Morphologies of PDMS-containing Diblock Polymers by Charlotte Stewart-Sloan Submitted to the Department of Materials Science and Engineering in partial fulfillment of the requirements for the degree of Master's of Science in Materials Science and Engineering at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY February 2012 © Massachusetts Institute of Technology 2012. All rights reserved. MASSACHUSETTS INSTl OFTCHNLOGY Author ...----------------RARIES Engineering and Science Department of Materials February, 2012 . ................................................. Edwin L. Thomas William and Stephanie Sick Dean of the George R. Brown School of Engineering Professor, Rice University Thesis Supervisor C ertified by ..... .... Accepted by ................... ............ .. .- - ..... . Christopher Schuh Chairman, Department Committee on Graduate Students Morphologies of PDMS-containing Diblock Polymers by Charlotte Stewart-Sloan Submitted to the Department of Materials Science and Engineering on February, 2012, in partial fulfillment of the requirements for the degree of Master's of Science in Materials Science and Engineering Abstract The morphologies of polydimethylsiloxane (PDMS)-containing diblock polymers are investigated as a function of volume fraction, segregation, processing procedure, and temperature. Strongly segregated polyisoprene-PDMS and polystyrene-PDMS diblocks are prepared according to standard procedures in the literature by anionic synthesis in the laboratory of Professor Apostolos Avgeropoulos at the University of Ioannina in Ioannina, Greece and their morphologies are investigated using small angle x-ray scattering and transmission electron microscopy. Good agreement is found between this work and other work on the structures of diblocks containing PDMS with a variety of complementary blocks and between this work and theoretical predictions for the morphologies of diblock polymers. Different processing treatments including casting from solvents with a range of preference for each block and a week-long anneal are tested to determine whether processing has a strong effect on final morphology; it is found that in most cases the morphology displayed after processing is consistent independent of the processing treatment, indicating that the morphologies are in equilibrium and fairly robust to preparation procedure. Finally, selected weakly segregated diblocks were studied at varying temperatures using synchrotron small angle x-ray scattering. The diblock samples appeared to be affected by the prior x-ray dose that the materials had received. With limited prior dose, the materials studied were ordered with little dramatic change in morphology throughout the temperatures investigated; under continual irradiation by a 1.371A (9.1 keV) beam for half an hour, the samples were damaged. The thesis concludes with a summary and suggestions for future work, including a discussion of experimental and theoretical work on the ways that equilibrium morphologies of block copolymers are perturbed when they are spatially confined to dimensions on the order of several times their repeat period. The small domain sizes achievable with and technological relevance of PDMS-containing diblocks make them ideal for use in microelectronics and information storage which provides a motivation for exploring this topic further. Thesis Supervisor: Edwin L. Thomas Title: William and Stephanie Sick Dean of the George R. Brown School of Engineering Professor, Rice University 2 Acknowledgments I would like to acknowledge the help that I have received from many during this thesis project. First, I'd like to thank my advisor, Professor Ned Thomas. His guidance has been invaluable; I am so lucky to have had the experience to work in his lab, travel to national facilities both with groupmates and on my own, write and submit proposals, and present my work at conferences. His gift of the freedom to guide my own project and pursue my own interests has been greatly appreciated. I would also like to acknowledge Professor Apostolos Avgeropoulos from the University of Ioannina in Ioannina, Greece. Without the careful synthesis work of his students, I would not have been able to study any of the materials described in this thesis. His guidance during the summers that we have worked together on TEM and SAXS have also been very helpful to my growth as a scientist. I would also like to thank all the members of the Thomas group for friendship and helpful discussions; I have enjoyed getting to know every one of you and look forward to more collaboration in the coming years. Finally, I would like to thank the MIT Presidential Fellowship and the NSF GRFP for financial support. 3 Contents . . . . . 19 19 20 23 27 29 43 and PI-PDMS Diblocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 45 45 47 48 48 54 54 56 56 56 62 1 Introduction and literature review 1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 1.2 Block Polymer Background . . . . . . . . . . . . . . . . . . 1.3 Characterization tools . . . . . . . . . . . . . . . . . . . . 1.4 Motivation for studying PDMS-containing block polymers 1.5 Prior work on PDMS-containing block polymers . . . . . . 1.6 Scope of Thesis Project . . . . . . . . . . . . . . . . . . . . 2 Strongly Segregated Morphologies of PS-PDMS 2.1 Introduction and motivation . . . . . . . . . . . 2.2 Materials synthesis . . . . . . . . . . . . . . . . 2.3 Experimental procedures . . . . . . . . . . . . . 2.4 R esults . . . . . . . . . . . . . . . . . . . . . . . 2.5 TEM results . . . . . . . . . . . . . . . . . . . . 2.6 SAXS results . . . . . . . . . . . . . . . . . . . 2.6.1 Morphology determination using SAXS . 2.6.2 D ata . . . . . . . . . . . . . . . . . . . . 2.7 D iscussion . . . . . . . . . . . . . . . . . . . . . 2.7.1 PS-containing diblocks . . . . . . . . . . 2.7.2 PI containing diblocks . . . . . . . . . . 3 Effect of Processing on Morphologies blocks 3.1 Introduction . . . . . . . . . . . . . . 3.2 Sample preparation . . . . . . . . . . 3.3 Results . . . . . . . . . . . . . . . . . 3.4 D iscussion . . . . . . . . . . . . . . . 3.4.1 PS-containing samples . . . . 3.4.2 PI-containing samples . . . . 4 Morphologies of low molecular as a function of temperature 4.1 Overview . . . . . . . . . . . 4.2 Experimental procedures . . 4.2.1 Sample descriptions . . . . . . . . . . . . . . . . . . . . . . -. . . . . . . . . . . . . . . . . . . . . . . of PS-PDMS and PI-PDMS Di. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 63 64 64 65 65 77 weight PS-PDMS and PI-PDMS diblocks 79 79 . . . . . . - -. . . . . . . . . . . . . .. 79 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4 4.3 4.4 80 83 87 . . . . 89 90 90 . . . . 92 . . . . 93 93 4.2.2 DSC procedures . . . . . . . . . . . . . . . 4.2.3 SAXS procedures . . . . . . . . . . . . . . DSC results . . . . . . . . . . . . . . . . . . . . . SAXS results . . . . . . . . . . . . . . . . . . . . 4.4.1 Minimal radiation exposure results . . . . 4.4.2 Initial protocol heating results . . . . . . . 4.4.3 Morphology change through the transition 4.4.4 Initial protocol cooling results . . . . . . . 4.4.5 Conclusions . . . . . . . . . . . . . . . . . 5 Conclusions and future work 5.1 Bulk morphologies . . . . . . . . . . . . . 5.2 Processing treatments . . . . . . . . . . . 5.3 Morphologies as a function of temperature 5.4 Outlook for the future . . . . . . . . . . . 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 . 112 . 113 . 114 . 115 List of Figures The repeat units of the three polymers used in this thesis are displayed ab ove. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 Schematic showing the structures of the various ordered diblock copolymer morphologies. "f" indicates the volume fraction of component A. Taken from [27]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3 A theoretically computed phase diagram for block copolymers which shows the expected morphology as a function of volume fraction of one of the blocks and XN. Q229 corresponds to BCC packed spheres and Q23 0 indicates the double gyroid phase. CPS corresponds to a close packed sphere structure (FCC or HCP) and is hard to distinguish experimentally from BCC packed spheres. The figure is taken from [28]. . . . . . . . . . . . . . 1-4 Experimentally determined phase diagram for PS-PI diblock copolymers taken from [16]. There are several differences between the data shown here and the morphologies which are theoretically predicted to occur: ODTs occur at lower temperatures, the phase boundaries are not symmetric around the 50:50 PS:PI composition, non-classical phases are present (e.g. HPL, which is the hexagonally perforated lamellar phase), and unanticipated Order-Order Transitions (OOT)s occur, among others. . . . . . . . . . . . 1-5 A schematic of a transmission electron microscope [30] (a) and a photograph of a transmission electron microscope [31] (b). On the right hand side of (a), the path that electrons take as they travel through a TEM in bright field imaging mode is compared to that of a beam of light travelling through an optical microscope . . . . . . . . . . . . . . . . . . . . . . . . . 1-6 A schematic of a small angle x-ray scattering setup (a) and an image of a SAXS instrument [32] (b). (a) shows the path of the x-rays from the source, through the sample, and on to the detector after scattering. . . . . 