Direct Imaging of Meniscus Formation in Atomic Force Microscopy

Aug 4, 2005 - Environmental scanning electron microscopy was used to image meniscus formation between an AFM tip and a surface. At high relative ...
3 downloads 0 Views 134KB Size
8096

Langmuir 2005, 21, 8096-8098

Direct Imaging of Meniscus Formation in Atomic Force Microscopy Using Environmental Scanning Electron Microscopy Brandon L. Weeks* and Mark W. Vaughn Department of Chemical Engineering, Texas Tech University, Lubbock, Texas 79409

James J. DeYoreo Lawrence Livermore National Laboratory, Livermore, California 94550 Received May 5, 2005. In Final Form: July 11, 2005 Environmental scanning electron microscopy was used to image meniscus formation between an AFM tip and a surface. At high relative humidity, 70%-99%, the meniscus formed is 100 to 1200 nm in height, orders of magnitude larger than predicted by the Kelvin equation using spherical geometry. The height of the meniscus also demonstrates hysteresis associated with increasing or decreasing relative humidity.

Transport through a liquid meniscus has been proposed as one of the possible mechanisms for ink transfer during dip-pen nanolithography (DPN).1,2,3 For this mechanism, the size of the meniscus is critical for predictable results. However, direct determination of the meniscus shape has not been performed experimentally. Most approaches to determining the meniscus shape have been indirect methods such as analytical solutions to the Kelvin equation4 or computer simulations.5 For capillary condensation, the Kelvin equation can be used to determine the radius of the meniscus formed in porous materials. Predictions of the height of the meniscus formed between an atomic force microscope (AFM) cantilever tip in contact with a surface using the same approach suggest that the maximum height of the meniscus formed is on the order of nanometers.6 However, one problem is that simple geometries have been used to determine the meniscus shape. For the AFM tip, the end is often treated as a sphere with a sub-20 nm radius.7 This simplified approach may be valid when the meniscus is much smaller than the radius of the tip; however, if the meniscus height is larger, then the approach breaks down. Monte Carlo simulations of meniscus formation at the interface between an AFM tip and a substrate have also been performed, primarily by Jang et al.8,9,10 Their work assumes thermodynamic equilibrium to determine the meniscus shape based on tip/sample separation and tip radius for a range of hydrophobic and hydrophilic tips. * To whom correspondence should be addressed. E-mail: [email protected]. (1) Piner, R. D.; Zhu, J.; Xu, F.; Hong, S.; Mirkin, C. A. Science 1999, 283, 661. (2) Haaheim, J.; Eby, R.; Nelson, M.; Fragala, J.; Rosner, B.; Zhang, H.; Athas, G. Ultramicroscopy 2005, 103, 117. (3) Ginger, D. S.; Zhang, H.; Mirkin, C. A. Angew. Chem.-Int. Ed. 2004, 43, 30. (4) Adamson, A. W.; Gast, A. P. Physical Chemistry of Surfaces; Wiley: New York, 1997. (5) Jang, J. Y.; Schatz, G. C.; Ratner, M. A. Phys. Rev. Lett. 2004, 92, 085504. (6) Sheehan, P. E.; Whitman, L. J. Phys. Rev. Lett. 2002, 88, 15. (7) Xiao, X.; Qian, L. Langmuir 2000, 16, 8153. (8) Jang, J.; Schatz, G. C.; Ratner, M. A. J. Chem. Phys. 2002, 116, 3875. (9) Jang, J.; Schatz, G. C.; Ratner, M. A. Phys. Rev. Lett. 2003, 90, 156104. (10) Jang, J. K.; Schatz, G. C.; Ratner, M. A. J. Chem. Phys. 2004, 120, 1157.

