We have measured the contact angle 0 of He on cesium-coated tungsten plates as a function of temperature.We find that 0 decreases to zero at T = 2.0 K in agreement with the wetting temperature found on bulk Cs. At T = 0 K the contact angle is 48~1, significantly larger than the predicted value of approximately 30 . The energy of the interface between Cs and liquid He has a large temperature dependence. This suggests that there are low-lying excitations on the liquid helium surface at this interface. Indeed it appears that liquid "He at this interface is similar to that at a free surface. PACS numbers: 67.70.+n, 68.10.Cr, 68.45.Gd Liquid He wets most materials and was thought until recently to be a universal wetting liquid. However, theoretical calculations [1,2] have predicted that He will not wet some alkali metals below a certain temperature called the wetting temperature T . Subsequent experiments have indeed shown that He does not wet Cs [3-5] and Rb [6]. This has been demonstrated by showing that the adsorbed helium film on Cs or Rb for T ( T is atomically thin and only becomes macroscopically thick for T ) T . This arises from the fact that the substrate surface is always attractive to He, but, if it is less attractive than the liquid-liquid attraction, then only a few atomic layers of He at most can form at T ( T . This thin film can disappear as T~0 as found for Cs [5].Another important property of a liquid that does not wet a substrate is that it forms drops of macroscopic size with a finite contact angle O. This angle is in general temperature dependent for T ( T and goes to zero at T and remains zero at all higher temperatures where the liquid wets the substrate. In fact, a nonzero contact angle is the usual indication of nonwetting, for example, water drops on solid hydrocarbons and mercury on glass.Besides this visual indication of wetting and nonwetting, the contact angle has an important connection to the free energies of the three interfaces involved.These are oi, o, , and o,I, which are, respectively, between liquid-vapor, substrate-vapor, and substrate-liquid. Young's equation connects these to 0' as sv O sl cosH = While oI can be readily measured and is indeed well known for He [7], the other interface free energies are not known, and so a measurement of O(T) allows a. , -o. ,~= Ao. to be determined and deductions to be made for a. , and o. ,~. The first theoretical estimate for O(0) for He on Cs was =95 [1]. This was revised to =30 [8] when T was measured to be =2 K [3,4], which was half that estimated theoretically. These estimates were based on the assumption that Ao. was temperature independent so that O(0) is given by cos8(0) = Ao. /crt (0) with Ao. = o. t (T ). Since then there have been calculations of the temperature dependence of Atr [9], especially o. ,t, which show that it is rather small compared to that of a. t, (T), and so the assumption above appears to be a good first approximation. However, there is recent experimental evidence that o,I is a strong function of temperature and t...

