Langmuir 1987, 3, 1027-1034 The lithography was done under fluorocarbon grease (Research Organic/Inorganic Chemical Corp. Poly-Fluor Laen grease). The lithography method was similar to that used by Becker et al. to write on germanium.1° To start, two images of a gold surface with a characteristic feature were taken to determine the drift rate. In this case, from the small mound in the lower left corner the drift rate 5 Next the tip was positioned could be seen to be ~ 0 . A/s. in the center of the scan area and scanning stopped. The bias voltage was increased from the 0.1 V used for imaging until the feedback voltage, which was applied to the z-piezo to keep the tunneling current constant, jumped suddenly. This usually occured at 53 V with the sample positive with respect to the tip. After this lithography, which took about 20 s, the bias voltage was reduced back to 0.1 V and the same area was imaged twice more. We were able to produce similar holes in five out of five trials. Using a nonpolar imaging fluid was the key to doing lithography out of a vacuum chamber. Our previous trials under air, water, and aqueous solutions were unsuccessful in producing nanometer-scale features. In these trials we were unable to turn the bias voltage above 2 V. At these low-bias voltages the tunneling electrons do not have enough energy to break chemical bonds. Images obtained under different fluids are of varying quality. In an attempt to determine if images are effected by the fluid used, we imaged both a thin gold film on a
1027
glass substrate and a GaAs sample 5 times under each fluid. These results indicated that most, if not all, of the variation between images is due to variations in tunneling tips and differences between different regions on the sample. We did not observe any systematic variation in images obtained with different oils and greases.
Conclusions In summary, a STM operated with its sample and tip covered with nonpolar fluids can (1)achieve atomic resolution, (2) provide images of air-sensitive materials, and (3) be operated with a high enough tip-to-sample voltage to do lithography. Acknowledgment. We thank R. Sonnenfeld for providing and helping us prepare the G d s surfaces, B. Drake for helping design and build our STM, R. Becker, J. Golovchenka, and B. Swartzentruber for their inspiration on lithography, and W. Kaska for suggesting and supplying the fluorocarbon grease. This work was supported by the National Science Foundation Solid State Physics Division under Grant No. DMR 86-13486. The STM imaging system was built, in part, with support from the Office of Naval Research. Registry No. GaAs, 1303-00-0; An, 7440-57-5; graphite, 7782-42-5.
Associated Complex Formation in the Liquid-Vapor Interface: The Water-Ethanol and Water-tert -Butyl Alcohol Systems E. Tronel-Peyroz,* J. M. Douillard, L. Tenebre, R. Bennes, and M. Privat U A 330 "Physicochimie des SystGmes PolyphasW, CNRS, 34033 Montpellier, Cedex, France Received January 14, 1987. In Final Form: April 17, 1987 The Gibbs surface excess rzland the ellipticity coefficient p were measured at the liquid-vapor interfaces for two binary mixtures: water-ethanol and water-tert-butyl alcohol. The surface composition fluctuations were deduced and explained by assuming the existence of an associated complex in the transition layer. These associated complexes are correlated in a crude manner with the contact angle measured at the triple contact line of liquid-glass-vapor.
I. Introduction The transition layer between a binary liquid mixture and its vapor has been the object of many experimental studies; many of which are based mainly on measurement of the surface tension.14 More recently, it appeared that the description of this transition layer deduced from these results can be improved upon by use of ellipsometric measurements.&1° (1) Randles, J. E. B.; Behr, B. J. Electround Chem. 1972,35, 389.
(2)Eriksson, J. C.Ark. Kemi 1966,26,49. (3)Guggenheim, E. A.; Adam, N. K. Proc. SOC.London, A 1933,139, 218. (4)Ter-Minassian-Saraga,L. J. Chim. Phys. (Paris) 1960,57, 10. (5)Rusanov, A. I. In Progress in Surface and Membrane Science; Danielli, J. F., Ed.; Academic: New York, 1971; Vol. 4,p 57. (6)Tenebre, L. J. Phys. (Les Ulis, Fr.) 1977,38C5,123. (7)Tenebre, L.; Lize, A. C. R . Seances Acad. Sci., Ser. C 1976,282, 995. (8)Lize, A.; Tenebre, L. C. R . Seances Acad. Sci., Ser. C 1973,276, 975.
0743-7463/87/2403-1027$01.50/0
Different approaches"J2 have shown how it is possible to correlate the relative surface excess determined from the Gibbs relation and adsorption studies, with the ellipticity coefficient p if one assumes, as in Drude's the0ry,I3that the surface phase heterogeneity, normal to the surface, is entirely responsible for p . However, the ellipticity coefficient cannot be reduced to the Drude component alone. It also includes the following two contributions: The first is a contribution PTM due to surface roughness14116caused by the thermal motions of the adsorbed (9)Engelsen, D. D.;de Koning, B. J. Chem. SOC.,Faraday Trans. 1 1974,70, 1603. (10)Mang, H.; Findenegg, G. H. Colloid Polym. Sic. 1980,258,428. (11) Beaglehole, D. J. Chem. Phys. 1980,73,3366. (12) Tronel-Pepoz, E.;Bennes, R.; Douillard, J. M.; Privat, M. C. R . Seances Acad. Sci., Ser. 2 1984,299,1313. (13)Drude, P. Pr6cis d'optique; Gauthier-Villars: Paris, 1912;Vol. 2, p 43. (14)Beaglehole, D. Physica B+C (Amsterdam) 1982, 112B, 320.