Their results qualitatively demonstrate that the meniscus width decreases with increasing tip curvature, and the tip/sample separation plays a large role in determining the onset of meniscus formation. Interestingly, they suggest that the temperature dependence has only a modest effect compared to the effect of varying the relative humidity (rh). Although the models presented for meniscus formation add insight into the initial understanding, there is still little information on the actual geometry of the meniscus formed at an AFM tip in contact with a surface. Here we use environmental scanning electron microscopy (ESEM) to obtain direct images of meniscus formation as a function of relative humidity.11 We show that the height of the meniscus is far greater than that predicted either by direct application of the Kelvin equation to a simplified spherical geometry or by the simulations performed to date. ESEM can image in various atmospheres (water vapor, nitrogen, air, etc.) at a relatively high pressure compared to standard scanning electron microscopy. The instrument works at these higher pressures by utilizing a special detector such that the gas molecules amplify the secondary electrons emitted from the surface being imaged. By using a Peltier cooling stage and varying the pressure of water vapor within the instrument, the relative humidity can be controlled and calculated precisely allowing images to be collected from 0 to ∼100% rh. The instrument used for these studies was an FEI XL30 field emission ESEM. The cantilevers imaged were supplied by Thermo Microscopes12 model MSCT-AUHW Sharpened Microlevers with tip “A”. The cantilevers were mounted on silicon or gold coated silicon substrates by gluing the glass chip directly on the surfaces with epoxy. Care was taken to be sure no adhesive was near the tip and was confirmed by ESEM micrographs. Images were collected with the cantilever parallel to the electron gun. This arrangement allowed direct imaging of the meniscus formation at the tip/substrate interface. Images were collected at 5°C and the rh was controlled by varying the water vapor pressure from 1.0 Torr (15% rh) to saturation (6.3 Torr). The saturation point is clearly visible as water (11) Schenk, M.; Fu¨ting, M.; Reichelt, R. J. Appl. Phys. 1998, 84, 4880. (12) Now Veeco Instruments, Santa Barbara, CA.

10.1021/la0512087 CCC: $30.25 © 2005 American Chemical Society Published on Web 08/04/2005

Letters

Langmuir, Vol. 21, No. 18, 2005 8097

Figure 1. Sequence of images collected at various relative humidity. All images were collected at 5 °C, 15.0 kV accelerating voltage, at 35 000×. The humidity was varied by decreasing the pressure of the water vapor from 6.4 Torr down to 1 Torr: 40% rh, 2 Torr; 60% rh, 3.2 Torr; 99% rh 6.4 Torr.

Figure 2. Measured meniscus height of a silicon nitride (SiN) cantilever in contact with both a silicon and gold surface. The difference in measured heights is likely due to the difference in the wettability of the silicon and gold surfaces.

Figure 3. Hysteresis measurements of the forward (0-99% rh) and retreat (99%-0 rh) for a SiN cantilever tip in contact with a silicon substrate. On the forward measurements, a meniscus is not observed until ∼70% rh, whereas on retreat, the meniscus is clearly observable until 40% rh. Images were allowed to equilibrate for a minimum of 10 min.

condenses over the entire surface of the sample being imaged. Figure 1 shows typical images taken at constant temperature while varying the pressure starting at 99% rh down to 15% rh (series from Figure 3). The meniscus heights were determined visually by measuring the perpendicular distance from the substrate surface to the condensate/cone contact line. The contact line is where the change in angle between the meniscus and the cantilever tip is observed. We also collected images by maintaining a constant pressure within the instrument and varying the temperature, but the time require to reach thermal equilibrium took over 10 min, much longer than by varying the pressure with a constant temperature. Results obtained by changing the temperature were similar to those collected by varying the pressure. Due to the longer equilibration time needed to change the temperature, results will be presented at constant temperature. Figure 2 shows the meniscus height vs relative humidity for both silicon and gold substrates. Data were collected by starting at 15% relative humidity and slowly increasing

