2846
Langmuir 1998, 14, 2846-2849
Role of Surface Perfection in Chemical Force Microscopy Rachel McKendry, Maria-Elena Theoclitou, Chris Abell, and Trevor Rayment* Department of Chemistry, Cambridge University, Lensfield Road, Cambridge CB2 1EW, U.K. Received April 11, 1997. In Final Form: February 26, 1998 In this paper we investigate the effect that the quality of the gold substrate has upon the distribution of adhesion force measurements in chemical force microscopy. Gold samples have been prepared by two protocols which give either predominantly single-crystal Au(111) or polycrystalline gold films on mica. Gold-coated tips and surfaces were functionalized with self-assembled monolayers terminating in carboxylic groups, and more that 500 force-distance plots were taken. Analysis of these plots revealed a 50% narrower distribution of adhesion forces between monolayers prepared on single-crystal surfaces and monolayers prepared on polycrystalline gold, when measurements were repeated at a single point. However when measurements were taken over a 1-µm2 area, the distribution of adhesion forces was similar for both samples. An explanation for this may lie in the size of the domains on the gold surface relative to the contact area of the tip.
Introduction Chemical force microscopy (CFM), which combines atomic force microscopy (AFM) with chemical sensitivity, has been developed over the last five years.1 The technique is feasible, since the force resolution of the AFM is sufficient that the intermolecular forces between chemical groups attached to the surface and the probe may be measured directly as a function of separation distance. Furthermore the surface topography and lateral friction forces may be imaged simultaneously. This enables image contrast to be interpreted in terms of the chemical interaction between specific functional groups on the tip and those on the surface. The principal potential advantage of this technique over others for measuring intermolecular forces is its speed combined with the ability to map the spatial arrangement of ligands on the surface. Frisbie and co-workers were the first to demonstrate the correlation between adhesive interactions and friction force images by the use of AFM with derivatized tips on patterned surfaces.1 They studied hydrophilic and hydrophobic interactions between tips and surfaces functionalized by self-assembly monolayers (SAMs) of alkanethiols terminating in either methyl or carboxyl groups. This work built upon studies using surface force apparatus2,3 and force-distance experiments in which the AFM tip and the surface were modified with alkyltrichlorosilanes of different lengths and the intermolecular forces were measured.4 This approach has since been applied to alkanethiols terminating in a range of different functionalities,5,6 polymers7,8 and biological systems,9,10 in* To whom correspondence should be addressed. (1) Frisbie, C. D.; Rozsnyai, L. F.; Noy, A.; Wrighton, M. S.; Lieber, C. M. Science 1994, 265, 2071-2074. (2) Israelachvili, J. Intermolecular and surface forces; Academic Press: New York, 1992. (3) Yoshizawa, H.; Chen, Y.-L.; Israelachvili, J. J. Phys. Chem. 1993, 97, 4128-4140. (4) Nakagawa, T.; Ogawa, K.; Kurumizawa, T.; Ozaki, S. Jpn. J. Appl. Phys., Part 2 1993, 32, L294-L296. (5) Han, T.; Williams, J. M.; Beebe, T. P. Anal. Chim. Acta 1995, 307, 365-376. (6) Sinniah, S. K.; Steel, A. B.; Miller, C. J.; Reuttrobey, J. E. J. Am. Chem. Soc. 1996, 118, 8925-8931. (7) Akari, S.; Schrepp, W.; Horn, D. Langmuir 1996, 12, 857-860. (8) Eastman, T.; Zhu, D. M. Langmuir 1996, 12, 2859-2862. (9) Browning-Kelley, M. E.; Wadu-Mesthrige, K.; Hari, V.; Liu, G. Y. Langmuir 1997, 13, 343-350. (10) Stuart, J. K.; Hlady, V. Langmuir 1995, 11, 1368-1374.
