PNIPAM Chain Collapse Depends on the Molecular Weight and

Langmuir 2006, 22, 4259-4266

4259

PNIPAM Chain Collapse Depends on the Molecular Weight and Grafting Density Kyle N. Plunkett,†,§ Xi Zhu,‡,§ Jeffrey S. Moore,† and Deborah E. Leckband*,†,‡ Department of Chemistry and Department of Chemical and Biomolecular Engineering, UniVersity of Illinois at Urbana-Champaign, 600 South Mathews AVenue, Urbana, Illinois 61801 ReceiVed NoVember 21, 2005. In Final Form: February 7, 2006 This study demonstrates that the thermally induced collapse of end-grafted poly(N-isopropylacrylamide) (PNIPAM) above the lower critical solution temperature (LCST) of 32°C depends on the chain grafting density and molecular weight. The polymer was grafted from the surface of a self-assembled monolayer containing the initiator (BrC(CH3)2COO(CH2)11S)2, using surface-initiated atom transfer radical polymerization. Varying the reaction time and monomer concentration controlled the molecular weight, and diluting the initiator in the monolayer altered the grafting density. Surface force measurements of the polymer films showed that the chain collapse above the LCST decreases with decreasing grafting density and molecular weight. At T > LCST, the advancing water contact angle increases sharply on PNIPAM films of high molecular weight and grafting density, but the change is less pronounced with films of low-molecular-weight chains at lower densities. Below the LCST, the force-distance profiles exhibit nonideal polymer behavior and suggest that the brush architecture comprises dilute outer chains and much denser chains adjacent to the surface.

Introduction Thermally responsive polymers such as poly(N-isopropylacrylamide) (PNIPAM) have a wide range of applications in biotechnology and medicine. PNIPAM, in particular, is insoluble and undergoes a conformational change above its lower critical solution temperature (LCST) of 32 °C. Below the LCST, the polymer chains swell in water. Above the LCST, the solvent quality changes and the polymer segments are thought to become more hydrophobic. Because the LCST of 32 °C is close to physiological temperature, this polymer has enormous potential in technology and in biomedical applications. PNIPAM coatings have been used for the programmed adsorption and release of proteins and cells from surfaces.1-3 A new stationary-phase chromatographic material used grafted PNIPAM coatings.4 The temperature-dependent retention time of steroids on these columns was proposed to result from changes in the hydrophobic/ hydrophilic character of the PNIPAM above and below the LCST.4 Investigations of thermally induced changes in the properties of PNIPAM coatings primarily measured the conformational change of polymer chains5 or characterized the interfacial properties.6 For example, AFM measurements showed that the swollen thickness of grafted PNIPAM films decreased by a factor of ∼2 when the temperature increased from 25 to 40 °C.7 The * To whom correspondence should be addressed. Phone: 217-244-0793. Fax: 217-333-5052. E-mail: [email protected]. † Department of Chemistry. ‡ Department of Chemical and Biomolecular Engineering. § These authors contributed equally to the work. (1) Huber, D. L.; Manginell, R. P.; Samara, M. A.; Kim, B.-I.; Bunker, B. C. Science 2003, 301, 352-354. (2) Akiyama, Y.; Kikuchi, A.; Yamato, M.; Okano, T. Langmuir 2004, 20, 5506-5511. (3) Xu, F. J.; Zhong, S. P.; Yung, L. Y. L.; Kang, E. T.; Neoh, K. G. Biomacromolecules 2004, 5, 2392-2403. (4) Kanazawa, H.; Yamamoto, K.; Matsushima, Y. Anal. Chem. 1996, 68, 100-105. (5) Zhou, S.; Wu, C. Macromolecules 1996, 29, 4998-5001. (6) Liang, L.; Feng, X.; Liu, J.; Rieke, P. C.; Fryxell, G. E. Macromolecules 1998, 31, 7845-7850. (7) Kidoaki, S.; Ohya, S.; Nakayama, Y.; Matsuda, T. Langmuir 2001, 17, 2402-2407.

