Contact Angle Hysteresis. IV. Contact Angle Measurements on

---

1507

CONTACT ANGLEHYSTERESIS

absence of any hydrogen-containing reactant, spurious HCN was formed. Similar difficulty was encountered by McElcheran, Wijnen, and Steacie13 in the photolysis of CO(CN)2, and it would appear to be quite a general problem which must be overconie before further progress can be made towards absolute rate parameters. Although CF, is relatively unreactive compared to CN, we did not think the indirect estimation of the extent of reaction 13 as ( ~ R c , H , RCHEN RcF~cH~), together with the indirect estimation of reaction 16 already proposed, was worthwhile a t this stage.

+

Contact Angle Hysteresis.

+

IV.

Preparation of ICN Solid 1 2 plus excess solid Hg(CX)z are sealed in an evacuated Pyrex tube. The tube is submerged in hot water until the color of iodine disappears (about 0.5-1.5 hr.s). Half the tube is then raised out of the water, when pure ICX needles deposit on the cool parts. This preparation is much faster and much more convenient than the conventional methods using solvents. The method also works for RrCN. (13) D. E. McElcheran, M. H. J. Wijnen, and E. W. R. Steacie, Can. J . Chem., 36, 321 (1958).

Contact Angle Measurements on

Heterogeneous Surfaces'

by Robert H. Dettre and Rulon E. Johnson, Jr. Contribution No. $16 f r o m the Experimental Station, Organic Chemicals Department, E . I . d u Pont de Nemours and Company, Wilmington 98, Delaware (Received September 6 , 1964)

The wettability of a heterogeneous surface was studied as a function of coverage by regions of different intrinsic wettabilities. The observed behavior shows good agreement with that of an idealized heterogeneous surface.

Introduction In a previous paper2 we reported the results of a computer study of the effect of surface heterogeneity on the wettability of an idealized heterogeneous surface. The surface was one having regions of different intrinsic wettabilities, each region being larger than molecular dimensions. The analysis of this model surface showed t.hat many metastable liquid-drop configurations are possible, each having a different contact angle and each separated from the next by an energy barrier. These energy barriers hinder the attainment of a configuration of minimum free energy. I t was suggested that an observed configuration would depend on whether the drop periphery is advancing over the solid surface or receding from it and also on any vibration which

could cause the periphery to niove. I t was also suggested that, for a heterogeneous surface of two components, the observed contact angle will usually be different from the angle, 4,given by Cassie's3 equation COS

4=

u1 COS

el

+

u2 COS

e2

(1)

where 4 is the contact angle for the configuration of minimum free energy, c1 is the fraction of the surface having intrinsic contact angle 81, and u2 is the fraction having intrinsic angle e2. The advancing angle will (1) Presented in part a t t h e Gordon Research Conference on Chernistry a t Interfaces, July 1961.

(2) R. E. Johnson, Jr., and R. H. Dettre, J . Phys. Chem., 68, 1744 (1 964). (3) A. B. D. Cassie, Discussions Faraday Soc., 3 , 11 (1948).

Volume 69, Number 5

M a y 1966

ROBERTH. DETTREAND RULONE. JOHNSON, JR.

1508

be larger than 4 and the receding angle smaller. The difference between the advancing and receding values gives the magnitude of the contact angle hysteresis. The study also showed that, whereas the advancing angle is relatively insensitive to coverage by highcontact-angle regions once coverage has exceeded about 40%, the receding angle increases rapidly as coverage changes from about 80 to 100%. Conversely, a t low ($overage the advancing angle is sensitive and the receding angle insensitive to coverage. There are, therefore, rcllatively narrow ranges of coverage over which contact angle hysteresis can change drastically. This effect should be of considerable importance in problems of liquid repellency, flotation, flow through porous niedia, or any other phenomenon which is affected by contact angle hysteresis. Sone of the studies reported in the literature on the effect of surface coverage on wettability by pure l i q ~ i d s ~has - ~ included both receding and advancing contact angle data and a description of the nature and magnitude of variations in surface coverage. In the present papw we describe the results of advancing and receding con tact angle nieasurenients on heterogeneous surfaces ha1 irig regions of different intrinsic wettabilities. The observed wettability behavior is compared with that of an idealized heterogeneous surface.2 We also attempt to interpret other wettability studies reported in the literature, using the concepts developed here and in rpf. 2 .