0 1987 American Chemical Society
1028 Langmuir, Vol. 3,No. 6,1987
molecule^.^^-^^ This contribution is difficult to evaluate,20 and as a first approximation we will neglect it in this paper. The second is a contribution p A resulting from the anisotropy of the surface This anisotropy is caused by the mean orientation of the molecules displaying different polarizabilities along their principal axes. Consequently, the transition-layer optical dielectric constant is anisotropic, and p A cannot be neglected. Unfortunately, it is difficult to calculate this term a priori as it requires a knowledge of the exact orientation of each molecule in the surface layer. Since this information is not directly available, we resort to an alternative approach. In this article, then, we evaluate the order of magnitude of p A and ita variation with the composition for two liquid binary mixturevapor systems: water-ethanol at 25 "C and water-tert-butyl alcohol at 26 "C. I t is well-known that adsorption studies lead to the determination of adsorption isotherms and of composition fluctuation^.^^ Hence it is possible to calculate the Drude component of the ellipticity coefficient (pD) and to deduce the anisotropic contribution @ A ) by subtracting p D from p and assuming, as above, that PTM is negligible. We show the existence of correlations between the structures of the interphase and one wetting transition at the triple contact line of liquid-gas-vapor on one hand and between the latter and changes in the orientation of the adsorbed molecules on the other. The above systems were chosen for two reasons: first, many reliable data on surface tension and bulk activity coefficients are available; second, many studies of bulk properties of these liquid mixtures have been done, and the results can be compared with those corresponding to the surface phases. 11. Expression of the Ellipticity Coefficient Taking into Account the Anisotropy of the Molecules For a rectilinearly polarized monochromatic light of wavelength A, coming from the vapor phase and reflected by the liquid surface, the ellipticity coefficient /s at the Brewster angle is given by20
where d is the thickness of the transition layer, 7 the optical dielectric constant of the bulk liquid binary mixture, c,(z) and &) the components of the optical dielectric constant of the transition layer parallel to the x axis and z axis, and z the distance of any point in the transition layer from an arbitrary zero. This zero corresponds to the following surface excess definition: 1 ri = -
d
ni(z)dz NA ri is equal to the number of molecules of species i in the entire transition layer of thickness d between the binary liquid mixture and its ni(z)is the number of (15) Zielinska, B. J. A.; Bedeaux, D.; Vlieger, J. Physica A: (Amsterdam) l981,107A,91. (16) Beaglehole, D. J. Phys. (Les Ulis, Fr.) 1983,44C10, 147. (17) Mandelstam, L.Ann. Phys. 1913,41,609. (18) Bouchiat, M. A.;Langevin, D. J.Colloid Interface Sci. 1978,63, 193. (19) Bouchiat, M. A.;Meunier, J. J 2hy.s. (Les Ulis, Fr.) 1971,32,561. (20) Beaglehole, D. Physica B+C (Amsterdam) 1980, IOOB, 163. (21) Abeles, F.J. Phys. (Les Ulis, Fr.) 1977,38C5, 67. (22) Sivukhine, D.V. Zh. Eksp. Teor. Fiz. 1951,2, 367. (23)Tenebre, L.Ph.D. ThBse, Montpellier, 1971. (24) Defay, R.;Prigogine, I. Surface Tension and Adsorption; Longmans: London, 1966;(a) p 28, (b) p 161,( c ) p 160,(d) p 90.
Tronel-Peyroz et al.
n. -L
1
"
I
-
A '
h
5: g d n i ( z ) d z
I
I
transition layer
4
vapor
Figure 1. Schematic view of the transition layer between a binary liquid and its vapor for an adsorbed constituent i.
molecules of species i per unit volume in the transition layer (Figure l), and NA is the Avogadro number. If we choose for the variation of t,(z) and E&) with the composition ni(z) a relation of the same type as the one used for the bulk s o l u t i ~ nwe , ~obtain ~~~~ with k = x , z
t k ( z )- 1 = 4xCcuikni(z)
(3)
L
Equations 1 and 3 and integration yield
p=-
A
x
(7
+ 1 ) 1 / 2 (7 - €")(e" 7-1
-
')d
+
tu
i.e., p = pD + PA. Here xiu is the mole fraction of species i in the surface phase (characterized by a):
Vuis the volume of 1 mol of the mixture in phase u and its optical dielectric constant:12
tu
To evaluate the term P = CiNAxiu(a;" - at),which corresponds to the change of orientation of transition-layer molecules with composition, it is necessary to calculate first pD: This implies that the composition of the surface mixture is known. 111. Determination of the Composition of the Surface Mixture 1. Useful Relations. The composition of the surface mixture is determined from the surface excesses rl and r2and the relative surface excess of substance 2 with respect to substance 1, rZl(2 is the constituent with the is evaluated from the lowsmallest surface tension). r21 ering of the surface tension, dy, and the variation of the bulk chemical potential of species 2, dr," (the exponent a indicates the bulk phase):24c
As stated by P r i g ~ g i n e ?the ~ ~area of 1 mol of the mixture in the surface phase is a homogeneous first-degreefunction of the surface composition: A" = A1"Xi" + AZ'XZ" (8) ~
~
~
~
~~
~~~
(25) Hasted, J. B. Aqueous Dielectrics; Chapman and Hall: London, 1973;p 176.
