Langmuir 1992,8, 369-371
369
Scanning Tunneling Microscopy on Pressed Powders 2. Zhang,M. M. Lerner: V. J. Marty,and P. R. Watson* Department of Chemistry and Center for Advanced Materials Research, Oregon State University, Corvallis, Oregon 97331 -4003 Received September 16,1991. In Final Form: December 9, 1991 Low- and high-resolutionSTM images have been obtained in air on pelletized graphiteand TaSz powders. Low-resolution images show different topographies,which includebroad, flat areassimilarto those observed on highly-oriented or single-crystalsamples. High-resolution images show, in atomic detail, the trigonal or honeycomb pattern for the graphite surfaces and a hexagonal pattern for TaS2.
I. Introduction Scanning tunneling microscopy (STM) has become a reliable method of characterizingsurfaces. A large number of reports have now been published on STM analyses of layered materials such as graphite and transition metal chalcogenides. Studies of the surfaces of intercalated graphite and metal disulfides have shown STM to be a valuable tool in the characterization of physical and electronic structures of intercalation compounds derived from layered host structures.l+ The great majority of STM analyses on graphite to date have been performed on highly-orientedpyrolytic graphite (HOPG). Work on other layered materials has similarly focused on single or quasi-singlecrystals. This is especially true in work including high-resolution, i.e. atomic scale, images. While single crystals offer the obvious advantage of possessing macroscopically planar and highly ordered surfaces, they are neither common nor readily prepared forms of most solids. In addition, due to the slow diffusion processes through highly crystalline solids, preparative chemistry involving these materials often results in inhomogeneous products. For these reasons it would often be advantageous to perform surface analyses such as STM on microcrystalline powders. A few groups have described STM analyses of graphite in forms other than HOPG. These reports include work on Kish (single crystal) g r a ~ h i t e ,flakes: ~ , ~ fibers,lOJ1and graphite-containing composites.12 In addition, glassy carbon and other carbon films have been examined.13J4 We are unaware, however, of any work which details results of STM analyses on pellets pressed from free-flowing (1)Olk, C.; Heremans, J.; Dresselhaus, M.; Speck, J.; Nicholls, J. J . Vac. Sci. Technol., B 1991,9 (2),1055. (2)Gauthier,S.;Rousset, S.;Klein, J.; Sacks, W.; Belin, M. J . Vac.Sci. Technol., A 1988,6 (2), 360. (3)Anselmetti, D.; Geiser, V.; Overney, G.; Weisendanger, R.; Guntherodt, H. Phys. Rev. B: Condens. Matter 1990,42,1848. (4)Tanaka, M.;Mizutani, W.; Nakashizu, T.; Morita, N.; Yamazaki, S.; Bando, H.; Ono, M.; Kajimura, K. J . Microsc. 1988,152 (11, 183. (5) (a) Kelty, S.; Lieber, C. J . Phys:Chem. 1989,93,5983.(b)Kelty, S.;Lieber, C. J. Vac. Sci. Technol., B 1991,9(21,1068. (c) Lieber, C.; Wu, X. Acc. Chem. Res. 1991,24,170,and references therein. (6)Giambattista,B.;Slough,C.; McNairy, W.; Coleman, R. Phys. Reu. B: Condens. Matter 1990,41,10082. (7)Morita, S.;Tsukada, S.; Mikoshiba, N. J. Vac. Sci. Technol., A 1988,6 (2),354. (8)Koga, Y.; Miyazaki, Y.; Nakagiri, N. Jpn. J . Appl. Phys., Part 2 1988,27(6),L976. (9) Kim, D.; Labes, M.; Siperko, L. Mater. Res. Bull. 1990,25 (121, 1461. (lo)Magonov, S.;Cantow, H.; Donnet, J. Polym. Bull. 1990,23 (6), 555. (11)Hoffman, W.; Hurley, W.; Liu, P.; Owens,T. J . Mater. Res. 1991, 6 (8),1685. (12)Wang, J.; Martinez, T.;Yaniv, D.; McCormick, L. J . Electroanal. Chem., 1990,286,265. (13)Siperko, L. J . Vac. Sci. Technol., B 1991,9 (21,1061. (14)Elinga, V.; Wudl, F. J . Vac. Sci. Technol., A 1988,6 (2),412.
0743-7463/92/240&0369$03.00/0
Figure 1. Low-resolution STM images (750A X 750 A) of the surface of a pellet pressed from SP1 graphite powder: tip bias = -86 mV; current = 2.5 nA. The image displayed is of raw data.
powders. As noted above, it is in this form that many materials are readily available in high purity and from which homogeneous intercalation products may easily be prepared. The pelletized powders of a large number of layered compounds are conducting or semiconducting and will therefore permit the flow of a tunneling current. Pellets may be readily formed with macroscopically flat and smooth surfaces. As is well-known from X-ray powder diffraction studies, the particles within such pellets commonly exhibit a strong preference toward orientation with basal planes parallel to the pellet surface. Microcrystalline powders with mean particle sizes in the 1-100 pm range contain crystalline faces with ordered regions that are certainly very large on the scale of the STM experiment. For these reasons, there seemsto be no reason why STM images cannot be readily obtained on pellets pressed from these powders. This communication describes results of STM analyses of pellets pressed from graphite and TaS2 powders.
