Study on the rapid deswelling mechanism of comb-type N-isopropylacrylamide gels

Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

Study on the rapid deswelling mechanism of comb-type N-isopropylacrylamide gels M. Annakaa,∗ , T. Matsuurab , E. Yoshimotoa , H. Taguchia , S. Sasakia , M. Sugiyamac , Y. Harab , T. Okanod a

d

Department of Chemistry, Kyushu University, 6-10-1, Hakozaki, Higashi-ku, Fukuoka-shi, Fukuoka 812-8581, Japan b Department of Ophthalmology, Nara Medical University, 840, Shijyo-cho, Kashihara-shi, Nara 634-8522, Japan c Research Reactor Institute, Kyoto University, Kumatori, Sennan-gun, Osaka 590-0494, Japan Institute of Advanced Biomedical Engineering and Science, Tokyo Women’s Medical University, 8-1, Kawada-cho, Shinjyuku-ku, Tokyo 162-8666, Japan Received 12 October 2003; accepted 30 March 2004

Abstract The shrinking mechanism of comb-type grafted poly(N-isopropylacrylamide) gel was investigated by fluorescence spectroscopy and small˚ is developed in the angle X-ray Scattering (SAXS). The SAXS reveals that the microdomain structure with characteristic dimension of 460 A comb-type grafted poly(N-isopropylacrylamide) gel during the shrinking process. Fluorescence spectroscopy together with SAXS observation suggests that the freely mobile characteristics of the grafted chains are expected to show the rapid dehydration to make tightly packed globules with temperature, followed by the subsequent hydrophobic intermolecular aggregation of the dehydrated graft chains. The dehydrated grafted chains created the hydrophobic cores, which enhance the hydrophobic aggregation of the networks. These aggregations of the NIPA chains contribute to an increase in void volume, which allow the gel having a pathway of water molecules by the phase separation. © 2004 Elsevier B.V. All rights reserved. Keywords: Poly(N-isopropylacrylamide) gel; Comb-type grafted gel; Shrinking kinetics; Fluorescence spectroscopy; Small-angle X-ray scattering

1. Introduction Biological materials research has to deal with three classes of materials: colloids (as the cell cytoplasm), liquid crystals (as the lipid protein bilayer of all membranes) and macromolecular networks (as the extracellular matrix). The biological macromolecules are well defined and beautifully exhibiting a manifold of fascinating and novel physical properties. For that reason they are of interest as models of polymers with unconventional and controllable properties and mode of behavior. The structure and dynamics of the macromolecules ∗

Corresponding author. Tel.: +81 92 642 2594; fax: +81 92 642 2607. E-mail address: [email protected] (M. Annaka).

0927-7765/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2004.03.021

and networks can be manipulated in many ways by making use of the tricks in nature. Natural polymers comprise the whole range of flexibilities: from random coil to semi-rigid threads. One of the most important prototypes is supramolecular assembly of semi-rigid and highly flexible filaments are often used by nature. The most prominent example is the proteoglycans, the basic building units of cartilage, as prototype of a comb-type copolymer from a semi-flexible core protein and flexible side chains. The main core is hyaluronic acid to which smaller core proteins are coupled [1]. Another example is the vitreous body of the eye. It is composed of the highly flexible hyaluronic acid. The fluid gel is interwoven by semi-rigid collagen threads, which serve the mechanical stabilization of the body. Recently we observed the collagen

