Adsorption studies of polystyrene on silica I. Monodisperse adsorbate

Adsorption studies of polystyrene on silica I. Monodisperse adsorbate

Adsorption Studies of Polystyrene on Silica I. Monodisperse Adsorbate CHARLES VANDER LINDEN AND ROBERT VAN LEEMPUT Universit~ Libre de Bruxelles, Facu...

765KB Sizes 0 Downloads 31 Views

Adsorption Studies of Polystyrene on Silica I. Monodisperse Adsorbate CHARLES VANDER LINDEN AND ROBERT VAN LEEMPUT Universit~ Libre de Bruxelles, Facultk des Sciences, Chimie Macromol(culaire, C.P. 206/3, Bd du Triomphe, 1050 Bruxelles, Belgique Received November 30, 1977; accepted May 23, 1978 Adsorption measurements in a 0 solvent (cyclohexane) and in carbon tetrachloride have been carried out at 35°C on amorphous silica with polystyrene samples (6 × l02 < M < 2 × l06) having a narrow molecular weight distribution. Data were obtained by means of infrared spectroscopy analysis for the adsorbance (q), the fraction of the segments adsorbed (p), and the relative number of sites occupied (0). The surface excess values (O/p) are in good agreement with those obtained independently from adsorbancy and site density data. A similar dependence on M is observed in both solvents, however, with the q and 0 values in CCI~ reaching only half those measured in cyclohexane. The molecular weight dependence of q initially follows a root mean square relationship, but becomes vanishingly small for M > 10s; p decreases rapidly toward a limiting value of 0.20-0.25; 0 equals unity in cyclohexane except for the shortest chains. INTRODUCTION

tails depend on the length and flexibility of the chain; they will undergo changes according to the thermodynamical conditions resulting from the mutual interactions between the three constituents present: the polymer, the solvent, and the adsorbent. The segment density distribution along a perpendicular to the interface displays various profiles, among which an exponential decrease from the surface may be taken as a crude approximation. The width of such a distribution would depend on the quality of the solvent, whereas the number of anchored segments would be ruled mostly by the preferential segment-surface interactions (1). Besides the results obtained by means of computer simulation, theories have progressed along two main axes: On the one hand, using a mono- or multilayer model, the interface has been analyzed with recourse to the Flory-Huggins theory in its zero-order approximation (1-4). On the other hand, the conformation of the isolated chain has been described in a more rigorous

Despite numerous practical applications of polymer adsorption phenomena at the liquid-solid interface in various fields, the structure of the adsorbed layer is not yet well established, except possibly from a theoretical point of view, where some progress has been made in recent years. This situation arises from the unique behavior displayed by macromolecules interacting with a surface, the interaction being governed by the intramolecular equilibrium between successive parts of the chain, some of which are tightly anchored to the surface (bidimensional trains), while others are maintained in the vicinity of the surface by one or both ends, with their segments extending more or less into the liquid phase as tails or loops. Such a model for the adsorbed chain accounts for the fact that the total number of segments associated with one surface site, i.e., the surface excess, is usually larger than unity. The relative sizes of loops, trains, and 48 0021-9797/78/0671-0048502.00/0 Copyright © 1978 by Academic Press, Inc. All rights of reproduction in any form reserved.

Journal of CoUoid and Interface Science, Vol. 67, No. 1, October 15, 1978

ADSORPTION

STUDIES

OF POLYSTYRENE

manner by means of a statistical mechanics formalism (5). However, in the latter case, the solution of the set of equations is so involved, or even not accessible in a suitable analytical form, that some approximation must be called upon in reducing the real value of the model. Basing his theoretical approach on self-avoiding walk results, Silberberg (6) took into account the excluded volume effect, but was forced to introduce an independent conformation parameter (TaTs) whose physical meaning is obscure. In fact, quantitative agreement is not to be expected between the theories, but nevertheless, some similarities in their predictions are worth emphasizing. Experimental verification remains a difficult task, because of the possible combinations of various controlling factors: polymer, adsorbent, solvent, and temperature, not to mention some inaccuracy which always affects the physicochemical characterization of the adsorbate and of the adsorbent. Each of these elements may affect, in some way, the conclusions one might draw about the shape of the isotherm currently described by means of a Langmuir relationship (7). Indeed, it must be stressed that the surface of the adsorbent seldom behaves in a unique fashion as far as the energy of interaction is concerned, which would lead one to postulate a differential occupation of the surface sites as a function of the adsorbance. Moreover, the molecular weight distribution can also play a role and induce selective adsorption. Thus, the isotherm is governed by a large number of factors, but would seemingly be little sensitive to the detailed structure of the interface; besides, most theories predict a limiting plateau (1, 2, 6) or sometimes a slight slope (8) in the low concentration range. The adsorbance (q), commonly expressed in milligrams of polymer per unit weight (or better, by unit surface) of adsorbent, is the relevant quantity, which is taken in conjunction with either the fraction of the segments