kinetics of swelling and drying the PNIPAM brushes above and below the LCST were also measured by interferometry.5 Upon stepping the temperature from 25 °C to T > LCST, the drying occurred in a three-stage process: a fast initial gel shrinking followed by a plateau and finally by another drying process.5 Liang et al. also reported that the water meniscus height in a capillary tube coated with PNIPAM changed by 7 mm in a 2-mm diameter capillary tube when the temperature increased from 20 °C to about 40 °C.6 Despite numerous reports of substantial PNIPAM changes at the LCST, such dramatic changes are not observed in all cases. The molecular weight and grafting density may also play important roles.8 Pelton and co-workers investigated the effect of molecular weight on the kinetics of surface tension lowering of water by soluble PNIPAM homopolymers above and below the LCST. The kinetics were not very sensitive to temperature for the lower-molecular-weight (13 100) PNIPAM, in contrast to the higher-molecular-weight (547 000) polymer.8 In a neutron reflectivity study, Yim et al. determined the conformation of PNIPAM chains with molecular weights of 33 000 and 220 000 end-grafted on silicon in D2O and in deuterated acetone.9 They found no temperature-dependent conformational change with the lowest-molecular-weight PNIPAM (33 000), and only a slight temperature-dependent thickness change with the higher-molecular-weight polymer (220 000). At much higher surface density and molecular weight, the thickness change above the LCST was much more pronounced.10 To describe the phase behavior of end-grafted polymers exhibiting an LCST, Halperin developed a scaling model based on the de Gennes n-cluster model.11 Halperin predicted that the phase behavior of these brushes would depend on the molecular weight and grafting density of the chains.12 The model also predicts a discontinuous segment density profile normal to the (8) Zhang, J.; Pelton, R. Colloids Surf. 1999, 156, 111-122. (9) Yim, H.; Kent, M. S.; Huber, D. L. Macromolecules 2003, 2003, 52445251. (10) Yim, H.; Kent, M. S.; Mendez, S.; Balamurugan, S.; Balamurugan, S. S.; Lopez, G. P.; Satija, S. Macromolecules 2004, 37, 1994-1997. (11) Halperin, A. Eur. Phys. J. B 1998, 3, 359-364. (12) Baulin, V.; Halperin, A. Macromol. Theory Simul. 2003, 12, 549-559.

10.1021/la0531502 CCC: $33.50 © 2006 American Chemical Society Published on Web 03/25/2006

4260 Langmuir, Vol. 22, No. 9, 2006

Plunkett et al.

Figure 1. Strategy used to graft PNIPAM from alkanethiol monolayers on gold. A mixed monolayer of SAM-Br (an ATRP initiatorterminated alkanethiol) and SAM-OH self-assembled on gold. The ATRP initiator surface concentrations were varied between 5 and 100 mol%. The polymerization of NIPAM from the surface was accomplished using the catalyst system of CuBr and PMDETA in an aqueous methanol solution.