202

M. Annaka et al. / Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

is strongly coupled to Na-hyaluronate in the vitreous body from the observations of the dynamics of light scattered by the calf vitreous body. The vitreous body is, therefore, considered to be a kind of comb-type copolymer gel. The vitreous body maintains its lucidity for well over 100 years, which requires however, a continuous maintenance by metabolism. Metabolic defects or high age may lead to a slight turbidity due to local aggregation of collagen fibers. This is one example of the close correlation between diseases (or age), and local phase separation. The comb-type architecture of polymer or polymer gel is of interest model to investigate the correlation between the physical properties of living materials and their biological functions from the physicochemical point of view. A gel system swells and shrinks in response to the environmental stimuli [2–6]. Since the process is diffusion-limited, the swelling and shrinking kinetics is strongly dependent of the size of gel. On the base of the collective diffusion of polymer network in a medium, the characteristic time of gel swelling and shrinking is described as τ ≈ R2 /D, where R and D are the size of the gel and the cooperative diffusion coefficient of the polymer network, respectively [7]. For typical polymer gels, D is on the order of 10−7 –10−6 cm2 /s, depending on polymer concentration, cross-linking density, etc. Recently Okano et al. introduced comb-type NIPA polymer chains to NIPA gel network, and found the comb-type grafted NIPA gel (GG) exhibited drastic acceleration of the shrinking kinetics compared to conventional normal-type NIPA gel (NG) [6,8,9]. The freely mobile grafted chains are expected to show the rapid dehydration, which enhances the dehydration of network chains of the gel. The strong aggregation force may induce the local phase-separated structure during volume change, which contributes an increase in void volume within a gel, resulting in a rapid release of water. Fluorescence spectroscopy and small-angle X-ray scattering (SAXS) may provide powerful tools for studying these phenomena at the molecular level. In this study, we carried out fluorescence spectroscopy and SAXS to elucidate the temporal change in the microscopic structure of the comb-type grated NIPA gel due to temperature jump across its volume transition temperature.

Table 1 Feed composition for preparation of NIPA normal-type gel (NG) and combtype grafted NIPA gel (GG) Sample code NIPA monomer (g) NIPA macromonomer (g)a DANSAEP (g) BIS (g) TEMED (␮l) APS (g) Water (ml)

NG 15.60 – 0.043 0.266 48 0.008 100

GG 10.92 4.68 0.043 0.266 48 0.008 100

a M = 7340, M = 5170, M /M = 1.42. The molecular weight distribuw n w n tion of NIPA macromonomer was determined by gel permeation chromatography apparatus (TOSOH SC-8020 system) in DMF with 10 mM LiCl. Calibration was carried out with monodisperse poly(ethylene glycol) standards purchased from TOSOH Corp.

2.2. Sample preparation N-isopropylacrylamide (NIPA; Kohjin) was recrystallized from the mixture of toluene and n-hexane. N,N -methylenebis(acrylamide) (BIS; Kanto), N,N -tetramethylethylenediamine (TEMED; Kanto) and ammonium peroxide (APS; Kanto) were used as received. NIPA macromonomer was synthesized by the esterification of acryloyl chloride with semi-telechelic NIPA polymer with a terminal hydroxyl end group, which was prepared by radical telomerization of NIPA monomer using 2-hydroxyethanethiol as a chain transfer agent [8]. The total amount of NIPA monomer unit (NIPA monomer + NIPA macromonomer) was kept constant, and the weight ratio of NIPA macromonomer to NIPA monomer was chosen to be 30% (w/w). The feed compositions are listed in Table 1. NG and GG (Fig. 1) were prepared by radical polymerization in distilled deionized water at 5 ◦ C for 24 h initiated by APS/TEMED in capillaries of inner diameter 0.5 mm (=do ). After gelation was completed, gels were removed from capillaries, and were cut into small cylinders of length 10 mm. Dansyl labeled NIPA hydrogel was prepared according to the same method described above except addition of N-[2-[[[5-(N,N-dimethylamino)-1-naphthalenyl]sulfonyl]-amino]ethyl]-2-propen-amide (DANSAEP) [10]. 2.3. Shrinking kinetics of gels

2. Experimental 2.1. Materials N-isopropylacrylamide (NIPA; Kohjin) was recrystallized form the mixture of toluene and n-hexane. Tetrahydrofurane (THF, Kanto), diethyl ether (Kanto), cyclohexane (Kanto) and acryloyl chloride (Aldrich) were purified just before use according to standard procedure. 2-Hydroxyethanethiol (HESH, Tokyo Kasei), benzoyl peroxide (Nakarai), N, N -methylene-bis-acrylamide (BIS, Kanto), N,N,N ,N -tetramethyl-ethylenediamine (TEMED, Kanto), and ammonium peroxide (APS, Kanto) were used as received.

Cylindrical gel with do = 0.5 mm was used for the measurement of shrinking kinetics experiments. The gel sample was placed in a thermostated cell filled with distilled deionized water, the temperature of which was controlled to within 0.1 ◦ C of the desired temperature. The temperature-jump (Tjump) experiments were carried out by exchanging the circulating water which temperatures were set to desired temperatures. The time required for T-jump was about 30 s. The swelling degree of gel, d/do , was obtained by measuring the diameter of the cylindrical gel, d, under a microscope. The measurement was repeated at least three times and its average was used as the value of d.