ON SILICA

49

adsorbed (p) or the thickness of the adsorbed layer. However, the surface excess (F2) is actually the quantity of interest, owing to its molecular and fundamental character, but it is seldom measured because some value for the mean surface area occupied by one segment is required. The relatively simple case of gas-solid interface already shows much scatter in the adsorption data for small molecules on various substrates (9). Moreover, a strict correspondence with such monomeric analogs must be precluded on the grounds of differences in configurational freedom between an isolated segment and a chain segment. Strictly speaking, another route may be sought, provided some form of interaction between segments and surface sites is postulated. Then only the nature and the density of the latter must be established. In this respect, nonporous silica behaves as a suitable substrate with surface groups whose density may be controlled, allowing a sensitive response in the infrared region to any interaction. Only sparse data have been gathered at the present time on the adsorption of polymers with a narrow molecular weight distribution. Therefore, this work has been carried out on anionic samples, and some parameters characterizing the adsorbed layer have been measured as a function of molecular weight. The data will be compared with the theoretical predictions made by Roe (1) and by Silberberg (6), A parallel monolayer model (2) has not been considered here, the magnitude of the surface excess precluding such a structure for the adsorbed layer. Comparisons will be made on different quantitative levels, depending on the theory considered. The set of equations put up by Roe has been dealt with by means of a computer program to perform the needed iterations within a reasonable time interval and computer memory occupancy. As for the analysis along Silberberg's lines, far Journal of Colloid and Interface Science, Vol. 67, No. 1, October 15, 1978

50

VANDER LINDEN AND VAN LEEMPUT

more exacting in these respects, his own published results have been used. Comparison will then be established for this case, assuming a macromolecular chain model with an appropriate choice of 0.1 for the flexibility parameter (YBYS). Strong interaction has been postulated between segments and surface sites, i.e., the preferential segment-surface interaction parameter Xs = ~, so that solvent molecules are completely excluded from the surface, the fraction of the surface sites occupied (0) being thus equal to unity. This model will be considered from a thermodynamical point of view in two boundary situations: (i) a pseudo-ideal solution for which the polym e r - solvent interaction parameter X = 0.5; this is the case of a polymer dissolved in a 0 solvent (25); (ii) an athermal solution, where X = 0. The systems under examination fall between these two limiting cases. MATERIALS

AND

METHODS

Materials Linear polystyrene samples with a narrow molecular weight distribution were used throughout this study; they are listed in Table I. Some were purchased from Pressure Chemicals Inc. (PC); others were prepared in this laboratory by anionic polymerization methods in vacuo described elsewhere (10) (L). One of the latter, designated by the letter f, was fractionated. Cyclohexane and carbon tetrachloride, of spectrometric grade, were used without further purification, being satisfactorily free of water and unsaturated compounds. Nonporous silica Aerosil 130 and Aerosil OX50 (Degussa) have been selected, with measured specific surface areas (BET) of 141 and 67 mZ/g, respectively. They were heated at 250°C for 18 hr under vacuum (p < 10-3 Torr) and subsequently stored under nitrogen. This ensures an average surface site density close to three OH groups per 100 /~2 (11). Titration with methylmagnesium iodide disclosed, on Journal of CoUoid and Interface Science, Vol. 67, No. 1, October 15, 1978

TABLE I Polymer CharacterizationData Manufacturer ( x 10 3) Sample PS2000 PS655 PS300 PS97 PS67 PS23 PSI0 PS2 PS0.6

This work (GPC) ( x 10-a)

Source

Mn

Mw

M~

Mw

PC PC L,f L L L PC PC PC

~1581F 640 -97 66.6 23.2 9.6 2.1 0.52-0.58

2050 678 ---21.1 ~10.6~ ~2.3a ~ 0 . 5 7 - 0 . 6 4a

-555 272 94 65 -9.4 1.77 0.56

-760 318 103 71 -10.5 2.08 0.67

Calculated from the

Mw/M~ value reported.

average, 17% fewer active sites for the Aerosil OX50, with respec t to the Aerosil 130 sample.