Experimental Xaterials. A polymeric and a nonpolymeric organic titanate were used to obtain titania coatings on glass. The polydibutyl titanate was Du Pont TyzorB P B organic titanate; solutions of 1.1 and 7.1 wt. yo in reagent grade (.arbon tetrachloride were used. The tetraisoprop,yl titanate was Du Pont Tyzor@PE organic titanate; a 0.6 wt. yo solution in reagent grade carbon tetrac*hloridcb\vas used. All coatings were made on soft, soda-hie glass microscope slides (Thomas Red Label brand; 2.5 X 7 6 cni.). Trimethyloc~tadecylammoniumchloride was obtained from Armoiir Industrial Chemical Co. as -4rquada 18-.iO, a 30(> solution in isopropyl alcohol (36%) and water (14TcI : aqueous solutions containing 0.5 to 2% of the quaternary aniiiioniuni salt were used (pH 5 to 6). Sodiimi stearate, prepared from high purity stearic actid (f.p 611.8"), was recrystallized twice from water; a saturated aqueous solution was used a t 65". Deionized water having a surface tension of 71.5 dynes cni. a t '25" was used in all solutions and contact angle iiieasureinents. The methylene iodide was The Journal of Physical Chemistry

Eastman Grade; it had a surface tension of 50.1 dynes/ cm. a t 24". Surface Preparation. Microscope slides were coated by withdrawigg them from the organic titanate solutions at a constant rate of 5 cm./niin. using a FisherPayne Dip-Coater. After evaporation of the carbon tetrachloride solvent, the slides were heated to 500550" for 1.5 to 2 hr. in order to assure complete conversion of the titanate coating to TiOz. This procedure was repeated for multiple coatings. X-Ray and electron diffraction studies of the coated surfaces and of abraded surface material indicated that the coating was principally anatase, the low temperature form of TiO,. Uncoated slides were given a similar heat treatnient. After cooling t o room temperature, each slide gave a zero advancing contact angle with water, indicating that the surface was free of organic material. Immediately after measuring the water angle, each slide was immersed in a sodium stearate solution for 3 hr. or in a solution of triniethyloctadecylaninionium chloride for a t least 0.5 hr. followed by rinsing with water for 10 to 1 5 sec. in order to rpmove all but the first adsorbed monolayer. Very est ensive rinsing can result in removal of some of the first nioriolayer, as indicated by changes in wettability, but there was no indication that the brief rinsing used here significantly affected the monolayer. Surfaces which were not dewetted after rinsing were dried in air prior to contact angle measurement. Water contact angles on coated surfaces increased with time of immersion in the sodium stearate solution; they remained relatively constant for ininiersion times greater than 2 hr. Contact angles were essentially independent of immersion time in the quaternary salt solution except for samples with inore than four coating treatments with 7.1% of polydibutyl titanate. On such surfaces the water advancing angle decreased by about 10" when immersion time was increased from 0.5 to 21 hr. Uncoated glass surfaces showed no immersion time dependence. A different slide was used for each multiple coating treatment with the exception of 20, 30, and 40 coating treatments with 1.1% of polydibutyl titanate. Here the same slide was used for all three measurements, but the adsorbed material was first renioved from the (4) W. Philippoff, S. R. B. Cooke, and D. E. Cadmell, Mining Eng., 4, 283 (1952).

S.Bartell and It. J. Ruch, J . Phys. ('hem.. 60, 1231 (1956). (6) L. S. Bartell and R. J. Ruch. ibid.,63, 1045 (1959). (7) J. W. Shepard and J. 1.' Ryan. ibid., 63, 1729 (1959). ( 8 ) G. L. Gnines. J r . , J . Colloid Sci., 15, 321 (1960). (9) L. 0. Brockway and It. L. Jones, Advances in Chemistry Series, No. 43, American Chemical Society. Washington, D. C., 1964, p. 275.

(5) L.