M. Annaka et al. / Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

2.4. Measurements of photophysical properties The fluorescence emission spectra were recorded on a Hitachi F-4010 fluorescence spectrophotometer with a thermostated cell filled with distilled deionized water, the temperature of which was controlled to within 0.1 ◦ C of the desired temperature. The excitation wavelength was set at 345 nm. The measurement of the temporal shift in the maximum emission wavelength, λem , was carried out within the wavelength range of λem of the initial temperature and that of destination temperature. The scanning rate was set to 10 nm/s, and the period for scanning the whole range was about 25 s. The measurement was repeated at least three times and its average was used as the value of λem . Fluorescence anisotropy ratio, r, was calculated from four polarized fluorescence spectra by using r= G=

(IVV − GIVH ) (IVV + 2GIVH ) IHV IHH

(1)

(2)

203

where I is the fluorescence intensity and the subscripts represent the orientation of polarizers (V is vertical, and H is horizontal), which are located for incident light (the first subscript) and for emitted light (the second subscript). The G value was used for the correcting the depolarization characteristics of the grating-type monochrometer. The anisotropy, r, were averaged for emission from 400 to 650 nm. 2.5. Small-angle X-ray scattering The SAXS experiment was carried out with the BL10C installed at Photon Factory (Tsukuba, Japan). An incident X-ray beam from the synchrotron orbital radiation was ˚ The scattered X-ray was demonochromatized to 1.49 A. tected by a one-dimensional position-sensitive proportional counter positioned at 1 m from the sample: the magnitude of ˚ −1 . the observed scattering vector ranged from 0.008 to 0.15 A The gel samples were sealed in a cell, the temperature of which was controlled to within 0.1 ◦ C of the desired temperature. The intensities were accumulated for 600 s in order to assure sufficient statistical accuracy without degrading the gel samples by X-ray irradiation. The scattered intensities were

Fig. 1. Preparation scheme of NIPA macromonomer and comb-type NIPA gel.

204

M. Annaka et al. / Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

Fig. 3. Temporal change in λem of dansyl probe for NG and GG after T-jump from 10 to 35 ◦ C. Fig. 2. The variation of the degrees of swelling, d(t)/d(0), of the NG and GG after T-jump from 10 to 35 ◦ C. The solid line is the fit with Eq. (3) and the evaluated values of the collective diffusion coefficient of the gel network, D for NG is 6.9 × 10−7 cm2 /s. [Reproduced with permission from Polymer, 44, 4405–4409 (2003). Copyright 2003 Elsevier].

corrected for the cell scattering and absorption, and then normalized with the thickness of the sample and irradiated time.

3. Results and discussions 3.1. Shrinking kinetics The kinetics of swelling of gel was treated by Tanaka and Fillmore [7]. The time variation of the gel size, e.g., the gel diameter, could be described by a single-exponential function
t d(t) − d(∞) 6 d(t)/d0 − d(∞)/d0 ≈ 2 exp − = (3) d0 − d(∞) d0 /d0 − d(∞)/d0 π τ where d(t) is the gel diameter at time t and τ is the characteristic time for swelling. Fig. 2 shows the variation of the degrees of swelling, d(t)/d(0), of the NG and GG after Tjump from 10 to 35 ◦ C. In the case of NG, the swelling ratio is satisfactory reproduced by Eq. (3). The evaluated values of the collective diffusion coefficient of the gel network, D for NG is 6.9 × 10−7 cm2 /s. In the case of GG, however, the analysis with a single-exponential function (Eq. (3)) found to be no more applicable. This is due to the fact that the theory is limited to the linear regime and the collective diffusion constant is assumed to be invariant upon shrinking/swelling [7]. It has been confirmed Eq. (3) was also valid for the shrinking kinetics as far as the shrinking did not accompany phase separation. Therefore we define the time required for halfshrinking, t1/2 , as follows [11]: d(t1/2 )/d0 − d(∞)/d0 1 d(t1/2 ) − d(∞) = = d0 − d(∞) d0 /d0 − d(∞)/d0 2

(4)