Techniques o f Measurements Three sets of measurements were performed on a Perkin-Elmer Model 457 spectrometer, equipped with two variablepath length cells. The following quantities were measured: (i) the total amount of polymer adsorbed, i.e., the adsorbance, and the equilibrium concentration of the polymer in the supernatant solution; (ii) the fraction of the surface sites occupied; and (iii) the fraction of segments adsorbed. Adsorbance. Seven to forty grams of a polystyrene solution, in the concentration range from 1 to 5 mg/g, was put in contact with a known amount of silica (0.1 to 0.5 g) in a Teflon tapered vessel. Special care was taken in choosing the technique of agitation; finally, a gentle magnetic stirrer was selected for its efficiency. Equilibrium was attained within 1 hr, a much shorter time interval than that required by a tumbling action. The supernatant solutions have been checked by gel permeation analysis for possible degradation (12). Fraction o f the surface sites occupied (0). Part of the gel, including the silica, the adsorbed polymer, and the supernatant solution, was put into a cell, the thickness

ADSORPTION STUDIES OF POLYSTYRENE ON SILICA of which was set at a rough value of 10-4 m. The infrared spectrum (Fig. 1) disclosed the growth of a band near 3580 cm -1 with a consequent reduction of the band located at 3700 cm -1, which is assigned to the isolated silanols. Thus, both the number of free sites (3700 cm -~) and the n u m b e r of sites occupied by segments (3580 c m - 0 were available, provided that the respective extinction coefficients were known. H o w e v e r , the amount of gel introduced into the measuring cell must be precisely evaluated in order to follow the band intensity changes as a function of the surface coverage. Casting the adsorbent in the form of a disk would not o v e r c o m e the difficulties, because the surface area available to the polymer falls so low as to prevent an accurate spectrometric m e a s u r e m e n t . T h e r e f o r e , since a frequency shift from 3700 to 3580 cm -~ suggests an interaction similar to that observed between O H groups and benzene (13), spectra were taken for silica disks immersed successively in carbon tetrachloride and in benzene (Fig. 2). This was conducted in a cell of the type described by L o w and Hasegawa (14). The result was that both coefficients were much alike. 100

t

I

|

i

I

I

51

The total n u m b e r of surface sites was thus given by the properly weighted sum of the absorbances at 3700 and at 3580 cm -1. The fraction of the surface sites covered then followed from the absorbances at both frequencies, measured for the silica in the gel before and after adsorption of the polymer had taken place. Fraction of the segments adsorbed (p). Infrared spectra of polystyrene were taken on a gel portion isolated by centrifugation, according to a procedure described by H e r d et al. (15). The measuring cell contained polymer chains, adsorbed as trains with loops and tails embedded in a solution of unattached macromolecules. The reference cell was filled with the solvent or with a solution of a known concentration of the polymer under study. The interaction between segments and the surface sites brings about a frequency shift from 698 to 702 cm -1 of the adsorption band, attributed to the out-of-plane C - C ring bending of the phenyl group. The very small contribution of the foot of the band peaking at 702 cm -1 to the band at 698 cm -t was neglected. The fraction of the segments adsorbed is readily obtained if one

I

I

I

I

I

I

i

80

60 u aO E ,,n ¢,o Ip-

g

20

~

a 0

o

b I

l

I 3800

3600

°, i

,

3800

Wave

number t

,I 3600

C ,

I i

I 3800

I

I[ 3600

cm-'

FIG. 1. Infrared spectra of Aerosil 130 in cyclohexane, in the presence of polystyrene samples of various molecular weights. (a) PS0.6; (b) PS2; (c) PS10. Journal of Colloid and Interface Science, Vol. 67, No. 1, October 15, 1978

52

VANDER LINDEN AND VAN LEEMPUT 100

I

I

a 80

60 0

E c

40

F-

20

f,i

o 4000

,~

II,

I 3500

3000

Wave n u m b e r ,

cm

.I

FIG. 2. Infrared spectra taken on a silica disk Aerosil 130) immersed in (a) carbon tetrachloride and (b) benzene. knows the silica content of the gel. This was done with special care by taking into a c c o u n t the unavoidable expulsion o f solution on compressing the gel when adjusting the path length. By weighing the centrifugate, an initial value was obtained for the amount of silica dispersed in the polystyrene solution. An aliquot of the gel was evenly distributed in the measuring cell and the optical path was adjusted, avoiding compression; after careful balance for the solvent contribution by means of the reference cell, the spectrum was taken between 4000 and 3400 cm -1, a region where polystyrene does not interfere. An apparent value of the extinction coefJournal of Colloid and Interface Science, Vol. 67, No. 1, October 15, 1978

ficient was thus obtained from the band at 3580 cm -1 for instance, and subsequently used for the measurements at shorter optical path lengths. The obvious source o f scatter in the results arises from the sampling of the gel when filling the measuring cell. EXPERIMENTAL RESULTS AND DISCUSSION