1-1 6 19 20 21 21 25 26 1-7 Three composite plots showing previously observed morphologies in PDMScontaining diblocks as a function of XN and PDMS volume fraction. Eighty six points are plotted in total; these include multiple plots of the same sample on two sides of an observed phase transition. It is clear that the locations of experimentally observed phases are not predicted perfectly by the phase boundaries computed using SCFT. It appears that the range of volume fractions at which the morphologies occur are broader than those predicted and that those regions interpenetrate. This may be due to more advanced factors such as conformational asymmetry, polydispersity, or be the result of preparation procedures which produced metastable samples. Graph (a) shows morphologies found at weak segregation; graph (b) shows morphologies found at intermediate segregation; graph (c) shows morphologies found at strong segregation. The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of xN outside the range computed in those works, the boundaries between phases were assumed to be independent of XN (vertical phase boundaries.) . . . . . . . . . . . . . . . . . . . . . . . . 30 1-7 Three composite plots showing previously observed morphologies in PDMScontaining diblocks as a function of xN and PDMS volume fraction. Eighty six points are plotted in total; these include multiple plots of the same sample on two sides of an observed phase transition. It is clear that the locations of experimentally observed phases are not predicted perfectly by the phase boundaries computed using SCFT. It appears that the range of volume fractions at which the morphologies occur are broader than those predicted and that those regions interpenetrate. This may be due to more advanced factors such as conformational asymmetry, polydispersity, or be the result of preparation procedures which produced metastable samples. Graph (a) shows morphologies found at weak segregation; graph (b) shows morphologies found at intermediate segregation; graph (c) shows morphologies found at strong segregation. The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of XN outside the range computed in those works, the boundaries between phases were assumed to be independent of XN (vertical phase boundaries.) . . . . . . . . . . . . . . . . . . . . . . . . 31 7 1-8 1-8 1-8 Plots showing previously observed morphologies in PDMS-containing diblocks as a function of xN and volume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS DG (c), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres (i), and disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of XN outside the range computed in those works, the boundaries between phases were assumed to be independent of XN (vertical phase boundaries.) Note that many phases are observed significantly outside of the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proximity to each other and in some cases the phase boundaries appear to interpenetrate. . . . . . . . . . . . . . . . . . . 32 Plots showing previously observed morphologies in PDMS-containing diblocks as a function of XN and volume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS DG (c), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres (i), and disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of xN outside the range computed in those works, the boundaries between phases were assumed to be independent of xN (vertical phase boundaries.) Note that many phases are observed significantly outside of the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proximity to each other and in some cases the phase boundaries appear to interpenetrate. . . . . . . . . . . . . . . . . . . 33 Plots showing previously observed morphologies in PDMS-containing diblocks as a function of XN and volume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS DG (c), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres (i), and disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of xN outside the range computed in those works, the boundaries between phases were assumed to be independent of xN (vertical phase boundaries.) Note that many phases are observed significantly outside of the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proximity to each other and in some cases the phase boundaries appear to interpenetrate. . . . . . . . . . . . . . . . . . . 34 8 1-8 1-8 2-1 2-2 Plots showing previously observed morphologies in PDMS-containing diblocks as a function of XN and volume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS DG (c), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres (i), and disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of XN outside the range computed in those works, the boundaries between phases were assumed to be independent of xN (vertical phase boundaries.) Note that many phases are observed significantly outside of the regions of volume fraction and values of xN that are predicted theoretically; different phases also exist in close proximity to each other and in some cases the phase boundaries appear to interpenetrate. . . . . . . . . . . . . . . . . . . 35 Plots showing previously observed morphologies in PDMS-containing diblocks as a function of XN and volume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS DG (c), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres (i), and disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of XN outside the range computed in those works, the boundaries between phases were assumed to be independent of XN (vertical phase boundaries.) Note that many phases are observed significantly outside of the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proximity to each other and in some cases the phase boundaries appear to interpenetrate. . . . . . . . . . . . . . . . . . . 36 Updated phase diagrams (compare to figures 1-8 and 1-7) of PDMS-containing diblocks reflecting the work done in this thesis. Extended data for the samples are available in Table 2.1. The SAXS and TEM data used for this table was that taken after the sample underwent a one week cast from a 5 wt% solution and then a one week anneal at 150 0C under vacuum. (a) shows only the samples characterized in this thesis; the samples that have PS as the complementary block are indicated by square markers while those that have PI as the complementary block are indicated by circular markers. (b) shows the samples characterized in this thesis plus those with values of XN between 0 and 150 investigated in other works. The samples which were studied here have larger markers. . . . . . . . . . . . . . . . . . . . . . . . 50 TEM images of the PS-PDMS samples which were characterized using this 0.37 technique. The PS cylinder morphology occurred when <pPDMS low molecular weight for the strongly segregated samples and for the only sample. Lamellae occurred when <5PDMS > 0.43. See Table 2.1 for more inform ation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 9 2-2 TEM images of the PS-PDMS samples which were characterized using this technique. The PDMS sphere morphology occurred when 2-3 2-4 2-4 2-5 2-5 2-6 #PDMS < 0.19. PDMS cylinders were present when #PDMS= 0.23. Table 2.1 for more inform ation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TEM images of the PI-PDMS samples which were characterized using this technique. See Table 2.1 for more information. . . . . . . . . . . . . . . . SAXS data from the PS-PDMS samples which are studied in this thesis. See Table 2.1 for more information. . . . . . . . . . . . . . . . . . . . . . SAXS data from the PS-PDMS samples which are studied in this thesis. See Table 2.1 for more information. . . . . . . . . . . . . . . . . . . . . . SAXS data from the PI-PDMS samples which are studied in this thesis. See Table 2.1 for more information. . . . . . . . . . . . . . . . . . . . . . SAXS data from the PI-PDMS samples which are studied in this thesis. See Table 2.1 for more information. . . . . . . . . . . . . . . . . . . . . . SAXS data from all of the low molecular weight samples investigated in this thesis. See chapter 4 for more information. . . . . . . . . . . . . . . Scattering from PS-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . 3-1 Scattering from PS-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . 3-1 Scattering from PS-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . 3-1 Scattering from PS-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . 3-1 Scattering from PS-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . 3-2 Scattering from PI-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . . 52 . 53 . 57 . 58 . 59 . 60 . 61 3-1 10 . 68 . 69 . 70 . 71 . 72 . 73 3-2 Scattering from PI-containing samples studied in this thesis after different processing treatments. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. See Tables 3.2 and 3.3 for more information. . . . . . . 