Associated Complex Formation
water ethanol tert-butyl alcohol (I
Langmuir, Vol. 3, No. 6, 1987 1029
Table I. Parameters Used T o Calculate the Superficial Compositiona Pi0 ti0 ti d?, A Vi", cm3/mol 3.6 X IO4 1.7740 0.81 6.0 22.3 8.25 X lo4 1.8463 0.85 15.7 69.0 9.75 x 10-4 1.9150 0.85 17.5 111.8
A;", cm2/mol 3.7 x 108 4.4 x 108 6.4 X lo8
See section 111.2 in the text.
where Aiu defines the partial molar area of constituent i in the a phase.'lz4" From (5) and (8) and the definition A" = 1/(rl+ I',) we obtain
AzT2 = 1 (9) From (7) and (9) and given values for Alu and ABuit is easy to evaluate rl and rz and thus xiu = f ( x i a ) . The thickness of the transition layer, d, will be evaluated using the relation d = V'/Au with Vu = CixiuV; where Vi. is the partial molar volume of constituent i in the surface phase.2 2. Evaluation of A;. As is generally done, we first shall assume that the partial molar areas and partial molar volumes of species i in the surface phase are independent of the comp~sition.'-~~~ So Ai" N Aid and Vi. N Vid with, by definition, Aid = Vid/di, where di is the thickness of the transition layer at the pure liquid component i-vapor interface. We know that the density of this transition layer is less than the corresponding bulk density;1*26'n so the ratio ti = Vid/Vid, which indicates the average density shift inside the surface phase,12 is less than 1. The ratio ti and the thickness di are not independent; they are related to the ellipticity coefficient of pure liquid substances, pi. If, as a first approximation, it is assumed that the anisotropy contribution of the surface layer to pi can be neglected, we obtain from eq 412 AlTl+
(10) A reasonable value of ti for water (tl 0.81) corresponds to an equivalent thickness of two layers of water at the pure liquid water-vapor i n t e r f a ~ ei.e., , ~ ~2dl ~ ~N~ 6 A. The magnitude of ti for the two alcohol molecules (component 2) is evaluated from the relation given by EyringP
-
1 -=l+f
(
1+:i)
where f is a constant, T the temperature, and Tci the critical temperature. From (11)we can write m
1 cl
So with tl N 0.81, T , = 647.3 "C (for water), Tc = 513.9 "C (for ethanol), and T , = 508.2 "C (for tert-butyl alcohol) we obtain t2 0.85 for both ethanol and tert-butyl alcohol. With these values we can now evaluate dz from eq 10 and then calculate ABd.
-
(26)Khabarov, G. N.; Rusanov, A. I.; Kochurova, N. N. Kolloidn Zh. 1976, 37,92. (27)Conway, B. E. Adu. Colloid Interface Sci. 1977,8,91. (28)Partyka, S.;Rouquerol, F.; Rouquerol, J. J. Colloid Interface Sci. 1979,68,21: (29)Lu, W. C.;Jhon, M. S.; Ree, T.; Eyring, H. J.Chem. Phys. 1967, 46, 1075.
The values of the different parameters used to calculate the surface composition xiu are given in Table I. It must be underlined that Aid, which is the partial molar area of specie i, is quite different from the geometrical area often ~ ~ e d . ~ p ~ ~ ~ ~ ~ ~
IV. Materials and Experimental Procedures The ethanol and tert-butyl alcohol used were commercial products obtained from Merck Chemical Co., their purity being controlled by chromatography. The water was purified by reverse osmosis and distilled twice (once on potassium permanganate). The surface of the binary mixtures was cleaned by suction, and the surface tension was determined from the maximum vertical pull on Wilhelmy plates and platinum stirrups, the Wilhelmy plates being made of rough platinum or rough mica. Measurements were performed with a Guastalla's tensiometer% built in the laboratory. This apparatus consists essentially of a cabinet, a water-jacketed Pyrex cell (inner diameter 8 cm), a balance with a horizontal torsion wire, and a vertical translator which serves to raise or lower the plate. The setup has temperature control within hO.1 "C. The advancing contact angles were measured on glass plates (length 1.8 cm) by two methods. For the Guastalla's the pulling force was determined with the tensiometer described above. For the capillary height we used a n apparatus consisting of a temperature-controlled box and a cathetometer (accuracy of 0.01 mm), the glass plate being suspended from a bracket. All the glass dishes and plates were cleaned by soaking in concentrated H8O4containing potassium dichromate, washing with water purified by reverse osmosis, and rinsing with distilled water. For contact angle measurements, the glass plates were dried in a clean oven. The activity coefficients of the alcohols in the bulk mixture were measured by using the chromatographic technique described by Mohilner e t al.36 The ellipticity coefficients at the Brewster angles were measured by using a n ellipsometer built by one of us23 with a He-Ne laser (A = 6328 A) and a modulated compensator plate, followed by a synchronized detector. The experimental results are reported in Tables I1 and 111.