cluding biotin and streptavidin,11-17 cell adhesion proteoglycans,18 oligonucleotides19,20 and nucleotide bases.21 Statistical methods have been developed to analyze adhesion force data5,22 based on the assumption that this force arises from the rupture of a discrete number of chemical bonds, and the distribution of such measurements will follow a Poisson distribution. As part of a study of intermolecular interactions via CFM, we chose to confirm our methodology by repeating several of the experiments of Frisbie1,23 involving the tip and the surface functionalized with alkanethiols terminating in methyl or carboxyl groups. While the pattern of our results was in agreement with Frisbie’s data, it was found that the distribution of mean adhesion force measurements was much smaller. The only discernible difference between the experimental procedures appeared to be in the preparation of the gold substrates. Frisbie’s gold samples were prepared at room temperature under the condition of moderate vacuum and were described as “polycrystalline” whereas our samples tend to be gold single crystals. It is well-known that a good substrate for scanning probe microscopy (SPM) studies should be atomically flat over (11) Lee, G. U.; Kidwell, D. A.; Colton, R. J. Langmuir 1994, 10, 354-357. (12) Florin, E. L.; Moy, V. T.; Gaub, H. E. Science 1994, 264, 415417. (13) Moy, V. T.; Florin, E. L.; Gaub, H. E. Colloids Surf., A 1994, 93, 343-348. (14) Moy, V. T.; Florin, E. L.; Gaub, H. E. Science 1994, 266, 257259. (15) Ludwig, M.; Moy, V. T.; Rief, M.; Florin, E. L.; Gaub, H. E. Microsc. Microanal. Microstruct. 1994, 5, 321-328. (16) Allen, S.; Davies, J.; Dawkes, A. C.; Davies, M. C.; Edwards, J. C.; Parker, M. C.; Roberts, C. J.; Sefton, J.; Tendler, S. J. B.; Williams, P. M. FEBS Lett. 1996, 390, 161-164. (17) Hlady, V.; Pierce, M.; Pungor, A. Langmuir 1996, 12, 52445246. (18) Dammer, U.; Popescu, O.; Wagner, P.; Anselmetti, D.; Guntherodt, H. J.; Misevic, G. N. Science 1995, 267, 1173-1175. (19) Lee, G. U.; Chrisey, L. A.; Colton, R. J. Science 1994, 266, 771773. (20) Florin, E. L.; Rief, M.; Lehmann, H.; Ludwig, M.; Dornmair, C.; Moy, V. T.; Gaub, H. E. Biosens. Bioelectron. 1995, 10, 895-901. (21) Boland, T.; Ratner, B. D. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 5297-5301. (22) Willliams, J. M.; Han, T.; Beebe, T. P., Jr. Langmuir 1996, 12, 1291-1295. (23) Noy, A.; Frisbie, C. D.; Rozsnyai, L. F.; Wrighton, M. S.; Lieber, C. M. J. Am. Chem. Soc. 1995, 117, 7943-7951.
S0743-7463(97)00374-0 CCC: $15.00 © 1998 American Chemical Society Published on Web 04/26/1998
Role of Surface Perfection in Chemical Force Microscopy
a large area and that production of a smooth gold surface is critically dependent on the nature of underlying substrate and the precise conditions of preparation. There have been many publications describing the conditions necessary to form atomically flat, high-quality Au(111) films.24-26 The present consensus is that mica is the best substrate and that good epitaxial growth of gold on mica is promoted by evaporation onto a heated substrate under UHV conditions, followed by annealing the samples at high temperatures. The latter is an important step to allow gold islands to coalesce to form large atomically flat domains. The structure of the SAM is strongly dependent on the crystallographic face of gold in contact with the thiol solution. It has been shown that alkanethiols form a hexagonal structure on Au(111) and a square lattice structure on Au(100).27-29 The density of defects in SAMs has also been found to depend on the thickness of the gold film but is constant for films greater than 40-nm thick.30 This paper reports our studies on the role that the quality of the gold substrate has upon the distribution of adhesion force measurements in CFM experiments by comparing