surface due to the coexistence of a dense inner “phase” and a dilute outer “phase”. A more recent study used self-consistent field theory to describe the first-order transition of PNIPAM brushes at the LCST.13 The latter study predicted that the chain collapse and transition temperature would depend on both the grafting density and the molecular weight. In this study, we systematically investigated the interfacial properties of end-grafted PNIPAM as a function of the chain grafting density, molecular weight, and temperature. The polymer was grafted from the surface of an alkanethiol monolayer on gold containing the initiator (BrC(CH3)2COO(CH2)11S)2 (SAM-Br). The polymer was synthesized by atom transfer radical polymerization (ATRP). Extensive chain collapse occurred at the highest grafting density and molecular weight, but the change in the film thickness decreased with decreasing grafting density and molecular weight. The LCST was the same within (1 °C for all films. The force profiles between the PNIPAM brushes and a second surface measured below the LCST further suggest a one-dimensional phase separation within the polymer brush. Materials and Methods Materials. High-purity 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine (DPPE) and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE) were purchased in powder form (purity > 99%) from Avanti Polar Lipids, Inc. (Alabaster, AL). N-Isopropylacrylamide (NIPAM), 11-mercapto-1-undecanol, 2-bromo-2methylporpionyl bromide, CuBr, and CuBr2 were from Aldrich. 1,1,4,7,7-Pentamethyldiethylenetriamine (PMDETA) was purchased from Acros. NIPAM monomer was recrystallized from hexane. CuBr was purified by dissolving in 48% HBr and precipitating with the addition of water. All inorganic salts were high purity (>99.5%) and were purchased from Aldrich (Milwaukee, WI). All aqueous solutions were prepared with ultrapure water purified with a Milli-Q UV-Plus water purification system (Millipore, Bedford, MA). The water had a resistivity of >18 MΩ cm-1. HPLC grade methanol and chloroform purchased from Mallinckrodt (St. Louis, MI) were used to prepare lipid solutions. High-purity silver shot (99.99%, Aldrich, Milwaukee. (13) Mendez, S., Curro, J. G., McCoy, J. D., Lopez, G. P. Macromolecules 2005, 38, 174-181.

