pubs.acs.org/Langmuir © 2010 American Chemical Society
Imaging Ice-Like Structures Formed on HOPG at Room Temperature Omar Teschke* Laborat orio de Nanoestruturas e Interfaces, Instituto de Fı´sica, UNICAMP, 13083-970, Campinas, SP, Brazil Received May 24, 2010. Revised Manuscript Received September 16, 2010 In this work, ice was viewed at the nanoscale by scanning an atomic force microscopy tip over a highly oriented pyrolytic graphite (HOPG) surface in air. At low scan velocities, the tip exhibited stick-slip motion with a period of 0.13 nm corresponding to the scanner step; at higher velocities, the HOPG lattice and the periodicity of the ice were visible. A hexagonal structure with a 0.45 ( 0.04 nm periodicity was observed in which the distance between the second neighbors of the HOPG coincided with the distance of the first neighbors for the ice-like arrangement. Small water clusters were also nucleated with an ice-Ic structure (0.34 ( 0.03 nm), and thus, the ice layers consisted of extensive sets composed of arrangements of hexamers and tetramers.
1. Introduction The structural and chemical properties of thin water films adsorbed on different substrates have been the subject of theoretical and experimental investigations1,2 because they are useful for controlling the formation of nanoparticles, clarifying the mechanisms of surface catalysts and understanding the formation of water-surrounded proteins. Ice has rarely been viewed at the nanoscale;3,4 here, we report the interfacial structure of ice-like deposits on highly oriented pyrolytic graphite (HOPG) in contact with air at room temperature (25 °C). Ice nucleation on a surface presents an opportunity to observe heterogeneous ice nucleation unfold at the molecular scale. Thus far, the clearest insight into the molecular details of the initial stages of ice nucleation have been provided by experimental surface science techniques.5 HOPG provides a useful surface for atomic force microscopy (AFM) studies due to its flat cleavage and inert nature, which makes it possible to obtain images in air, liquid, and other environments with atomic resolution.6,7 Recently, Jinesh and Frenken8 using ultralow scan velocities and modest relative humidities reported a periodicity of 0.38 nm, but no images were observed. Here, for the first time, we report two-dimensional ice images generated by tip scanning water films on a HOPG substrate at a relatively high velocity (∼100 nm/s) and at low applied forces (∼4 nN). The images have sufficient lateral resolution to identify the atomic structure of the various ice-like deposits. We also discuss the need for high scanning velocities in high-resolution ice imaging. 2. Experimental Section AFM can be conducted in a constant-height or constant-force mode. In both modes, the height or the force applied to the sample is kept constant by a feedback loop that regulates the position of *E-mail:
[email protected]. (1) Thiel, P. A.; Madey, T. E. Surf. Sci. Rep. 1987, 7, 211–213. (2) Henderson, M. A. Surf. Sci. Rep. 2002, 46, 1–4. (3) Carrasco, J.; Michaelides, A.; Forster, M.; Haq, S.; Rava, R. Nat. Mater. 2009, 8, 427–428. (4) Teschke, O.; Valente Filho, J. F.; de Souza, E. F. Chem. Phys. Lett. 2010, 485, 133–137. (5) Michaelides, A.; Morgenstern, K. Nat. Mater. 2007, 6, 597–599. (6) Colton, R. J.; Baker, S. M.; Driscoll, R. J.; Youngquist, M. G.; Baldeschwieler, J. D.; Kaiser, W. J. J. Vac. Sci. Technol., A 1988, 6, 349–352. (7) Terashima, K.; Taniguchi, Y.; Yamaguchi, N.; Takamura, Y.; Yoshida, T. Thin Solid Films 1999, 345, 146–149. (8) Jinesh, K. B.; Frenken, J. W. M. Phys. Rev. Lett. 2008, 101, 036101–036108.
16986 DOI: 10.1021/la103227j
the sample during scanning. The scanning velocity is limited by the time constant (∼50 ms) of the piezoelectric translator and feedback circuits.4 In this work, all of the feedback loops were opened, and the image was obtained by plotting the deflection of the cantilever versus its lateral and up-and-down position. Consequently, there were no velocity limitations imposed by the piezoelectric translator or the feedback system. The detector voltage was registered as the output signal. The lateral force mode, which detects the tangential force along the scanning direction, and the internal sensor mode, which registers the force normal to the scanning direction, were recorded simultaneously; in the case of atomically flat surfaces, both images showed the same periodicity, and in this report, the image with the better signal-to-noise ratio is displayed. We used an atomic force microscope (model TMX2000, TopoMetrix, Veeco) in which a silicon nitrite (Si 3 N4) tip (Microlevers, Veeco, model MSCT-AUHW) with a spring constant of about 0.03 N/m was scanned over HOPG and mica surfaces. The radius of curvature of these AFM tips was approximately 5 nm. Each map of a sample surface consisted of 300 300 grid points. The scan velocity was optimized to obtain the best signal-to-noise ratio, resulting in a value of 200 nm/s for HOPG in air. The constant normal load was also optimized; a value of ∼4.0 nN was used throughout the experiments. The patterns shown herein correspond to the best signal-to-noise ratio images of various HOPG samples scanned at various regions and after a large number of cleavages. The displayed images were selected from a group of more than 100 images. HOPG and mica were cleaved with an adhesive tape in air, and then, the sample was immediately placed in the chamber in which the AFM head was enclosed. Samples with dimensions of 11 cm2 that were typically several tenths of a millimeter thick were used without any previous treatment. The AFM contact and lateral force images indicated that the surface was atomically flat without any cleavage steps over a few micrometers. The contrast in the lateral force images showed no variations, indicating that the freshly cleaved surface was free of contamination. The AFM head was enclosed in a plastic box. The relative humidity (RH) in the box was increased by evaporating water from a beaker or decreased by inserting a desiccant material and by flowing dry nitrogen gas. Because it is rather difficult to fix the RH at a given value (except at the ambient RH, which was ∼65% in our laboratory), we let the RH drift from an initial value by introducing a desiccant material. Generally, this drift was slow, on the order of tens of minutes. After reaching a new RH value with a variation of less than 5%, the imaging experiments were begun, and each took about 1 h.
