Microcalorimetric detection of lower critical solution temperatures in

May 1, 1990 - Cation-Induced Hydration Effects Cause Lower Critical Solution Temperature Behavior in Protein Solutions. Olga Matsarskaia , Michal K. B...
1 downloads 10 Views 629KB Size
J . Phys. Chem. 1990, 94, 4352-4356

4352

Microcalorimetric Detection of Lower Critical Solution Temperatures in Aqueous Polymer Solutions Howard G. Schild and David A. Tirrell* Polymer Science and Engineering Department, University of Massachusetts, Amherst, Massachusetts 01 003 (Receioed: October 16, 1989)

A sensitive solution microcalorimeter was used to study lower critical solution temperatures (LCSTs) of aqueous polymer solutions. Endotherms with enthalpies on the order of the strength of hydrogen bonds were observed at temperatures concurring with LCSTs detected by classical cloud-point measurements for solutions of poly(N-isopropylacrylamide) (PNIPAAM), poly(viny1 methyl ether) (PVME), poly(propylene glycol) (PPG), and hydroxypropylcelluloe (HPC). The LCST of PNIPAAM was found to be dependent on the chain length, generally increasing with decreasing &tnin the range A?, = 1.6 X los to 5.4 X IO3. Treatment of the calorimetric endotherm according to a two-state transiton model afforded cooperative unit sizes of the order of the chain length for PNIPAAM, PVME, and PPG. The apparent cooperative unit for the demixing transition of HPC was found to be larger than a single chain, consistent with previous observations of aggregation of HPC below the LCST. Depression of the LCST by added salts was observed.

Introduction Lower critical solution temperatures (LCSTs) are found in many polymer solutions characterized by strong hydrogen bonds.'S2 The formation of hydrogen bonds between solutes and solvents lowers the free energy of solution; however, the specific molecular orientations required by these bonds lead to negative entropy changes and positive contributions to the free energy. This phenomenon is particularly important in aqueous media where a further negative entropy change is contributed by the hydrophobic effecta3 Precipitation (or a coil-to-globule transition in very dilute solution^)^-^ is observed in such systems above an LCST, when the enthalpic contribution to the free energy is dominated by the growing entropic component at temperatures below the boiling point. Systems that behave in this fashion' can be created by incorporating hydrophobic units into water-soluble polymers through polymer modification or copolymerization. For the most part, experimental characterization of these LCST phenomena has been limited to simple cloud-point measurements,s-8 an approach complicated by variations in precipitated aggregate sizes and settling of precipitate^.^,^^ A more powerful technique that may be applied to the study of the LCST is solution calorimetry;5*7 8 11.12 this method provides thermodynamic parameters that lend insight into the forces responsible for the transition. Use of a sensitive scanning microcalorimeter developed to study structural transitions in proteins, lipids, and nucleic acids'517 allows the LCST transition 8

(I) (2) (3) 1973. York, (4)

5

Walker, J . A.; Vause, C. A. Sci. A m . 1987, 253, 98. Tager, A. Physical Chemistry of Polymers; Mir: Moscow, 1972. (a) Tanford, C. The Hydrophobic Effect, 2nd ed.; Wiley: New York, (b) Ben-Naim, A. Hydrophobic Interactions; Plenum Press: New 1980.

Williams, C.; Brochard, F.; Frisch, H. L. Annu. Rec. Phys. Chem.

1981, 32, 433. (5) Yamamoto, I . ; Iwasaki, K.; Hirotsu. S . J . Phys. SOC.Jpn. 1989, 58,

?in _.".

( 6 ) Fujishige, S.; Kubota, K.; Ando, I . J . Phys. Chem. 1989, 93, 3311.