V. Discussion 1. Adsorption Isotherms and the Activity of the Adsorbed Molecules. The adsorption isotherms obtained have the aspect observed generally (Figure 2). The differences observed betwen the isotherms we obtained and those of Eriksson2 concerning the water-ethanol (W-E) system are due to the given values of the partial molar areas we have used. Those used by Eriksson are geometrical values. The activities of species 1and 2 in the superficial phase (aiu)have been calculated by using the relation pi" = pcLp for the chemical potentials, which expresses the thermodynamic equilibrium condition between phases a and CY. (30)Bennes, R. J. Electroanal. Chem. 1979,105, 85. (31)Tronel-Peyroz, E. J. Phys. Chem. 1984,88,1491. Mohilner, D. M.; Mohilner, P. R. J. Phys. Chen. (32)Nakadomari, H.; 1976,80,1761. (33)Lumen-Reynders, E. H.In Progress in Surface and Membrane Science; Cadenhead, D. A., Damolli, J. F., Eds.; Academic: New York, 1976;Vol. 10, p 268. (34)Guastalla, J. J. Colloid Interface Sci. 1956,11, 623. (35)Neumann, A. W. Adu. Colloid Interface Sci. 1974,4 , 105. (36)Mohilner, D. M.: Bowman, L. M.: Freeland. S.J.: Nakadomari. H.J. Electrochem. Soc.. 1983,120, 1658.' (37)Butler, J. V. A.; Wightman, A. J. Chem. SOC.1932,2089.
1030 Langmuir, Vol. 3, No. 6, 1987
Tronel-Peyrot et al.
Table 11. Experimental Data Concerning the Water-Ethanol System z 1.7740 1.7750 1.7760 1.7770 1.7775 1.7787 1.7795 1.7810 1.7825 1.7910 1.8005 1.8135 1.8200 1.8280 1.8375 1.8440 1.8485 1.8545 1.8570 1.8580 1.8575 1.8555 1.8515 1.8463
Za
0 0.0025 0.0050 0.0075 0.01 0.0125 0.015 0.0175 0.02 0.04 0.064 0.1 0.12 0.15 0.2 0.25 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0 8.25 x 10-3 1.65 X 2.475 X 3.30 x lo-* 0.0412 0.0495 0.0577 0.066 0.132 0.208 0.305 0.347 0.397 0.454 0.497 0.525 0.575 0.622 0.676 0.741 0.817 0.902 1.0
p x 104 3.6 3.65 3.7 3.75 3.8 3.84 3.89 3.94 3.97 4.47 4.83 5.51 5.79 6.12 6.50 6.81 7.05 7.43 7.71 7.94 8.1 8.2 8.23 8.25
Y,*
"em-'
rzlx W0, mol/cm2
4 deg
72.0 68.4 65.6 62.3 60.8 59.5 58.1 56.9 55.4 48.0 39.5 36.5 34.0 32.1 29.8 28.5 21.5 26.2 25.4 24.6 23.9 23.3 22.6 22.0
0 1.05 1.80 2.35 2.77 3.07 3.37 3.60 4.45 5.35 6.18 6.76 7.01 7.02 6.69 6.28 5.56 4.90 3.82 3.39 2.91 2.74 2.32
0 13.2 22.2 27.5 31.0 32.5 34.5 36.0 39.5 41.5 42.0 40.0 36.0 32.5 26.0 0
'Values interpolated from the data of ref 3 and 37. bValues interpolated from the data of ref 3.
Figure 2. Adsorption isotherms of ethanol (0) and tert-butyl alcohol ( 0 )at the liquid-vapor interface calculated with parameters grouped in Table I. Curve a is Eriksson's isotherm for ethanols2
Formulating p; as done by Eriksson? this condition leads to the relation aiu = ai" ex.[
Aiuo(Y- Y?)
RT
]
(13)
where aip is the activity of constituent i in the CY phase and yp the surface tension of constituent i in the pure liquid state. The results obtained are represented in Figure 3. For the two systems studied the activity curve of the alcohol (the constituent with the smallest surface tension) is located below the Raoult line for values of x2u smaller than 0.25 (W-E) and 0.20 (W-B). According to M c G l a ~ h a n this , ~ ~ behavior indicates that molecules of (38) McGlashan, M. L.
J. Chem. Educ. 1963, 40, 516.
Figure 3. Activities of species i in the adsorbed layer versus the surface mole fraction of alcohol molecules xZr: a and a' are respectively the plots of tert-butyl alcohol and water for the water-tert-butyl alcohol system; b and b' are the plots of ethanol and water for the water-ethanol system.
different species have the tendency to associate. For x Z u > 0.25 (W-E) and x2u > 0.20 (W-B), the activity curve of the alcohol is located above the Raoult line; this behavior reflects an association of identical molecules. Finally, for xpu> 0.8, the activities of all the constituents tend toward ideality and the surface mixing of water and alcohol takes place without any particular interaction between the molecules. The variation of the interfacial excess free energy of mixing, AGEu,can be calculated starting from the relation AGEu = R T X x i uIn f i u i
where fiu is the activity coefficient of species i in the superficial phase: f L u = aiu/xiu.
