Langmuir 1991, 7, 665-671
665
Interaction of Poly(N-isopropylacrylamide)with Sodium n-Alkyl Sulfates in Aqueous Solution Howard G. Schild and David A. Tirrell* Polymer Science and Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003 Received August 24, 1990 Cloud point, calorimetric, and fluorescence probe methods have been combined to examine aqueous mixtures of poly(N4sopropylacrylamide) (PNIPAAM) and a series of sodium n-alkyl sulfates of alkyl chain lengths (n)in the range 1-16. Surfactants of chain lengths 14 depress the lower critical solution temperature (LCST) of PNIPAAM and exhibit no evidence of enhanced aggregationin PNIPAAM solutions. This pattern of behavior is characteristic of the simple salt Na2S04 as well. Surfactants of intermediate chain lengths ( n = 5-10) depress the LCST at low surfactant concentration but cause an increase in the LCST at concentrations that exceed the critical aggregation concentration (CAC). Sodium n-dodecyl sulfate ( n = 12) elevates the LCST even at low concentrations and forms aggregates in PNIPAAM solutions at a concentration 10-foldlower than the critical micelle concentration (cmc). Sodium n-hexadecyl sulfate forms aggregates in PNIPAAM solutions at temperatures below its Krafft temperature in water. The formation of polymer-bound micelles below the cmc is reported by pyrene, 1-pyrenecarboxaldehyde,and sodium 2-(N-dodecylamino)naphthalene-6-sulfonate but not by 1-benzoylacetone. The theory of Nagarajan and Ruckenstein, and in particular the treatment of Ruckenstein and co-workers,serves to rationalize the observed aggregation behavior as a consequence of screening of surfactant-water interaction by PNIPAAM chain units at the micellar surface.
Introduction Investigation of the interactions between nonionic polymers and surfactants in aqueous solutions is motivated by biological, industrial, and theoretical concerns. Goddard' has published the most recent review, which summarizes many of the systems and methods investigated to date. Numerous investigations indicate that nonionic polymers such as poly(ethy1ene oxide) (PEO) and poly(vinylpyrrolidone) (PVP) form complexes with anionic surfactants, particularly with sodium n-dodecyl sulfate (SDS).' The presently accepted model of the resulting structure is that proposed by Cabane and Duplessix on the basis of neutron scattering studies.2 Through selective deuteration, it was established that isolated polymer chains bind a number of miniature surfactant micelles along the chain contour. Other experiments have focused on determining the critical aggregation concentration (cac) a t which these attached micelles form and on the associated changes in polymer conformation.' Our studies have sought an understanding of aqueous mixtures of the nonionic polymer poly(N-isopropylacrylamide) (PNIPAAM, 1) and diverse cosolutes. T h e
interactions between PNIPAAM and surfactants such as SDS will have important implications for the emerging practical uses of the Furthermore, whereas (1) Goddard, E. D. Colloids Surf. 1986, 19, 255. (2) Cabane, B.; Duplessix, R. J . Phys. 1982, 43, 1529. (3) Monji, N.; Hoffman, A. S. Appl. Biochem. Biotechnol. 1987, 14, 107. (4) Taylor, L. D.; Cerankowski, L. D. J. Polym. Sci., Part A: Polym. Chem. 1975, 23, 2551.
many studies of polymer-surfactant complexes focus on the change in the critical micelle concentration (cmc) of the surfactant,' fewer studies have examined the corresponding change in polymer conformation. Because the binding of anionic surfactant micelles converts nonionic polymers into charged species, polyelectrolyte behavior may be anticipated. Indeed, characteristic polyelectrolyte viscosity behavior'vs has been observed in such systems.8 PNIPAAM exhibits a lower critical solution temperature (LCST) in water above which it precipitates upon heating.9 Since attached, charged surfactant aggregates create an electrostatic barrier that opposes polymer collapse and aggregation, one would expect enhanced polymer solubility in surfactant solutions as detected by an elevated LCST.' This phenomenon would not be observed in aqueous solutions of PEO and PVP, since their critical temperatures lie above the boiling point of the solution.'O Comblike polymers1' of PEO do exhibit LCSTs in water, as do poly(propy1ene oxide)12 and partially hydrolyzed poly(viny1acetate)s;13J4each of these polymers exhibits elevated cloud point temperatures upon surfactant addition. We report herein a systematic investigation of the interaction of PNIPAAM with a series of sodium n-alkyl sulfates. Solution microcalorimetry (DSC) and cloud point methods were used to study the LCST of PNIPAAM. At the same time we have applied fluorescence techniques to (5) Bae, Y. H.; Okano, T.; Hsu, R.; Kim, S. W. Makromol. Chem., Rapid Commun. 1987, 8, 481. (6) Hoffman, A. S. J. Controlled Release 1987, 6, 297. (7) Okahata, Y.; Noguchi, H.; Seki, T. Macromolecules 1986,19,493. (8) Gilanyi, T.; Wolfram, E. In Microdomains in Polymer Solutions; Dubin, P., Ed.; Plenum: New York, 1985. (9) Heskins, M.; Guillet, J. E. J. Macromol. Sci., Chem. 1968, A2 (8), 1441. (10) Molyneux, P. Water Soluble Polymers: Properties and Behaoior, CRC Press: Boca Raton, FL, 1983. (11)Nwankwo, I.; Xia, D. W.; Smid, J. J. Polym. Sci., Part B: Polym. Phys. 1988,26, 581. (12) Pletnev, M. Yu.; Trapeznikov, A. A. Kolloidn. Zh.1978,40,948, as reported in ref 1. (13) Saito, S. J. Polym. Sci., Part A-1 1969, 7, 1789. (14) Tadros, Th. F. J . Colloid Interface Sci. 1974, 46, 528.