(7) Taylor, L. D.;Cerankowski, L. D.J . Polym. Sci., Part A : Polym. Chem. 1975, 13, 2551. (8) Wolf, B. A. Pure Appl. Chem. 1985, 57, 323. (9) Schild. H. G . Ph.D. Dissertation, University of Massachusetts, 1990. ( I O ) Cole, C.; Schreiner, S. M.; Priest, J. H.; Monji, N.; Hoffman, A. S. ACS Symp. Ser. 1987, 350, 245. (11) Maderek, E.; Wolf, B. A. Polym. Bull. 1983, IO, 458. (12) Heskins, M.; Guillet, J . E. J . Macromol. Sci., Chem. 1968, A2, (8), 1441. (13) Krishnan, K. S.; Brandts, J. F. Methods Enzymol. 1978, 49, 3. ( I 4) Mabrey-Gaud, S. In Liposomes: From Physical Structure to Therapeutic Application; Knight, C. G . , Ed.; Elsevier/North-Holland Biomedical Press: Amsterdam, 1981. (15) Hinz, H . J . Methods Enzymol. 1986, 130, 59. (16) Battistel, E.: Luisi, P. L.; Rialdi, G. J . Phys. Chem. 1988, 92, 6680.

0022-3654/90/2094-4352$02.50/0

parameters to be determined with precision. Our main concern is the LCST observed in aqueous solutions of poly(N-isopropylacrylamide)(PNIPAAM) and the perturbation of this transition by added c o s o l ~ t e s . ~This report presents cloud-point and microcalorimetric data for PNIPAAM samples of various molecular weights and polydispersities, in pure water and in aqueous salt solutions. These techniques are then extended through application to several other polymers that exhibit LCSTs in aqueous solution, including poly(viny1 methyl ether) (PVME), poly(propy1ene glycol) (PPG), and hydroxypropylcellulose (HPC).

Experimental Section Materials. The synthesis and properties of the PNIPAAM samples used in this work are summarized in Table I. Two samples (N1 and N12) were obtained from Professor Allan S . Hoffman of the University of Washington. The other samples were synthesized from N-isopropylacrylamide obtained from Eastman Kodak Co., which was recrystallized (mp 64-66 "C) from a 65/35 mixture of hexane and benzene (Fisher Scientific Co.). The salts used in the redox polymer synthesis were obtained from Fisher except for magnesium sulfate (Mallinkrodt) and ammonium persulfate (J. T. Baker). Tetramethylethylenediamine (TEMED) was used as received from Kodak; chloroform, tetrahydrofuran (THF), and methanol were purchased from Aldrich Chemical Co. and were of HPLC grade. Wet cellulose dialysis tubing was no. 6, 1000 molecular weight cutoff, from Spectrum Medical Industries, Inc. Acetone (HPLC grade) was obtained from Fisher and azobis(isobutyronitri1e) (AIBN) from Alfa Chemical Co. The latter was recrystallized from methanol (Aldrich), avoiding decomposition by maintaining the temperature below 40 "C. PVME was purchased as 50 wt % ' solutions in toluene or in water (Aldrich). The toluene solution (100 g) was further diluted to 600 mL with toluene and precipitated in 2 L of n-heptane (Fisher). The resulting rubbery solid was vacuum dried and recovered in 35% yield; GPC (THF): ,%fw = 155000; I@, = 83 OOO; Mw/,%f,= 1.9. GPC (DMF) gave a peak molecular weight (M,) of 140000. PVME in aqueous solution (190 g) was precipitated by heating and then azeotropically dried by redissolving the polymer in benzene (300 mL) and restripping the solvent three times. The yield upon a final vacuum drying to constant weight was 77%. GPC (DMF): broad and bimodal with the major peak corresponding to M p = 70000. HPC (average molecular weights: 100000, 300000 and lOOOOOO), PPG (average molecular weight: lOOO), polyethylene oxide (PEO) (average molecular weight: 100 000), polyacrylamide (PA) (average molecular weight: (17) Jackson, W.; Brandts, J. Biochemistry 1970, 9, 2294.

0 1990 American Chemical Society

Lower Critical Solution Temperatures in Aqueous Polymers

The Journal of Physical Chemistry, Vol. 94, No. 10, 1990 4353

TABLE I: Synthesis and Properties of Poly(N-isopropylacrylamide)Samples samole M W O MJMn M,b cloud point, "C N1 530 000 146 000 3.7 35.7 2.8 580000 32.2 AI 440 000 160 000 RI 400 000 73 000 5.5 580000 33.2 R2A 1 60 000 49 000 3.2 33.9 11 o o o g 6.9 34.2 R2B 76 000 N12' 14 000 5 400 2.3 34.3

AT, ,!