11. Experimental Section SP1grade graphitepowder (UnionCarbide, 99.99996,average particle diameter = 100 pm), TaS2 (Aesar, 99.896, particle diameter < 44 pm), and MoS2 (Aldrich, 99%, particle diameter < 1 pm) were used as received and pellets were pressed in a standard hydraulic press. Teflon or highly-polished stainless steel diskswere placed between the metal die facesand the powder to produce smooth pellet faces. When some pellet surfaces were imaged at high resolution, the tip did occasionallycrash into the surface, but this was not a significant obstacle when care was taken to first locate a flat and smooth area under low resolution. Overall, no significant experimental difficulties associated with 0 1992 American Chemical Society
370 Langmuir, VoZ. 8, No. 2,1992
Letters
Figure 3. High-resolution STM image (25 A X 25 A) of an SPl pressed pellet surface showing the hexagonal pattern. Scan conditions were tip bias = -45 mV and current 2.4 nA. The image is software-zoomed but not filtered.
Figure 2. High-resolution STM images (50 A X 50 A) of (a) an SPl pellet, and (b) HOPG. Scan conditions for both images were tip bias = -45 mV and current = 1.8 nA. The images correspond to software-zoomed, but unfiltered, data.
a rough sample surface were observed. HOPG samples (Union Carbide, Advanced Ceramics Division, Z9B grade) were also imaged for comparative purposes. An HOPG sample was generally imaged immediately before and after the pressed powders to ensurethat the tip was not damaged or altered during the experiment. STM analyses were performed on a Nanoscope I (Digital Instruments,Inc., Goleta,CA)that had been modified to operate with an in-house data collection and image processing system. CommercialPt/Ir alloytips (DigitalInstruments)wereemployed. Experiments were all run in air on untreated surfaces. As the data obtaineddo not appear time dependent,and similar results are obtained on freshly cleaved HOPG surfaces,it was concluded that surface contamination was not a limitation in these experiments. X-ray diffraction data were obtained using a Siemens D5000 powder diffractometer at slow scan rate (0.6O 26/min) and an Enraf-NoniusGuinier camera. 111. Results and Discussion Figure 1 shows a topologically complex region imaged by STM at low resolution. The image is representative of those obtained and contains smooth regions suitable for high-resolution imaging. The dimensions and nature of the atomic structure near the grain boundaries are of considerable interest and currently under investigation but are not further considered in this report. Highresolution images of smooth areas on the pressed graphite powder and HOPG (for comparison) are shown in Figure 2. As can be seen, the images obtained on the pressed
Figure 4. (a) Low-resolution (750 A X 750 A) and (b) highresolution(50 A X 50 A) imagesof a TaS2pellet. Scan conditions were (a) tip bias = -50 mV, current = 0.8 nA and (b) tip bias = -50 mV, current = 1.3 nA. The images are software-zoomed but not filtered.
graphite powder are similar to those from HOPG at high resolution. These images are similar to those obtained by previous workers for HOPG15and show the trigonal pattern (15)Liu, C.; Hsiangpin, C.; Bard, A. Langmuir, 1991, 7, 1138, and references therein.
Letters typically observed on the graphite surface. The alternate honeycomb pattern described previously by others,16where all atoms within the graphite layer are imaged, has also been observed in both high and low resolution on the graphite pellet surface. A high-resolution image of the honeycomb pattern is shown in Figure 3. The degree of preferred orientation in these pressed graphite powder samples can be observed by the relative intensities of the (hkO)and (001) reflections. The results obtained by X-ray diffraction (graphite: 1m2/11m, theoretical value = 29, experimental value = 600) indicate that nearly all the crystallites in the pressed pellets are oriented with the c axis normal to the pellet surface. The lattice plane imaged by STM therefore correspondsto the basal plane as in the case of HOPG. Images from the TaS2 powder at low and high resolution are shown in parts a and b of Figure 4, respectively. Figure 4a reveals a broad flat region similar to those observed on single crystal surfaces, and Figure 4b shows, in atomic resolution, the hexagonal array of surface S atoms in the basal plane. As in the case of the graphite pellets, X-ray diffractometer data confirm that the vast majority of crystallites are aligned with the c axis normal to the pellet face. The spacing between S atoms in the image is near 3.1 A, close to the 3.3 A which has been determined crystallographically in bulk TaS2.17 X-ray powder data obtained by both Guinier and diffractomtter methods (16) Buckley, J.; Wragg, J.; White, H.; Bruckdorfer, A.; Worcester, D. J. Vac. Sa. Technol., B 1991, 9, 1079.
Langmuir, Vol. 8, NO.2, 1992 371 indicate that the sample is comprised primarily of the 2H-TaSz phase (a = 2 X 6.04 A, c = 3.32 A), which does not exhibit charge density waves at ambient temperat ~ r e .The ~ ~long-range, ~ periodic changes in tunneling current, which have been associated with charge density waves: were not observed in these experiments. Pellets pressed from MoSz powder could not be imaged by these STM methods. It is presumed that this is due to the relatively large surface resistance (>lo MQ) exhibited by these pellets.
IV. Conclusions The data obtained demonstrate that STM images may readily be obtained from pressed powders of graphite and TaS2. Work is now ongoing to image other pressed powders, including those of other layered transition metal sulfides, and intercalation compounds of some of these layered hosts. In addition, the nature of the grain boundaries observed in these pressed pellets is being explored. Acknowledgment. M.M.L. acknowledges the support of an NSF Starter Grant (CHE-9020375)and PRW funds for purchase of the STM from the NIH Small Instrumentation Grant Program. Registry No. TaSz, 12143-72-5; graphite, 7782-42-5. (17) Gamble, F.J. Solid State Chem. 1974,9, 368.