the humidity up to 99%. No meniscus was observed until ∼70% rh with a gold substrate and 80% rh with a silicon substrate. Even though a meniscus was not observed below 70%, we cannot determine from this work whether it exists. Very small menisci would not be observable since the resolution limits of the ESEM under our imaging conditions are approximately 50 nm. Between 70% and 95% rh, the height of the meniscus appears to increase exponentially with humidity, but currently, we have no model to explain this dependence. The difference in the observed height for the gold and silicon substrates is likely due to the differences in wettability of the gold and silicon surfaces.13-15 One interesting observation is that the meniscus height exhibits hysteresis associated with growth (15-100% rh) or retreat (100-15% rh). Figure 3 shows the results from a tip in contact with a silicon surface for the growth followed by retreat of the meniscus. Each micrograph was collected such that the height of the meniscus was stable for at least 5 min. Although the growth of the meniscus shows an exponential dependence on rh, the retreat is linear over the range recorded. In addition, a stable meniscus could be observed at 40% rh. There is a strong possibility that the meniscus height will continue to decrease, but no changes were observed over a 10 min period. This suggests that if any changes were occurring they were very slow. Hysteresis has been observed in other systems related to the condensation and evaporation in confined geometries and is likely controlled, in part, by the radius of curvature of the meniscus.16,17 In the literature on DPN, the meniscus has been a source of controversy, particularly with regard to whether a large enough volume of water is formed to allow for the observed rates of molecular transport by bulk diffusion alone.6,18,19 Our results suggest that at high rh there is certainly enough water for the meniscus to play a role depending on the ink and substrate. Although the results leave the situation at low rh unresolved, they are none-the-less, quite consistent with the general trend observed in patterning experiments over a wide range of humidities. Such experiments have shown that at rh from ∼0-50% there is very little increase in the patterning rate of mercaptohexadeconoic acid (MHA) as the humidity is increased.20,21 However, as the humidity is increased above (13) Schneegans, M.; Menzel, E. J. Colloid Interface Sci. 1982, 88, 97. (14) Yakubov, G. E.; Vinogradova, O. I.; Butt, H. J. J. Adhesion Sci. Technol. 2000, 14, 1783. (15) Vinogradova, O. I. Langmuir 1998, 14, 2827. (16) Donohue, M. D.; Aranovich, G. L. J. Coll. Interface Sci. 1998, 205, 121. (17) Maeda, N.; Isrealachvili, J. N. J. Phys. Chem. B 2002, 106, 3534. (18) Rozhok, S.; Piner, R.; Mirkin, C. A. J. Phys Chem. B 2003, 107, 751. (19) Schwartz, P. V. Langmuir 2002, 18, 4041. (20) Weeks, B. L.; Noy, A.; Miller, A. E.; De Yoreo, J. J. Phys. Rev. Lett. 2002, 88, 255505.

8098

Langmuir, Vol. 21, No. 18, 2005

50%, the patterning rate increased significantly. At 72%, the rate increases by 2-fold, whereas at 99% rh, the patterning rate increased by over an order of magnitude compared to the low humidity data (0-50%).21 More importantly, the observed hystereses in meniscus size is critical to interpreting the patterning rates observed in DPN experiments. For example, Peterson et al. showed that the patterning rate at a given humidity was less if data was collected when rh is increased from 0 to 100% than if collected at the same humidity as the rh is decreased from 100%-0 (see Figure 4, ref 21). Moreover, Weeks et al. found a substantially higher ink transfer rate in the 40-60% rh range than observed by some others.20 However, their data was collected by reducing rh from an initially high value. The results presented here provide a possible explanation for the differences in transport rates obtained while increasing rh opposed to those obtained while rh is decreasing. In conclusion, we have demonstrated that ESEM can directly image a meniscus formed at an AFM tip in contact (21) Peterson, E. J.; Weeks, B. L.; De Yoreo, J. J.; Schwartz, P. V. J. Phys. Chem. B. 2004, 108, 15206-15210.

Letters

with a substrate. The height of the meniscus formed at high humidity is orders of magnitude larger than calculated using either the Kelvin equation or Monte Carlo modeling and displays significant hysteresis. Although the initial formation of the meniscus may be due to capillary condensation, when the cantilever tip is assumed to be a sphere, this alone does not predict the overall height. Other physical phenomena must be occurring to account for the additional water condensed. Whatever the source of this apparent anomaly as a consequence the meniscus appears to be large enough to participate in the transport process during DPN. Moreover, the hysteresis provides an explanation for previously observed differences in DPN transfer rate during experiments performed with variable humidity. Acknowledgment. The authors thank Jose Saleta (UCSB) for ESEM assistance. This work was performed under the auspices of U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under Contract No. W-7405-Eng-48. LA0512087