CONTACT ANGLEHYSTERESIS

slide before each series of coating treatments. All coated surfaces were specular. Contact Angle Measurements. Contact angles were measured at 25 f 1" on profiles of sessile drops using a microscope fitted with a goniometer eyepiece'O; niagnification was 20x. Unless otherwise noted, relative humidity was not controlled. Advancing angles were measured after the drop volume was increased and the periphery advanced over the surface; receding angles were measured after the volume was reduced arid the periphery receded. Readings were take11 within 10 t o 15 sec. after nioveriierit of the drop periphery, and care was taken to niininiize vibration and drop distortion during volume changes. Average drop size was 0.05 nil. Angles were measured on four to six different regions of each surface and averaged. Reproducibility for any given region was + 2 " . The lengths of the 1 ertical lines on the points in the graphs and the f values in Table I represent the magnitudes of the standard deviations. Coating Thzckness. The thicknesses and refractive iiidicrs of several of the coatings were determined after contact angle measurements, using the ellipsonietric method of S7a&Eek.11The iristrunient used was a spectrometer (Gaertner Scientific Corp.) provided with polarizer and a rialyzer Sicol prisms, a Babinet Solei1 conipensator, and a sodiuni arc light source. Incident light was plane polarized at 45 or 135" to the plane of incidence, and nieasurenients were made at angles of incidence of BO and 70". Thicknesses were not determined for all coatings, but readings were taken on every coated slide to ascertain that each coating and heating treatnient resulted in deposition of Ti02 on the surface. The average refractive index, determined for coatings between 40 and 100 nip thick, was 2.3 f 0.1 at 5893-A. wave length. The value for an anatase single crystal is 2.53 at 6000 a.I2 The smaller value for the coating has been previously observed by Haas,I2 who attributed it to a lower density for the coating material. As will be more apparent later, this low value could also be a result of incomplete coverage by titania. Electron Microscopy. Replicas of several of the slides were prepared after contact angle measurements. Each surface was giveii several thin coatings of nitrocellulose using an amyl acetate solution. The nitrocellulose impression was stripped from the slide by imniersion in water and coated by vaporization with a thin carbon film. The nitrocellulose was then dissolved in amyl acetate, and the remaining carbon film was shadowcast with Pt-Pd alloy at an angle of about 13". To verify results by this technique an alternative method was used whirh involved coating the surface directly

1509

with the carbon, shadow-casting, and stripping in aqueous HF. Both methods gave the same results. Electron micrographs were obtained with an RCA Type EAIU-2D electron niicroscope. Li+-Na + I o n Exchange. After contact angle measurements, the surfaces of some of the slides were covered with a eutectic niixture of KS03 and L i s 0 3 (60 mole yo of KNOJ at 200-220" for 3-6 hr.. cooled to room temperature, and then iniinersed in water.13 Under the above conditions the lithium ions replace the sodiuni ions in the glass, and, because of the sniall size of the lithium ion, the glass surface tends to shrink thereby producing in the surface a pattern of nunierous intersecting crack^.'^ The density of the crack pattern increases as the temperature and/or time of exchange increase and, therefore, presumably as the extent of exchange increases. The slides werr examined under an optical microscope, and the densities of the crack patterns were compared as a function of the number of coating treatments and the concentration of organic titanate in the coating solution.

Results Wettability Studies. Water contact angles on titania-coated glass surfaces after treatment with triniethyloctadecylamnioniuni chloride solutions are given in Figures 1 to 3 as a function of the number of coating

\

\

\

\ \ \ I

45

0

NUMBER OF COATING TREATMENTS

Figure 1. IVater wettability of titania-coated glass after treatment with trimethyloctadecylammoniurn chloride solution plotted as a function of the number of coating treatments with 1.1% polydibutyl titanate ~~

(10) (a) W. C . Bigelow, D. L. Pickett, and TV. A. Zisman, J . Colloid Sci., 1, 513 (1946); (b) H . R . Fox and TV. A. Zisman, ihid.. 5 , 514 (1950). (11) (a) A. V a E e k , J . Opt. SOC.Am., 37, 145 (1947): (h) A . TaiiEek, Phys. Rer., 57, 925 (1940)

(12) G. Haas, V a c u u m , 2 , 331 (1952). (13) F. M. Ernsherger. Proc. Roy. Soc. (London), A257, 213 (1960)

Volume 69. .YurnbeT 5

.Mau 1965

1510

ROBERT H. DETTREAND RCLONE. JOHNSON, JR.

treatments with organic titanate solutions. Figure 4 gives the water wettability of titania-coated glass after treatment with a saturated sodium stearate solution. hlethylene iodide contact angles on some of the titania-coated surfaces before and after treatment with solutions of the quaternary ammonium salt are given in Table I. (ba and are advancing and receding contact angles, respectively, measured at 50% relative humidity. The advancing contact angle of 54" on uncoated glass i n Table I is slightly lower than the 59" reported by Levirie and Zisniani4 for a monolayer obtained by retraction of a glass slide from a dilute aqueous solution of trimethyloctadecylammonium chloride. This difference could be due to differences in glass composition.

Table 1: Methvlene Iodide Contact Anales on Glass and Titania-Coated Glass before and after Treatment with Trimethyloctadecylammonium Chloride Solutions -Before

treatmentn-

@a

Uncoated glass 5 coatings with 1.1% polydibutyl titanate 1 coating with 7.17, polydibutyl titanate 4 coatings with 7.17, polydibutyl titanate 7 coatings wit.h 7.17, polydibutyl titmate

41 f 1'

9r

-.4fter

RECEDING ANGLES

0

I

2 3 4 5 NUMBER OF COATING T R E A W E N T S

6

7

Figure 2. Water wettability of titania-coated glass after treatment with trimethyloctadecylammonium chloride solution plotted as a function of the number of coating treatments with 7.1% polydibutyl titanate.