Tmn,: OC

35.5 32.4 33.5 34.0 34.1,35.1* ca. 35

oc

0.4 0.8 0.5 0.4 0.78,3.3h >4

AHe

10 (1.1) 13 (1.5) 12 (1.4) 15 (1.7) 12 (1.4) ca. 3 (0.3)

ACJ 17 13 15

24 6,3h 2

"Determined by GPC; molecular weights calculated on the basis of polystyrene standards in THF (error *lo%). bPeak molecular weight determined by GPC; calculated on the basis of polystyrene standards in DMF (error *lo%). CTemperatureof peak maximum of microcalorimetric endotherm. Width at half-height of microcalorimetric endotherm. CCalorimetricenthalpy of endotherm (cal/g of polymer). Values in parentheses are kilocalories per mole of monomer repeating units. /Calorimetric peak height (cal/"C.g) of polymer). EBimodal molecular weight distribution observed. hEstimated for each component of the transition illustrated in Figure 2. 'Light scattering determined as 10700 g/mol.

5000 000-6000 000) and ACS-reagent-grade sodium bromide, sodium sulfate, and sodium thiocyanate were used as received from Aldrich. Molecular weights quoted for HPC, PPG, and PEO are those stated by the supplier. Distilled water was analyzed (Barnstead Co., Newton, MA) to contain 0.66 ppm total ionized solids (as NaCI) and 0.17 ppm total organic carbon (as C). Synthesis. The aqueous redox polymerization of NIPAAM was adapted from a procedure reported by Hoffman and coworkers.'O The solvent was a 15 mM phosphate buffer in normal saline (600 mL of distilled water, 0.6914 g of Na2HP04,0.9094 g of NaH,P04, and 5.09 g of NaCI; titrated with ca. 50 mL of 0.1 N NaOH to pH 7.4). After 22.2 g of monomer, 2.3 of ammonium persulfate in 5 mL of water, and 12 mL of TEMED were added, the reaction mixture was stirred for 15 h under nitrogen. Precipitation was carried out by dropwise addition of the polymerization mixture to 800 mL of methanol. The resulting polymer was dissolved in 200 mL of distilled water and dialyzed against regularly freshened distilled water for 5 days. The polymer was precipitated in an equal volume of methanol and vacuumdried. The polymer was then dissolved in chloroform, dried (MgSO,), and precipitated in hexane. The 2.62 g of polymer obtained after vacuum-drying was designated R2A. Anal. Calcd for C 6 H I I N O :C, 63.7%; H, 9.8%; N , 12.4%. Found: C, 63.5%; H, 9.9%; N , 12.2%. ' H N M R (200 MHz, DzO) 6: 1.0 (CH3, 6 H), 1.2-2.1 (-CH2CH-, 3 H), 3.7 (CH, 1 H). No vinyl protons were detected. IR (CHC13 cast film), cm-I: 3300, 2960, 2925, 2860, 1635, 1530, 1455, 1375, 1390, 1170, 1130, 750. Absent were the 1620-cm-I ( C = C ) , 1410-cm-I (CH,=), and C-H vinyl out-of-plane bending vibrations observed in the spectrum of the = 49000; Aw/An = 3.2. monomer. GPC: A, = 160000; A,, A second sample, designated R2B, was recovered from the CHCl,/hexane filtrate by evaporation and then precipitated and dried as described above to provide 1.70 g of PNIPAAM of A?, = 76000; A,= 1 1 000; kfw/&fn = 6.9 Sample R1 was synthesized through an identical procedure except that distilled water was substituted for the buffer. Spectral and elemental analyses were identical; however, the molecular weight distribution differed: A, = 400000; A, = 73000; M,/M,, = 5.5. A final sample, designated A l , was prepared in the following manner. A solution of 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 at 49 OC for 22 h under a positive nitrogen pressure, the solvent was evaporated. The resulting crude solid was vacuum-dried and crushed. After dissolving it in acetone (47 mL), the product was recovered through precipitation by dropwise addition to hexane (600 mL). Upon filtering and drying, 3.62 g (76% yield) of polymer was obtained. Elemental, infrared, and N M R analyses were in agreement with those above; GPC: A?w = 440000, A?, = 160000, M , / M , = 2.8. Preparations. Samples for cloud-point and microcalorimetric measurements were prepared by 10-fold dilution of 4.00 mg/mL stock solutions of polymers dissolved at room temperature in distilled water with 0. I % sodium azide as a bactericide. PPG required refrigeration to obtain clear solutions. For some studies, smaller quantities of polymer stock solutions were diluted to a total volume of 2.0 mL with distilled water.

