Chain length dependence of the mean eld temperature in poly(oxyethylene)poly(oxybutylene) diblock copolymers

J. Patrick A. Fairclough,a Anthony J. Ryan,a Simon Turner,a Ian W. Hamley,b Shaomin Mai,c Colin Boothc and Richard C. Dennyd a b c d Department of Chemistry, University of Sheffield, Sheffield, UK S3 7HF Department of Chemistry, University of L eeds, L eeds, UK L S2 9JT Department of Chemistry, University of Manchester, Manchester, UK M13 9PL CCL RC Daresbury L aboratory, Daresbury, W arrington, UK W A4 4AD

Received 18th December 1998, Accepted 5th March 1999

The mean eld behaviour of a series of short poly(oxyethylene)poly(oxybutylene) diblock copolymers with 50/50 volume fraction is discussed. The mean eld temperature (T ) is determined from small angle X-ray MF scattering (SAXS). It is found to increase linearly with r , the volume normalised degree of polymerisation. V

Introduction
Block copolymers comprise two or more chemically distinct polymeric species joined by a covalent bond. Most commonly the component chains are incompatible and will phase separate on a length scale comparable to the radius of gyration of the polymer. They are said to microphase separate. Segregation is the term used to describe the three regions of behaviour within block copolymer melts. At strong segregation (low temperature) the chains are appreciably stretched and the interface between the blocks is sharp. At intermediate segregation the interface adopts a sigmoidal shape and the chains become less stretched. With increasing temperature the interface breaks down and the chains mix. For weak segregation the chains adopt a Gaussian conformation with chains intimately mixed and mean eld behaviour is exhibited. Leibler1 developed a mean eld description of the behaviour of block copolymer melts for weak and intermediate segregation. We have studied a series of diblock copolymers of poly(oxyethylene)poly(oxybutylene) (E B ) in the semicrysn m talline solid phase,2h4 ordered melt phases and disordered melt.5 Here we report further on the melt states and in particular the behaviour above the order to disorder transition (ODT) and the transition to mean eld behaviour. Theoretical and experimental aspects of microphase separation have been reviewed and the reader is referred to these for a full description of the associated phenomena.6h9 For a block copolymer chain of N total monomer units comprising N , A monomer units and N , B monomer units A B the non-combinatorial Gibbs energy can be expressed as * G \ kT (sr )/ / (1) mix nc V A B / / are the volume fractions of the A and B blocks, respecA B tively. Here r \ N ] N (l /l ) (2) V A B A B is a volume weighted degree of polymerisation, for v and v A B the respective monomer volumes. The function s, the FloryHuggins interaction parameter, has empirically been observed to exhibit a temperature depen This work was carried out at the CCLRC, Daresbury Laboratory, Daresbury, Warrington, UK WA4 4AD.

dence such as s\ a ]b T (3)

Where a and b are system dependant parameters, related to the enthalpy and entropy of mixing, respectively. We have previously calculated the temperature dependence of s based on a value of 10.5 at the ODT for symmetric volume fraction copolymers.5 To take account of uctuation eects this has been recalculated using Fredrickson and Helfands expression for a uctuating melt.10 This gives a value of sr B 15.5 at the V ODT. Allowing for uctuation eects, the following parameterisation has been obtained for poly(oxyethylene) poly(oxybutylene) diblock copolymers. s\ 75.6 [ 0.0929 T (4)

The high value of s for the EB system leads to short polymer molecules microphase separating and an expected deviation from mean-eld behaviour. The purpose of this study is to evaluate the eect of chain length on the mean eld behaviour within this system. The mean eld temperature, T , is the MF temperature below which deviation from the Leibler form of the small angle scattering intensity, I(q), is expected i.e. deviation from I(q)~1 P T ~1. This deviation is primarily due to the formation of density uctuations and chain stretching. Recent work by Maurer et al.,11 indicates that the form of s describing block copolymers is not sufficient to describe the behaviour of blends. The authors ascribe this to the eects of chain stretching and propose two separate forms of s for block copolymers and blends.