,

I

I

1

treatment-

&a

30 =t2" 54 f 2'

9,

4312'

..

...

4 8 1 2 2 5 3 ~ 8

..,

...

5111

275~3

8 f2

0

2 7 f 2

1113

8 f1

0

19+2

3 1 2

.

ADVANCING ANGLES

Surface was previously rinsed with water and dried in air a t 50y0humidity. 01 0

Timmons and Zisman have recently reported that the wettability of platinuni by methylene iodide depends to some extent on water adsorption. We have observed a similar effect on glass. Uncoated glass surfaces which had been heated to 550" and then kept a t 1&i0 for several days had advancing and receding methylene iodide angles of 40 f 1 and 15 -f 7" when measured immediately after cooling. On standing (50% relative humidity) the advancing angle remained constant, but the receding angle increased, approaching the value given in Table I for the waterrinsed surface. This behavior was observed only when the glass had been kept at high temperature for several days. The sensitivity to water resembles that observed by Bartell and Bristoli6 for acetylene tetrabroiiiide on glass and quartz. Li+-Na+ Ion Exchange. Under identical ion exchange conditions, the density of the crack pattern The Journal of Physical Chemistry

2

4 6 8 IO NUMBER OF COATING TREATMENTS

12

Figure 3. Water wettability of titania-coated glass after treatment with trimethyloctadecylammonium chloride solution plotted as a function of the number of coating treatments with 0.67, tetraisopropyl titanate.

decreased with increasing nuniber of coating treatments and with increasing concentration of titanate in the coating solution. If we assume that the titania is a barrier to the ion exchange, then the above decrease in density suggests increasing coverage by titania. Ion exchange studies conibined with electron microscopic examination of rubbed and unrubbed titaniacoated glass surfaces, discussed later, indicate that the above assumption is a reasonable one. (14) 0. Levine and W. A. Zisman, J . Phys. Chem., 61, 1068 (1957). (15) C . 0.Timmons and W. A. Zisman, ibid., 68, 1336 (1964). (16) F. E. Bartell and K. E. Bristol, ibid., 44, 86 (1940).

CONTACT ANGLEHYSTERESIS

L

1511

80

ADVANCING ANGLES

i

LI

-1

RECEDING ANGLES

I

..

-A

I O

K

k

-

a

T 6 NUMBER

I

1

8

10

I

T

It

12

14 4

6

OF COATING TREATMENTS

Figure 4. Water wettahility of titania-coated glass after treatment with sodium stearate solution plotted as a function of the number of coating treatments with 1.1% polydihutyl titanate.

Electron Microscopy. Electron micrographs indicated that the titania was present on the glass surface in discrete patches. They also showed that coverage increased with 1 he number of coating treatments with the titanate solutions. Examination of many micrographs showed that coverage was about 6% for one, 11%, for two, and 25% for five coating treatments from 1.170polydibutyl titanite. The titania patches had an arerage height of 25 nip arid average widths from 60 nip for one mating to 160 nip for five coatings. The more concentrated titanate solution gave larger patches and greater coverage per coating treatment. For example, one coating treatment from 7.1y0 polydibutyl titanate gave about 50% coverage. The widths of the coated regions ranged from 0.5 to 10 p, and the heights appeared to be only slightly greater than those of coatings obtained from the less concentrated solutions. Electron niicrographs of titania coatings from 0.6% tetraisopropyl titanate indicated a coverage per coating treatment internlediate between those of the two polydibu t yl titanat e solutions. I t was possible to alter the titania coverage on a coated surface by rubbing the surface with a firepolished glass rod. This was accomplished without scratching by first treating the surface with a suitable surfactant such as dimethyloctadecylamine. The rubbed surface was still specular, but electron niicrographs revealed that the titania patches had been deformed and spread over the glass surface thereby increasing coverage. For example, electron micrographs of a coated surface obtained from one coating treatment with 7.1'7, polydibutyl titanate showed that

