3 (1971) 1215; N.D. Lang, W. Kohn, Phys. Rev. B 1 (1970) 4555. [23] H.L. Skriver, N.M. Rosengaard, Phys. Rev. B 46 (1992) 7157. [24] P.O. Gartland, S. Berge, B. Slagsvold, J. Phy. Rev. Lett. l (1972) 738. [25] A.W. Dweydari, C.H.B. Mee, Phys. Status Solidi. A 17 (1973) 247; A.W. Dweydari, C.H.B. Mee, Phys. Status Solidi. A 27 (1975) 223. [26] C. Kittel, Introduction to Solid State Physics, seventh ed., John Wiley & Sons, Inc., 1996, pp. 590. [27] N. Saito, Y. Inoue, J. Phys. Chem. 106 (2002) 5011.