Experimental
Five diblock copolymers of symmetric volume fraction were prepared by sequential anionic polymerisation. The polydispersity of all samples (M /M ) was below 1.07 as evaluated by w n gel permeation chromatography (GPC). The composition was checked by NMR (see Table 1). These are the values used in subsequent calculations. The presence of approximately 2% triblock copolymer was not felt to be sufficiently signicant to warrant inclusion in the analysis. Phys. Chem. Chem. Phys., 1999, 1, 20932095 2093

Table 1 Characteristics of polymer studied Composition E B E56B27 60 29 E B E74B37 85 45 E B 96 47 Volume fraction (/ ) E 0.52 0.52 0.51 0.50 0.52 r V 107 115 148 170 185

Fig. 1 Typical plot of scattering vector (q) versus intensity versus temperature for polymer E B . The temperature range during this 74 37 run is reduced to show the ODT and the DOT.

The mean eld temperature was determined from the temperature dependence of the intensity of the correlation hole peak in the disordered melt via small angle X-ray scattering (SAXS). The SAXS experiments were carried out on Station 8.2 of the CCLRC Daresbury Synchrotron Radiation Source (SRS). The optics, beamline geometry and data acquisition are described elsewhere.12 The sample container comprises a Linkam single pan aluminium DSC pan, with a hole cut in the lid and base covered with thin mica sheet to allow the transmission of X-rays.5 The data were collected on a quadrant linear detector, as the scattering from the samples was isotropic. The data were corrected for scattering from the sample pan and camera, for changes in transmission and decay in the beam and nally for detector non-linearity. The data were converted to an absolute intensity scale by comparison with a secondary standard. The SAXS data were electronically correlated with the DSC data (sample temperature and input heating power). The sample was heated at a rate of 50 C min~1 to above its ODT and then cooled at 10 C min~1 to 70 C, then heated at 10 C min~1 to 40 C above the ODT to a maximum of 220 C. This was found to be necessary as signs of degradation were observed at 240 C. The scattering from the melt phase was tted with a Pearson VII function by the Collaborative Computing Project 13 (CCP13) t program.13 This function is useful because it allows the shape of the peak to vary from a Gaussian to a Lorentzian prole. This is measured by a shape parameter, low values indicating a Lorentzian prole, high values a Gaussian. On cooling through the disorder to order transition (DOT) there is a sharp transition from the weak Lorentzian uctuation peak to a strong, sharp Gaussian characteristic of the instrument resolution function. The data are collected as a series of time frames and the tting routine takes the results of the previous t as a rst step in tting the next data set. As the data are electronically correlated, temperature versus intensity plots can readily be created.

Results and discussion

Fig. 1 shows the scattering as the ODT is crossed. The scattering peak transforms within 2 C from a sharp Gaussian, in the lamella ordered phase to a weak Lorentzian in the disordered melt. During this transition the intensity, width and shape of the peak change dramatically. The peak centre does not change. This is a characteristic of the transition as it is a critical phenomenon, whereby on cooling from the disordered melt a single dominant wavelength is selected from the range of uctuation wavelengths above the ODT. This becomes the domain period within the ordered structure. From the Leibler theory,1 the scattering from a diblock copolymer can be described as S(q) \ N F(x) [ 2sN (5)

Fig. 2 Typical data plot for T ~1 versus I(q)~1 showing the mean eld temperature for copolymer E B . Insert shows the variation in 74 37 d spacing (d \ 2p/q*) where q* is the position of the peak centre.