rubbing increased titania coverage from 45 f 10% to at least 95y0. Water contact angles measured after treatment with the quaternary salt solution decreased with rubbing. In the above example, the advancing angle decreased from 58 f 3 to 28 f 2" and the receding angle from 15 f 6 to 0". The absence of any crack pattern on the rubbed surface after a Li+-Sa+ ion exchange treatment is consistent with very high coverage. The unrubbed surface under the same ion-exchange conditions had a dense crack pattern. Coating Thicknesses. Table I1 gives the average thicknesses of some of the titania coatings used in this study. The thickness increment per coating treatment is essentially constant for 0.6% tetraisopropyl titanate solutions, decreases slightly with increasing for 7.1% polydibutyl titanate, and fluctuates somewhat for 1.1% polydibutyl titanate. For the last, the apparently smaller increment below n = 10 is consistent with average thicknesses calculated from the coverages and titania patch heights obtained from electron micrographs. For example, the average thickness calculated from electron micrographs for n = 5 is 6 to 8 mp. It was not possible tb calculate thicknesses and refractive indices using ellipsometric measurements on surfaces prepared from 1.1% polydibutyl titanate solution when n was less than 7 . This may be due to the low titania coverage (less than 357,) on these surfaces. ~

~~

~

~~

Table I1 : Average Thicknesses of TiO, Coatings on Glass" From 0.6%

tetraisopropyl titanate

From 7.170

-From n

10 13 15

1.1% polydibutyl titanatet

n

t

34fll16 9654 5 4 5 4 18 1 1 9 i 3 68 5 4 20 135 i 4

polydibutyl titanate n

t

2

50iS

4 9 0 i 8 6 117 i 3

n

i

1 7=t2 3 1 5 i 3 6 35 i S

'n = number of coating treatments with organic titanate solution; t = thickness in millimicrons.

Discussion Interpretation of Wettability Data. 1 . Adsorption Properties and Effect of Surface Roughness. A glass surface is negatively charged in water and strongly adsorbs long-chain cations from aqueous solutions." The extent of the adsorption depends on the connposition of the glass. The decreasing contapt angles (17) L. Ter-Minassian-Saraga, Advances in Chemistry Series, KO.43, American Chemical Society, Washington, D. C., 1964, p 232

Volume 69, h'rrmber 6

M a y 1965

ROBERTH. DETTREAND RULON E. JOHNSON, JR.

1512

of Figures 1 and 2 , combined with the evidence that titania coverage increases with the number of coating treatments, suggest that either the titania does not adsorb long-chain cations from aqueous solution or, if it does, it does not retain them after rinsing with water. The large increase in advancing contact angle observed in going from zero to two coating treatments in Figure 4 indicates that titania adsorbs and retains long-chain anions and glass does not. This behavior of glass is quite similar to that observed for mica and silica surfaces.l8 The observed wettability behavior shown in Figures 1 to 3 and in Table I cannot be due to an increase in surface roughness as the number of coating treatments increases. If, as we have suggested, the titania patches do not adsorb long-chain cations, then these patches constitute completely wettable surface asperities and, as such, should have the same effect on contact angle as if they were absolutely flat. However, if the patches do adsorb long-chain cations, thereby becoming less wettable, one would expect the advancing angles in Figures 1 to 3 and in Table I to increase with the number of coating treatment^.'^^^^ Since they do not, we conclude that our suggestion is valid and that surface roughness has a negligible effect in the studies involving treatment with the quaternary ammonium salt solution. An estimate of the effect of surface roughness when the titania patches are hydrophobic can be obtained from Figure 4 where the slight increase in advancing angle with the number of coating treatments is probably due to both increased surface roughness and increased coverage by hydrophobic regions. The magnitude of this change suggests that the effect of roughness is relatively small. However, the large standard deviations in Figure 4 are probably due to surface roughness combined with nonuniform coverage by the titania patches. Asperity heights (25 nip average and 75 nip maximum), estimated from electron micrographs, indicate that the magnitude of the energy barriers which control hysteresis on rough surfaceslg should be small. Also, paraffin wax impressions of coated surfaces were specular and showed no difference in water wettability (+a = 110 =t l o , = 103 i 2') from that of impressions of uncoated glass previously reported20 for the same sample of paraffin wax. We conclude, therefore, that the effects observed are due primarily to variations in coverage by high- and low-contact-angle regions and that surface roughness has, a t most, a second-order effect. &. W'ettability us. Coveyage. If, for each coating treatment, the increase in coverage by organic titanate +?