74 4-1 Calculated values of XN(T) at different temperatures for the three samples studied in this section. Values of X were computed based upon the work in [74]. The annealing temperature and anticipated ODTs for each sample are also shown; anticipated ODTs are based upon the figure in [15]. . . . . The temperature programs for the S-64LMW samples described in Table 4.4. The temperature profiles used for S-64LMW-A and B were selected in order to observe scattering at a wide and complementary range of temperatures and investigate whether or not the scattering was dependent upon thermal history. The temperature profile for S-64LMW-C was selected in order to observe scattering at many different and closely spaced temperatures and to detect where morphological transitions in this range occur with fine precision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The temperature programs for the S-45LMW samples described in Table 4.4. These temperature profiles were chosen in order to sample a wide range of temperatures at which morphology transitions could occur and to provide comparative data for morphology transitions upon heating and cooling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The temperature programs for the I-39LMW samples described in Table 4.4. The two temperature profiles were selected to investigate structure in I-39LMW at temperatures both above and below room temperature and to investigate morphology transitions as a function of temperature, thermal history (i.e. heating vs. cooling), and time. . . . . . . . . . . . . . . . . . . Data from DSC scans of S-64LMW (a), S-45LMW (b), and I-39LMW (c). The scans were performed at 5m nte and began at room temperature. The samples were heated to 180 0 C, held for three minutes, cooled to -80 0 C, held for 3 minutes, and then heated back to room temperature. The morphology transitions observed in SAXS can be seen to varying degrees in the DSC data. For S-64LMW, the endothermic peak at approximately 50 0 C to 70 0 C corresponds somewhat with the morphology transition observed in SAXS between 80 0 C and 165 0 C. For S-45LMW, the endothermic peak at approximately 30 0 C to 50 0 C is at the low end of the 40 0 C to 80 0 C range where the morphology transition was observed to occur with SAXS. There do not appear to be any thermal transitions in I-39LMW. The thermal transitions are only observed during the heating cycles, consistent with the faster kinetics of structural rearrangement observed during SAXS. The large vertical lines which occur at 25 0 C are due to the rapid change in 4-2 4-3 4-4 4-5 80 84 85 86 temperature of the DSC pans to the initial experimental temperature. . . . 95 11 4-5 Data from DSC scans of S-64LMW (a), S-45LMW (b), and I-39LMW (c). The scans were performed at 5m ate and began at room temperature. The samples were heated to 180 0 C, held for three minutes, cooled to -80 0 C, held for 3 minutes, and then heated back to room temperature. The morphology transitions observed in SAXS can be seen to varying degrees in the DSC data. For S-64LMW, the endothermic peak at approximately 50 0 C to 70 0 C corresponds somewhat with the morphology transition observed in SAXS between 80 0 C and 165 0 C. For S-45LMW, the endothermic peak at approximately 30 0 C to 500 C is at the low end of the 40 0 C to 80 0 C range where the morphology transition was observed to occur with SAXS. There do not appear to be any thermal transitions in I-39LMW. The thermal transitions are only observed during the heating cycles, consistent with the faster kinetics of structural rearrangement observed during SAXS. The large vertical lines which occur at 250 C are due to the rapid change in temperature of the DSC pans to the initial experimental temperature. . . . 96 Images recorded on the marCCD detector for a sample with the same characteristics as sample II-D where 40 second exposures were continuously taken throughout the duration of the experiment; this sequence of images obviates the need for the revised protocol and suggests that during x-ray irradiation some event such as cross-linking which promotes the formation of locked-in non-equilibrium morphologies or some event such as depolymerization of the PDMS occurs which irreversibly changes the equilibrium morphology occurs. (a) shows the image taken at 25 0 C before any temperature treatment has been performed. (b) shows the scattering from this sample at 90 0 C after raising the temperature. (c) shows data taken after holding the sample at 90 0C for 40 minutes. (d) shows the image taken after a quench to 60 0 C. (e) shows the data obtained after holding at 60 0 C for 18 minutes. (f) shows data taken after the sample was then cooled to 25 0 C and held there for one hour and 15 minutes. (g) shows data taken after this sample was removed, translated a small distance, and re-exposed to the x-ray beam. This sequence shows that sample exposure to the x-ray beam strongly affects its structure and the resulting data obtained. . . . . 97 4-7 Scattering data and caption taken from [79] showing scattering from a PEO-PDMS and D3 blend. Compare this to the scattering shown in Figure 4-8. It is clear that the images obtained after x-ray irradiation look qualitatively similar to those shown above, providing evidence for the hypothesis that the x-ray beam causes depolymerization of the PDMS to form D 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 8 4-6 12 4-8 4-9 4-9 4-10 4-10 Images of the scattering patterns recorded by the CCD camera for selected temperatures for samples S-64LMW-A and S-64LMW-B. In these images, the integrated intensity presented in Figure 4-10 does not fully capture the x-ray scattering that occurred due to the development of orientation during the experiment. S-64LMW-A at 80 0 C is shown in a; S-64LMW-B at 90 0 C in b; S-64LMW-A at 105 0 C in c; S-64LMW-B at 110 0C in d; S-64LMWA at 120 0 C in e; S-64LMW-B at 125 0 C in f; S-64LMW-A at 150 0 C after 30 minutes in g and after approximately 70 minutes in h; S-64LMW-B at 165 0 C in i. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scattering data from samples S-45LMW-D and I-39LMW-C following the revised protocol described. (a) shows scattering from sample S-45LMW-D and (b) - (d) show scattering from I-39LMW-C. It appears that there is very little change with temperature in the scattering data for S-45LMW-D for the temperature range investigated (25-80 0C.) For I-39LMW-C, there appear to be one prominent and one secondary peak from 10-80 0 C, which becomes one peak which steadily loses intensity through 200 0 C. . . . . . . Scattering data from samples S-45LMW-D and I-39LMW-C following the revised protocol described. (a) shows scattering from sample S-45LMW-D and (b) - (d) show scattering from I-39LMW-C. It appears that there is very little change with temperature in the scattering data for S-45LMW-D for the temperature range investigated (25-80 0 C.) For I-39LMW-C, there appear to be one prominent and one secondary peak from 10-80 0 C, which becomes one peak which steadily loses intensity through 200 0 C. . . . . . . Scattering from S-64LMW-A, B, and C as a function of temperature. In all cases, the data from the last exposure taken at a particular temperature during the heating cycle is plotted. S-64LMW-A data is shown in red; S64LMW-B in green; S-64LMW-C in green. For all samples, it is clear that in the region of 70 0 C-165 0 C the low temperature high q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in each sample, indicating that there is both a time and temperature effect in the morphological transition that is occurring; the behavior, however, is the same in all cases. . . . . . . . . . . . . . . . . Scattering from S-64LMW-A, B, and C as a function of temperature. In all cases, the data from the last exposure taken at a particular temperature during the heating cycle is plotted. S-64LMW-A data is shown in red; S64LMW-B in green; S-64LMW-C in green. For all samples, it is clear that in the region of 70 0 C-165 0 C the low temperature high q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in each sample, indicating that there is both a time and temperature effect in the morphological transition that is occurring; the behavior, however, is the same in all cases. . . . . . . . . . . . . . . . . 13 99 100 101 102 103 4-10 Scattering from S-64LMW-A, B, and C as a function of temperature. In all cases, the data from the last exposure taken at a particular temperature during the heating cycle is plotted. S-64LMW-A data is shown in red; S64LMW-B in green; S-64LMW-C in green. For all samples, it is clear that in the region of 70 0 C-165 0 C the low temperature high q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in each sample, indicating that there is both a time and temperature effect in the morphological transition that is occurring; the behavior, however, is the same in all cases. . . . . . . . . . . . . . . . . 104 4-11 Scattering from S-45LMW-A, B, and C as a function of temperature. In all cases, the data from the last exposure taken at a particular temperature during the heating cycle is plotted. S-45LMW-A data is shown in red; S-45LMW-B in green; S-45LMW-C in green. For all samples, it is clear that in the region of 40 0 C-85 0 C the low temperature high q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in each sample, possibly indicating that longer annealing times are necessary in this regime to achieve equilibrium morphologies; the behavior, however, is the same in all cases. . 