The small water droplets which adhere the material surface cause corrosion and photo catalytic reaction. So it is very important to study the behavior of water droplet. The relation between the humidity and the thickness of water layer is studied by using elipsometry1), but the detail of the water layer (shape ,thickness and distribution) is unknown by using elipsometry. Recently Hu2) et al. succeeded to observe the droplet of KOH solution by SPFM (Scanning Polarization Force Microscopy) method where the metal coated tip is located about 40 nm from the specimen and the voltage is applied to induce the electrostatic force. Modifying this method, I succeeded to observe the droplet of pure water by non-contacting AC AFM mode. The average size of water droplet was about 20 nm in width and 1.5 nm in height. The water droplets and water film ware observed on both graphite and mica. The behaviors of these water droplets and film are different from large size of water and these water droplets do not evaporate more than several days.

In order to get the image of very thin film or small droplets the specimen surface should be very clean and flat. Graphite and mica satisfy these conditions, so we chose graphite (hydrophobic) and mica (hydrophilic) as the test materials. The test was done at 296 k and humidity of 40% in air. In order to distinguish the dust in air from water droplets, following four methods were used for graphite specimen and two methods (1 and 3) for mica specimen: 1. Observation at the surface freshly created by peeling off the surface, 2. Observation at the surface peeling off the surface and sprayed by HFC134a for removing dust, 3. Observation at the surface peeling off the surface and pouring pure water (deionized and distilled water) and sprayed by HFC134a. 4. Observation at the surface peeling off the surface and pouring pure water and dried naturally (since graphite repels water, no water can be observed before drying). The tests were done repeatedly at least 3 times to confirm the reproducibility. The AFM (Atomic Force Microscopy) observation was done within 5 minutes after peeling off the surface to avoid the adhesion of dust in air. The AFM device used was SP3700(SEIKO) and the scan area was 500 nm to 5000 nm and the scan speed was 1 mm/s to 2.5 mm /s. The AFM tip was gold coated Si tip, the lever constant is 1.8 nN/m and the oscillation frequency is about 25 kHz.

The AFM mode employed was AC non-contacting mode where the oscillation width was about 10 nm. The effect of bias on the force curve at graphite is shown in Figure 1. The average decay of the amplitude at non-contact region becomes large as the bias increases. The accuracy of the observation increases as the decay increases. This means that the accuracy of AC non-contacting AFM observation is low when no bias is applied while the accuracy increases by applying the bias. The decay of amplitude becomes large as the distance form the sample approaches to zero. We used the bias of 0.5 V to 2.5 V and the distance from the sample few nm to get the image.

Mono-layer and multi-layer step were observed on the AFM image of the graphite surface freshly created by peeling off the surface, but no dust was observed. The similar images were observed on the graphite surface freshly created by peeling off the surface and sprayed by HFC134a. This indicates that the spraying does not make any droplet on the graphite surface. Figs. 2 show the AFM image of the graphite surface peeling off the surface and pouring pure water. In Fig. 2(a) the surface was sprayed by HFC134a and In Fig. 2(b) and Fig. 2(c) the surfaces were dried naturally. In these cases we could not get a good image without applying bias, so bias of 0.5 V to 2.5 V was applied. When surface was sprayed by cleaner gas the density of water droplets depended on the time spent for pouring water and the delay for spraying. A number of water droplets composed of smaller water droplets ( may be water cluster) were observed and they tended to gather at step when the density of water droplets was low. The size of water droplets adhered at step were usually smaller than the other droplets. While when the surface was dried naturally the density of water droplets was very high. Sometimes the surface was covered with water film of about 0.6 nm in thickness as shown in Fig. 2(c). These water droplets were very stable and did not evaporate more than several days. Fig. 3 shows the formation of very thin water film of about 0.3 nm in thickness (may be mono layer) at slow scanning (0.5 mm /s) area. The mechanism of formation of the water layer at slow scanning area is unknown. Once water film is produced the shape of water film was always changed while the location of water droplets did not change. These water droplets were easily moved. Figs. 4 show the movement of water droplets by contact mode scanning. The force applied during contact mode scanning was 1.8 nN. The contact mode scannings were done only in the middle region of the image. The water droplets were moved by scanning and the small water droplets join together to become a big droplets. The water film could not removed perfectly by this scanning as shown in Fig. 4(b). The sizes of small water droplets (may be water cluster) were between 15 nm and 25 nm in width and 1 nm and 2.5 nm in height as shown in Figs. 2. The tip used was the gold coated one and its radius of curvature R was around 20 to 30 nm. If we assume the shape of the water droplet is hemisphere, the expanding ratio of image by the tip is expressed as (2R/r+1)1/2. Where r is the radius of curvature of water droplet. Using this equation by assuming the tip radius of curvature is 25 nm, the expanding ratio of width becomes 5.1 when the height of water droplet is 2 nm. This indicates the hemisphere shape of water droplet of 2 nm in height is observed as the water droplets of 20 nm in width. From this result there is a possibility that the width of these water droplets is around 4 nm.

In the case of mica no step and no dust were observed on the surface freshly created by peeling off the surface. Figs. 5 show the AFM image of the mica surface peeling off the surface and pouring pure water and sprayed by HFC134a. In most cases surfaces were covered by water film of about 1 nm thickness. The size of water droplets on mica was almost the same as that on graphite. The difference of morphology from graphite was the existence of thick water film.