The Journal of Physical Chemistry

(and, consequently, TiOs) is proportional to the fraction of the surface still uncovered, then the fraction covered after n treatments, fn, is given by n

fn = f l

a=1

(1 - f l ) a - l

(2)

wherefl is the fraction covered after the first treatment. Extrapolation of the advancing contact angle curve of Figure 1 to zero yields n = 50, the number of coating treatments necessary for essentially complete coverage by titania. If is assumed to be 0.99 i 0.005, then eq. 2 requires that fl be 0.08 to 0.09. Similar treatment of the data in Figure 2 gives a zero advancing angle at n = 8 and a value of 0.50 for fl when fs is 0.99. These fl values and the coverage variation given by eq. 2 are in reasonably good agreement with the results of electron microscopy. Figure 5 gives the wetkability data of Figures 1, 2 , and 4 as a function of titania coverage determined by using the above fl values in eq. 2 . The previously discussed result for the rubbed, coated surface is in good agreement with the dotted curve of Figure 5 . Applying the above coverages to the data in the last two columns of Table I gives the dashed curves in

'4

D :so

'

l

_---ADVANCING ANGLES /

PER CENT COVERAGE BY Ti02

Figure 5. Water wettability of titania-coated glass after treatment with surfactant solutions plotted as a function of coverage by titania. Solid curves: using contact angle curves of Figure 1 and titania coverage calculated for fi = 0.08 in eq. 2. Dotted curves: using contact angle curves of Figure 2 and coverage calculated for .fI = 0.50 in eq. 2. Dashed curves: using contact angle curves of Figure 4 and coverage calculated for f i = 0.08 in eq. 3. (18) D. J. O'Connor and J. V. Sanders, J . Colloid Sci., 11, 158 (1956). (19) R. E. Johnson, Jr., and R. H. Dettre, Advances in Chemistry Series, No. 43, American Chemical Society. Washington, D. C., 1964, p . 112.

(20) R. H. Dettre and R. E. Johnson, Jr., i t ~ i d . p. , 136.

CONTACT ANGLEHYSTERESIS

O

20

40

L

60

1513

100 L

P E R CENT COVERAGE BY TI02

Figure 6. Methylene iodide wettability of titania-coated glass after treatment with trimethyloctadecylammonium chloride solution. From data of Table I and titania coverage calculated using eq. 2. Solid curve is given by = 54' and 62 = 0". eq. 1 with

Figure 6. The solid curve is given by eq. 1 using = 54" and %i = 0". The zero value for O2 was chosen because the data in the second and third columns of Table I indicate that titania-coated glass at 100% coverage should be completely wetted by methylene iodide. The wettability curves shown in Figures 5 and 6 agree qualitatively with those obtained from the computer study of an idealized heterogeneous surface (Figure 4 of ref. 2 ) . In view of this good agreement and considering the various features of the idealized surface,* we submit that contact angle hysteresis on sniooth, nondeforniable surfaces requires the existence of surface regior s of different intrinsic wettability or surface energy as previously suggested by Pease2I and Ruch and Bartell.** The magnitude of the hysteresis is determined by such factors as the relative proportions and surface distribution of high- and lowcontact-angle regions, their sizes and surface energies, the surface tension of the wetting liquid, and the vibrational energy associated with movement of the liquid-solid-air interfacial boundary. It can be seen froin the analysis of an idealized heterogeneous surface2 that the absolute values of the energy barriers controlling hysteresis are proportional to the sizes of the different regions and the surface tension of the liquid drop. The hysteresk observed on uncoated glass after treatment with the quaternary aninionium salt solution is an indication of iriconiplete coverage by the adsorbed monolayer; L e . , there are regions on the glass surface, larger than molecular dimensions, where essentially no adsorption takes place. The existence of adsorbed monolayers consisting of discrete patches is well

known.9*23-29It is possible that much of the wettability data reported in the literature for adsorbed monolayers may have been obtained on surfaces with less than complete coverage. In inany instances, however, the reported advancing contact angles are probably very close to the values one would obtain for a close-packed, complete monolayer. This conclusion is based on the observed insensitivity of advancing contact angle to coverage by high-contactangle regions once coverage exceeds about 50%. This insensitivity depends on relatively large highcontact-angle regions and small vibrational effects. When contact angles are measured as a function of surface coverage using techniques in which vibration is deliberately introduced, observed wettability should follow eq. 1. This has been observed for water on composite surfaces of Lucitea and glass.4 3. Interpretation of Other Wettability Data. There are a number of wettability studies reported in the literature which deserve re-examination and interpretation in terms of our present concepts. -4 study by Bartell and Bristo1"j of the effect of water adsorption on the wettability of glass and quartz by acetylene tetrabromide showed that the dynamic advancing angle is insensitive to water vapor pressure at pressures approaching saturation, but the dynamic receding angle increases drastically, approaching the advancing angle as the pressure approaches saturation. The increase starts at vapor pressures corresponding to the beginning of multilayer adsorption; at lower vapor pressures, the change in receding angle is relatively small. We interpret these data as evidence of adsorption of water, at the lower pressures, in multilayer patches to give a heterogeneous surface consisting of high-contact-angle water regions and low-contactangle regions where little or no water is adsorbed. As the vapor pressure increases, the water patches coalesce to a coniplete multilayer film. The existence of "clusters" of adsorbed water on glass has been reported by Frazer. 30 (21) D. C. Pease, J . Phys. Chem., 49, 107 (1945). (22) R. J. Ruch and L. S.Bartell, ibid., 64, 513 (1960) (23) J. Karle, J . Chem. Phys., 17, 500 (1949). (24) H. T. Epstein, J . Phys. Chem., 54, 1053 (1950). (25) J. P . Ryan and J. W. Shepard, ibid., 59, 1181 (1955). (26) (a) H. E. Ries, Jr., and W. A. Kimball, ibid., 59, 94 (1955); (b) H. E. Ries, Jr., and W. A. Kimball, "Proceedings of the Second International Congress of Surface Activity," Vol. I , Butterworth and Co. Ltd., London, 1957, p. 75. (27) I. N. Plaksin, ibid.,Vol. 111, p. 355. (28) R. T. Mathieson, Nature, 183, 1803 (1959). (29) G. D. Cheever and E. G. Bobalek, Ind. Eng. Chem., Fundumentuls, 3, 89 (1964). (30) J. H. Frazer, Phys.