1.20 0.90 0.60

0.30 0.00 30 32 34 36 38 40 TEMPERATURE ("C) Figure 1 . Cloud-point curves for two samples of PNIPAAM (0.40 mg/mL) in water. Samples N12 and A1 are identified in Table I. 26

28

Measurements. Infrared spectra were obtained on films cast from chloroform on NaCl plates with a Perkin-Elmer Model 1320 infrared spectrophotometer. N M R spectra were obtained on a Varian Model XL-300 spectrometer. Gel permeation chromatography (GPC) was performed with a Waters M45 solvent pump coupled to a R410 differential refractometer and a HewlettPackard Model 3380A digital integrator. Degassed tetrahydrofuran (THF, Aldrich, H P L C grade) was eluted at 1.1 mL/min through four Waters pStyragel columns (lo6, lo5, lo4, lo3 A). N,N-Dimethylformamide (DMF, Aldrich, HPLC grade) was eluted at 1.O mL/min through three Waters pBondagel columns (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. Optical density (OD) measurements were done at 500 nm on a Beckman Model DU-7 spectrophotometer with a water-jacketed cell holder coupled with a Lauda Model RM-6 circulating bath. Temperatures were manually ramped at rates of ca. 0.5 OC/min and monitored by an Omega Model 450-ATH thermistor thermometer. Cloud points were taken as the initial break points in the resulting optical density versus temperature curves and were independent (to within f0.5 "C) of slight fluctuations in the heating rate. Calorimetric (DSC) scans were obtained on a Microcal, Inc., Model MC-1 scanning micr~calorimeterl~ at a heating rate of 15 "C/h unless indicated. Samples were degassed and transferred to the sample cell with a calibrated syringe. Polymer-free solutions of the same solvent composition were similarly placed in the reference cell. Calibration was achieved by supplying a precisely known current to the reference cell of the calorimeter. LCSTs and transition widths are accurate to within fO.l "C; enthalpies (AH) and relative peak heights (AC,) are reproducible to within f 2 units (cal/g of polymer for AH and cal/(OC-g) of polymer for AC,). Results and Discussion PNZPAAM. Samples. Table I summarizes the synthesis and properties of the PNIPAAM samples used in this study. Six samples varying in molecular weight (An) from approximately 5000 to 160000 were examined. No deliberate attempts to fractionate these samples were made; M w / M nvaried from 2.3 to 6.9 within the sample set.

4354

The Journal of Physical Chemistry, Vol. 94, No. 10, 1990

Schild and Tirrell

I

I

I 32

5 28

0

k

24 20

600 900 1200 1500 1800 CONCENTRATION (mM) Figure 3. LCST of PNIPAAM sample A1 (0.40 mg/mL) as a function of concentrations of added salts. Filled symbols refer to cloud points, open symbols to calorimetric transition temperatures: (.,0) Na2S04; (&A) NaBr; (.,a) NaSCN.

0

300

a

TEMPERATURE ("C) Figure 2. Microcalorimetric endotherms for various samples of PN IPAAM (0.40 mg/mL). Temperatures of endothermic peak maxima are listed in Table I; temperature scale is given in lower right portion of Figure.