Adsorbance Measurements were carried out in cyclohexane at 35°C to ensure 0 conditions for the polymer in solution (25). The isotherms display the typical shape, with a limiting plateau starting at a low equi-

ADSORPTION STUDIES OF POLYSTYRENE ON SILICA librium concentration ( < 5 × 10 -4, w / w ) ( T a b l e I I ) . F i g u r e 3 s h o w s , as a f u n c t i o n o f molecular weight on a log-log scale, the dependence of the amount of polymer ad-

53

s o r b e d (q), c o n s i d e r e d a t a f i x e d - w e i g h t f r a c t i o n o f 2 × 10 -3 f o r t h e p o l y m e r in t h e s u p e r n a t a n t s o l u t i o n . T h e r e is a d e f i n i t e leveling off on the high molecular weight

TABLE II Adsorption Data for the System Polystyrene/Aerosil 130/Cyclohexane at 35°C Equilibrium concn (mg/g)

Adsorbance q (mg/m~)

0.74 0.83

2.70 2.67

1.97

2.68

PS655

1.13 1.87

2.33 2.41

PS300

0.27 0.37 0.50 0.95

1.85 1.80 2.02 2.20

Polymer

PS2000

PS97

PS67

PS23

PSI0

PS2

PSO.6

Fraction of segments adsorbed p

Fraction of sites occupied 0

(0.25) c

(1.00)

0.21 0.30

1.00

3.8

4.6

1.00

3.7

4.1

1.00

3.6

3.2

0.28

Surface excess 0/p~

Fzb

(4)

5.1

1.75

2.15

2.22

2.14

0.38

0.45 1.06

1.56 1.66

0.31 0.35

1.36

1.66

0.19

1.91

1.68

0.36

0.45 0.49 1.18 1.87

1.43 1.60 1.59 1.66

0.31 0.33 0.29 0.26

1.00

3.4

3.1

0.46 0.97 2.01 2.05 2.74

1.07 1.14 1.19 1.22 1.20

0.28 0.31 0.29 0.37 0.34

1.00

3.0

2.3

0.42 0.98 1.56 1.99 2.32

0.71 0.75 0.68 0.70 0.67

0.48 0.46 0.42 0.40 --

1.00

2.3

1.4

0.30 0.38 0.50 1.71

0.22 0.28 0.34 0.31

0.71 0.87 0.92 0.80

0.64

0.8

0.6

1.82 2.05

0.12 0.14

(1.00)

0.36

0.4

0.3

a Value of p deduced from the best fit on experimental data (Fig. 4). a Calculated form Eq. [2], with noH = 3" 10TM OH groups/m 2. c Values in parentheses are estimated (Fig. 4). Journal of CoUoid and Interface Science, Vol. 67, No. 1, October 15, 1978

54

VANDER LINDEN AND VAN LEEMPUT I

I

I

i

fo

/

10 0 o

E r~ c'qg

0

<~

/ 10

o//

.1

I

I

10 3

10 4

I 10 5

Molecular

I 10 6

weight

FIG. 3. Logarithmic plot of the adsorbance of polystyrene vs molecular weight. Adsorbent, Aerosil 130; solvent, cyclohexane; temperature, 35°C; equilibrium concentration, 2 mg/g.

side, the trend of which might be better assessed by means of a relationship of the form: q = K M a.

[1]

For M < 104, the exponent approaches a value of 0.5 and the adsorption is proportional to the root mean square of the molecular weight. From thereon, the exponent decreases steadily toward zero, reaching this value for M > 2 × 106.

Fraction of the Segments Adsorbed The scatter of the data obscures any trend along the isotherm, but the results plotted in Fig. 4 leave no doubt that the fraction of segments anchored to the surface (p) decreases with M toward a limiting value of 0.20-0.25, and at a faster rate for M < 4 × 104. The agreement with the theory of Silberberg seems, at first sight, excellent, but one should not draw conclusions too hastily about the validity of the theoretical parameters, as this might be only Journal o f Colloid and Interface Science, Vol. 67, No. 1, O c t o b e r 15, 1978

fortuitous. We shall only maintain that the shape of the dependence of p conforms to the predictions put forward by this model. One would be less convinced by the values predicted from the theory of Roe: 1 Depending on the value selected for the parameter Xs, P falls toward a limit situated between 0.33 and 0.46, in any case significantly higher than the experimental asymptote.

Fraction of the Surface Sites Occupied Surface coverage is complete only when the polymer molecular weight is above 104. For shorter chains, the number of anchoring segments per chain decreases; whereas the probability of desorption of the chain taken as a whole increases, the solvent competes more effectively with the polymer for the surface. The fraction of the segments adsorbed has not been considered explicitly by Roe, and was accordingly calculated by means of the ratio ( ~ - ~p*)/~ (~od - ~*) [for symbols, see Ref. (1)].