Published on Web 10/08/2010
Langmuir 2010, 26(22), 16986–16990
Teschke
Article
Figure 2. Contact AFM image of HOPG scanned in air (35% RH at 25 °C), displaying the lateral force between a Si3N4 tip and a graphite substrate scanned over a distance of 5 nm at v = 250 nm/s; the graphite lattice periodicity of 0.24 is visible. The top inset shows a Fourier transform power spectrum of the image. Spots surrounded by circles correspond to a periodicity of 0.24 ( 0.03 nm. The right lower inset corresponds to an unfiltered image.
Figure 1. (a) Contact AFM imaging of mica scanned in air (35% RH at 25 °C) over a distance of 4 nm at a velocity of 250 nm/s, used as a calibration pattern for vertical and horizontal scans. (b) The same pattern shown in part a rotated by 12°. The effect of the noncoincidence of the lattice axis and the scanner direction, after 12° rotation (indicated by the direction of the lines in parts a and b), is shown by the FFT spectrum variation.
2.1. Calibration Procedure. One of the objectives of this work was to determine the crystalline structure of ice-like films, which requires a resolution on the order of ∼0.01 nm because the ice-like periodicity is on the order of a few tenths of a nanometer. The calibration of the vertical and horizontal piezoelectric transducers was carried out as follows. Topographic measurements of a standard calibration grid with known pitch size were performed, in our case, using a silicon test grid. After the sample was scanned and a topographic image was acquired, the calibration was performed by using the analysis tools to measure the grid periodicity. The new voltage applied to the scanner was calculated as follows: measured length Vnew ¼ Vold known length Then, recalibration was accomplished through direct editing of the scanner system file by resetting the x, y, piezo voltages based on the actual versus measured values.
3. Results and Discussion The standard calibration procedures described in the previous section were followed by a finer scale calibration procedure using a freshly cleaved mica sample, which was scanned in air at 35% RH. The internal sensor mode was used. The lattice parameters of mica are 0.52 and 0.9 nm; the scanners were recalibrated by comparing these values to the measured periodicity in the vertical and horizontal directions. After the recalibration procedure, a new image was registered. The result is shown in Figure 1. Using the vertical profile image (VPI), the lattice parameters of mica were measured to be 0.54 and 0.94 nm, showing errors of 4.4% for the vertical dimension (0.52 nm) and 4.64% for the horizontal dimension (0.9 nm). The distribution of measured values of the lattice parameter may also be evaluated by fast Fourier transform (FFT) techniques, as shown in the inset of Figure 1a. In order to evaluate the FFT calculation subroutine, the mica crystalline structure was scanned after rotation by an angle of ∼12° with respect to the Langmuir 2010, 26(22), 16986–16990
direction indicated by a full line in Figure 1a. The result is shown in Figure 1b. Observe that a similar structure is visible, except that it is rotated by ∼12° (the angle between the lines in Figure 1a and b). In the FFT pattern shown in the inset, the symmetric distribution of dots was altered. This difference is explained as follows: the FFT spectrum acquisition of the AFM images uses a different technique than that used for X-ray crystallography of bulk crystals, where an X-ray beam incident upon the crystal, typically more than a few millimeters thick and wide, generates a diffraction pattern formed by the reciprocal lattice vectors, which are experimentally determined and correspond to the dots in the inset of Figure 1b. Here, we measured the spatial periodicity of the crystal along the scanned direction using the measured profiles for each line; the power spectrum was derived by performing a FFT on each line and then normalizing the results of all of the lines. Consequently, the periodicity registered by the FFT is not equal to the real dimensions of the crystal when the crystalline structure and the scanning direction are not aligned. For a surface atomic distribution with a thickness of a few atoms, the projection of the atomic period along the scanned line the alignment of crystalline rows with the scanning direction is a difficult task because the crystalline axis varied along the sample (shown in Figure 3), which had a highly stressed structure, so the FFT spectrum was used to measure the lattice parameter only for image filtering. For an RH of 65%, water films formed on mica have been previously investigated.4 When the RH was decreased to values