The values of the time required for half-shrinking, t1/2 , for NG and GG after T-jump from 10 to 35 ◦ C are 4000 and 100 s, respectively. The shrinking kinetics for GG compares approximately 40 times accelerated with that for NG. The rapid shrinking of the comb-type grafted NIPA gel is expected to

originate in a microphase-separated structure during volume change. 3.2. Temporal changes in λem and r for T-jump process The temporal change in microenvironment of dansyllabeled NG and GG gel were investigated by means of fluorescence spectroscopy. Fig. 3 shows the temporal changes in λem for gels after T-jump from 10 to 35 ◦ C. Observed timevariation behavior of λem are similar as that of gel size and analysis with a single-exponential function is not applicable. We defined the time required for half- shift, t1/2 , of λem as follows: 1 λem (t1/2 ) − λem (∞) = (5) λem (0) − λem (∞) 2 The values of the time required for half-shift, t1/2 , for NG and GG T-jumped form 10 to 35 ◦ C are 150 and 14.3 s, respectively. The change in the microenvironment for GG is accelerated by an order of magnitude comparing with that for NG. The grafted chains maintain their mobility, while NIPA networks are anchored at several points by cross-linking on each chain restricting the mobility. The λem of dansyl probe decreases initially very rapidly with time while the diameter of the microgel changes only slightly. Dansyl probe is expected to act as a molecular probe sensing the microscopic environment. The position, shape and intensity of the emission are sensitive to molecular mobility, solvation, and polarity of the microscopic environment around the chromospheres. The temporal change in the diameter and λem , is related to the several factors; variation of dehydration of chains, local viscosity, structural relaxation, and so forth. To understand the microscopic factors that occur during the gel shrinkage, we carried out the fluorescence anisotropy measurements. Fig. 4 shows the temporal changes in anisotropy, r for gels after T-jump from 10 to 35 ◦ C. Observed time-variation behavior of r are similar to that of λem . We defined the time required for half-change, t1/2 , of r as follows: 1 r(t1/2 ) − r(∞) = (6) r(0) − r(∞) 2 The values of the time required for half-change, t1/2 , for NG and GG T-jumped form 10 to 35 ◦ C are 240 and 108 s, re-

M. Annaka et al. / Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

205

fact that the shrinking of NG is controlled by the collective diffusion of network, therefore dansyl locally probes both viscosity and hydrophobicity. It should be mentioned that the characteristic time, t1/2 values of λem and r are much smaller than that of diameter change, which are by a factor of 10 for NG, and by a factor of 100 for GG. The microscopic change is complete faster than macroscopic one. This is likely related to the structural relaxation of gel network and the change in the microscopic structure of gels during the shrinking process, though we cannot give clear interpretation here. Fig. 4. Temporal change in r of dansyl probe for NG and GG after T-jump from 10 to 35 ◦ C.

spectively. The values of the time required for half-change in the microviscosity show following order; GG < NG. GG exhibits smaller t1/2 value for λem than that for r. These observations suggest that the rapid thermal response of the comb-type grafted NIPA gels microscopically attributes to the following mechanism; the rapid dehydration of the freely mobile NIPA graft chains due to the increase in hydrophobicity, and the subsequent hydrophobic intermolecular aggregation of dehydrated graft chains enhances the hydrophobic aggregation of the NIPA networks because of the increase in microviscosity. In the case of NG, however, the temporal change in r for NG almost coincides with that in λem . This is due to the

3.3. Small-angle X-ray scattering (SAXS) We carried out, therefore, the SAXS experiment to confirm the microscopic structure of the gels. Fig. 5 shows the SAXS intensity profiles for NG and GG equilibrated at 10 and 35 ◦ C. When cross-links are introduced to polymer solution, the concentration fluctuations are perturbed. Geissler et al. proposed that the scattering intensity from an uncharged gel could be described by a sum of dynamic and static components [12–15]. I(q) = Idyn (q) + Ist (q) =

IL (0) + IG (0) exp[−Ξ 2 q2 ] 1 + ξ 2 q2 (7)

Fig. 5. SAXS intensity profiles and reconstructed spectra for NG and GG. The dotted curve represents the Lorentzian behavior with the correlation length ξ, the broken curve is the contribution of the solid-like (static) nonuniformity with the mean size Ξ, and the full curve is total intensity: (a) NG at 10 ◦ C, (b) GG at 10 ◦ C, (c) NG at 35 ◦ C, (d) GG at 35 ◦ C. The fitting parameters are summarized in Table 2. [Reproduced with permission from Polymer, 44, 4405–4409 (2003). Copyright 2003 Elsevier].