WI) used for the preparation of silver films on the mica was from Alfa Aesar (Ward Hill, MA). Synthesis of (HO(CH2)11S)2 (SAM-OH).14 11-Mercapto-1undecanol (500 mg, 2.45 mmol) was dissolved in dichloromethane (20 mL) and 10% potassium hydrogen carbonate (3 mL). A solution of bromine (0.2 g, 1.23 mmol) was added slowly to the well-stirred mixture. The reaction was stirred for 20 min. The organic phase was separated, and the resulting aqueous phase was extracted with dichloromethane (2 × 15 mL). The organic layers were combined, dried over MgSO4, and concentrated to give 428 mg (86.5%). The sample was then recrystallized with 3:1 hexane/ethanol to yield 370 mg of a white solid (74.8%). 1H NMR (400 MHz, CDCl3): δ 1.21 (bs,16 H), 1.49 (s, 1 H) 1.56 (pent, J ) 6.6, 2 H), 1.66 (pent, J ) 7.0, 2H), 2.65 (t, J ) 6.7, 4 H), 3.64 (t, J ) 6.5, 4 H). m/z (FD) ) 406.0 Synthesis of (BrC(CH3)2COO(CH2)11S)2 (SAM-Br).14 2-Bromo2-methylpropionyl bromide (0.475 g, 0.255 mL, 2.06 mmol) was added dropwise to a solution of the (HO(CH2)11S)2 (0.35 g, 1.72 mmol) and triethylamine (0.89 g, 1.22 mL, 8.6 mmol) in 20 mL dichloromethane at 0°C. The reaction was stirred for 1 h at 0 °C and 2 h at room temperature. The solvent was extracted with 2 N sodium carbonate (20 mL) saturated with NH4Cl. The organic layer was dried with MgSO4, concentrated, and purified using silica gel chromatography (13:1 hexane/ethyl acetate) to produce 400 mg (66.1%). Rf (13:1 hexane/ethyl acetate) ) 0.30. 1H NMR (400 MHz, CDCl3): δ 1.27-1.4 (bs, 16 H), 1.67 (m, 4 H), 1.92 (s, 6 H), 2.67 (t, J ) 7.4, 2 H), 4.16 (t, J ) 6.5, 2 H). m/z (FD) ) 704.0. Surface-Initiated Polymerization of N-Isopropylacrylamide. Gold-coated slides were prepared by soaking microscope slides in an acid-peroxide bath (66% HCl, 33% H2O2) for 30 min at 60 °C. The slides were washed with water and dried under a stream of nitrogen. They were then placed in a thermal evaporator and a 10Å-thick layer of chromium followed by a 400-Å-thick layer of gold were deposited at LCST (Table 3). The collapse is reversible. Reducing the temperature to 26 °C restored the initially measured force profiles (Figure 6). Again, the PNIPAM adhered to the mica, and the disks jumped out of contact at 2280 ( 100 Å with a pull-off force of -0.48 ( 0.06 mN/m. Forces were also measured at a lower grafting density of 480 ( 10 Å2/chain, and with a PNIPAM molecular weight of 238 000. At 26 °C, the range of the repulsive force was 2132 ( 50 Å. The onset of the repulsion occurred at a shorter distance, compared to the range at the higher grafting density (229 ( 5 Å2/chain) and slightly higher molecular weight (238 000). At high grafting density, each chain occupies a smaller area and extends away from the surface to form a brush.23 Although there was some hysteresis upon separation of the PNIPAM and mica, the surfaces did not adhere. After increasing the temperature to 36 °C, the range of the repulsion decreased to 1568 ( 50 Å (Table 3). Finally, forces were measured at the third and lowest grafting density of 1930 ( 40 Å2/chain at both 26 and 36 °C (Figure 7). The molecular weight of the PNIPAM was somewhat lower at 207 000. Upon approach, the forces were repulsive at both 26 and 36 °C, with a range of the steric repulsion at 1209 ( 25 and 836 ( 25 Å, respectively (Figure 7, Table 3). There was a slight hysteresis between advancing and receding curves at both temperatures (Figure 7), but there was no adhesion. At this grafting density, the film thickness decreased by 26% at 36 °C (Table 3). Force Measurements between Low-Molecular-Weight PNIPAM Brushes and Mica. To further investigate the molecular weight dependence of the chain collapse, SFA measurements were carried out at 26 and 36 °C with PNIPAM brushes with lower-molecular-weight chains but at the same grafting densities described above (see Table 1). Figure 8 shows the force profile measured between mica and a PNIPAM brush with Mw ) 74 000 chains at 229 ( 25 Å2/ chain. Upon approach, the surfaces repelled at both temperatures and there was no adhesion. The onset of the repulsion was 1007 ( 30 Å at 26 °C and 830 ( 30 Å at 36 °C. Above the LCST, the thickness decreased by ∼18%. A similar repulsive force profile was measured at the intermediate grafting density of 480 ( 10 Å2/chain and with an average molecular weight of 51 000. In this case, the onset of the repulsion was at 561 ( 20 Å at 26 °C and 492 ( 20 Å at 36 °C (Table 3). At the lowest grafting density of 1930 ( 40 Å2/chain (Mw ) 60 000), the force profiles measured at 26 and 36 °C were very similar. The brush thickness was 383 ( 20 Å

Figure 7. (a) Normalized force versus distance between mica and a PNIPAM brush polymerized at a grafting density of 1930 ( 40 Å2/chain and a molecular weight of 207 000 g/mol. The filled triangles and open triangles indicate the force measured during approach and separation, respectively, at 26 °C. The filled and open squares correspond to force curves measured during approach and separation, respectively, at 36 °C. The filled and open circles indicate the force curves measured during approach and separation, respectively, after the temperature was restored to 26 °C. (b) Fit of the force profile at 26 °C to eq 2 (line).

at 26 °C and 380 ( 20 Å at 36 °C (Table 3). There was some hysteresis between the advancing and receding curves but no adhesion. Comparison of the Force Profiles Measured at 26 °C with Simple Polymer Theory. As shown in Table 1, the values of s/2RF in all cases were much smaller than 1, and the polymers were in the brush regime.16 To test whether PNIPAM behaves as a simple polymer below the LCST, the force profiles were compared to the Alexander-de Gennes model, which describes the forces between end-grafted brushes in good solvent.23-25 Equation 2 gives the theoretical normalized force-distance profile between brushes on opposed crossed-cylinders.