105 4-11 Scattering from S-45LMW-A, B, and C as a function of temperature. In all cases, the data from the last exposure taken at a particular temperature during the heating cycle is plotted. S-45LMW-A data is shown in red; S-45LMW-B in green; S-45LMW-C in green. For all samples, it is clear that in the region of 40 0 C-85 0 C the low temperature high q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in each sample, possibly indicating that longer annealing times are necessary in this regime to achieve equilibrium morphologies; the behavior, however, is the same in all cases. . 106 4-12 Scattering from I-39LMW-A and B as a function of temperature. In all cases, the data from the last exposure taken at a particular temperature during the heating cycle is plotted. I-39LMW-A data is shown in red; I39LMW-B in green. For both samples, in the range of 30 0 C-50 0 C the low temperature high q peak is suppressed and a high temperature low q peak is form ed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 4-13 SAXS data taken from diblocks undergoing structural evolution at a constant temperature. Ten sequential three minute exposures from samples transitioning between ordered and disordered structures are shown and the growth of the high temperature correlation hole peak at the expense of the low temperature ordered structure peak is evident. (a) shows S-64LMW-A at 80 0 C; (b)shows S-45LMW-A at 50 0 C; (c) shows I-39LMW-B at 40 0 C. . 108 14 4-13 SAXS data taken from diblocks undergoing structural evolution at a constant temperature. Ten sequential three minute exposures from samples transitioning between ordered and disordered structures are shown and the growth of the high temperature correlation hole peak at the expense of the low temperature ordered structure peak is evident. (a) shows S-64LMW-A at 80 0 C; (b)shows S-45LMW-A at 50 0 C; (c) shows I-39LMW-B at 40 0 C. . 109 4-14 SAXS data taken from all samples during their cooling runs. Note the lack of structure development for any of the samples which received only a 30 minute anneal at the colder temperatures. I-39LMW-B shows morphology development at longer temperatures due to the longer two hour anneal at each temperature that it received. (a) shows data from S-64LMW-A and B; (b) shows data from S-45LMW-A and B; (c) shows data from I-39LMW-A and B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 4-14 SAXS data taken from all samples during their cooling runs. Note the lack of structure development for any of the samples which received only a 30 minute anneal at the colder temperatures. I-39LMW-B shows morphology development at longer temperatures due to the longer two hour anneal at each temperature that it received. (a) shows data from S-64LMW-A and B; (b) shows data from S-45LMW-A and B; (c) shows data from I-39LMW-A and B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111 15 List of Tables 1.1 A chart detailing the morphologies found in the literature for PDMScontaining diblocks (it does not include triblocks or other multiblocks also described in the text); it contains the data displayed in figures 1-7 and 1-8. Some data found in the literature discussed was purposefully excluded if there were significant reasons to believe that the data presented was either not sufficient to establish the morphology claimed, e.g due to large amounts of obvious homopolymer contamination in the TEM images, etc. Note that because much of the relevant information was not available in the papers, in some cases calculations were made to arrive at the volume fraction and XN values. If necessary, volume fractions were computed from mole or weight fractions using commonly available densities. X values were taken from [74] if available or calculated from solubility parameters found in [75]; in the latter case, the reference volume was assumed to be 100". The temperature used was that at which the last thermal treatment was done (e.g. annealing, crosslinking, etc.) or room temperature if the sample was not thermally treated after casting. For works in which several samples were examined at a large variety of temperatures, the chart displays multiple entries for each transition temperature ±50 C: the lower temperature for the phase present below the transition temperature and the upper temperature for the phase present above the transition temperature. . . . . 43 16 2.1 The list of samples analyzed in this thesis and their relevant characterization data. These materials were synthesized in the laboratory of Professor Apostolos Avgeropoulos at the University of Ioannina in Ionannina, Greece. The molecular weights and volume fractions of the samples were also determined there; gel permeation chromatography (GPC) was used to establish the former and nuclear magnetic resonance (NMR) the latter (except where noted that the characterization was performed using GPC.) The x-ray measurements and TEM measurements were performed at the Institute for Soldier Nanotechnologies at the Massachusetts Institute of Technology in Cambridge, Massachusetts. - means that the relevant measurement was not performed. N/P means that there were not sufficient peaks in the SAXS data to assign the value with certainty. * Indicates that toluene was the casting solvent and no annealing treatment was performed. ** Indicates that cyclohexane was used as the casting solvent and the annealing treatment was that described in the text. . . . . . . . . . . . 49 3.1 Solubility parameters for the polymers and solvents studied in this thesis. 64 . . ... .. .. . . .. .. All values are taken from [75]. . .. . . . .. . Summary of sample morphologies obtained by SAXS after different casting and annealing treatments for the samples whose morphology was discussed in Chapter 2. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week at 150 0 C. Morphologies were determined by the relative q values of the positions of the peaks in the scattering spectrum, which are characteristic of different morphologies. S indicates a spherical morphology, C indicates a cylindrical morphology, and L indicates a lamellar morphology. - indicates that the experiment was not performed and N/P indicates that there were insufficient peaks in the SAXS data to assign a morphology. . . . . . . . . 66 Summary of d-spacing values obtained by SAXS after different casting and annealing treatments for the samples whose morphology was discussed in Chapter 2. The diblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one week 3.2 3.3 at 150 0 C. D-spacings are obtained from the relationship d = g. In some cases, the ratios of q at which the main peaks appear indicate a morphology change, so the changes in peak spacings are attributable to changes in both the sizes of the domains and the distances between the first planes from which scattering appears. See Table 3.2 for more information about the latter. - indicates that the experiment was not performed and N/P indicates that there were insufficient peaks in the SAXS data to assign a m orphology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 17 4.1 4.2 4.3 4.4 The composition and anticipated ODT's of the low molecular weight diblocks studied in this section. The molecular weights were determined by GPC and the volume fractions were determined by NMR. Estimated values of XN at the ODT were estimated based on the figures in [15] and the value of x was determined from the equations presented in [74]. . . . . . . . . . . Physical properties of the chains that form the diblocks investigated in this thesis. The entanglement molecular weights and radii of gyration are taken from [77]; note that these assume infinitely long chains and so may overestimate the radii of gyration. The first number given for radius of gyration is the ratio of the radius of gyration to the square root of sample molecular weight; the latter are the calculated values for the relevant samples. The glass transition data is taken from [75]; the measured values were performed on longer chains and so may overestimate the values of Tg. The second procedure used in SAXS studies at BNL. . . . . . . . . . . . . Summary of different sample preparation procedures used during this experiment. The initial scattering pattern displayed at 25 0 C was consistent across prior preparation methods. The thermal programs for samples receiving continuous x-ray exposure can be found in Figures 4-2, 4-3 and 4-4 while those for the samples not receiving continuous x-ray exposure can be found in Table 4.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 81 82 87 88 Chapter 1 Introduction and literature review 1.1 Overview A-B diblock polymers are materials which are composed of two homopolymers which are chemically linked at a junction point. The combination of chemical incompatibility between the homopolymers and their topological linkage forces spatial segregation between the blocks into predictable, ordered structures with typical periodicities in the tens of nanometers. These structures have been of interest to researchers who work on developing materials where such features are important, including as templates for deposition of magnetic material [1] and as photonic crystals [2-8], as well as to scientists studying fundamental questions about the effects of confinement on material properties [9-13]. Because of block copolymers' relevance to many different research efforts, many studies have focused on the factors which determine block copolymer structure [14-20] and ways in which the equilibrium morphologies of these materials can be tuned [21-26]. The focus of this thesis is an investigation into the structures formed by two types of diblock copolymers which contain polydimethylsiloxane (PDMS): polystyrene (PS)-PDMS and polyisoprene (PI)-PDMS (see Figure 1-1 for the repeat units of these materials). PDMS-containing diblocks have received much research interest due to the combination of their ability to form structures with small length scales due to strong spatial segregation between the component blocks at comparatively low molecular weights and the difference in chemical reactivity between PDMS and other polymers with a variety of materials which allows for removal of one of the blocks while the other retains its initial morphology; these PDMS PS P1 Si Figure 1-1: The repeat units of the three polymers used in this thesis are displayed above. 