Fig. 3 Plot showing the variation in the mean eld temperature (T ) with chain length. The right hand axis shows the product sr for MF V s calculated at T using eqn. (4). MF

where F(x) is a correlation function depending on the number of segments, N is the total number of segments (N ] N ) and A B the copolymer architecture and x \ q2R2 , R being the radius g g of gyration. Thus s is proportional to S(q)~1 and hence I(q)~1, as such the reciprocal of intensity should be proportional to reciprocal temperature in the mean eld region. A plot of reciprocal intensity versus reciprocal temperature can be seen in Fig. 2 with an indication of the mean eld temperature. This is the temperature below which the linear relationship no longer holds.14 Characteristically at this temperature (T ), MF the peak centre also shows a discontinuity in the temperature dependence as indicated in Fig. 2. Other authors have reported that d ln(q*)/dT changes faster than d ln R /dT below the g mean eld temperature.14,15 This indicates that the block copolymer is undergoing an increasing amount of stretching

2094

Phys. Chem. Chem. Phys., 1999, 1, 20932095

as the temperature is lowered. Sakimoto and Hashimoto14 observe the gradient of the d spacing increasing discontinuously on heating at T thus indicating the onset of chain MF stretching. The range of mean eld temperatures exhibited, for our EB copolymer is plotted in Fig. 3. This shows a linear variation with chain length (r ). Plotting T versus sr proV MF V duces no discernible trend with chain length, indicating a mean eld value of sr of approximately 10.2, this indepenV dence is a consequence of assuming a T ~1 dependence of s from eqn. (4).

3 4 5 6 7 8 9 10 11 12 13 14 15

Conclusions
Temperature correlated SAXS data have been used to calculate the transition to mean eld behaviour in a series of block copolymer melts. This is seen to vary linearly with r , the V volume normalised degree of polymerisation.

Acknowledgements
Funding was provided for JPAF and RCD under EPSRC Contract GR/K05982 and BBSRC Contract 28/X04460, respectively. The authors are grateful for the help of B.U. Komansheck for the setting up of the SAXS experiment.

References
1 2 L. Leibler, Macromolecules, 1980, 13, 1602. A. J. Ryan, J. P. A. Fairclough, I. W. Hamley, S-M. Mai and C. Booth, Macromolecules, 1997, 30, 1723.

S-M. Mai, J. P. A. Fairclough K. Viras, P. A. Gorry, I. W. Hamley, A. J. Ryan and C. Booth, Macromolecules, 1997, 30, 8392. I. W. Hamley, M. L. Wallwork, D. A. Smith, J. P. A. Fairclough, A. J. Ryan, S-M. Mai, Y-W. Yang and C. Booth, Polymer, 1998, 39, 3321. S-M. Mai, J. P. A. Fairclough, I. W. Hamley, M. W. Matsen, R. C. Denny, B-X. Liao, C. Booth and A. J. Ryan, Macromolecules, 1996, 29, 6212. I. W. Hamley, T he Physics of Block Copolymers, Oxford University Press, Oxford, 1998. F. S. Bates and G. H. Fredrickson, Annu. Rev. Phys. Chem., 1990, 41, 525. R. A. Brown, A. J. Masters, C. Price, X-F. Yuan, in Comprehensive Polymer Science, ed. C. Booth and C. Price, Pergamon, Oxford, 1989, ch. 6. M. Matsen and F. S. Bates, Macromolecules, 1996, 29, 1091. G. H. Fredrickson and E Helfand, J. Chem. Phys., 1987, 87, 697. W. W. Maurer, F. S. Bates, T. P. Lodge, K. Almdal, K. Mortensen and G. F. Fredrickson, J. Chem. Phys., 1998, 108, 2989. W. Bras, G. E. Derbyshire, A. J. Ryan, G. R. Mant, A. Felton, R. A. Lewis, C. J. Hall and G. N. Greaves, Nucl. Instrum. Methods Phys. Res., 1993, 29, 1091. The program is available from the Fibre and Polymer Diraction Collaborative Computing Project, CCP13, web site at http :// wserv1.dl.ac.uk/SRS/CCP13/software.html. N. Sakamoto and T. Hashimoto, Macromolecules, 1995, 28, 6825. J. H. Rosedale, F. S. Bates, K. Almdal, K. Mortensen and G. D. Wignall, Macromolecules, 1995, 28, 1429.

Paper 8/09870K

Phys. Chem. Chem. Phys., 1999, 1, 20932095

2095