Langmuir, Vol. 3, No. 6,1987 1031
Associated Complex Formation
A 0.1
. !E : RT
Figure 4. Variation of the ratio AGEu/RT versus xza: (0) water-ethanol and ( 0 )water-tert-butyl alcohol systems.
The variation of AGEuwith the composition (Figure 4) shows that for both systems studied AGEu is similar to the one obtained by Eriksson2 The thermodynamical analysis of the bulk properties of these systems leads to large positive values for the excess free energy of mixing.% With AGEu expressing the divergence from ideality, one can deduce that the surface mixture is less nonideal than the bulk one; so the surface phase would therefore appear to be further from an eventual possibility of immiscibility than the bulk phase.& However, a more precise analysis concerning AGEu needs studies of adsorption at different temperatures. 2. Fluctuations of the Surface Composition. The surface layer is characterized by the absolute temperature (T), the pressure (P),the composition of the system, and either the area (A) or the surface tension (7)of the adsorbed layer. In the bulk adjacent phases, all the thermodynamic variables fluctuate around their mean value.4l This must be the same for the thermodynamic variables of the surface phase under consideration, as indicated by the work of Mandelstam, for example.42 As we have it is possible to apply Landau's theor9l to obtain explicit equations of the surface thermodynamic variables fluctuations in the adsorbed layer of a binary liquid in equilibrium with its vapor. For the surface composition fluctuations, we obtain
where Nu is the total number of molecules 1and 2 in the f l u ~ t u a t i o nand ~ ~ ( ( 6 ~ ~ "the ) ~mean ) square values of the composition fluctuation. The fluctuations of the composition F2" show two distakes place at a tinct maxima. The first maximum, FZU1, composition of xZu N 0.1-0.2; it is smaller than the one in the region of more concentrated solutions (Fzu1I) (Figures 5 and 6). These local fluctuations arise from the intermolecular interactions and can be considered to yield in(39) Franks, F.; Ives, D. J. G. Q.Rev., Chem. SOC.1966,20, 1. (40) Mohilner, D. M.Bioelectrochem. Bioenerg. 1978, 5 , 185. (41) Landau, L.;Lifshitz, E. Statistical Physics; Pergamon: London, 1958. (42) Mandelstam, L.Ann. Phys. (Leipzig) 1913, 41, 609. (43) Bennes, R.;Douillard, J. M.; Privat, M.; Tronel-Pepoz, E. J. Colloid Interface Sei. 1987, 117, 574.
Figure 5. Variation of the surface concentration fluctuation versus xZufor the water-ethanol system: (0) experimental points; (-) calculated values for different associated complex models; (- - -) theoretical values expectedc for an ideal binary mixture.
0.5
xs
Figure 6. Variation of the surface concentration fluctuation versus x2u for the water-tert-butyl alcohol system: (0) experimental points; (-) calculated values for different associated complex models; (- - -) theoretical values expected for an ideal binary mixture.
formation which is useful for the understanding of the mixing in surface mixtures from a molecular point of view. For the bulk phase, instead of considering directly intermolecular interactions, Fujiyama and co-workershave introduced the concept of an associated complex to in(44) Iwasaki, K.; Katayanagi, Y.; Fujiyama, T. Bull. Chem. SOC.Jpn. 1976,49, 2988.
(45) Kato, T.; Fujiyama, T. J. Phys. Chem. 1976, BO, 2771. (46) Iwasaki, K.; Fujiyama, T. J.Phys. Chem. 1977, 81, 1908.
1032 Langmuir, Vol. 3, No. 6, 1987
Tronel-Peyroz et al.