1

ADSORPTION STUDIES OF POLYSTYRENE ON SILICA I

1.0

~ X \ \

0.8

"X

!

I

".. -

•x ' x ,

\ / • \ ~L¢' ~.

0.6 tO 123

I

55

/

0.4

/

k • -

..............

,

- - \ .~..,. •__._ It/ ............

0.5

0.2

I

I

I

10 3

I0 4

10 5

Molecular

I 10 6

weight

FIG. 4. Semilogarithmic plot vs molecular weight of the experimental relative n u m b e r of sites occupied (©; 0) and of s e g m e n t s adsorbed (O; p). System: polystyrene/Aerosil 130/cyclohexane at 35°C. Theoretical d e p e n d e n c e predicted for 0 conditions: ( ..... ) by Silberberg theory; ( . . . . . . . ) by Roe theory for typical values o f the preferential s e g m e n t - s u r f a c e interaction p a r a m e t e r equal to 0.5, 1.0, and 5.0, respectively.

Surface Excess This quantity, F2, accounts for the total number of segments adsorbed (trains and dangling parts as well) per site and is given by: r~ -

q "NA

- - , n o n • Mu

[2]

where Mu is the molecular weight of the monomeric unit, non the site density, and NA the Avogadro number. All the same, this quantity is also given by the ratio O/p. Keeping in mind that the determination of 0 and p is entirely independent from the absolute magnitude of noH, two totally independent ways are at hand to measure the surface excess. The values of F2 deduced from Eq. [2] and those obtained from the ratio O/p are plotted on a log-log scale in Fig. 5 as a function of molecular weight;

comparison is also made with the theory of Silberberg, for some relevant equilibrium concentrations. One must remember that the latter model is only valid for degrees of polymerization higher than 100, thus pertaining to the region for M > 104. The present results somehow establish a link between chains of "finite" and "infinite" sizes. Considered from a quantitative point of view, the agreement appears excellent for the O/p values, the experimental and theoretical dependence showing the same trend (this was already apparent from Fig. 4, since 0 = 1). The surface excess values calculated using the theory of Roe are also given; the discordance observed here suggests two remarks, which will be commented on before we proceed further. The first is concerned with the geometry of the surface area and dwells on the obvious disparity between theoretical models, makJournal of Colloid and Interface Science, Vol. 67, No. 1, October 15, 1978

56

VANDER LINDEN AND VAN LEEMPUT I

I ~,-- 10 " 2 j¢--- ! 0 " 3

~-1o-4

J

J/ / f

/

// / // //

/

,/,,~/~, //. //

/"

.Y/

u

/

0.5

/

1



/

/" I"

LU 0 . 9

rr"

. .........................

/

bJ

Z) u')

1_,_

,, ~"

/

tO

o

5 .............

/'

tO

X lid

/e

;//

0.8 0.7

i

I"

i

0.6

i i

0.5

i i

0.4

0.3

/

i i i i I

/i

I I I

i i 10 4

MOLECULAR

I

I

10 5

10 6

WEIGHT

FIG. 5. Logarithmic plot of experimental surface excess values vs molecular weight, obtained: (0) by means of Eq. [2], F2 = qNA/(nonMu);(O) from the ratio O/p. System: polystyrene/Aerosil 130/cyclohexane at 35°C. Dependences predicted for 0 conditions: ( . . . . . . . ) from Roe theory for different values of the preferential segment-surface interaction parameter, Xs; (..... ) from Silberberg theory for X~ = ~ and some typical values of polymer fraction ~2. ing u s e o f a plane and an infinite surface, and the actual situation e n c o u n t e r e d in t h e s e e x p e r i m e n t s . S u r e l y e n o u g h , the surface is by no m e a n s infinite, but i n d e e d is so d i v i d e d that the s i z e o f the silica particles lies Journal of Colloid and Interface Science, Vol. 67, No. 1, October 15, 1978

c l o s e to the a v e r a g e m o l e c u l a r d i m e n s i o n s o f the p o l y m e r in solution. H o w e v e r , the incidence of those experimental conditions is p r o b a b l y m i n o r (16), as m a y be a s s e s s e d f r o m the m a g n i t u d e s o f the a d s o r b a n c e s