CFM data obtained from gold films prepared at room temperature with those from films prepared by heating prior to and after deposition. Experimental Section The atomic force microscope used in these studies was manufactured by East Coast Scientific, Cambridge, U.K. All force-distance measurements and images were taken in a liquid cell under ethanol (HPLC grade, Aldrich, U.K.). We used silicon nitride cantilevers with a square pyramidal tip, which had nominal spring constants of 0.03 N/m (Park Scientific Instruments). The spring constant of the modified tips was determined by measurement of their resonant frequencies,31,32 and it was found to be 0.04 N/m. Mica substrates were cut to an appropriate size, typically 1.5 cm2, and cleaved with a razor blade to produce a fresh surface just prior to placement on a heated stainless steel plate within the vacuum chamber. Gold films were prepared by thermal evaporation in a Balzers UTT 400 UHV thin film preparation system for which the base pressure is less than 10-8 mbar. Prior to gold deposition the chamber was baked at 160 °C for at least 12 h, to ensure high-quality operational vacuum. High-purity gold wire (99.99% purity, Advent Chemicals, U.K.) was evaporated at pressures less than 10 -7 mbar. All AFM tips and mica substrates were coated with 1000 Å of gold at a rate of 10 Å/s. The thickness and deposition rate of the gold films were monitored by a quartz crystal oscillator. Gold films were prepared according to two different protocols. In the first, mica substrates were placed in the evaporation chamber and heated to 340 °C overnight. Samples were then coated with 1000 Å of gold and annealed at this temperature for 5 h after deposition. Analysis of these samples by AFM showed the surface had a rms roughness of 5 Å, over a scan area of 1 µm2. Transmission XRD shows that material deposited under these conditions is predominantly single-crystal Au(111).33 (24) Hwang, J.; Dubson, M. A. J. Appl. Phys. 1992, 72, 1852-1857. (25) Golan, Y.; Margulis, L.; Rubinstein, I. Surf. Sci. 1992, 264, 312326. (26) Derose, J. A.; Lampner, D. B.; Lindsay, S. M.; Tao, N. J. J. Vac. Sci. Technol., A 1993, 11, 776-780. (27) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546-558. (28) Camillone, N.; Chidsey, C. E. D.; Liu, G.; Scoles, G. J. Chem. Phys. 1993, 98, 4234-4245. (29) Ulman, A. An Introduction to Ultrathin Organic Films from Langmuir-Blodgett to Self-Assembly; Academic Press: Boston, 1991. (30) Zhao, X. M.; Wilbur, J. L.; Whitesides, G. M. Langmuir 1996, 12, 3257-3264. (31) Sarid, D. Scanning Force Microscopy with Applications to Electric, Magnetic and Atomic Forces; Oxford University Press: New York, 1994. (32) Sader, J. E.; Larson, I.; Mulvaney, P.; White, L. R. Rev. Sci. Instrum. 1995, 66, 3789-3798.
Langmuir, Vol. 14, No. 10, 1998 2847 Table 1 average resolution adhesion standard (percentage force (nN) deviation deviation)
sample Annealed Gold measurement at single point measurement over surface Polycrystalline Gold measurement at single point measurement over surface Frisbie et al.23
2.3 2.1
0.4 0.6
17 29
2.1 2.1
0.9 0.8
43 38
2.3
0.8
35
In the second method for preparing gold films, the mica samples were not heated at any stage of preparation but were otherwise treated exactly the same as annealed gold. Analysis by AFM showed a rougher, textured surface with a rms roughness of 12 Å. Transmission XRD scans showed these samples were polycrystalline. Levers were coated with 1000 Å of gold at room temperature, since heating was found to cause these bimetallic strips to curl. All freshly coated tips and mica substrates were removed from the evaporator and immersed in a 1 mM solution of HS(CH2)15COOH in ethanol overnight. The next day they were removed from solution, rinsed in excess ethanol, dried under nitrogen, and used immediately.