206

M. Annaka et al. / Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

Table 2 Scattering parameters from NIPA normal-type gel (NG) and comb-type grafted NIPA gel (GG) in equilibrium state at 10 and 35 ◦ C Sample

Temperature (◦ C)

˚ ξ (A)

NG

10 35

41.0 ± 1.0 46.1 ± 2.7

181 ± 1.4 159 ± 0.7

GG

10 35

59.3 ± 1.2 59.1 ± 5.4

176 ± 1.6 160 ± 0.4

˚ Ξ (A)

where Ξ is the mean size of the solid-like (static) nonuniformity, and ξ is the correlation length. The Ornstein–Zernike expression (first term) in Eq. (7) corresponds to the thermal fluctuation in a semi-dilute polymer solution, where ξ is the mean distance between contact points [16]. The second term describes static concentration fluctuation, such as elastic constraints frozen-in by the cross-links or other structural features present in a particular gel. As shown in Fig. 5, at swelling equilibrium at 10 and 35 ◦ C, the scattered intensities for NG and GG monotonously decrease with the scattering magnitude q, and indicate the absence of microphase separated structure. The parameters deduced from the fits of the SAXS spectra to Eq. (7) are summarized in Table 2. In the swollen state at 10 ◦ C, it is found that Ξ for NG and GG are almost identical, which indicates that the introducing of the graft chains does not induce the extra static inhomogeneity, namely the static inhomogeneity basically results in the distribution fluctuation of the crosslinkers in the pre-gel polymer solution. In contrast to Ξ, the difference in correlation length, ξ is probably due to the freely mobile characteristics of the NIPA graft chains. As the grafted chains are structurally separated from the backbone crosslinked network, they could behave the free polymer chains as in the polymer solution, which makes the dynamic fluctuation

of the polymer density for the GG larger than that for NG. As shown in Fig. 5(c) and (d), in the collapsed state at 35 ◦ C, remarkable decrease in the dynamic component of the scattering intensity was observed for NG and GG. With regard to the dynamic component due to the scattering of the solution like structure in the gel, it is reasonable that its contribution to the scattering intensity is neglected. Fig. 6(a) shows the SAXS intensity profiles, illustrating the trends observed as the time after T-jump from 10 to 35 ◦ C. The scattered intensities for NG monotonously decrease with the scattering magnitude q. It is remarkable for GG that a scat˚ −1 at 10 min after T-jump. tering peak appeared at q = 0.014 A The intensity was increased with time, and was confirmed to reach a maximum at 420 min after T-jump, and the peak position unchanged until the GG reached the equilibrium state. The Cahn–Hilliard theory is based on the linearization of a generalized diffusion equation and is considered to applicable to the early stage of spinodal decomposition [17,18]. In this theory, the time dependence of the static structure factor after the system is quenched into the two-phase region should follow an exponential growth as I(q, t) = I(q, 0) exp[2R(q)t]

(8)

with the I(q,t) as the time dependent static structure factor. Where q = (4π/λ) sin (θ/2) is the magnitude of the scattering wave vector, with the incident beam wavelength λ and scattering angle θ. The growth rate R(q) is given as R(q) = −Dq2 (1 − ξ 2 q2 )

(9)

where D(= −(∂2 f/∂φ2 )0 Λ(q)) is the effective diffusion constant. (∂2 f/∂φ2 )0 is the second derivative of the free energy at the initial state and Λ(q) is a Onsager coefficient. When

Fig. 6. . (a) Time evolution of SAXS intensity profiles for (a) NG and (b) GG after T-jump from 10 to 35 ◦ C. Solid lines in (b) bare the results of curve fitting with Eq. (10). [Reproduced with permission from Polymer, 44, 4405–4409 (2003). Copyright 2003 Elsevier].