19 Figure 1-2: Schematic showing the structures of the various ordered diblock copolymer morphologies. "f" indicates the volume fraction of component A. Taken from [27]. two features make PDMS-containing systems ideal block copolymers for pattern transfer applications. 1.2 Block Polymer Background Because block copolymers have received substantial interest from engineers interested in applying their unique physical properties to technical problems, much is known about their structure and the way that they behave. The simplest class of block polymers (and the class that this thesis focuses on exclusively) are diblock copolymers, which are composed of two homopolymers which are chemically linked at one end to each other. At thermodynamic equilibrium, diblocks form five known structures depending on the volume fraction of the minority block: a disordered and phase-mixed structure, a body centered cubic array spheres composed of the minority block chains in a matrix composed of the majority block chains, an array of hexagonally packed cylinders composed of the minority block chains in a matrix composed of the majority block chains, a structure with two interpenetrating gyroid networks composed of the minority block chains in a matrix composed of the majority block chains, and alternating lamellae composed of chains from each block (see Figure 1-2). This has been verified theoretically for idealized diblocks [15] (see Figure 1-3) which are Gaussian chains (not valid for shorter chain polymers), have a X which is independent of composition, and have known polydispersity (often assumed to be one) and ratio between segment size (often assumed to be one.) The morphology phase diagram has also been studied experimentally for the widely used diblock consisting of polystyrene (PS) and polyisoprene (PI) [16] (see Figure 1-4.) These microdomain phases are also present in other diblock systems, although the phase diagram boundaries have received less interest, as it has been assumed that they are similar to the PS-PI case. Theoretically, the thermodynamics of block copolymers is well understood; the structure at equilibrium is that with the lowest free energy and is typically calculated using self-consistent mean field theory (SCFT). This procedure involves numerically minimizing the free energy of the system while assuming that each chain is subject to a mean field instead of computing individual chain-chain interactions. While this requires extensive computational power and the results are not readily apparent before the calculations are performed, a basic understanding of the factors influencing equilibrium structure can be easily explained and is also quite illuminating. 20 100 W 4 80 ti L H 60 XN 40 20 2Y CPS CPS DIS 0 0.2 0.4 f 0.6 0.8 1 Figure 1-3: A theoretically computed phase diagram for block copolymers which shows the expected morphology as a function of volume fraction of one of the blocks and xN. Q229 corresponds to BCC packed spheres and Q230 indicates the double gyroid phase. CPS corresponds to a close packed sphere structure (FCC or HCP) and is hard to distinguish experimentally from BCC packed spheres. The figure is taken from [28]. IfxIA LAM Hx 40 1011 Disordered 0 0 0.1 - - 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Figure 1-4: Experimentally determined phase diagram for PS-PI diblock copolymers taken from [16]. There are several differences between the data shown here and the morphologies which are theoretically predicted to occur: ODTs occur at lower temperatures, the phase boundaries are not symmetric around the 50:50 PS:PI composition, non-classical phases are present (e.g. HPL, which is the hexagonally perforated lamellar phase), and unanticipated Order-Order Transitions (OOT)s occur, among others. 21 For two dissimilar polymers which are unattached, i.e. an A-B homopolymer blend, there are two factors which influence their tendency to demix: the loss of entropy upon confining each polymer to specified regions of space and the gain in enthalpy achieved by minimizing unlike monomer-monomer contacts. Both of these can be determined by the Flory-Huggins theory, which states that the configurational entropy of mixing per lattice site is given by ASmix = -k (1 N1 In 01+ n4 N2 (1.1) 2 and that the sum of the resultant changes in entropy and enthalpy due to increased spatial proximity between unlike monomers per lattice site is (1.2) AHmix = k - T - X-#1-42 In equations 1.1 and 1.2, N is the degree of polymerization for each polymer, # is the volume fraction, k is Boltzmann's constant, T is temperature, x is the interaction energy resulting from having unlike monomers in close proximity scaled by Boltzmann's constant multiplied by temperature. X can be computed by x = C - [eAB - 2 (eAA + (1.3) EBB)] where C is a constant accounting for the number of monomer-monomer contacts associated with the lattice and Exy is the pairwise interaction energy between monomers X and Y. As in all materials, the equilibrium morphology is that which minimizes the Gibbs free energy of the system. This occurs when the chemical potential of each component is equal in all phases; for this to be true, 9 (where nA is the number of moles of A in phase 1) does not vary with phase. The chemical potential in any mixed phase is given by p 1 =jtp 1 o+k-T,-n#+ (1-41- 1- + (1 - #1) (1.4) where 1 is the chemical potential of species 1 in a mixed phase and piLo is the chemical potential of species 1 in its pure phase. This equation can be plotted to yield the binodal curve which defines at which volume fractions and values of x, NA and NB phase separation occurs. For block copolymers, the Flory-Huggins polymer blend theory is modified by the presence of a junction linking the homopolymers together. Because of this, the size of phase-separated domains is limited, increasing the contacts between unlike monomers in the microphase separated state. The attached homopolymers are also stretched away from the junction point, which decreases the configurational entropy. These two effects both raise the free energy of the phase separated state. For example, in a 50:50 blend of chains of equal length, phase separation occurs at x -(NA + NB) of 2, if the two chains are attached the x -(NA + NB) at which phase separation occurs is raised to ~ 10.5. For other compositions, the required x - (NA + NB) is increased further. Boundaries between morphologies in diblocks occur when the sum of chain stretching energy and surface energy 22 becomes minimized for a different microdomain structure. Experimentally, it became possible to perform precise investigations into block copolymer structure after anionic polymerization was discovered by Szwarc in 1956 [29]. Before development of this technique, polymers were synthesized through either free-radical polymerization or through step growth polymerization, both which resulted in highly polydisperse products. In the former, initiators decompose to form two free radicals which react with monomers to form long chains; in this case, polymerization is terminated when two propagating radicals meet and there are chains of all lengths present during the polymerization process. In step growth polymerization, chemical reactions occur between monomers which result in the formation of ester, amide, or other bonds; the reaction proceeds until equilibrium is reached. Neither of these two methods is well suited for creating model polymers: the choice of monomers is limited and there is no ability to precisely control their sequence on the chain and form ideal diblock chains. In anionic polymerization, a living anion is formed which is coordinated to a cation. When this anion reacts with monomers in the reaction vessel, the chain grows while maintaining a negatively charged end group. If initiation occurs at the same time, all of the chains in the solution grow at approximately the same rate and the composition of the chain can be tuned by changing the composition of monomers present during the reaction. Because of this, precisely defined chains with uniform lengths and compositions can be achieved, allowing for the creation of model materials whose structural properties can be studied as a function of volume fraction/composition. This thesis investigates diblocks composed of pairs of the following homopolymers: PS, PI and PDMS. PS is a polymer with a backbone composed of carbons connected by a single bond; to half of these backbone carbons, a phenyl ring is attached. PS has a glass transition of approximately 100 0 C. PI has a backbone made up of an alternating pattern of two single bonds and one double bond; a methyl group is attached to one of the double-bonded carbons. Both it and PDMS are rubbery at room temperature. PDMS has a backbone with alternating silicon and oxygen atoms; two methyl groups are attached to each silicon atom. See Figure 1-1 for more detail about the molecular structures of these monomers. 