ADSORPTION STUDIES OF POLYSTYRENE ON SILICA measured on Aerosil OX50 and on Aerosil 130, respectively, whose particle diameters are in the ratio 2:1; if we take into account the difference between the site densities of both substrates, the adsorbances lie in the same range (Fig. 6). Second, the theory of Roe is seen to systematically underestimate the surface excess. The discrepancy would be less pronounced if one would accept an additional contribution, not dealt with in the original theory, such as a phase separation, promoted by the macromolecular layer, but not under the direct action of the surface. In this connection, Silberberg has made an estimate for this kind of contribution (17). Using a crude model, he considered the influence of an accumulation of macromolecules near the surface but free of any contact with it. According to his calculations, a polymer having, for instance, a degree of polymerization of 106 and taken as an infinite chain would meet the thermodynamical I

57

requisites for a critical point to be reached, at least in a 20 ° interval b e y o n d the 0 points, encompassing quite well the experimental conditions leading to the results reported here. This contribution should be molecular weight dependent and, accordingly, could not account for the vanishing experimental dependence depicted in Fig. 3. This is perhaps not a decisive argument, owing to the numerous factors involved; but nevertheless, this assumption should be checked on a system comprising a thermodynamically better solvent, such as carbon tetrachloride. The behavior of this globular solvent molecule toward silica follows that observed for cyclohexane, noting, for instance, that the O H frequencies in the infrared are shifted by the same amount when silica is put into contact with both solvents (13). When the molecular weight of the polymer is low, the specific adsorption measured in carbon tetrachloride (Table III I

C) e') X

o

E

o

w" Z < ,,a nO (/} ,<

0

l

2

ADSORBANCE, mg/m2 130 FIG. 6. Correlation diagram between adsorbance data obtained for polystyrene/cyclohexane at 35°C on two silicas with different particle sizes: Aerosil 130, (b = 16 nm; Aerosil OX50, (b = 4 nm. Journal of ColloM and Interface Science, Vol. 67, No. 1, October 15, 1978

58

VANDER LINDEN AND VAN LEEMPUT T A B L E III Adsorption Data for the S y s t e m Polystyrene/Aerosil 130/Carbon Tetrachloride at 35°C Equilibrium conch (mg/g)

Absorbance, q (mg/m2)

0.20 0.44 0.88 0.92

1.20 1.22 1.21 1.21

0.16 0.19 0.18

PS97

0.40 0.92

1.09 1.06

PS 10

1.25 1.46

PS2

1.64 1.90

Polymer

PS300

Fraction of segments adsorbed p

Fraction of sites occupied 0

Surface excess O/p ~

F2b

0.52

2.9

2.3

0.21 0.18

0.52

2.5

2.1

0.64 0.67

0.18 0.16

0.49

1.5

1.3

0.36 0.42

0.37 0.31

0.32 0.35

0.8

0.8

-

-

Value o f p d e d u c e d from the best fit on experimental data (Fig. 8). b Calculated from Eq. [2], with noH = 3' l0 TM O H g r o u p s / m 2.

and Fig. 7) (for a polymer weight fraction of 10-3) is of the same order of magnitude as that in the pseudo-ideal solution. For longer chains, however, there is a significant difference and the adsorption amounts to one-half that in cyclohexane for M > 105. A decrease with enhanced quality of the solvent is, on the whole, also predicted by the theories. I

c~

_E

The fraction of the segments adsorbed, p (Fig. 8), shows an asymptotic behavior reminiscent of that observed in cyclohexane, but located at a slightly lower level. In compliance with the enhanced quality of the solvent, the surface coverage (Fig. 8) does not reach the magnitude observed in cyclohexane; these observations are in close accordance with some fragmentary

i

I

I

_____¢____--o--

10 °

E

U t2 IO I/I "1o

.<

10 "1

I

1o 3

I

1o 4 Molecular

I

1o 5

I

10 6

weight

FIG. 7. Logarithmic plot of the a d s o r b a n c e of p o l y s t y r e n e vs molecular weight. A d s o r b e n t , Aerosil 130; solvent, carbon tetrachloride; temperature, 35°C; equilibrium concentration, 1 mg/g. Journal of Colloid and Interface Science, Vol. 67, N o . 1, O c t o b e r 15, 1978

ADSORPTION S T U D I E S OF P O L Y S T Y R E N E ON SILICA I

I

I

59

1

0.8

0.6 ¢3. 5O ¢D

t')

ri

9 w

• •

O.4

0.2

8 0

I

I

I

I

10 3

10 4

10 5

10 6

Molecular

weight

FIG. 8. Semilogarithmic plot vs molecular weight of the experimental relative number of sites occupied ((3; 0) and of segments adsorbed ( 0 , p). System: polystyrene/Aerosil 130/carbon tetrachloride at 35°C.