Binary Solutions. Figures 1 and 2 illustrate typical cloud-point and microcalorimetric results for the PNIPAAM samples examined in this work. Each of these polymers yields well-defined cloud points and calorimetric endotherms at the LCST, though substantial variation in the shape of the transition is apparent. For example. the first four samples listed in Table I afford relatively sharp calorimetric endotherms (AT,,2I 0.8 "C), whereas samples R2B and N I2 are characterized by broader cloud-point transitions and endothermic peaks several degrees in width. These differences in transition behavior are almost certainly a result of variation in chain length; the molecular weights of R2B and N12 are estimated to be significantly (i.e., 4-30-fold) lower than those of the other four samples. Although Fujishige and cc-workers6 report chain-length-independent cloud points for PNIPAAM, their experiments were confined to molecular weights 2 5 X IO4. Our samples N12 and R2B, on the other hand, are characterized by number-average molecular weights of 5-1 1 OOO, where chain length effects would be expected to be more significant. The more puzzling anomaly in Table I is the relative high LCST (35.7 "C) of sample N 1 ; the onset of phase separation of this polymer occurs at a temperature 3.5 "C higher than that observed for sample AI, despite the similarity in the molecular weights of these samples. We have no explanation for this difference in behavior. Study of the endotherm associated with the LCST of sample A1 shows that peak shape and temperature are independent of heating rate over a range from 3.3 to 30 "C/h. Transition kinetics are thus unimportant on the time scale of our experiments, and thermodynamic parameters may be obtained from our calorimetric measurements. Table I shows heats of transition of ca. 1.5 kcal/mol of repeating units for each sample with the single exception of the very low molecular weight N12. This transition enthalpy is similar to that reported by Fujishige6 and is consistent with the loss of ca. 1 hydrogen bond/repeating unit upon phase separation.'s Similar results were obtained at polymer concentrations varying from 0.40 to 4.00 mg/mL, consistent with the rather "flat" phase diagrams typical of aqueous polymer solutions in such narrow concentration This is to be expected if the LCST phenomenon involves a coil-globule transition with subsequent aggregation as proposed.6 The calorimetric endotherms associated with the LCST can be analyzed by a two-state model to provide a measure of the cooperativity of the t r a n s i t i ~ n . ' ~The , ~ ~size of the cooperative unit is defined as the ratio of the van't Hoff enthalpy to the (18) Israelachvili, J. N. Intermolecular and Surface Forces; Academic Press: London, 1985. (19) Privalov, P. Pure Appl. Chem. 1976, 52, 479. (20) Mabrey, S.; Sturtevant, J. M. Methods Membr. Biol. 1978. 9, 237.

No Salt

Na SCN

A 225 mM

989 mM

r y

0.8 "C

TEMPERATURE PC)

Figure 4. Microcalorimetric endotherms for aqueous PNIPAAM sample AI (0.40 mg/mL) in the presence of various salts: (a) no salt; (b) NaBr; (c) Na2S04;(d) NaSCN. Temperatures listed are those of maxima in Xp.The corresponding plot of transition temperatures is in Figure 3. Temperature scale given at lower right.

calorimetric enthalpy calculated on the basis of the area under the observed peak. The van7 Hoff enthalpy can be obtained from either the peak height (AC,)l9 or the transition width By either method, we obtain a value of ca. 430 repeating units for the size of the cooperative unit associated with the LCST of sample A I . Given the uncertainty of our estimated molecular weights and the polydispersitiesof our samples, a cooperative unit of this magnitude can be interpreted only to mean that the process reported by microcalorimetry involves long segments, and perhaps the entire length, of the PNIPAAM chain. It is interesting to note that sample R2B, which is characterized by a bimodal molecular weight distribution, displays two overlapping calorimetric transitions: a relatively sharp ( A T l j 2= 0.7 "C) component centered at 34.1 "C and a second broader endotherm (AT1j2= 3.3 "C) at 35.1 "C. It is reasonable to suggest that the sharper, lower temperature transition is associated with the phase separation of the higher molecular weight component of the PNIPAAM population and the broad, higher-temperature endotherm with shorter chains. Effects of Added Salts. Several investigators have demonstrated a correspondence between the Hofmeister series2' (which ranks the tendency of various salts to perturb the denaturation temperatures of proteins) and the effects of those same salts on the LCSTs of synthetic polymers, including PE02' and PVME.22 Figures 3 and 4 summarize our observations regarding the sensitivity of the LCST of PNIPAAM to Na,SO,, NaBr, and NaSCN. Sulfate salts out PNIPAAM more effectively than does bromide; similar results have been reported for PEO" and (21) Ataman, M. Colloid Polym. Sci. 1987, 265, 19. (22) Horne, R. A,; Almeida, J. P.; Day, A. F.: Yu,N-T. J . Colloid Interface Sri. 1971, 35, 77