Table 111. Experimental Data Concerning the Water-tert -Butyl Alcohol System azU 7 p x 104 7,mN-m-' rZ1x W o , mol/cm2
XZU
2.625 X 1.476 X 0.625 X 7.31 x 10-4 1.476 X 2.625 X 6.195 X 1.237 X 1.96 X loT2 3.24 X 0.07 0.1035 0.2135 0.3 0.4 0.4473 0.5 0.53 0.6 0.66 0.7 0.8 0.9
1.7740 1.7740 1.7740 1.7745 1.7750 1.7760 1.7790 1.7850 1.7930 1.8070 1.8360 1.8510 1.8800 1.8930 1.9020 1.905 1.907 1.908 1.910 1.9115 1.912 1.9135 1.914
0 1.78 x 10-3 3.162 X 0.01 0.01778 0.0316 0.07464 0.149 0.236 0.3902
0.615 0.675 0.685 0.690 0.702 0.724 0.770 0.840 0.905
3.6 4.7 4.95 5.25 5.20 4.9 3.6 2.7 2.7 3.2 4.35 5.30 7.80 8.7 9.15 9.3 9.4 9.5 9.55 9.6 9.65 9.70 9.75
terpret concentration fluctuation results. The associated complex formation in binary mixtures can be divided into three groups: In the first group, associations occur between the same species; the concentration fluctuation, F, is larger than the concentration fluctuation for an ideal mixture, Fp, throughout the whole concentration range. In the second group, associations take place between different species; here, F is smaller than Fp. Finally, associations occur simultaneously between the same and different species; here, both of the above situations may take place for F with respect to Fp. We have applied these ideas to calculate the adsorbed phase composition fluctuations reported in this work. For the two systems studied, there exist three concentration ranges (with W = water, A = ethanol (or tert-butyl alcohol), and x2" = mole fraction of ethanol (or tert-butyl alcohol)): In the first range, 0 6 x2" < 0.25-0.35. In this dilute range of concentration, Fzuis smaller than Fzup,so associations take place between different species. However, water molecules are probably in an associated form, so the binary mixture of W and A is composed of W,, A, and W,A, with the dissociation equilibrium
W,A
W,
+A
(15)
In this case, the composition fluctuations are given b y 5 Fz" = xZ"(l - ~ 2 " ) [ (+ l (JJ- 1)~2")'- 4RpXz"(l - X Z " ) ] ~ / ~ (16) where R = K/(1 + K ) and K is an equilibrium constant of the dissociation equilibrium (15). We found that the interfacial mixture is composed of Wz, A, and an associated complex molecule WzA with a dissociation equilibrium constant K 3-5 for waterfor water-tert-butyl alcohol systems. ethanol and K In the second range, 0.25-0.35 6 x2" < 0.65-0.85. The surface water-alcohol mixtures can consist of few associated complex molecules W,A and a lot of A molecules, which are associated with each other in a "polymer-like" form A,. This latter complex mainly determines the variation of Fzuwith the composition in this concentration range, obtained by45
-
Fz" = l/p[(p
-
Q)
+ l ) ~ 2 "- I][(s - 1) + [ ( p + I) - s ] x ~ ~ J (17)
72.0 70.4 68.9 64.8 60.5 56.6 48.0 40.9 35.7 30.3 24.4 23.0 22.2 22.0 21.6 21.4 21.3
1.749 0.857 1.33 2.17 2.82 3.36 4.16 4.69 4.98 5.21 5.25 5.25 5.12 4.88 4.65 4.5 4.36 3.99 3.72 3.08 2.72 1.95 1.45
21.0 20.7 20.4 20.2 -0.15
8, deg 0 11.8 13.0 17.2 21.3 25.4 33.5 38.8 39.4 34.5 0
-0.55
Figure 7. Summary of the associated complexes found to explain the surface composition fluctuations (see text).
(For the values of s found in this concentration range, see Figure 7.) In the last range, xzu > 0.65 (WE) or 0.85 (WB). The surface mixture is composed of W, and A,,,, and Fz"is given by45 F2" = x2"(l - x2")[m+ (p - m)x2"] (18) We found that p is always 1 for both systems and m decreases from 3 to 1 when x2" goes from 0.65-0.85 to 1. The theoretical curves of F2" corresponding to these associated complexes are represented in Figures 5 and 6 (full line). Figure 7 sums up the different associated complexes used versus the surface-phase compositions. It is important to underline the part played by the associations of the polymer type A,. For the two mixtures studied, the appearance of these structures in the surface layer coincides with the composition beyond which the surface tension reaches a value very close to that of the pure alcohols utilized. This conclusion has to be checked for other surface-activesubstances. Such a study has been undertaken for the mixture water-2-butoxyethan01.~~ 3. Changes in the Polarizability within the Transition Layer. The knowledge of the composition of the surface mixture (19)
makes it possible to calculate successively the thickness of the transition layer (d = V"/A"),e' (given by eq 6 assuming that a: = &, the mean polarizability of the mol(47) Bennes, R.;Privat, M.; Douillard, J. M. C . R. Seances Acad. Sci., Ser. 2 1983,296, 537.