0743-7463/91/2407-0665$02.50/0 0 1991 American Chemical Society
666
Langmuir, Vol. 7, No. 4 , 1991
m o n i t o r c h a n g e s i n the c m c of e a c h s u r f a c t a n t , i n o r d e r t o relate t h e association behavior of the s u r f a c t a n t s to changes i n the LCST. The m e c h a n i s m s of polymer-surf a c t a n t interaction a r e discussed in terms of the models of N a g a r a j a n and Ruckenstein.
Experimental Section Materials. N-Isopropylacrylamide was obtained from Eastman Kodak Co. and recrystallized (mp 64-66 "C) from a 65/35 mixture of hexane and benzene (Fisher Scientific Co.). Acetone (HPLC grade) was also obtained from Fisher. Azobis(isobutyronitrile) (AIBN) from Alfa Chemical Co. was recrystallized from methanol (Aldrich Chemical Co.); decomposition was avoided by maintaining the temperature below 40 "C. Surfactants were obtained as samples of the highest purity available from the suppliers listed below: sodium n-dodecyl sulfate (J. T. Baker), sodium n-decyl sulfate (Kodak), sodium n-octyl sulfate (Aldrich), and sodium sulfate (Aldrich, anhydrous), sodium salts of n-hexadecyl, n-heptyl, n-hexyl, n-pentyl, and n-butyl sulfate (Lancaster Synthesis, Ltd.), sodium ethyl sulfate (Pfaltz and Bailer, Inc.), and sodium methyl sulfate (American Tokyo Kasei, Inc.). Carbon and hydrogen analyses were within 0.3$1 of theoretical for all surfactants. Na analyses were within 1.0% of theoretical for all surfactants; S analyses were found to be lower than theoretical as a result of the fact that combustion yields a poorly soluble product. P y e n e , 1-benzoylacetone, and 1-pyrenecarboxaldehyde were used as received from Aldrich Chemical Co. The sodium salt of 2-(N-dodecylamino)naphthalene-6-sulfonicacid (C12NS) was used as received from Molecular Probes, Inc. Hydrocarbons for surface tensiometry were pentane (HPLC grade, Aldrich), hexane (OmniSolv, E M Science), heptane (HPLC grade, Fisher Scientific Co.), octane (99+ , Aldrich), and decane (99+R, Aldrich). Distilled water was analyzed (Barnstead Co., Newton, MA) to contain 0.66 ppm total ionized solids (as NaCl) and 0.17 ppm total organic carbon (as C). Sodium azide (Fisher) was employed as a bactericide (0.10 (w/v) ' 1 ) in stock polymer solutions. Synthesis. N-Isopropylacrylamide ( 5 g), dissolved in 40 mL of benzene with 1 mol recrystallized AIBN, was degassed through three cycles of freezing and thawing. After polymerization by stirring in an oil bath a t 49 "C for 22 h under a positive nitrogen pressure, the solvent was evaporated. The crude solid was vacuum-dried, crushed, dissolved (acetone, 47 mL), and precipitated by dropwise addition to hexane (600 mL). After being filtered and dried, 3.62 g (76'1 yield) of polymer was obtained. Anal. Calcd for C ~ H I ~ N OC,: 63.7; H, 9.8; N, 12.4. Found: C, 63.5; H , 9.9; N, 12.2. 'H NMR (200 MHz, D2O) d 1.0 (CH?, 6 H), 1.2-2.1 (-CH2CH-, 3 H), 3.7 (CH, 1 H). No vinyl protons were detected. IR (CHC13 cast film) 3300, 2960, 2925, 2860,1635,1530,1455,1375,1390,1170,1130,750cm-1.Absent were the 1620 cm-' (C=C), 1410 cm-I (CH2=), and C-H vinyl out-of-plane bending vibrations observed in th? spectrum of the monomer. Gel p_ermeation chromatography: M , = 440 000, M, = 160000, M,/M,, = 2.8. S a m p l e P r e p a r a t i o n . The PNIPAAM concentration in all measurements was 0.40 mg/mL. Stock solutions of concentration 4.00 mg/mL were prepared through dissolution of the polymer in distilled water with 0.1 w/v c( sodium azide for several days. This polymer solution (0.20 mL) was diluted to 2.00 mL with distilled water or with salt or surfactant stock solutions. The salt and surfactant solutions (except that of sodium n-hexadecyl sulfate) were prepared by addition of known volumes of water to weighed amounts of surfactant. Concentrations are thus millimolal (mm) which are equivalent to millimolar (mM) a t lower concentrations. Samples that precipitated a t room temperature were refrigerated for several hours to allow for redissolution. Microliters of methanolic stock solutions of sodium n-hexadecyl sullate were evaporated in vials to which diluted polymer stock solution was added. Stock solutions for fluorescence measurements were prepared as described above, with slight modification for probe incorporation. Where pyrene or 1-pyrenecarboxaldehyde was used, microliters of a millimolar acetone solution were added to sample vials with subsequent evaporation of the acetone prior to the addition of the polymer solution. Sufficient probe was incor-