Results Between 500 and 600 force-distance plots were taken consecutively over approximately a 3-h time period on each sample. Adhesion force measurements were taken at one particular point on the COOH-functionalized surface, for both annealed and polycrystalline gold. For comparison, force-distance plots were taken randomly over an area of 1 µm2. The distributions of adhesion forces for the different samples can be seen in Figure 1. Statistical analysis of measurements taken at one point on the surface showed the average adhesion force for annealed gold was 2.3 ( 0.4 nN. This corresponds to an experimental adhesion force resolution of 17%. Whereas polycrystalline gold samples gave an average adhesion force of 2.1 ( 0.9 nN, which is a resolution of 43% (Table 1), CFM images were taken over an area of 1 µm2 (Figure 2) and adhesion force measurements were taken at random points over this surface. The average COOH/COOH adhesion force on annealed gold was 2.1 ( 0.6 nN, which is a 29% resolution, and that on polycrystalline gold was 2.1 ( 0.8 nN, a 38% resolution. These results show that if measurements are made at a single point on annealed gold, the distribution of adhesion forces is very narrow compared with that of polycrystalline gold. However what is surprising is that if measurements are made over a 1-µm2 area, the distribution from both samples is quite similar. The reason for this is unknown, but a likely explanation for the difference may lie in the crystalline order present in the SAM for both samples. It is known that the domain size on single-crystal gold is approximately 10 nm34 but is much smaller in polycrystalline gold, possibly 1-2 nm.35 These figures should be compared with the contact area between the tip and the surface, which is estimated using the JKR theory to be about 3 µm2.23 SAMs are known to form tilted close-packed layers28 with different orientations in neighboring domains. The effect of relative orientations of the SAMs on the tip and the surface upon the adhesion force is not known but may be a significant factor. (33) Ramadan, A. R. An X-ray Diffraction Study of the Solid-Liquid Interface. Ph.D. Thesis, Cambridge, U.K., 1997. (34) Poirier, G. E.; Tarlov, M. J. Langmuir 1994, 10, 2853-2856. (35) Rayment, T.; Hammond, D. Unpublished work.
2848 Langmuir, Vol. 14, No. 10, 1998
McKendry et al.
Figure 1. Histograms to show the distribution of COOH/COOH adhesion forces derived from force-distance measurements: (a) annealed gold, measurements taken at a single point on the surface; (b) annealed gold, measurements taken over a 1-µm2 area; (c) polycrystalline gold, measurements taken at a single point on the surface; (d) polycrystalline gold, measurements taken over a 1-µm2 area.
Figure 2. CFM images with COOH on the tip and the surface where the underlying substrate is (a, left) annealed gold or (b, right) polycrystalline gold. The scan area is 1 µm2.
We suggest that a narrow distribution at a single point is caused by the interaction between the tip and one domain upon the surface. However if measurements are made at a number of points, then the tip will sample a range of different orientations and hence different adhesion forces. In the case of polycrystalline gold the domain size is comparable with the contact area and so thermal
drift alone would be sufficient to effect a large change in the area sampled by the tip and hence lead to a large fluctuation in the measured adhesion forces while measuring at apparently a single point. Another contributing factor to the fluctuation of adhesion forces is the surface morphology. The rms roughness is not large for either sample, being 12 Å for polycrystalline
Role of Surface Perfection in Chemical Force Microscopy
gold and 5 Å for annealed gold surfaces. However polycrystalline gold has a “cobbled” appearance and the maximum slope of the surface is much greater for these rougher samples. This means that the surface will approach the tip at different angles and therefore sample a different part of the tip. More work will be required to investigate the dominant factors which determine the distribution of adhesion force measurements. We have applied statistical methods used by other workers5,22 to analyze sets of adhesion data taken at single points on the surface to our distributions. A plot of the variance versus the mean adhesion force, for 11 data sets of 50 measurements, produces a correlation coefficient of 0.7. If we follow this method of analysis, we can obtain estimates of single-bond rupture forces of 0.1-0.2 nN, thus involving 14 to 18 bonds. However at this stage we are not sure how meaningful this analysis is in view of the fact that we do not observe a Poisson distribution; see Figure 1. Conclusions Our studies show that improving the surface quality of a gold substrate gives a marked increase in the reproduc-
Langmuir, Vol. 14, No. 10, 1998 2849
ibility of adhesion measurements in CFM studies. It suggests that more attention should be given to the nature of the substrate if CFM is to become a useful routine tool in the study of intermolecular forces. This is exemplified in our recent work, where we have shown that it is possible to discriminate between simple chiral enantiomers patterned on a surface using a chiral tip.36 This highly subtle discrimination was only possible on annealed gold surfaces.
Acknowledgment. We acknowledge support from Zeneca for a studentship (R.M.) and EPSRC for postdoctoral support (M.-E.T.) and provision of the AFM. We wish to thank Mark Welland and Sean O’Shea, Cambridge, ECS Ltd, and Department of Engineering, Cambridge, for fruitful discussions and technical assistance. LA970374E (36) McKendry, R.; Theoclitou M.-E.; Rayment, T.; Abell, C. Nature 1998, 391, 566-568.