Wyatt and Klier reply: In our Letter [1] we drew attention to the fact that the metastability of thick helium films on quench-condensed Cs substrates was not resolved. The situation is more complicated than the metastability expected for a first order wetting transition on a perfect substrate. This is clear from the fact that dry holes can be created in the helium films by a laser pulse and these holes do not expand causing the Cs to dry [2]. This shows that the thermal nucleation problem is irrelevant and that the pinning of the contact line is crucial. We also addressed the surprising phenomenon that these same Cs surfaces show a memory effect; that is, once the surface has been flooded with liquid helium and then the surface has apparently dried, a subsequent flooding only flows out to the previous flooded boundary. We proposed that stable micropuddles on a rough Cs surface could explain both the strong pinning of the contact line and the memory effect.We did not comment on the liquid on liquid system as it has no bearing on the problem of quench-condensed Cs. These Cs substrates are rough compared to a liquid, and it is the roughness that causes the pinning and makes their behavior not ideal. We showed that micropuddles on a rough surface can be thermodynamically stable. This is important for a long term memory effect. Stability is achieved by having pressure differentials between the vapor and the liquid in the micropuddle due to surface tension and van der Waals forces. The pressure differential across the free liquid surface is given by p 2 p 0 2s ly ͞R, and there is no conflict as suggested by Bonn et al. [3]; on the contrary, it is consistent and necessary.We drew an explicit distinction between cavities that can be filled by flooding and by capillary condensation. One of the new aspects in our paper was that micropuddles can form only by flooding. We envisage that cavities on thin Cs films are shallow and do not fill from the vapor phase. There is no evidence or suggestion that capillary condensation occurs on these Cs surfaces.Bonn et al. [3] suggest that the quench-condensed Cs surface is so extremely rough that it is "a spongelike porous medium" with connected liquid channels. We had not thought that most quench-condensed Cs were as bad as this. If it is, then the contact angle measurements made on them are profoundly modified by the surface topography and have little resemblance to the contact angle on a smooth surface. It is very misleading of Bonn et al.[3] to use these low values of the contact angle in their estimates of the pinning energies. Also such a surface invalidates the applicability of any theory that assumes a plane surface, such as the one referred to in their second paragraph. It seems unlikely to us that a connected filamentary superfluid film on a rough two-dimensional substrate can explain the memory effect. However, if such a film becomes disconnected, then it is similar to an extended micropuddle. The liquid isolation of the micropuddle is necessary for 279602-1 0031-9007͞01͞ 87...

No abstract

The influence of substrate roughness on the wetting scenario of adsorbed van der Waals films is investigated by theory and experiment. Calculating the bending free energy penalty of a solid sheet picking up the substrate roughness, we show that a finite roughness always leads to triple-point wetting reducing the widths of the adsorbed solid films considerably as compared to that of smooth substrates. Testing the theory against our experimental data for molecular hydrogen adsorbed on gold, we find quantitative agreement. DOI: 10.1103/PhysRevLett.88.055702 PACS numbers: 64.70.Hz, 67.70. +n, 68.08.Bc, 68.35.Rh Wetting of a solid substrate, exposed to a gas in thermodynamic equilibrium, is an ubiquitous phenomenon, with both fundamental aspects [1,2] and important applications [3 -5]. Microscopically, substrate wetting by a liquid film is caused by a strong substrate-particle attraction mediated by van der Waals forces. At present, an almost complete microscopic understanding of wetting on flat solid substrates is available [1,2,6] predicting the thickness of the liquid film as a function of the substrate-particle and interparticle interactions for given thermodynamic parameters such as temperature and pressure. The following basic theoretical predictions were confirmed by experiments using, e.g., noble gases [1] on different substrates: (i) For fixed thermodynamic conditions, the thickness of the wetting layer grows for increasing substrate-particle attraction. (ii) Complete wetting (i.e., a diverging thickness of the liquid layer) occurs if the substrate-particle attraction is stronger than the interparticle interaction and the thermodynamic conditions approach liquid-gas coexistence. The latter condition can be achieved only if the system temperature T is above the triple temperature T 3 . For T , T 3 , on the other hand, a solid film shows up near the sublimation line. Various experiments have shown [7][8][9][10] that the width of the solid layer always remains finite when approaching gas-solid coexistence. It is only near the triple point that a liquid layer on top of the solid sheet is formed, with a diverging width as the triple point is approached. This universal behavior is called "triple-point wetting."One major difference between a liquid and solid wetting layer is that a solid cannot relax the elastic compression caused by the substrate attraction as embodied in the (reduced) wall-particle Hamaker constant R. This fact is the basic ingredient in the traditional Gittes-Schick theory [11] of solid adsorption on flat substrates. It predicts that, for a particular value R R 0 of the substrate attraction, complete wetting is possible, while for R . R 0 , in contrast to liquid wetting, the thickness of the solid film ᐉ s decreases with increasing R. In this Letter we show that the key parameter governing adsorption of solid films is the substrate roughness rather than the elastic deformation caused by the substrate attraction. As a result of our theoretical analysis, a finite substrate roughness leads inev...

No abstract