Schild and Tirrell porated to prepare a ca. 1 @ Maqueous solution. For l-benzoylacetone, the PNIPAAM stock solution was diluted with surfactant stock solutions prepared in 62 pM aqueous l-benzoylacetone instead of distilled water. For C12NS, samples were prepared by diluting the PNIPAAM stock solution with replacement of 0.10 mL of distilled water with 0.10 mL of a 12.6 @M C12NS stock solution (final concentration 0.63 WM). All samples were vortexed prior t o spectral measurements. Measurements. Infrared spectra were obtained on films cast from chloroform on NaCl plates with a Perkin-Elmer 1320 infrared spectrophotometer. NMR spectra were obtained on a Varian XL-300 spectrometer. Gel permeation chromatography (GPC) was performed with a Waters M45 solvent pump coupled to an R410 differential refractometer and a Hewlett-Packard 3380A digital integrator. Degassed tetrahydrofuran (THF, Aldrich, HPLC grade) was eluted at 1.1 mL/min through four Waters @Styragelcolumns (lo6, lo5, lo4, lo3 A). N,N-Dimethylformamide (DMF, Aldrich, HPLC grade) was eluted a t 1.0 mL/ min through three Waters @Bondage1columns (E-1000, E-500, E-125). Polystyrene standards (Polysciences) were used for calibration; molecular weights are thus estimated as those of polystyrenes of equivalent elution volume. Cloud points were determined to within f0.5 "C by measuring the abrupt, temperature-dependent increase in optical density a t 500 nm with a Beckman DU-7 spectrophotometer coupled to a Lauda RM-6 circulating bath. Temperatures were manually ramped a t rates of ca. 0.4 "C/min and monitored by an Omega 450-ATH thermistor thermometer. The same temperature control system was used for scattering studies, which employed a Perkin-Elmer MPF-66 fluorescence spectrophotometer. These measurements utilized 500 nm as both the excitation and emission wavelength and monitored the change in scattering intensity observed a t 90". Calorimetric transition temperatures were obtained to within f O . l "C with a Microcal, Inc., MC-1 scanning microcalorimeter (DSC) a t a scanning rate of 15 "C/h. Samples were degassed and transferred to the sample cell with a calibrated syringe. For surfactant concentrations greater than ca. 100 mm, a polymer-free surfactant solution was placed in the reference cell; a t lower concentrations, a distilled water reference was used. The former solutions provided improved baseline control with no other effect on results. Calibration was achieved by supplying a precisely known current to the reference cell of the calorimeter. Samples were scanned only once, as subsequent scans could suffer from ambiguity associated with slow redissolution of the polymer.9 Spectra for 1-benzoylacetonewere obtained in the wavelength range 200-350 nm with a Beckman DU-7 spectrophotometer. Pyrene, C12NS, and 1-pyrenecarboxaldehyde emission spectra were obtained with a Perkin-Elmer MPF-66 fluorescence spectrophotometer exciting a t 337, 303, and 365.5 nm, respectively. Slit widths were 3 nm for pyrene and 5 nm for the other probes. All measurements were done a t 24.5 A 0.5 "C, unless indicated, by coupling the spectrophotometer with a Lauda RM-6 circulating bath. Calculations were done by using the T K Solver program on a Macintosh S E computer. Interfacial tensions were measured by the Wilhelmy plate method with a Rosano surface tensiometer (Biolar Corp.) and by the du Nouy ring method with a Fisher Model 21 Tensiomat following ASTM method@ a t room temperature (ca. 24.5 "C).
Results and Discussion Synthesis of PNIPAAM. Free-radical polymerizat i o n of N-isopropylacrylamide i n benzene with AIBN as initiator results i n precipitation of PNIPAAM. The conditions used i n t h i s work yielded a h i g h molecular weight, polydisperse s a m p l e (&Iw = 440 000; M , = 160 000; based on polystyrene s t a n d a r d s ) that was used t h r o u g h o u t t h e s e s t u d i e s at a fixed concentration of 0.40 m g / m L . Aqueous Solutions of PNIPAAM. O u r previous studies16 established that t h e coupling of microcalorime t r y with cloud p o i n t measurements c o n s t i t u t e s a productive a p p r o a c h to exploration of LCST p h e n o m e n a in a q u e o u s PNIPAAM solutions. Figure 1 shows typical (15)Annual Book of ASTMStandards; American Society for Testing and Materials: Philadelphia, PA, 1986; Vol. 8.02, pp D971 and D1331. (16) Schild, H.G.; Tirrell, D. A. J . Phys. Chem. 1990,94,4352.