Lower Critical Solution Temperatures in Aqueous Polymers

The Journal of Physical Chemistry, Vol. 94, No. 10, 1990 4355

TABLE 11: Prooerties of Various Polvmer Svstems with LCSTs in Aaueous Solution' ~

sample

concentration, mg/mL

A. PVME B. PVME C. PPG D. HPC E. HPC F. HPC

0.40 0.40 4.15 4.0 4.0 4.0

molecular weight

cloud point, O

70O0Ob I40 OOOb

33.8 33.8

1 OOOd 100 OOOd 300 OOOd 1000 OOOd

C

46.0 42.9 41.9

C

T "

A T i p OC 0.9 0.3 11.8 4.0 4.9 5.3

OC 36.4 36.5 40.9 48.4 45.6 44.1

AH 16 24 24 5 5 5

'Parameters defined in Table I. bDetermined by GPC; calculations on the basis of polystyrene standards in DMF (error &IO%). Figure 6. Molecular weights as stated by Aldrich Chemical Co.

1'

*CP

ACP 14 35 2

(0.9) (1.4) (1.4) (1.8) (1.8) (1.8)

1 1

1

'Very broad: see

B.

A

A . 2 "C

32.8 *C

I

I

36.4 'C

I 29.3 'C

I

I

-1 23

"

I

29

'

I

32

'

1

I

35

38

,

I

8

41

44

8

I

47

*

I

50

'

53

TEMPERATURE ("C)

Ti

I

I

I

1

26

Figure 6. Cloud-point curve and microcalorimetric endotherm for PPG (4.15 mg/mL). Sample C identified in Table 11.

40.1)

32.9'C

8

I 36.5'C

I

I

I

40.1 "C

TEMPERATURE ("C) Figure 5. Microcalorimetric endotherms for PVME (0.40 mg/mL). Samples A and B are identified in Table 11. Temperature scale given in lower portion of figure.

PVME.22 Thiocyanate elevates the LCST at some concentrations in accord with previous studies of other polymer^.^^^^^ Figure 4 shows the endotherms themselves. The LCSTs derived from calorimetry concur with the cloud points except in those cases where the DSC peaks are very broad. The greater LCST depression by sulfate anion vis-a-vis bromide is accompanied by a greater broadening of the endotherm and a reduction in peak height. Although one might then conclude that the perturbation of the endotherm of PNIPAAM by thiocyanate should be minor (based on the small depression of the LCST), large effects on peak shape are observed. Thus, the physical origins of the LCST depression must differ in each of these salt solutions. PVME. Only one extensive study of aqueous PVME solutions has been reported: Horne and co-workers22 have observed the effects of a wide range of cosolutes (including salts and alcohols) on the LCST. We have recently compared the sharply contrasting behavior of PVME and PNIPAAM in aqueous methanol mixt u r e ~ here, ; ~ we present comparative results for methanol-free aqueous solutions. Although the LCSTs of the two PVME samples listed in Table I1 are virtually identical, the shapes of the calorimetric endotherms (Figure 5) are quite different: as with PNIPAAM, the lower molecular weight species affords a shorter and broader transition endotherm. Transition enthalpies for both samples are again comparable to hydrogen bonding strengths, and comparison of the van't Hoff and calorimetric enthalpies suggests (23) Saito, S.; Kitamura, K. J . Colloid Interface Sci. 1971, 35, 346. (24) Saito, S.;Yukawa, M. J . Colloid Interface Sci. 1969, 30, 21 1.