M. Annaka et al. / Colloids and Surfaces B: Biointerfaces 38 (2004) 201–207

the thermal fluctuation is included, Eq. (8) is modified and becomes: I(q, t) = Ix (q) + [I(q, 0) − Ix (q)] exp[2R(q)t]

(10)

and Ix (q) =

−q2 Λ(q) R(q)

(11)

Solid lines in Fig. 6(b) are the results of curve fitting for GG at the time after T-jump with Eq. (10). The amplitude of the static concentration fluctuation is very small, therefore the contribution of the solid-like (static) nonuniformity to I(q,t) is negligible. The theoretical curves reasonably reproduce the observed scattering intensity functions. From the fitting results, the correlation length of the concentration fluctuation ˚ at each time after T-jump. This is evaluated to be ξ = 49.5 A value is almost the same as the correlation length of the dynamic fluctuation for GG at equilibrium state summarized in Table 2. This indicates that the concentration fluctuation is mostly associated with individual network segment, and graft chains due to its freely mobile characteristics, and its amplitude increases by T-jump. Contrary to GG, the scattered intensity for NG is still monotonous decreasing function of the scattering magnitude q, and the increase in the dynamic fluctuation is observed. The observed microphase separated structure that developed in GG during the shrinking process well explains the drastic acceleration of shrinking of GG compared to NG. These observations and results of fluorescence spectroscopy suggest that the freely mobile characteristics of the grafted chains are expected to show the rapid dehydration to make tightly packed globules with temperature, followed by the subsequent hydrophobic intermolecular aggregation of the dehydrated graft chains. The dehydrated grafted chains created the hydrophobic cores, which enhance the hydrophobic aggregation of the networks. These aggregations of the NIPA chains contribute to an increase in void volume, which allow the gel having a pathway of water molecules by the phase separation.

4. Conclusion The shrinking mechanism of comb-type grafted poly(Nisopropylacrylamide) gel was investigated by fluorescence spectroscopy and small-angle X-ray Scattering (SAXS). Flu-

207

orescence spectroscopy and SAXS observations suggest that the freely mobile characteristics of the grafted chains are expected to show the rapid dehydration to make tightly packed globules with temperature, followed by the subsequent hydrophobic intermolecular aggregation of the dehydrated graft chains. The dehydrated grafted chains created the hydrophobic cores, which enhance the hydrophobic aggregation of the networks. These aggregations of the NIPA chains contribute to an increase in void volume, which allow the gel having a pathway of water molecules by the phase separation.

Acknowledgments This work has been supported by Grant-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture. The SAXS experiments were performed under the approval of the Photon Factory Advisory Committee (proposal No.2000G242).

References [1] L. Stryer, Biochemistry, fourth ed., Freeman, New York, 1995. [2] M. Shibayama, T. Tanaka, Adv. Poly. Sci. 109 (1993) 1, and references therein. [3] T. Tanaka, Phys. Rev. Lett. 40 (1978) 820. [4] Y. Hirokawa, T. Tanaka, J. Chem. Phys. 81 (1984) 6379. [5] J. Kiler, A.B. Scranton, N.A. Peppas, Macromolecules 23 (1990) 4944. [6] R. Yoshida, K. Uchida, Y. Kaneko, K. Sakai, A. Kikuchi, Y. Sakurai, T. Okano, Nature 374 (1995) 240. [7] T. Tanaka, D.J. Fillmore, J. Chem. Phys. 70 (1979) 1214. [8] Y. Kaneko, K. Sakai, A. Kikuchi, R. Yoshida, Y. Sakurai, T. Okano, Macromolecules 28 (1995) 7717. [9] Y. Kaneko, S. Nakamura, K. Sakai, A. Kikuchi, T. Aoyagi, Y. Sakurai, T. Okano, Polym. Gels Networks 6 (1998) 333. [10] M. Annaka, H. Noda, R. Motokawa, T. Nakahira, Polymer 42 (2001) 9887. [11] H. Hirose, M. Shibayama, Macromolecules 31 (1998) 5336. [12] S. Mallam, A.M. Hecht, A.R. Renniie, E. Geissler, Macromolecules 23 (1991) 543. [13] F. Horkey, A.M. Hecht, S. Mallam, E. Geissler, A.R. Renniie, Macromolecules 24 (1991) 2896. [14] E. Geissler, F. Horkey, A.M. Hecht, Phys. Rev. Lett. 645 (1993) 71. [15] F. Horkey, W. Burchard, E. Geissler, A.M. Hecht, Macromolecules 26 (1993) 1296. [16] P.G. de Gennes, Scaling Concepts in Polymer Physics, Cornell University Press, Ithaca, 1979. [17] M. Okada, C.C. Han, J Chem. Phys. 85 (1986) 5317. [18] Y. Li, G. Wang, Z. Hu, Macromolecules 28 (1995) 4194.