Langmuir, Vol. 7, No. 4, 1991 667
Mixtures of PNIPAAM and Sodium n-Alkyl Sulfates 1.80
7'
1.50 1.20
*r
0.90
0.60 0.30 0.00 28
29
30
31
32
33
34
35
36
Temperature ("C)
Figure 1. Cloud point curve ( 0 ) and microcalorimetric endotherm for PNIPAAM (0.40 mg/mL) in water. 2.w
8.70 mm
>
I
H 0.6 O C
Temperature ("C) Figure 3. Microcalorimetric endotherms for PNIPAAM (0.40 mg/mL) with added SDS. Temperatures of peak maxima are plotted in Figure 4. a
0
10
20
30
40
50
60
70
SDS (W) Figure 2. Change in optical density upon heating PNIPAAM (0.40 mg/mL) through the LCST (ca. 25-35 "C) with varying amounts of SDS. No further increase in optical density occurs in the clear solutions up to the boiling point. calorimetric and turbidimetric results, in which the LCST is manifested through an abrupt change in optical density a t 32.2 "C and a calorimetric endotherm centered a t 32.4 "C. The LCSTs reported by these two methods are insensitive to a ca. 10-fold variation in heating rate. Effects of SDS on the LCST of PNIPAAM. Eliassaf reported in 1978l' that no precipitation was observed upon boiling a PNIPAAM solution containing 1%(ca. 35 mM) SDS. Indeed, our standard cloud point method (cf. Experimental Section) reports a change in optical density upon heating through the LCST only in solutions that contain less than 10 pM SDS, i.e., fewer than ca. six SDS molecules per polymer chain (Figure 2).18 In seeming contradiction, the microcalorimetric results shown in Figure 3 support the persistence of the LCST up to considerably higher surfactant concentrations. The temperature and width of the demixing transition increase rapidly with increasing [SDS]; the calorimetric enthalpies and peak heights decrease. The discrepancy between the cloud point and calorimetric results is resolved by using a more sensitive scattering method; measurement of 90" scattering (cf. Experimental Section) yields well-defined cloud points a t least up to 5 mM SDS. Apparently, the precipitated particles are too small to cause visible turbidity a t high surfactant concentrations.19 Retarded intermolecular aggregation can be rationalized as a consequence of interparticle repulsions resulting from introducing charged micelles along the polymer chain. The scattering cloud points agree very well with the calorimetric LCSTs (Figure 4). The rapid rise in the LCST with increasing SDS concentration is almost certainly (17) Eliassaf, J. J. A p p l . Polym. Sci. 1978, 22, 873. (18) 0.40 mg/mL PNIPAAM = 3.5 mM repeat units = 1018 chains/L (DP = ca. 2000) versus 10 HM SDS = 6 X 10'8 molecules/L. (19)A similar observation has been reported by Hoffman and co-workers: Cole, C.; Schreiner, S. M.; Priest, J. H.; Monji, N.; Hoffman, A. S. ACS Symp. Ser. 1987, No. 350, 245.
0
5
10
15
20
25
30
35
40
45
50
55
60
Concentration (mm) Figure 4. LCSTs of PNIPAAM (0.40 mg/mL) in aqueous solutions with added SDS ( 0 , O )and sodium n-decyl sulfate (m, 0).Filled symbols refer to cloud points; open symbols refer to calorimetric transitions. associated with surfactant binding and intermicellar repulsion; above ca. 17 mM SDS, no LCST is observed. Effects of Other Sodium n-Alkyl Sulfates on the LCST. Solution Turbidity. The surfactant concentration below which PNIPAAM visibly precipitates above the LCST increases as the n-alkyl chain length decreases. For sodium n-decyl sulfate and sodium n-octyl sulfate, these concentrations are ca. 3 and 100 mM, respectively. For all shorter hydrocarbon chain lengths and for sodium sulfate, PNIPAAM solutions are visibly turbid above the LCST a t any sulfate concentration. Thus the precipitated particle sizes are smallest in solutions of surfactants of longest alkyl chain lengths. LCST Trends. Cloud point and calorimetric results for aqueous mixtures of PNIPAAM with sulfates of chain length 0 (Na2S04) to 12 are summarized in Figures 4-7. With increasing concentrations of surfactant, the LCST The initially decreases except for chain lengths (n)2 minimum value of the LCST is reduced a t short chain lengths, and for n < 5, a monotonic decrease in the LCST is observed. For n = 5-8, the LCST passes abruptly through a minimum and then rises to temperatures above its surfactant-free value. The rise in the LCST is shifted to higher sulfate concentrations for smaller n. A plateau (20) The initial decrease in the LCST observed in solutions of sodium n-decyl sulfate is within the experimental error in our measurements.
Schild and Tirrell
668 Langmuir, Vol. 7, No. 4, 1991 64 56 h
48
5
40
32 24 0
0
1000
500
1500
3000
2500
Zoo0
Concentration (mm) Figure 5. LCSTs of PNIPAAM (0.40 mg/mL) in aqueous solutions with added sodium n-alkyl sulfates: ( , 0 )n = 8; (A, A) n = 7; ( 0 ,0)n = 6; (m, 0)n = 5 ; (m, 0)n = 4. Filled symbols refer to cloud points; open symbols refer to calorimetric tran-
600
300
900
1200
1500
1800
Concentration (mm) Figure 7. LCSTs of PNIPAAM (0.40 mg/mL) in aqueous solutionswith added sodium n-alkylsulfates and sodium sulfate: ( 0 , O ) n = 4; (A,A) n = 2; (m, 0)n = 1; (e, 0) Na2S04. Filled symbols refer to cloud points; open symbols refer to calorimetric transitions.
sitions. 35 _I
1.90
I
1.70
I1 I,
1.50
1.30
IO0 294 0
.
I
25
.
I
I
50
75
.
I
100
.
I
125
.
I
150
.
I
175
. 1
sitions.
value of the LCST is subsequently reached upon further addition of surfactant. As the n-alkyl surfactant chain length decreases, this plateau temperature decreases and the onset of the plateau shifts to greater concentrations. Beyond this apparent saturation plateau, the LCST is observed to decrease for n < 10. It is reasonable to suggest that the plateau coincides with saturation of the polymer by surfactant and that free micelles formed in solution behave in the manner of simple salts, which serve to reduce the LCST of PNIPAAM.16 All of our experiments were done a t concentrations below those of any mesomorphic transitions in the surfactants.21 The expanded curves shown in Figure 6 focus on the initial depression of the LCST that occurs on adding surfactants with alkyl chain lengths less than or equal to CS. Divergence of the cloud point and calorimetric results for n = 7 and 8 is attributed to the breadth of the calorimetric peaks, which precludes precise identification of the temperature of the maximum in ACp. This phenomenon is observed only near the minimum in the LCST vs surfactant concentration curves for n = 5-8. For n < 5 , the LCST is never elevated (Figure 7 ) ;it is depressed below the freezing point of the mixture a t even higher concentrations. Sodium sulfate is most effective in depressing the LCST; even if one accounts for the higher ionic strength of these solutions, it is a stronger LCST depressant than any of the alkyl sulfates. As the shortest (21) Attwood, D., Florence, A. T., Eds. Surfactant Systems; Chapman
and Hall: London, 1983.