a cooperative unit of size comparable to that of the chain. Thus, the behavior of PVME in water is similar to that of PNIPAAM, despite the opposite effects on the LCST caused by added methan~l.~ PPG. Despite its isomeric relationship to PVME, PPG is water soluble only in oligomeric form.25 Indeed, even this limited solubility has been suggested to result from spiral folding of the chain into tightly coiled disks in aqueous solution.26 Figure 6 shows the broad cloud-point and microcalorimetric transitions for PPG in water. The width of the calorimetric transition is nearly 12 OC, presumably as a consequence of the short chain length of the PPG sample used. Nonetheless, the transition enthalpy remains at 1.4 kcal/mol of repeating units, similar to the values observed for PNIPAAM and PVME. Interestingly, the size of the cooperative unit is still on the order of the chain length, and thus the transition breadth results from the very low molecular weight (ca. 1000) of the PPG sample used in this work. We attempted to observe microcalorimetric endotherms in aqueous solutions of PEO, a hydrophilic homologue of PPG, but the LCST is too close to the boiling point to allow direct measurement. Therefore, we "salted" the polymer out with either sodium sulfate or bromide; the cloud points we detected were similar to those previously r e p ~ r t e d . ~ ' Nevertheless, ,~~ no endotherms were observed by microcalorimetry at the concentration (4.0 mg/mL) of PEO used. Evidently, the transition is too broad to allow calorimetric detection. HPC. The Winniks and their c ~ - w o r k e r s ~ have ~ - ~ ' reported a series of interesting studies of the solution properties of modified HPC. Figure 7 and Table I1 summarize our observations for a set of HPC homopolymers. The endotherms are broad, considering that these samples are among the highest molecular weight polymers studied, and the LCST drops from ca. 46 to ca. 42 OC (25) Molyneux, P. In Water, A Comprehensiue Treatise; Franks, F., Ed.; Plenum Press: New York, 1975; Vol. 4. (26) Brackman, J. C.; van Os, N. M.; Engberts, J. B. F. N. Langmuir 1988, 4 , 1266. (27) Hines, P. M.; Boucher, E. A. J . Polym. Sci., Part B: Polym. Phys. 1976, 14, 2241. (28) Winnik, F. M.; Winnik, M. A,; Tazuke, S. J . Phys. Chem. 1987, 91, 594. (29) Winnik, F. M.: Winnik, M. A.; Tazuke, S.; Ober, C. K. Macromolecules 1987, 20, 38. ( 3 0 ) Winnik, F. M. Macromolecules 1987, 20, 2745. (31) Winnik, F. M. Macromolecules 1989, 22, 734.

J . Phys. Chem. 1990, 94, 4356-4363

4356

are consistent with HPC aggregation below the LCST.31 Application of similar techniques to PNIPAAM reveals no such aggregation? a result consistent with the smaller cooperative unit calculated for the PNIPAAM transition on the basis of calorimetric measurements.

T

*CP

l 38.4 "C

l

l

l

43.4 "C

l 48.4

"c

l

l

l

53.4 'C

TEMPERATURE ("C) Figure 7. Microcalorimetric endotherms for HPC (4.0 mg/mL). Samples D. E, and F are identified in Table 11.

as the molecular weight rises from IO5 to lo6. The enthalpy and width of the transition appear to be chain length independent. In each case, the cloud point agrees more closely with the temperature of the onset of the endotherm than with the peak value. The size of the cooperative units calculated for these samples increases with increasing molecular weight from ca. 3 to 8 to 35 polymer chains, suggesting an important role for chain aggregation. Winnik has reported results of nonradiative energy-transfer experiments that