1.3 Characterization tools This thesis primarily relies upon two characterization tools to understand block copolymer structure: transmission electron microscopy (TEM) and small angle x-ray scattering (SAXS). These two techniques provide complementary information about sample structure: TEM provides two-dimensional projections of real-space of selected small areas of the structure while SAXS experiments yield information about the characteristic feature sizes and spacings over volumes of cubic millimeters. The use of both of these techniques together allows for determination of the size and shape of microdomain features (and hence microdomain structure) in the material of interest and their orientation with respect to each other. TEM studies involve electrons which have either been transmitted (forward scattered) through a sample (bright field) or scattered at a defined angle (dark field). Because the 23 polymers studied in this thesis were amorphous at room temperature and thus do not have a crystalline diffraction pattern, only bright field imaging was used. In this case, regions of the sample through which many electrons have been transmitted remain bright while areas through which many electrons were scattered appear darker. Because the scattering power of an atom is dependent upon its atomic number, there is significant inherent contrast between the carbon-based blocks and the PDMS blocks, removing the need for any heavy metal stain. While the TEM can readily produce images which appear to show true structure, understanding the actual morphologies of the block copolymers being examined has more subtlety: the data that is being recorded is a two-dimensional projection of a selected location of the sample and so the image can be affected by the angle of the grain with respect to the electron beam, the thickness of the sample, the number of unit cells present in the projection, etc. Moreover, as samples are typically prepared by cutting very thin slices (less than 100nm) from a piece of bulk material, it is possible that additional artifacts may be introduced into the image due to deformation during cutting or that the area under examination may not be representative of the larger overall sample. For this reason, it is necessary to examine multiple locations in the slice, corroborate TEM-based structure assignment with SAXS data, and be aware of how commonly found structures typically look in projections at different angles. The microscope must also be appropriately focused in order to accurately measure distances and generate the most representative image. Despite these limitations, it is the best way to perform real space imaging of structures whose characteristic length scales match those of block copolymers. See Figure 1-5 for a schematic of how images are produced in a TEM using bright field imaging and a picture of a TEM instrument. SAXS data is obtained by irradiating a piece of bulk material (typically 1 mm thick) with an x-ray beam and recording the intensity of scattered radiation at a set distance (typically 1-2m) from the sample. If there are ordered, periodic variations in the electron density of a material, x-rays will be scattered at a characteristic angle given by nA = 2dsin6 (1.5) where n is any integer, A is the wavelength of the incident radiation, d is the spacing between of the periodic electron density variation in the sample, and 6 is half of the scattering angle. This is due to the constructive interference of waves when the path difference between each plane is an integer multiple of the x-ray wavelength. In small angle scattering, 6 is on the order of several degrees and q is given by q = A- sin0 = 2-r d (1.6) For a d-spacing of 35 nm and a wavelength of 1.54 A (which is the typical setup in laboratory x-ray diffractometers which use Cu-Ka radiation), 6 is approximately .002 radians and q is approximately 0.18 nm- 1 . Therefore, knowledge of the first location in q of the maximum scattered intensity allows for a direct calculation of the spacing between the planes in the material with identical electron density distributions which have the largest spacing. Peaks in the scat24 i0 W, Jn Figure 1-5: A schematic of a transmission electron microscope [30] (a) and a photograph of a transmission electron microscope [31] (b). On the right hand side of (a), the path that electrons take as they travel through a TEM in bright field imaging mode is compared to that of a beam of light travelling through an optical microscope. tering intensity at higher q values are present when there are additional scattering angles at which constructive interference occurs. This happens when n > 2 in equation 1.3 and when planes other than those responsible for the lowest q scattering scatter x-rays. The ratio of the q's of these higher-order peaks is determined by the morphology of the sample: both the geometry of the features themselves and the lattice type that they occupy affect the positions and intensities of scattering angle maxima. The influence of feature shape on the scattering pattern is called the form factor; at different angles, different geometries will have different intensities of radiation scattered due to interference of the waves that are scattered from each part of the object. The influence of the lattice that the features are ordered on is called the structure factor; it describes how the minima and maxima in the intensity of the scattered radiation depend on the interference of the radiation scattered by each object. The product of these two terms describes the total intensity. Also, as the perfection of the lattice increases, additional higher order peaks appear and the peaks become sharper; thus, SAXS data can corroborate the morphologies established by TEM images and provide insight into structural uniformity across macroscopic length scales. See Figure 1-6 for a schematic of how diffraction patterns are obtained using SAXS and an image of a laboratory-scale instrument. 25 LiW source beam stop sample and scattered beams direct pinholes a 1 meter detector Figure 1-6: A schematic of a small angle x-ray scattering setup (a) and an image of a SAXS instrument [32] (b). (a) shows the path of the x-rays from the source, through the sample, and on to the detector after scattering. 26 1.4 Motivation for studying PDMS-containing block polymers Due to their unique properties, PDMS-containing block polymers have been used for many different applications. These include both the development of new technologies and the creation of materials which can be used to measure scientific properties not readily accessible by other means. In the former category, works which take advantage most directly of the different chemistries of PDMS and carbon-backbone polymers typically involve removal one of the component blocks after structure formation. PDMS can be removed from a material by an etch in hydrofluoric acid or tetrabutylammonium fluoride; this has been used to create many different advanced materials. Materials with PS serving as the matrix have been made which contain either spherical or gyroid shaped voids [33]. Cross-linked PI has also been of interest for use as a matrix material because its shape after removal of the PDMS can be tuned by crosslink density; PI with cylindrical voids can retain its initial structure when highly crosslinked [34] but will have voids which are collapsed until swollen with a solvent when lightly crosslinked [35]. Polybutadiene (PB), another elastomer, has also been used to make a matrix containing voids which are gyroid shaped or performated lamellar shaped [36]. In all of these cases, the etching rate of the PDMS by the fluorinated chemical is sufficient for preparation of bulk porous samples which could be used to create large-area applications. Other work has focused on exploiting the morphologies found in thin film PDMScontaining block copolymers. Investigators have typically used oxygen plasma to simultaneously remove the carbon-containing block and convert PDMS into a silicon oxy carbide ceramic for this application due to its higher safety compared to fluorinated etchants and because the smaller distances that the etchant must penetrate allows the use of slower and less aggressive methods. When PDMS is the majority phase of the block copolymer being treated, oxygen plasma treatment yields a thin film nanoporous structure similar to those described above for bulk materials; when another olefin is placed under it and etched, a topographic pattern is formed where material is removed from the areas the pores covered [37]. When PDMS is the minority phase, the ceramic formed is left free-standing on the substrate where the PDMS was located initially; dots [38] as well as other structures have been formed this way. In some cases, the locations where the carbon-backbone polymer was located can be in-filled with a technologically relevant, multifunctional material such as Chromium [1]. Perturbing the block copolymer structure from