'
IO
IO
lo4
Concentration (mm)
200
Concentration (mm) Figure 6. Expanded plot of LCSTs of PNIPAAM (0.40 mg/ mL) in aqueous solutions with added sodium n-decyl, n-octyl, and n-heptyl sulfates from cloud point and microcalorimetric data: (A,A ) n = 10; ( 0 , O ) n = 8; (m, 0)n = 7. Filled symbols refer to cloud points; open symbols refer to calorimetric tran-
10
Figure 8. 11/13of pyrene (ca. 1 MM)solubilized in increasing concentrationsof sodium n-alkyl sulfates at 24.5 "C: (0) n = 12; ( 0 )n = 10; (0) n = 8; (m) n = 7 ; (A) n = 6; (A) n = 4.
n-alkyl chain length surfactants used here do not form micelles, it is reasonable to attribute the sudden increase in the LCST for n = 5-8 to micellar binding to PNIPAAM. This point is discussed further below. Critical Surfactant Concentrations. Pyrene was applied as a polarity probe to detect the cmc's of the sodium n-alkyl sulfates. An abrupt drop in the ratio (11/13) of the intensities of the 0,O (373 nm) and 0,2 (385 nm) bands in the emission spectrum occurs as pyrene is solubilized in the more hydrophobic environment of the micellar s u r f a ~ t a n t . ~The , ~ ~presence - ~ ~ of pyrene a t a concentration of 1 pM was found not to affect the LCST of PNIPAAM as determined by cloud point and microcalorimetric methods; furthermore, 1 1 / 1 3 was found to have the same value (1.87 f 0.03) in water and in aqueous solutions of PNIPAAM (0.40 mg/mL). Turro and co-workers have shown that pyrene does not perturb the cmc's of anionic surfactants as determined by surface tension measurements.25 Typical data for micelle formation in polymer-free solution are illustrated in Figure 8. In solutions of increasing surfactant concentration, 1 1 / 1 8 remains a t its surfactant-free value until the cmc, where it abruptly decreases to a lower plateau value. Table I summarizes the cmc's calculated from the inflection points of such curves; in general, we find excellent agreement with (22) Kalyanasundaran, K.; Thomas, J. K. J . Am. Chem. Soc. 1977,99, 2039. (23) Turro, N. J.; Baretz, B. H.; Kuo, P. L. Macromolecules 1984, 17, 1321.
(24) Winnik, F. M.; Winnik, M. A.; Tazuke, S. J. Phys. Chem. 1987,
91, 594.
(25) Ananthapadmanabhan, K. P.; Goddard, E. D.; Turro, N. J.;Kuo,
P. L. Langmuir 1985, 1, 352.
Langmuir, Vol. 7, No. 4, 1991 669
Mixtures of PNIPAAM and Sodium n-Alkyl Sulfates Table I. Properties of Sodium a-Alkyl Sulfates in Aqueous PNIPAAM Solutions8 cmc
n
1it.b
12 10 8
33
7 6 5
cac expt
8.3
expt
-AC,e cal/mol
Cmlnd
calcdc
7.1 0.79 0.73f 0.18 0-0.75 33 7.5 6.2f 1.3 0-6.0 110 64 393~6 40 260 190 1 1 0 f 10 140 580 560 3003Z 30 350 1000 1000 670f60 800 4500 4500 f g h h f g h h f g
130 260 520 1040
4 2 1
1300 880 320 190 20 0 0 0 0
All concentrations in millimolal (mm). References 25-27. Calculated values for critical aggregation concentrations as discussed in text. d Surfactant concentration a t which the LCST reaches its minimum value. e AG = -RT In (cac/cmc). f Cannot fit model with physically meaningful 8. R Monotonic depression of LCST to below solution lreezing point. No decrease in 1 1 / 1 3 with increased sodium n-alkyl sulfate concentration. 0
1M
*.w
l
I1 I3
.
1.60 1.40
1.20
8
0
h
7
-
: & CAC
1.00 10.'
.
-4
CMC
'
loo
.....JTr
' " " " ' ~
10 I
lo2
'
'
_
'
lo3
[SDS]mm Z1/Z3 of pyrene (ca. 1pM) solubilized in water (0) and aqueous PNIPAAM (0,0.40 mg/mL) solutions with increasing concentrations of SDS a t 24.5 "C.