F(R) 16kTπL L 5/4 D 7/4 ≈ 7 +5 - 12 3 R D L 35s

[( )

()

]

The theoretical brush thickness, L, is

L)s

() RF s

5/3

(3)

Here s refers to the average distance between the grafting sites, k is the Boltzmann constant, and T is the absolute temperature.

Table 3. Measured Thickness of PNIPAM Films above and below the LCST monomer concn and polymerization time

grafting density (Å2/chain)

3.9 M and 20 min

229 ( 5 480 ( 10 1930 ( 40 229 ( 5 480 ( 10 1930 ( 40

2.0 M and 1 min

(2)

polymer thickness (Å) 26 °C 36 °C 2690 ( 50 2132 ( 50 1209 ( 25 1007 ( 30 561 ( 20 383 ( 20

2074 ( 64 1568 ( 50 836 ( 25 830 ( 30 492 ( 20 380 ( 20

% change 30 27 26 18 12 0

Poly(N-isopropylacrylamide) Chain Collapse

Langmuir, Vol. 22, No. 9, 2006 4265 Table 4. Polymer Brush Parameters

monomer concn and polymerization time

grafting density (Å2/chain)

s/2RF

measured thickness (Å)

model calcd thickness (Å)

fitted thickness (Å)

R2

prefactor

3.9 M, 20 min

229 ( 5 480 ( 10 1930 ( 40 229 ( 5 480 ( 10 1930 ( 40

0.016 0.021 0.046 0.033 0.053 0.096

2690 ( 50 2132 ( 50 1209 ( 25 1007 ( 30 561 ( 20 383 ( 20

3181 2458 1344 894 526 390

2737 ( 100 2175 ( 120 1300 ( 90 1131 ( 90 580 ( 80 400 ( 70

0.98 0.94 0.97 0.96 0.95 0.98

5.5 ( 0.5 1 ( 0.1 0.4 ( 0.05 1.2 ( 0.05 0.8 ( 0.2 0.2 ( 0.02

2.0 M, 1 min

Since eq 2 is a scaling relationship, the prefactor is not exact. The prefactor should, however, be of order unity.26 In our study, forces were measured between a polymer brush and a hard wall (mica surface) instead of between two polymer brushes. Since the model assumes that the brushes are impenetrable, we replaced one polymer layer with a hard wall, and 2L (twice the polymer thickness) was replaced with L in eq 2. In fits of the data to eq 2, both the prefactor and the brush thickness (L) were allowed to vary. One point worth noting is that the mica and the polymer brush adhere, and the theory does not address this. Nevertheless, comparing the data to eq 2 gives a semiquantitative indication of how closely PNIPAM below the LCST approximates an ideal polymer in good solvent. Figures 6b, 7b, and 8b show the fitting results for all force profiles measured at 26 °C. At the highest grafting density (229 ( 25 Å2/chain) and molecular weight (>200 000), the force profiles agree with theory at all distances measured (Figure 6b) (Table 4). At all other grafting densities and molecular weights studied, there was good agreement with theory at large distances, but the data deviated from the theory at D < 250 Å. We therefore fit only the force profile at large distances to eq 2 and obtained the value for L. The experimentally determined polymer thickness was assumed to roughly coincide with the onset of steric repulsion. The model-determined thickness was calculated with eq 3. Table 4 summarizes the results of these three different determinations of the polymer extension.

Figure 8. (a) Normalized force vs distance between mica and a PNIPAM brush polymerized at a grafting density of 299 ( 5 Å2/ chain and Mw of 74 000. The filled and open triangles indicate the force measured during approach and separation, respectively, at 26 °C. The filled and open squares show the force curves measured during approach and separation, respectively, at 36 °C. The filled and open circles indicate the force curves measured during approach and separation, respectively, after the temperature was restored to 26 °C. (b) Fit of the force profile at 26 °C to eq 2 (line).