Conclusions Calorimetry has been successfully applied to study LCST phenomena in polymer blend^;^^,^^ its extension to polymer solutions is facilitated through application of very sensitive instrumentation previously used to study phase transitions in proteins, lipids, and nucleic a ~ i d s . ' ~ - Good ~ ~ , ~agreement ~ - ~ ~ is typically found between transition temperatures and widths determined by cloud point and calorimetric methods; moreover, the availability of thermodynamic parameters permits a more thorough description of the aqueous solution behavior of these polymers and the perturbation of that behavior by ionic cosolutes. The transition enthalpies found for PNIPAAM, PVME, PPG, and HPC are all consistent with the loss of ca. 1 hydrogen bondlrepeating unit on phase separation. The sizes of the cooperative units of these transitions appear to be of the order of the size of the polymer chain, although the uncertainty of the estimated molecular weights and the polydispersities of the polymers preclude precise comparison. The exception is HPC, where the size of the cooperative unit exceeds that of the chain, consistent with preaggregation of HPC below the LCST. Acknowledgment. This work was supported by a National Science Foundation Predoctoral Fellowship to H.G.S. and by a grant from the U S . Army Research Office (DAAL03-88-K0038). Registry No. PVME, 9003-09-2; PPG, 25322-69-4; HPC, 9004-64-2; PNIPAAM, 25 189-55-3; Na2S04,7757-82-6; NaBr, 7647-1 5-6; NaSCN , 540-72-7. (32) Ebert, M.; Garbella, R. W.; Wendorff, J. H. Makromol. Chem., Rapid Commun. 1986, 7 , 65. ( 3 3 ) Percec, V.; Schild, H. G.; Rodriguez-Parada, J . M.; Pugh, C. J . Polym. Sci., Part A: Polym. Chem. 1988, 26, 935.

Dynamics of Electron-Hole Pair Recombination in Semiconductor Clusters M. O'Neil, J. Marohn, and G. McLendon* Department of Chemistry, University of Rochester, Rochester, New York 14627 (Received: September 8, 1989; In Final Form: December I, 1989)

The kinetics of radiative electron-hole pair recombination in CdS and Cd3As, clusters (where the radius of the cluster is smaller than the de Broglie wavelength of photogenerated excitons) were studied with picosecond photon counting luminescence decay measurements over wide temperature and energy ranges. The decay profiles were quantitatively examined with several models. The decays are composed of two distinct time regimes, each with very different temperature and emission energy dependence. The first (fast) regime is attributed to an unusually efficient thermal repopulation mechanism. The second (slow) component is well described by a distributed kinetic model. The kinetic behavior of wide (CdS) and narrow (Cd3As2) band gap materials was remarkably similar when composed of clusters in the quantum confined regime.

Introduction The study of the transition between molecular and bulk properties in semiconductors has been facilitated by the recent introduction of chemical synthesis of stable clusters with controllable mean diameter of less than 50 Numerous reports ( I ) Rossetti, R.: Elison, J.; Gibson, J.; Brus, L. J . Chem. Phys. 1984, 80, 4464. ( 2 ) Fojtik, A.; Weller, H.; Koch, U.;Henglein. A. Ber. Bunsen-Ges. Phys. Chem. 1984, 88,969. (3) Weller, H.; Koch, U.:Gutitrrez, M.; Henglein, A. Ber. Bunsen-Ges. Phys. Chem. 1984, 88, 649.

0022-3654/90/2094-4356$02.50/0

of the photophysical and optical properties of these materials have appeared recently, with particular emphasis paid to the quantum restriction effect on the static optical absorption and luminescence of cadmium sulfide and related material^.^-^ (4) Dannhauser, T.; O'Neil, M.; Johansson, K.; Whitten, D.; McLendon, G . J . Phys. Chem. 1986, 90,6074. (5) Brus, L. J . Chem. Pkys. 1983, 79,5566. (6) Brus. L. J . Chem. Phvs. 1984. 80. 4403. ( 7 ) Nedeljkovic, J.; Nenahovic, M'.; Micic, 0.; Nozik, A. J . Phys. Chem. 1986. 90. 12. (8) AIiv&tos, A. P.; Harris, A.; Levinos, N.; Steigerwald, M.; Brus, L. E. J . Ckem. Phys. 1988, 89,4001. (9) Mills, G.; Meisel, D. J . Colloid Interface Sci. 1987, I20, 540.

0 1990 American Chemical Society