Figure 9.
literature val~es.26-2~No micelles were formed from the ethyl and methyl species; Le., no depression of 11/13 was observed for these solutions. The observed cmc's were independent of temperature in the range of 20-40 "C as e~pected.~' Figure 9 shows that in solutions of 0.40 mg/mL PNIPAAM, the abrupt drop in 11/13 is shifted to lower surfactant concentration, e.g., for SDS, from 7.1 to 0.79 mm. The decrease in microenvironmental polarity in this case is attributed to the formation of polymer-bound micelles a t a critical aggregation concentration (cac) some 10-fold lower than the cmc. The minimum in 11/13 observed in the concentration regime where polymerattached micelles dominate (0.79-7.1 mm) suggests that the polymer-bound micelles are less polar than free SDS micelles. This contrasts with reports on mixtures of SDS with PEO or with PVP23-30 wherein 11/13 in the presence of polymer never drops below the value characteristic of (26) Aniansson, E. A. G.; Wall, S. W.; Almgren, M.; Hoffmann, H.; Kielmann, I.; Ulbricht, W.; Zana, R.; Lang, J.; Tondre, C. J . Phys. Chem. 1976, 80, 905. (27) Mukerjee, P.; Mysels, K. J. Critical Micelle Concentrations of Aqueous Surfactant S y s t e m ; NSRDA-NBS-36; U S . Government Printing Office: Washington, DC, 1971. (28) Birdi, K. S. In Micellization, Solubilization & Microemulsions, Mittal, K. L., Ed.; Plenum: New York, 1977; p 151. (29) The identification of the cmc with the initial break in the 1 1 / 1 3 curve, rather than with the inflection point, would appear to make more physical sense. On the other hand, Goddard and Turro and their coworkers have shown that the initial break consistently falls below the cmc determined by surface tension mea~urements.2~ The excellent agreement shown in Table I between our measured cmc's and those reported in the literature supports the use of the inflection point to estimate aggregation concentrations. We do not believe that any of our conclusionsare affected by the choice of this estimation method. (30) Zana, R.; Lianos, P.; Lang, J. J . Phys. Chem. 1985, 89, 41.
1,4 %
1.2010 -5
10.~
Concentration (M) Figure 10. Z1/13of pyrene (ca. 1 pM) solubilized in water (0) and aqueous PNIPAAM ( 0 , 0.40 mg/mL) solutions with increasing concentrations of SHS at 24.5 "C. The solubility limit of surfactant at this temperature is approached at the highest two concentrations used.
SDS micelles. The minimum observed in PNIPAAM solutions must be associated with the greater hydrophobic character of the polymer (PNIPAAM exhibits a lower LCST in aqueous solution than either PEO or PVP'O and is more effective in reducing the interfacial tension between water and hydrocarbons (vide infra)). Above the cmc, 11/13 averages to the micellar value as pyrene is partitioned in large part into free micelles. Goddardl has pointed out the puzzling observation that this convergence in 1 1 / 1 3 does not occur until one reaches ca. 10-fold higher SDS concentrations in solutions of PVP.22 As the chain length of the surfactant decreases, the gap between the cac and cmc narrows (Table I), and for n I 6, the cac and cmc are indistinguishable. The standard free energy relationship31(AG = -RTln (cac/cmc)) permits quantitative estimation of the excess free energy that stabilizes polymer-surfactant complexes versus free micelles. As shown in Table I, the magnitude of this stabilization decreases from 1300 to 20 cal/mol as the chain length is reduced from C12 to Cg. The role of the surfactant chain length is demonstrated most strikingly by the behavior of sodium n-hexadecyl sulfate (SHS,Figure 10). SHS has a Krafft t e m p e r a t ~ r eof ~ ~31, ~"C, ~ below which its solubility is less than the cmc (0.55 mM). Micellization is thus precluded. Nevertheless, in PNIPAAM solutionspolymer-bound micelles do form (and solubilize pyrene) a t 24.5 "C; in fact, 11/13 is depressed a t M SHS to an extent that requires 10-fold higher concentrations of SDS. Depression of the Krafft temperature of SHS by added PVP has been reported previously by Schwuger and Lange.34 Relationship between the CAC and the LCST. Table I lists for each surfactant the concentration ( G i n ) a t which the LCST of PNIPAAM reaches its minimum value, and Figure 11 shows typical plots of the variation of I l / I 3 and of the LCST with surfactant concentration (results shown for n = 6). The agreement between the values of Cminand the CAC is reasonably good, regardless of n, and suggests that it is the binding of micelles that causes elevation of the LCST.35 We propose that elevation of the LCST is a result of electrostatic repulsion between charged, polymer-bound micelles, which oppose polymer collapse and aggregation. (31) Chu, D.; Thomas, J. K. J . Am. Chem. SOC.1986, 108, 6270. (32) Schwuger,M.J. In StructurelPerformanceRelationshipsin Surfactants; Rosen, M. J., Ed.; American Chemical Society: Washington, DC, 1984. (33) Lange, V. H.; Schwuger, M. J. Kolloid-2. 1968, 223, 145. (34) Schwuger, M. J.; Lange, H. Proc. 5thInt. Congr. Surf. Act. Agents 1968, 2, 955, as reported in ref 1. (35) We believe that the systematic negative deviation of C,i, from the cac is a consequence of the relatively broad transition in11/13,coupled with our use of the inflection point (rather than the initial break) to define the cac. There can be little doubt that C,i, and the cac show similar dependences on surfactant chain length.
Schild and Tirrell
670 Langmuir, Vol. 7, No. 4, 1991 54.0 f -.\
1.90
48.6-
- 1.80
43.2-
- 1.70
h
3b
-1.60
1, -
11.50
I3
37.8-
32.41 1 27.0!. I 0 200
.
-
I
I
400
600
I . 800
I
I
- 1.40 11.30
loo0 1200
Concentration (mm) Figure 11. 11/13ofpyrene (m, ca. 1 pM) solubilized in aqueous PNTPAAM (0.40 mg/mL) with added sodium n-hexyl sulfate
superimposed on the LCSTs (0, 0 ) of these solutions. Filled circles refer to cloud points; open circles refer to calorimetric transitions.