Langmuir, Vol. 3, No. 6, 1987 1033
Associated Complex Formation P.102
t
lo
Figure 8. Variation of P (see text) versus xgUfor water-ethanol (0) and water-tert-butyl alcohol ( 0 )systems. ecules), and p D . One can then deduce from the experimental values of the coefficient p the contribution to p due to the anisotropy of the adsorbed molecules and the variations of P with the composition [Figure 81. P = CNAxi'((Yt - a?) = CNAxi'8ai (20) i
i
P has the dimension of a volume and can therefore be treated as a homogeneous function of the composition of the first degree, thus enabling the evaluation of 8a1and 6a2. In fact, as it is extremely difficult to estimate the anisotropy of the surface layer of a pure substance,22we have arbitrarily assumed this to be zero for the three pure substances utilized (see section 111.2). If this choice seems justifiable in the case of it is possibly not so for the alcohols. We therefore extrapolate the graphs P = f(x2') in order to obtain the value of the anisotropy of a component in mixture (hi) only in the region of high concentration of each species. 6ai must hence be assigned to the shift of the anisotropy of molecule i in the mixture relative to that of pure substance i. The variation of P suggest the following remarks. (1)The two systems studied do not behave identically: the differences observed for ellipticity coefficient variations with composition are found here again. For the waterethanol mixture there is one maximum. For x2" < 0.2, P(x2()is very nearly 0. This indicates that there is no important change in the structure of the solvent. When x2u is near to 0.2, one notes an increases in P, which indicates a change in the polarizability of the water (6al # 0). When xZu is greater than 0.5, one notes that 6az is almost 0, which means that the molecules of ethanol in the surface mixture have an anisotropy which is identical with that at the surface of the pure substance. For the water-tert-butyl alcohol mixture the variation of P(x23is complex, showing two maxima and a minimum. These numerous changes are difficult to interpret. Here again the water anisotropy can be extrapolated to that of the pure substance for x2u G 0.05. Similarly,the anisotropy of the tert-butyl alcohol in the surface layer can be extrapolated to that of the pure tert-butyl alcohol for x2u > 0.5. (2) The fact that the presence of a small quantity of alcohol molecules in u phase perturbs only slightly the superficial water structure (6a1 0) implies that only the polar heads of the alcohol molecules must penetrate the water network while the aliphatic chains remain free on the vapor side. When x Z u is near 0.05-0.2, this initial superficial structure of water shifts to an unknown new one, corresponding to a value of 6al different from 0. This has already been considered by one of us in an attempt to explain the variations in the ellipticity coefficient observed in the case of water-long-chain alcohol and (48) Orttung, W.
H.;Meyers, J. A. J. Phys. Chem. 1963, 67, 1905.
Figure 9. Variation of the advancing contact angle at the triple contact line glass-liquid-vapor for water-ethanol (0)and water-tert-butyl alcohol ( 0 )systems. water-long-chain acid systems.- In the two cases, it was assumed that the variation of the ellipticity coefficient with the composition of solute molecules was the result of two additional terms: an anisotropic contribution of the solute molecules and a constant term depending of the state of the surface water molecule, pw (which represents initial water with 8a1= 0 when x Z u < 0.15-0.2 and another type of surface water with 6a1# 0 when xaU > 0.15-0.2). This assumption seems also justified, a posteriori, by our result showing that the thickness of the surface layer varies sufficiently little in the very dilute concentration range under consideration. In contrast, for x2" > 0.7-0.8 the structure of the alcohol molecules is very similar to that of the pure substances in which the water molecules are dissolved without undue difficulties, as has been indicated by the activity coefficient curves (u2u 1). 4. Contact Angle. We have measured the advancing contact angle at the triple contact line glass-liquid-vapor for the two systems studied (Figure 9, Table 11, Table 111). In both cases, and only for surface mixtures with an alcohol mole fraction in phase u less than 0.3-0.4, the contact angle 6 is not zero. Under the experimental conditions and the technique of cleaning used, the glass plate is completely wetted by both pure water and alcohols. All angle variations can thus be assigned to a modification of the surface-layer properties and not to a modification of the properties of the solid surface. One notes from our results that the concentration range where 8 is different from zero corresponds to the one where there are concentration fluctuations of type I with F2" smaller than F2"ideal. We have no precise interpretation of this correlation; however, we speculate that they can serve to better understand the wetting phenomena. It is also seen that the non-zero contact angle corresponds to the formation of the W2A complex. It vanishes on the formation of the polymer-like complex (A8). If we make an analogy, the associated complex W2Awets glass like long n-alkanes while the polymer-like complex, A,, wets like the pure alcohol. This fact confirms that the aliphatic chains of the W2A complex stay toward the vapor side of the interface and suggests that the structure of the superficial water-concentrated alcohol mixture is nearly the same as the pure alcohol, with some polar heads standing on the gas side.
-
VI. Conclusion From all these results one can speculate about the structure of the interfacial water and the solvation of the
1034
Langmuir 1987,3, 1034-1044
absorbed alcohol molecules, both systems studied behaving in roughly the same manner. For very dilute solutions in alcohol, the structure of the surface water is very similar to the one at the interface of pure water-vapor; the activity coefficients and polarizability calculations strengthen the idea, brought up by some that only the polar head and probably the first carbon atom of the aliphatic chain are solvated. The paraffinic chains are probably perpendicular to the surface and directed toward the exterior, which would explain the existence of a non-zero angle of contact and, probably, the variation of 6a2while at the same time 6a1is nearly equal to zero. With an increase of the alcohol mole fraction (xzu > 0.1) (49) Bauer, E.; Guastalla, 3.; Guastalla, L. P.; Lize, A. J. Chim.Phys. (Paris) 1969, 66, 99.
the structures undergo a modification. The aliphatic chains of the alcohol draw closer together but are always directed toward the exterior since the contact angle increases continuously. This leads to a weakening of the water structure, permitting a progressive autoassociation of the alcohol molecules. Above the alcohol concentration where the contact angle becomes zero, there is a change in the solvent; the aliphatic chains are no longer directed toward the exterior, and the alcohol molecules associate together to form a network, close to the one of the pure compound surface, with the water molecules integrating themselves in the empty spaces of the alcohol structure.