Rev.,33, 97 (1929).

Volume 69, Number 6

M a y 1966

1514

The dynamic advancing angles reported by Bartell and Bristol are somewhat higher than the corresponding static values. This observation is consistent with our views of a heterogeneous surface since we would expect dynamic angles to be equivalent to static angles of very low vibrational energy. High sensitivity of advancing angles and relative insensitivity of receding angles to the treatment history of silica and glass surfaces were observed by Bartell and Wooley31for liquids such as acetylene tetrabroiuide and a-bronionaphthalene. The above systems are ones in which coverage by low-contact-angle regions is very high and one would expect, by analogy to the solid curve in Figure 5, slight variations in coverage by high-contact-angle regions ( L e . , water) to produce large changes in the advancing angle. Bartell and Ray32 found that the receding angles for water on surfaces of cellulose derivatives are strongly df.pendent on the mode of forniation of the surface, the angles being considerably lower on surfaces formed against glass than on surfaces formed in air. The corresponding advancing angles show little or no dependence on the mode of surface formation. The explanation offered by Bartell and Ray is that, in surfaces formed against glass, molecular orientation occurs which causes the surface to become more hydrophilic. In addition, we suggest that this results in an increase in the riuniber and/or size of the hydrophilic or low-contact-angle regions in the surface. Since it is very probable that coverage by such lowcontact-angle regions is sniall, we would expect slight variations in coverage by these regions to produce large changes i n the receding angle and sniall changes in the advancing angle. Bartell and Ruch5l6 have measured contact angles of water, hexadecane, and several other liquids on partially depleted monolayers of octadecylamine. Depletion can affect the nionolayer structure in two different ways. It can change the sizes and shapes of the high-contart-angle regions without changing their intrinsic wettabilities, or it can occur uniformly throughout these regions keeping their sizes and shapes constant but changing their intrinsic wettabilities. Each of these affects the wettability characteristics differently, and both probably occur in any real system. With the first mode of depletion the wettability rurve resembles those of Figure 5 , and with the second mode the :tdvancing angle decreases as depletion proreeds. The wet1 ability characteristics are sensitive not only to the geoinetry of the surface heterogeneity but they also depend on the wetting liquid. Since the energy barriers controlling hysteresis are proportional to the The ,Joltma1 o f Physical Chemistry

ROBERTH. DETTREA N D RULOSE. JOHNSON, JR.