observations made by Joppien ( l l ) . In carb o n t e t r a c h l o r i d e , 0 is j u s t half the amount m e a s u r e d in the 0 solvent. CONCLUSIONS

To sum up, there are n o t e w o r t h y correlations b e t w e e n the results obtained in both solvents; although the fractions of the segments adsorbed a m o u n t to the same values (the difference would not be taken as significant), the a d s o r b a n c e s fall to half their values in carbon tetrachloride c o n c o m i t a n t with a decrease in the surface o c c u p a n c y rates. A c o m p a r i s o n on a quantitative basis with the theoretical results of Silberberg could not be made in the absence o f reported data for X other than 0 or 0.5. On the other hand, the theory of Roe does not predict such a drastic fall of the surface excess with the t h e r m o d y n a m i c quality of the solvent. A contribution of one or m o r e extra layers of m a c r o m o l e c u l e s to those already attached to the surface does not s e e m needed to account for the respective adsorbances in a good or a bad solvent. Thus

we h a v e here s o m e reason to discard the aforementioned a s s u m p t i o n s o m e t i m e s put forward as a factor responsible for the large thickness values e n c o u n t e r e d for the a d s o r b e d layer. The close correlation found b e t w e e n the surface excesses m e a s u r e d by independent means (Fig. 9) strengthens the confidence in the p a r a m e t e r s deduced in either a good or a bad solvent. Data in the literature are sparse and the s y s t e m s under study are frequently ill defined, but a c o m p a r i s o n m a y be a t t e m p t e d with our results, and is in fact m o s t illuminating when restricted to adsorbents not too dissimilar to those used in this work. Our data are spread o v e r a most extended range of molecular weights, and the other data fit satisfactorily (Fig. 10); the differences are ascribable to different thermal treatments affecting the n u m b e r of active sites per unit of surface area. F o r a given molecular weight, the a d s o r b a n c e appears m a x i m u m in a bad solvent and grows lower at higher temperatures. The fractions of the Journal of Colloid and Interface Science,

Vol. 67, No. 1, October 15, 1978

60

VANDER LINDEN AND VAN LEEMPUT I

I

I

I

/

0

/ /

O / / / I~? 4

/0

3

/

U:I (/) LI.I

/ /

c) X

"~

00

/

2

/

hl

(J h nr" :3 U~

/0 I

/

O

/

,Jo (5 0

I

I

I

I

1

2

3

4

SURFACE

EXCESS,

8/p

FIG. 9. Correlation plot for experimental surface excess values, calculated by means of Eq. [2], F2 = qNA/(nonMu); and from the ratio O/p, respectively. System: polystyrene/Aerosil 130 at 35°C. Solvent: (©) cyclohexane; ( 0 ) carbon tetrachioride. I

I

I

I

10 0 9 8 7 0

6

Z

5 ~n n~

<

4

3

2

I

I

I

I

10 3

10 4

10 5

10 6

MOLECULAR

WEIGHT

F I G . 10. Logarithmic plot of adsorbance data vs molecular weight for polystyrene on amorphous silica at 25°C (or otherwise specified) in various solvents. Empirical fit of the present data: ( ) in cyclohexane; ( ..... ) in carbon tetrachloride. Circles represent cyclohexane at 35°C: (O) on aluminum oxide (18); ( 0 ) Ref. (19). Triangles represent carbon tetrachloride: (A) t = 29°C, F2 = 2.2, 0 = 0.5 - * p = 0.23 (11); (&) Ref. (20); (A) Ref. (15). Squares represent trichloroethylene: (IS]) Ref. (16); (11) Ref. (15) (IS]) t = 30°C (19). Crosses represent tetrachloroethylene (21). Journal of Colloid and Interface Science, Vol. 67, No. l, October 15, 1978

ADSORPTION STUDIES OF POLYSTYRENE ON SILICA I

I

I

--

....

I

.o°'**°°*°'°', ~ 2 . ...... .

° . . . . ° ~ i , ° o°°°

...;:::: ....... ul e,,

poe

10 "5 IE

9

(u u~

8

~

6

0 E

4

u ¢,ID

•prnma_

o

pbd

-fi/

3

/peo

0 <

~-

//:

7

/

pS

..-"