The behavior of short chain alkyl sulfates is consistent with this proposal. Methyl and ethyl sodium sulfates do not form micelles and PNIPAAM does not induce their formation. In solutions of sodium n-butyl sulfate, micelles appear to form a t the highest concentrations examined in polymer-free solution (Figure 8),but mixtures with PNIPAAM are turbid even upon freezing a t these and higher concentrations up to the limit of solubility of the surfactant. Thus, even if they are formed, polymerbound micelles of chain length C4 do not solubilize PNIPAAM. Modeling Polymer-Surfactant Complexation. Nagarajan and Ruckenstein have applied multiple equilibrium m 0 d e l s ~ 6 to - ~the ~ description of polymer-surfactant complexation. We adopt herein the treatment of Ruckenstein and c o - w o r k e r ~in, ~which ~ the free energy change that accompanies the formation of micelles of size i in the absence of polymer (2q.q) is written as
+
Api = APHC,~ ApC
+ a ( a - a,) - kT In (1- a,/a) +
'Felectr (1) where J p ~ c = l ~-(2.05 + 1.49n)hT, the free energy advantage conferred when the tail of the surfactant of chain length n is transferred from water to a liquid hydrocarbon medium. The second term, Apc = 40.50 0.24n)kT, corrects for the constraining of these tails by attachment to polar headgroups "tied" to the micellewater interface. The third term accounts for residual contact between the hydrocarbon core and water; u represents the hydrocarbon-water interfacial tension of 50 dyn/cm, a is the area per surfactant molecule a t the micellar interface, and a, is the area per surfactant molecule shielded from water by the polar headgroup (which for sulfates is estimated as ap = 17 A2, the polar headgroup area). The fourth term is due to surface exclusion of the headgroups caused by their finite size. The final term is a Debye-Huckel approximation to the electrostatic repulsion between the ionic headgroups
Lpelectr=
(me2P2/2cr)(l+ d a i ) / ( l + dai + air)
and is a function of the sodium counterior. radius, ai = 1 A, the electronic charge, e, the dielectric constant, e = 80, the micellar radius calculated by assuming a spherical ( 3 6 ) Nagarajan, R. Polym. Prepr. (Am. Chem. Soc.,Diu. Polym. Chem.) 1981, 22 (2), 33. (47) Nagarajan, R.; Harold, M. P. Solution Behaoior of Surfactants; Mittal, K. L., Fendler, J. H., Eds.; Plenum: New York, 1982; Vol. 2. (38) Nagarajan, R. Colloids and Surf. 1985, 13, 1. (39)Nagarajan, R. J . Chem. Phys. 1989,90,1980. (40) Ruckenstein, E.; Huber, G.; Hoffmann, H. Langmuir 1987,3,382. (41) Nagarajan, R.; Ruckenstein, E. J . ColloidInterface Sci. 1979, 71, 580. (42) Nagarajan, R. Adu. Colloid Interface Sci. 1986, 26, 205.
aggregate of m surfactant molecules, r, the degree of dissociation of the headgroups, @, and the reciprocal Debye length, d = (c1 + ~ , d d ) ' / ~ / ( 3 . 0 8X lo-') cm-', where Cadd is the added ionic salt concentration and c1 is the monomeric surfactant concentration. R u ~ k e n s t e i nsuggests ~~ that the primary role of the polymer is one of shielding of the headgroups and the hydrocarbon cores of the micelles from water. The third term of eq 1 (for the case where a, is less than the cross sectional area of the hydrocarbon tail) then becomes ( a - Aa)(a -a,)
+ up Aa,
where Au and AaP are the changes in the interfacial tension between the hydrocarbon core and water and between the headgroups and water, respectively, caused by polymer. Ruckenstein takes these two parameters to be equal, and simplifies this term to u(a - a,)
-
a Aa
+ 2a, A a
One then estimates A a as the difference of the interfacial tension between water and hydrocarbon and the interfacial tension between aqueous polymer solution and hydrocarbon. We have measured A u to be 33 f 2 dyn/cm, using 0.40 mg/mL PNIPAAM (as was used in all our experiments) and hydrocarbons ranging from pentane to decane. This value was independent of the hydrocarbon selected and of PNIPAAM concentration over 2 orders of magnitude. Similar behavior was reported for PE0,40 though Au was found to be only 20 dyn/cm. Following Ruckenstein, we calculated the cac for each of the sodium n-alkyl sulfates in 0.40 mg/mL PNIPAAM, using the P calculated from the experimental cmc data and the modified term that takes into account the interfacial properties of the polymer (Au = 33 dyn/cm). Table I lists the resulting calculated cac's. Agreement with the experimental cac's (and with Cmin) is excellent for the longest alkyl chain lengths, but for n < 8 the calculated cac is systematically lower than the experimental value. Nagarajan observed a similar poor fit with short hydrocarbon tails42in his analysis of mixtures of PVP with sodium n-alkyl sulfates. Because the assumptions of the model (in particular the assumption of a sharp water-hydrocarbon interface) are most likely to break down a t short chain lengths, the limited success of the model is to be expected. In fact, given that there are no adjustable parameters in the cac calculation (once is calculated from the cmc), the performance of the model is commendable. Results with Other Fluorescent Probes. 1-Pyrenecarboxaldehyde exhibits blue-shifted emission maxima in nonpolar media, and the blue shift can be used as a reporter of surfactant a g g r e g a t i ~ n . ~Turro ~ and cow o r k e r ~have ~ ~ suggested that results obtained with this probe are in closer agreement with those of surface tension measurements than results obtained by using pyrene. As shown in Figure 12,l-pyrenecarboxaldehydereports a cmc of 8.0 mm for SDS in water and a cac approximately 1 order of magnitude lower in solutions containing 0.4 mg/ mL PNIPAAM. The PNIPAAM/SDS complex appears (as before) to be less polar than free micelles between the cac and cmc, and we find a t higher concentrations the averaging effect found with pyrene. The origin of the increase in the apparent breadth of the transition to the micellar form of the surfactant in PNIPAAM solution is not clear. The position of the tautomeric equilibrium of 1-benzoylacetone shifts toward the enol form in hydrophobic (43) Kalyanasundaram, K.; Thomas, J. K. J . Phys. Chem. 1977, 81, 2176.