its equilibrium morphology can expand the range and utility of the morphologies of the resulting ceramic. This can be accomplished by creating templates using interference lithography into which the block copolymer can be deposited. The presence of surfaces in close proximity to each other raises the energy of defects and so forces the creation of highly aligned, nearly defect-free, and uniformly spaced structures. For example, cylinders [39] can be formed when long and narrow trenches are used to confine the material and concentric tori when shallow cylinders are [1]. When low molecular weight block copolymers are placed in these templates, free standing cylinders with very small periods (17 nm) and line widths (8 nm) are formed. Another way to tune the morphology formed prior to etching is to perform a 27 two solvent treatment, where the first step swells the PDMS as the film is being deposited and then the second promotes contraction so that rings are the final structure that is formed [40]. The high immiscibility of PDMS with other polymers has been another area which has driven application development. The low cohesive energy density of PDMS compared to other materials means that the chemicals that it is soluble in and can solvate overlap very little with those for the blocks it is commonly linked to. When the block copolymer is the solute in a preferential solvent, spherical and cylindrical micelles with a core comprised of the block which is insoluble are formed; these solutions can be dried and the previously solvated structures collapsed into tablets and ribbons [41]. When the block copolymer is the solvent, materials such as C60 [42] and gold nanoparticles [43] are localized in the non-PDMS block. In the latter two examples, the concentration of solute in the diblock is low enough that well-ordered and uniform morphologies are formed by the solutions, increasing the technological relevance of this patterning method. The differential solvation properties of block copolymers can also be used in instances where a uniform and ordered morphology is not required but the presence of two materials with different physical properties in close spatial proximity is desired. For these cases, small amounts of diblock copolymer can be added to a blend composed of the constituent homopolymers (i.e. PDMS and the block that it is linked to) to reduce the free energy of the interface between the homopolymers. This stabilizes homopolymer domains against coarsening and thus preserves the morphological integrity of the original material over its lifetime. PDMS has been combined in a diblock with PS [44-46], PI [47], and polyethylene oxide (PEO) [48] for this purpose. PDMS-containing diblocks can also lower the energy of free surfaces by forming an interfacial layer where the PDMS block is in contact with the air interface and the other block is in contact with the substrate. This property has been exploited by experimentalists to form stamps for nanoimprint lithography where a high surface energy polymethyl methacrylate (PMMA) tethers the diblock to the substrate while the low surface energy PDMS allows for facile mold-resist separation [49] as well as lubricating [50] and low energy [51] surfaces. Other uses of PDMS-containing block polymers where the primary effect of interest occurs due to the close spatial presence of two materials with differing properties are numerous. In the field of selective membranes, PDMS has been used as both the more permeant conductive block and the less conductive block; when paired with sulfonated polystyrene in a membrane [52], it increases the ratio of water conducted to ethanol conducted while when combined with PMMA it decreases the ratio of water conducted to methanol conducted [53]. The former occurs because PDMS retards both methanol and water permeation through the membrane by restricting the size of the more highly conductive pathways present; methanol is retarded more than water so the overall membrane is more selective. The latter occurs because both PMMA and PDMS are less conductive for these polar materials than sulfonated PS but the rubbery PDMS allows many more molecules to permeate than the glassy PMMA due to its larger free volume; as ethanol has a higher conductivity than water in PDMS, increasing the PDMS volume fraction in the membrane increases its preference for conducting ethanol over water. PDMS has also been used to increase the solubility of ionically conducting polypyrrole in order to 28 aid processability [54]. Other practical uses of PDMS-containing block polymers include flame retardant transparent materials when PDMS is attached to polycarbonate [55] and materials for external coatings for spacecraft when combined with polyimides due to their resistance to etching by the oxygen present in outer space and ability to couple mechanically with the commonly used underlying polymer layers [56]. In addition to being widely known for their use in practical and technical applications, PDMS diblocks have also received attention for their ability to be formed into materials on which scientific experiments about fundamental polymer physics can be performed. These are primarily investigations into the mechanical and physical-chemical properties of one of the blocks when under confinement. Examples include investigations into the rubbery dynamics of PDMS under soft confinement for PI-PDMS diblocks [57] and PDMS under hard confinement either tethered at one end in PS-PDMS diblocks [58] or at both ends in PS-PDMS PS triblocks [59]. Other work that PDMS has been used for in this area includes a research program dedicated to studying polymer dynamics under solvation and tethering; PDMS is an ideal block to attach to polymers under study because its low surface energy pins it to the surface of fluids which allows for the creation of model systems in which many different experimental factors such as surface grafting density, chain length, etc. can be independently varied [60]. 1.5 Prior work on PDMS-containing block polymers While many theoretical advances have been made in understanding block copolymer structure and how it depends upon composition, segregation strength between the component blocks, and other factors, experimental knowledge of practical systems is necessary for engineers who wish to make use of specific materials. Because of this, there have been several studies examining the bulk structures PDMS-containing block polymers. For these works, PDMS has been combined with many other polymers, including glasses, rubbers, or polymers commonly used in biological or electronic applications. In many cases, X is quite large for the resultant diblocks; the PDMS backbone is covered by methyl groups which have a lower cohesive energy density than many other arrangements of atoms and so the two components are often immiscible even for short chain lengths and high temperatures. Prediction of x for PDMS-containing diblocks has been treated theoretically in [61] and [62]. Another interesting feature of block polymers containing PDMS is the experimental observation of structures present at volume fractions which are substantially different than expected. Theoretical and experimental descriptions of the classical phase diagrams of diblock copolymers predict a symmetric phase diagram in which the four ordered phases occur at defined volume fractions independent of the chemical identity of the component blocks. However, this is not true for many actual systems. In Table 1.1 and figures 1-7 and 1-8, all of the bulk morphologies which have been observed in the literature and accompanied by reasonably supporting data are displayed as a function of the volume fraction of PDMS xN; these include morphologies discussed in the previous section as well as those treated below. Many authors have noted the presence of morphologies occurring in unexpected regions of the theoretical phase diagram; figures 1-7 and 1-8 show that 29 50 100 45 95 40 90 35 85 * 30 80 z25 75 20 70 15 65 10 60 5 55 0 so a 0 0.2 0.4 1 0.8 0.6 b 0 0.4 0.2 PDM S sphees a PDMScyhnders 4 PDMS DG I PDMIS perforated Lamena 1 PDMS volume fraction PDMS volume fraction a 0.8 0.6 U .ameeiJ 6 Oth b1 k peforatc d me. Oa Ot! blot k D, a Othrl biock vhldLts UOther block phtee 4 Diordered igure 1-7: Three composite plots showing previously observed morphologies in PDMS-containing diblocks as a function of nd PDMS volume fraction. Eighty six points are plotted in total; these include multiple plots of the same sample on two s f an observed phase transition. It is clear that the locations of experimentally observed phases are not predicted perfectly he phase boundaries computed using SCFT. It appears that the range of volume fractions at which the morphologies oc re broader than those predicted and that those regions interpenetrate. This may be due to more advanced factors such onformational asymmetry, polydispersity, or be the result of preparation procedures which produced metastable samp Graph (a) shows morphologies found at weak segregation; graph (b) shows morphologies found at intermediate segregat raph (c) shows morphologies found at strong segregation. The lines dividing the regions of expected morphologies w reated by plotting a smooth curve through the points found in [15] and [27]; for values of xN outside the range compute hose works, the boundaries between phases were assumed to be independent of xN (vertical phase boundaries.) 