...'" /

61

2

pnps I

I

10

I

10 2

Degree

10 3

of

I 10 4

polymerization

FIG. 1I. Logarithmic plot of adsorbance data vs molecular weight for different polymers on amorphous silica (at 25°C, or otherwise specified) in various solvents. Empirical fit of the present data as in Fig. 10. PMMA, polymethylmethacrylate: (©) CCldp = 0.55 (15); (0) benzene (15); (Q) trichloroethylene/p = 0.24 (15); (~) id./p = 0.29-0.37 (16); (@) id./24°C/p = 0.32 (22). PBD, poly(butadiene)/p = 0.30-0.45 (21). POE, poly(oxyethylene) (23). PEO, poly(ethylene o-phthalate)/CHCla/ 35°C/p = 0.29-0.40 (24). PNPS, poly(neopentylsuccinate)/CHC1J30°C (24). segments adsorbed, w h e n reported in the literature, are, for the m o s t part, in fair a g r e e m e n t with those r e p o r t e d here. N o further analysis will be t e m p t e d , b e c a u s e the latter quantities are always obtained under e x t r e m e experimental conditions and could possibly be subjected to systematic errors, depending on the s y s t e m under examination. Other p o l y m e r s were also taken for comparison, a m o n g which p o l y m e t h y l m e t h a crylate has b e e n the m o s t thoroughly studied, owing to the strong absorption band for the carbonyl in the infrared. Polymers such as poly(butadiene), poly(oxyethylene), and poly(ethylene orthophthalate) have also been included. One c o m m o n point has governed the selection made: Each mon-

omeric unit interacts more or less strongly with the O H groups on the surface, forming a hydrogen bond. All of the available data have b e e n plotted in Fig. 11, due normalization being m a d e on a degree o f polymerization scale. Notwithstanding the obvious differences in the flexibility and chemical nature of the chains, and in the interaction energy with the surface, all o f the p o l y m e r s depict a similar b e h a v i o r toward adsorption. REFERENCES 1. Roe, R. J., J. Chem. Phys. 60, 4192 (1974). 2. Slow, K. S., and Patterson, D., J. Phys. Chem. 77, 356 (1973). 3. A s h , S. G . , E v e r e t t D . H . , a n d F i n d e n e g g , G . H . , Trans. Faraday Soc. 64, 2 6 4 5 (1968); Trans. Faraday Soc. 66, 708 (1970).

Journal of Colloidand Interface Science, Vol. 67, No. 1, October 15, 1978

62

VANDER LINDEN AND VAN LEEMPUT

4. Bladon, J. J., Higuchi, W. I., and Malokhia, A. M., J. Colloid Interface Sci. 52, 14 (1975). 5. Chan, D., Mitchell, D. J., and White, L., Discuss. Faraday Soc. 59, 181 (1975) (and references cited therein). 6. Silberberg, A., J. Chem. Phys. 48, 2835 (1968). 7. Kipling, J. J., "Adsorption from Solutions of Nonelectrolytes." Academic Press, New York, 1967. 8. Hoeve, C. A. J., J. Chem. Phys. 44, 1505 (1966). 9. McClellan, A. L., and Harnsberger, H. F., J. Colloid Interface Sci. 23, 577 (1967). 10. Meunier, J. C., and Van Leemput, R., Makromol. Chem. 142, 1 (1971). 11. Joppien, G. R.,Makromol. Chem. 175, 1931 (1974). 12. Vander Linden, C., and Van Leemput, R., J. Colloid Interface Sci. 67, 63 (1978). 13. Curthoys, C., Davydov, V. Ya., Kiselev, A. V., and Kuznetsov, B. V., J. Colloid Interface Sci. 48, 58 (1974). 14. Low, M. J. D., and Hasegawa, M., J. Colloid Interface Sci. 26, 95 (1968).

15. Herd, J. M., Hopkins, A. J., and Howard, G. J., J. Polym. Sci. C 34, 211 (1971). 16. Thies, C., J. Phys. Chem. 70, 3783 (1966). 17. Silberberg, A., J. Colloid Interface Sci. 38, 217 (1972). 18. Burns, H., and Carpenter, D. K., Macromolecules 1, 384 (1968). 19. Howard, G. J., and Woods, S. J., J. Polym. Sci. A2 10, 1023 (1972). 20. Bogacheva, Y. K., Kiselev, A. V., Nikitin, Y. S., and Eltekov, Y. A., VysokomoL Soeyed. A10, 574 (1968). 21. Botham, R. A., and Thies, G., J. Colloid Interface Sci. 45, 512 (1973). 22. Thies, C., J. Polym. Sci. C 34, 201 (1971). 23. Howard, G. J., and McConnell, P., J. Phys. Chem. 71, 2974 (1967). 24. Peyser, P., Tutas, D. J., and Stromberg, R. R., J. Polym. Sci. AI 5, 651 (1967). 25. Flory, P. J., "Principles of Polymer Chemistry." Cornell University Press, Ithaca, N. Y., 1953.

Journal of Colloidand Interface Science, Vol.67. No. 1, October 15, 1978