Mixtures of PNIPAAM and Sodium n-Alkyl Sulfates 480
-
Langmuir, Vol. 7, No. 4, 1991 67 1 430 1
h
475v
470-
2 .d
465:
3w
460-
'2 W
455-
.e
450-
.
.
1
..I
F i g u r e 14. Wavelength maxima of ClzNS (0.63 pM) emission and aqueous PNIPAAM (0,0.40 mg/mL) solutions in water (0) with added SDS a t 24.5 "C.
observed. At the concentration (0.63 gM) used in these experiments, ClzNS itself has no effect on the LCST of PNIPAAM. At higher concentrations of ClzNS in aqueous PNIPAAM solutions, we observe the formation of polymerattached Cl2NS micelles, concurrent with an increase in the LCST. These results are reported elsewhere.48
Conclusions
environments, including micelles.44 As shown in Figure 13, this probe shows an abrupt increase in the intensity of the absorption associated with the enol form (at 313 nm) and an accompanying drop in that of the keto absorption (at 249 nm) with increasing addition of SDS. These transitions occur at the cmc of SDS. In contrast to the results obtained with the pyrene probes, these curves do not shift appreciably upon addition of PNIPAAM. Thus, 1-benzoylacetone does not appear to be solubilized in the polymer-attached micelles detected by the fluorescent probes. This observation is particularly interesting in view of previous studies that showed that fluorosurfact ant^^^ and nonionic poly(oxyethy1ene) surfactant^^^ are unable to solubilize the enol form of l-benzoylacetone. Finally, we chose the sodium salt of 2-(N-dodecylamino)naphthalene-6-sulfonic acid (C12NS) as an amphiphilic probe. The octadecyl analogue (ONS) has been applied as a polarity probe47 by Waggoner and Stryer. The emission maximum blue shifts with decreasing polarity: micromolar solutions of ClzNS in water, 50/50 water/ methanol, methanol, and ethanol emit a t 430, 421, 415, and 411 nm, respectively. Figure 14 shows the transition curves o b t a i n e d in SDS solutions and in PNIPAAM/SDS mixtures. The curves are similar to those obtained by using pyrene, although lower critical surfactant concentrations (a cac of 0.67 mM and a cmc of 7.1 mM) are (44) Meguro, K.; Shoji, N. In Solution Chemistry of Surfactants; Plenum: New York, 1979; Vol. 1. (45) Suzuki, T.; Esumi, K.; Meauro, K. J . Colloid Interface Sci. 1983, 93, 205. (46)Saito, Y.; Sato, T. J. Phys. Chem. 1985, 89, 2110. (47) Waggoner, A. S.; Stryer, L. Proc. Natl. Acad. Sci. U.S.A. 1970,67 (2), 579.
Codissolution of PNIPAAM and sodium n-alkyl sulfates in aqueous media is accompanied by marked changes in the solubility properties of both polymer and surfactant. In general, PNIPAAM promotes surfactant aggregation and enjoys enhanced water solubility (as reflected in elevation of the LCST) as a result of the binding of surfactant micelles. Both these effects become most apparent as the length of the surfactant tail increases and are lost a t chain lengths of 4 or less. A t surfactant concentrations less than those required for micellar binding, the LCST of PNIPAAM is depressed with respect to the LCST of the polymer in pure water. Surfactant aggregation is reported by the fluorescent probes pyrene, l-pyrenecarboxaldehyde, and sodium 2-(N-dodecylamino)naphthalene-6-sulfonate, but not by 1-benzoylacetone. The emission behavior of the pyrene and naphthalene probes suggests that polymer-bound micelles are somewhat less polar than the corresponding surfactant aggregates formed in polymer-free solution. The theory of Nagarajan and Ruckenstein, in a form that assigns the polymer the role of modulating the tension at the micelle-water interface, is successful in predicting the observed aggregation behavior of all but the shortest chain length surfactants.
Acknowledgment. This work was supported by a National Science Foundation Predoctoral Fellowship to Howard G. Schild and by a grant from the U.S. Army Research Office (DAAL03-88-K-0038). We thank Professor D. A. Hoagland for helpful suggestions and Professor R. Nagarajan for instructive correspondence. Registry No. 1 (homopolymer), 25189-55-3; H(CH2)w OS03H-Na, 151-21-3; H(CHz)loOS03H.Na, 142-87-0; H(CH2)sOS03H.Na, 142-31-4;H(CH2)T0SO3H.Na, 18981-98-1;H(CH2)eO S 0 3 H - N a , 2207-98-9; H ( C H 2 ) 5 0 S 0 3 H . N a , 556-76-3; H(CH2)40S03H-Na,1000-67-5; H(CH2)20S03H.Na,546-74-7; CH30S03H.Na, 512-42-5;C12NS.Na, 129985-62-2;sodium sulfate, 7757-82-6;pyrene, 129-00-0; 1-pyrenecarboxaldehyde, 3029-19-4. (48) Schild, H. G.; Tirrell, D. A. Langmuir 1990, 6 , 1676.