Adsorption of Soluble PNIPAM on OH-Terminated SAMs. To determine whether the PNIPAM adsorbs to the underlying SAM, the adsorption of soluble PNIPAM chains (Mw ) 100 000) on a SAM-OH layer was measured by SPR at 26 °C. An aqueous 10-mg/mL solution was injected into SPR cell. The effective optical thickness increased instantly to 2.6 ( 0.2 nm. After the signal stabilized, the cell was flushed with pure water. The optical thickness immediately dropped back to 0.0 ( 0.2 nm. The abrupt change in the resonance angle upon changing the solution indicated that the change is due to the refractive index change of the solution instead of to slower polymer adsorption. This result indicated that PNIPAM did not adhere to the -OHterminated SAM layer at T < LCST.

Discussion This systematic study of the temperature-dependent changes in the interfacial properties of surface-grafted PNIPAM chains above and below the LCST shows that the extent of chain collapse above the LCST depends on both the grafting density and molecular weight. Chain collapse occurred at high grafting density and molecular weight as expected. However, the collapse and corresponding changes in interfacial properties depend on the grafting density and molecular weight (Table 3). These findings are consistent with previous reports, which suggested that the extent of the change in PNIPAM at the transition temperature depends on the molecular weight.8,9 They also confirm qualitatively the predicted dependence on both chain length and grafting density.13 The synthetic procedures used in this study afforded good control of the grafting density and molecular weight of the PNIPAM. It was not possible to control the molecular weights more tightly on account of the polymerization characteristics. The film thicknesses increased rapidly at the beginning of polymerization reaction but leveled off after ca. 30 min (Figure 2). This is more characteristic of a conventional redox-initiated polymerization with chain termination via polymer coupling than of a controlled living polymerization. Huang and co-workers observed this polymerization behavior (asymptoting film thickness) for the ATRP polymerization of 2-hydroxyethyl methacrylate.27 In their system, the addition of 30 mol% CuBr2, which is the ATRP deactivating species, provided greater control of the film thickness over a longer reaction time. Unfortunately, the addition of CuBr2 to the NIPAM solution yielded thin films (ca. 2-3 nm) even after long polymerization times. Such films were not useful for these investigations of both low- and highmolecular-weight polymers. The polymerization conditions in the absence of CuBr2 did yield polymer brushes with controlled thicknesses on a 100% SAM-Br surface. They were therefore (22) Klein, J.; Luckham, P. F. Nature 1984, 308, 836-837. (23) Alexander, S. J. Physique 1977, 38, 983-987. (24) de Gennes, P. Macromolecules 1981, 14, 1637-1644. (25) de Gennes, P. AdV. Colloid Interface Sci. 1987, 27, 189-209. (26) Kuhl, T. L.; Leckband, D. E.; Lasic, D. D.; Israelachvili, J. Biophys. J. 1994, 66, 1479-1488. (27) Huang, W.; Kim, J. B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002, 35, 1175-1179.

4266 Langmuir, Vol. 22, No. 9, 2006

used to fabricate other polymer brushes on monolayers with varying amounts of ATRP initiator. Despite some variation in molecular weight for given reaction conditions, the differences are less than 15%. AFM measurements of the dry PNIPAM film thicknesses are considerably lower than those determined by ellipsometry (Table 1). Two other groups investigating PNIPAM on gold28 and on silicon29 observed similar differences in the results obtained with these two techniques. They suggested that the adhesion between the PNIPAM and AFM tip dampened the tip oscillation and increased the hysteresis in the force-distance curve. This could introduce systematic errors in the film thickness determinations. The steric thickness of the swollen PNIPAM films in water measured with the SFA is nevertheless more relevant to the brush properties in aqueous media. The latter is nearly an order of magnitude larger than the dried films measured by either AFM or ellipsometry. The contact angles on PNIPAM brushes with Mw > 47 000 sharply increase above ∼32 °C. However, the contact angles on the lower-molecular-weight brushes (