Acknowledgment. We express our thanks to the reviewers for their helpful comments. Registry No. EtOH, 64-17-5; t-BuOH, 75-65-0.
Penetration-ControlledReactions in Organized Monolayer Assemblies. 1. Aqueous Permanganate Interaction with Monolayer and Multilayer Films of Long-chain Surfactants Rivka Maoz and Jacob Sagiv* Department of Isotope Research, The Weizmann Institute of Science, 76100 Rehovot, Israel Received December 6, 1986. In Final Form: May 26, 1987 The extent of oxidation of unsaturated monolayer constituents by aqueous KMn04 is employed to probe the penetration of ions from an aqueous phase into organized monolayer assemblies of some long-chain acid and silane surfactants. The study compriees solid-supportedLangmu+Blcdgett (LB)and self-assembled (SA) monoIayers as well as a series of LB built-up (multilayer)films, the molecular architecture of which was planned such as to furnish evidence on the depth of penetration of the permanganate ion into the inner core of a compact film assembly. A combined analysis of the structural stability and the reactivity of the f i ,as revealed by FTIR-ATR spectroscopy and wettability observations, suggeats a defed-controlled mechanism for the passage of ions across tightly packed monolayers and multilayers of oriented long-chain surfactants. The barrier efficiency of the investigatedfilms is thickness independent, in the range between one to three superimposed monolayers, being determined by the structural perfection of the films and their stability under the action of the penetrating species, solely. Self-assembled monolayers are found to be more stable and less penetrable than each of the presently studied monolayer or trilayer LB assemblies. Wetting of a monolayer- or multilayer-covered polar solid by water or by the aqueous permanganate solution is shown to require establishment of direct contacts between the bulk liquid and the underlying solid surface and lateral diffusion of the liquid in the film-solid interface.
Introduction Organized monolayer structures produced on polar solid surfaces via spontaneous adsorption from organic solutions (self-assembling monolayers) are supramolecular organizates resembling, in some respecta, the well-known Langmuir-Blodgett (LB) built-up films while displaying other distinct and rather unique features.l-1° Apart from their (1) Bigelow, W. C.; Pickett, D. L.; Zisman, W. A. J. Colloid Sci. 1946,
I , 513. (2) Maoz, R.; Sagiv,J. J. Colloid Interface Sci. 1984, 100, 465. (3) (a) Netzer, L.; Iscovici, R.; Sagiv,J. Thin Solid F i l m 1983,99,235; (b) 1983, 100, 67. (4) (a) Nuzzo, R. G.; Allara, D. L. J.Am. Chem. Soc. 1983,105,4481. (b) Allara, D. L.; Nuzzo, R. G. Langmuir 1986, 1, 45; (c) 1986, 1, 52. (5) (a) Sandroff,C. J.; Garoff, S.; Leung, K. P . Chem. Phys. Lett. 1983, 96,547. (b) Garoff, S.; Hall,R. B.; Deckman, H. W.; Alvarez, M. S. Proc. Electrochem. Soc. 1986,85-88, 112. (6) For example, see: (a) Shafrin, E. G.; Zisman, W. A. Adu. Chem. Ser. 1969,87,20 and references cited therein. (b) Bewig, K. W.; Zisman, W. A. J.Phys. Chem. 1963,67,130. (c) Brockway, L. 0.; Jones, R. L. Ada Chem. Ser. 1964,43,275. (d) Bartell, L. S.; Betta, J. R. J. Phys. Chem. 1960, 64, 1075. (e) Gaines, G. L., Jr. J. Colloid Sci. 1960, 15, 321. (fj Young, 3. E. Aust. J. Chem. 1956, 8, 173.
0743-7463/87/2403-l034$01.50/0
relevance in the study of molecular self-organization, in general, much of the interest in self-assembling (SA) monolayers stems from their potential in a wide range of scientific and technological application^.^ So far, most of the published material on SA monolayers has focused on their formation and structure as derived by various physical methods.l-l0 The chemical properties and reactivity of self-assembling monolayers have only briefly been touched, mainly in relation to their binding to the solid s u r f a ~ e ~ , ~and ~ ' - 'to ~ their chemical modification, as required in the construction of planned multilayer structure^.^ The present series of papers is intended to (7) (a) Chapman, J. A.; Tabor, D. Proc. R . Soc. London, A 1967,242, 96. (b) Bowden, F. P.; Tabor, D. The f i c t i o n and Lubrication of Solids; Clarendon: Oxford, 1950; Chapter X. (8) Gun, J.; Iscovici, R.; Sagiv, J. J. Colloid Interface Sci. 1984,101, 201. (9) Boerio, F. J.; Chen, S. L. J. Colloid Interface Sci. 1980, 73, 176. (10) (a) Finklea, H. 0.; Robinson, L. R.; Blackburn, A.; Richter, B.; Allara, D.; Bright, T. Langmuir 1986,2,239. 03) Sabatani, E.; Rubinstein, I.; Maoz, R.; Sagiv, J. J. Electroanal. Chem. 1987,219, 365.
0 1987 American Chemical Society