surface tension of the wetting liquid, higher surface tension liquids will show greater hysteresis, and the advancing angle will tend to remain high at greater extents of depletion. If the molecules of the wetting liquid can be incorporated into the monolayer, the wettability will be less sensitive to depletion. These concepts are consistent with the observations of Bartell and R ~ c h . ~ , ~ However, their conclusion that zero advancing rontact angles are obtained at 30 to 4oy0 coverage by an adsorbed film is not consistent with our view which would require zero coverage for complete wettability. The desirability of comparing the results of the ellipsometric method, which they used to determine coverage, with electron micrographs of surfaces with incomplete coverage has already been suggested. A number of other wettability studies give results which are consistent with our observations. Shepard and Ryan,' using a radiometric technique to estimate coverage, reported that hexadecane advancing angles are independent of coverage by adsorbed perfluorooctanoic acid films with greater than 5oy0coverage on surfaces such as glass, platinum. and aluminum. RIoreover, finite angles (20-30") were reported for surfaces which had radiometric readings corresponding to coverages of 5 to 10%. Brockway and Jonesg recently reported wettability studies on partial monolayers of cerotic and behenic acids which had been adsorbed on glass in discrete patches. Coverage was estimated from electron niicrographs. From 10 to 80% coverage by the acids, methylene iodide advancing angles are slightly greater than those calculated from eq. 1 for O1 = 70" and 0 2 = 41", the angles corresponding to 100% and zero coverage, r e ~ p e c t i v e l y . ~From ~ 10 to about 35% coverage, hexadecane advancing angles on partial films of cerotic acid are slightly greater than those calculated from eq. 1 for O1 = 45" and e2 = 0". The deviations are larger than the estimated probable error in the nieasurenients. Rlethylene iodide advancing angles measured by Gaines* on partial monolayers of radiostearic acid on mica, platinum, and glass are larger than those calculated from eq. l for 81 = 71" and 0 2 = 41" for mica; O1 = 71" and O2 = 23" for platinum; el = 71" and O2 = 26" for c h r o m i ~ n i . ~ ~ ~ ~ ~

(31) F. E. Bartell and A. D. Wooley, J . Am. Chem. Soc., 55, 3518 (1933). (32) F. E. Bartell and B. R. Ray, ibid., 74, 778 (1952). (33) Here we have used advancing angles in eq. 1. If we had used receding angles for 8 2 , the positive deviations would have been even larger.

REVERSIBLE UNFOLDING OF RIBONUCLEASE AND

POLY-~-BENZYL-~-GLUTAMATE

We appreciate the assistance Of Kay, who made the electron micrographs, T' E' Beukelnian~ who carried Out the X-ray studies, and 0 . E. Schupp, Jr., for his electron diffraction measurements.

w. 8.

1515

(34) Methylene iodide contact angles, even when measured on clean, high energy surfaces such as glass and platinum, are subject t o some variation depending on the humidity conditions under which they are determined. For mica the angle was assumed to be the same as t h a t of glass (Table I ) ; for platinum the value reported by Timmons and Zisman" was used; the angle on chromium was measured in our laboratory. All values are for 50% relative humidity and 20-25'.

Influence of Pressure on the Reversible Unfolding of Ribonuclease

and Poly-7-benzyl-L-glutamate'

by S. J. Gill and Robert L. Glogovsky Department of Chemistry, University of Colorado, Boulder, Colorado

(Received September 63,1964)

The effect of pressure on the thermal transitions of ribonuclease aqueous solution a t pH 2.80 and of poly-y-benzyl-L-glutamate (PBG) in dichloroacetic acid and 1,Zdichloroethane has been investigated to 1400 atm. by optical rotation measurements. An increase in pressure enhances unfolding for both ribonuclease and PBG. The pressure dependence on the extent of unfolding can be used to calculate the volume change for assumed basic steps of the transition process. For ribonuclease the volunie change as calculated for a single-step reaction mechanism was found to be sufficiently smaller than that given by direct determinations of Holcomb and Van Holde, so that the reaction mechanism must consist of more than one reaction step. For the PBG transition an application of the helixcoil transition theory of Zinini, Doty, and Iso, along with the measured pressure dependence upon the extent of transition, showed that the volume change of the process is a t the experimental limit of direct determinations. The shift in the transition temperature with pressure for PBG was used to estimate a heat capacity change a t constant pressure of approximately 140 cal./mole-deg. for the helix to randoin coil transition.

Introduction Thermally induced transitions of niacroniolecular structures have been found for a variety of polynier materials. The reversible transitions of ribonucleases-6 and synthetic polypeptides'-'0 have been the subjects of experiniental and theoretical investigation. One of the primary purposes of such studies has been to gain insight into the mechanism of the transition process. In the case of ribonuclease, the strategy has been to interpret the extent of reaction as a function of temperature by assunied siniple reaction mechanisms.

The heat of the assunied reaction may then be found the dependence Of the extent Of reaction upon transition." Until recently,12direct calorimetric nieastl,F2rayE;;2wworted

by the National Science Foundation

(2) L. Afandelkern, ,8Crystallization of Polymers,,, R.lcGraw-HI1l Book CO., Inc., New York, N. Y., 1964. (3) H. Scheraga, "Protein Structure," Academlc Press, New York, N. Y . , 1961, PP. 270-287. (4) J. G . Foss and J. A. Schellman, J Phys. Chem., 63, 2007 (1959). (5) J. Hermans, J r . , and H. A. Scheraga, J . Am. Chem. Soc.. 83, 3283

(1961).

Volume 69, A-umber 5

M a y 1965