1500 1300 1100 z 900 700 500 300 100 0 0.2 0.4 0.6 0.8 1 PDMS volume fraction U PDMS phce U PDMScyhnde S PDNIS DG PDNIS eforIted lamdlae dl a Lamelie Other bicxk perfo ted aimele Othe biotk D6 a Othe block cihader a Othe blod Spheles a Doot r ed ontinued Figure 1-7: Three composite plots showing previously observed morphologies in PDMS-containing diblocks a unction of xN and PDMS volume fraction. Eighty six points are plotted in total; these include multiple plots of the sa ample on two sides of an observed phase transition. It is clear that the locations of experimentally observed phases ot predicted perfectly by the phase boundaries computed using SCFT. It appears that the range of volume fractions hich the morphologies occur are broader than those predicted and that those regions interpenetrate. This may be due ore advanced factors such as conformational asymmetry, polydispersity, or be the result of preparation procedures wh oduced metastable samples. Graph (a) shows morphologies found at weak segregation; graph (b) shows morphologies fo intermediate segregation; graph (c) shows morphologies found at strong segregation. The lines dividing the regions xpected morphologies were created by plotting a smooth curve through the points found in [15] and [27]; for values of utside the range computed in those works, the boundaries between phases were assumed to be independent of xN (vert hase boundaries.) PDMS Cylinders PDMS Spheres 1600 500 S 450 1400 400 1200 350 1000 300 800 z 250 Z 600 400 200 0 ao 200 KL 0.2 150 100 50 0 0.4 0.6 0.8 1 b0 8 PDMS suheie PDMS cyhndcr PM)IS DG PDMS pd.0oatd OamIah 0.8 0.6 0.4 0.2 1 PDMS volume fraction PDMS volume fraction a t amellae Other boc k perorated lanmlle Othe' blo k DG a Ot her Wok vlide * Othe block spheree a Disorder'd igure 1-8: Plots showing previously observed morphologies in PDMS-containing diblocks as a function of xN and vol action. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS DG DMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres (i), isordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve through oints found in [15] and [27]; for values of XN outside the range computed in those works, the boundaries between phases w ssumed to be independent of XN (vertical phase boundaries.) Note that many phases are observed significantly outsid he regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proxim o each other and in some cases the phase boundaries appear to interpenetrate. DG 300 PDMS HPL PDMS DG 100 90 250 80 70 200 60 z150 S50 40 100 30 20 50 10 0 0 0.2 C 0 0.4 0.8 0.6 1 d 0 0.2 PDMS volume fraction 8 PDMS B PDMS 8pre ybnlmder 9 PDMIS DG - PDMIS perforated Iamelej 0.4 0.6 0.8 1 PDMS volume fraction a Lamllae * Othef block perforte f d amellae Other Lock DG 0 Other block t yhnde 3 0ther bock 3 D 1Odered Wpe ontinued Figure 1-8: Plots showing previously observed morphologies in PDMS-containing diblocks as a function of XN olume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS ), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres nd disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve throu e points found in [15] and [27]; for values of xN outside the range computed in those works, the boundaries between pha ere assumed to be independent of XN (vertical phase boundaries.) Note that many phases are observed significantly outs the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proxim each other and in some cases the phase boundaries appear to interpenetrate. Other Block HPL Lamellae 300 50 45 250 40 35 200 30 z 150 z 25 20 100 15 10 50 5 0 e 0 0 0.4 0.2 0.6 0.8 1 f 0.2 0 PDMS volume fraction U PDMS sphece a PDMS ybdes 8 PDMIS DG PDMSperoiated lamelfaf 0.4 0.6 0.8 1 PDMS volume fraction a talnwlae 0 Other bock perfoted lamellae Other -o k D *ther block cybooers a Othelr bkxk phere. a Dordered ontinued Figure 1-8: Plots showing previously observed morphologies in PDMS-containing diblocks as a function of xN olume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS ), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres nd disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve thro e points found in [15] and [27]; for values of XN outside the range computed in those works, the boundaries between pha ere assumed to be independent of xN (vertical phase boundaries.) Note that many phases are observed significantly outs the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proxim each other and in some cases the phase boundaries appear to interpenetrate. Other Block DG Other Block Cylinders 50 250 45 40 200 35 30 150 z z25 20 100 15 10 so 5 0 g0 0 0.2 0.4 0.6 h0 0.8 0.2 PDMS volume fraction a PDMS sphere U PDMSt vbndef 0 PDMS DG PDMS plror aed lelma 0.4 0.6 0.8 1 PDMS volume fraction a .LIela a Othe blok k pirtorted lanea i OtIr Luo k DO a Other bOck ylmdels I a Other blot k 'pheres a Dsordored ontinued Figure 1-8: Plots showing previously observed morphologies in PDMS-containing diblocks as a function of xN olume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS ), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres nd disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve throu e points found in [15] and [27]; for values of xN outside the range computed in those works, the boundaries between pha ere assumed to be independent of xN (vertical phase boundaries.) Note that many phases are observed significantly outs the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proxim each other and in some cases the phase boundaries appear to interpenetrate. Other Block Spheres 200 50 180 45 160 40 140 35 120 30 z100 l25 80 20 60 15 40 10 20 5 0 0 i Q Disordered 0.2 0.4 0.6 0.8 1 D 0 0.2 0.4 PDMS volume fraction U PDMS pherv a PDMScyhtnder- U PDMS DG P)MS perfoated lamela 0.6 0.8 1 PDMS volume fraction a Lamellae * OtIot )k peffoahd aia1m Ot 1hr I DG U Other block vhlder s Other lotik pheices a DI, pdred ontinued Figure 1-8: Plots showing previously observed morphologies in PDMS-containing diblocks as a function of XN a olume fraction. Each graph shows the data for a distinct morphology: PDMS spheres (a), PDMS cylinders (b), PDMS ), PDMS HPL (d), lamellae (e), other block HPL (f), other block DG (g), other block cylinders (h), other block spheres d disordered (j). The lines dividing the regions of expected morphologies were created by plotting a smooth curve throu e points found in [15] and [27]; for values of xN outside the range computed in those works, the boundaries between pha ere assumed to be independent of xN (vertical phase boundaries.) Note that many phases are observed significantly outs the regions of volume fraction and values of XN that are predicted theoretically; different phases also exist in close proxim each other and in some cases the phase boundaries appear to interpenetrate. in many cases structures appear at volume fractions of PDMS closer to 0.5 than would be predicted theoretically. Reference [63] examined possible causes for the presence of PDMS-minority phases occurring at higher PDMS volume fractions than expected by considering the effect of the conformational asymmetry of PDMS and PS, which is the difference in the ratios between flexibility (measured by the radius of gyration of the homopolymer chains) and specific volumes of each chain and can sometimes account for differences in morphologies obtained; stiffer chains prefer to form structures which have less domain surface curvature. However, the calculated conformational asymmetry parameter was much too small to explain the pronounced shift that they observed. The presence of morphologies occurring in unexpected regimes is important both technically and scientifically: it is necessary for synthetic chemists to understand what compositions they should target in order to achieve their desired morphologies and it is interesting to polymer scientists to understand why unexpected behavior occurs. The first comprehensive study of PS-PDMS morphology was undertaken by the Register group in 1995 [64]; in this paper, the bulk morphologies of six diblocks of PS-PDMS were studied using TEM, SAXS, and Differential Scanning Calorimetry (DSC.) The microdomain structures observed were different than those which were expected based upon accepted PS-PI behavior: while samples with a PDMS majority behaved as anticipated, two samples with a slight PS majority which were predicted to fall inside of the lamellar region instead formed morphologies which consisted of PDMS cylinders in a PS matrix. The authors discussed the possibility that the casting solvent (toluene) might have preferentially swollen the polystyrene matrix during casting and that a non-equilibrium structure may have been locked in as the polystyrene went through its glass transition during solvent evaporation. They concluded, however, that this explanation may not be accurate because the Mark-Houwink exponents for PS and PDMS in toluene are very similar. Because this exponent describes scaling of the viscosity of a polymer with chain molecular weight and intrinsic viscosity is directly proportional to the volume of a chain in solution, using its value to as a way to compare swelling and solvent quality is quite reasonable. Moreover, similar results were found in a study that analyzed the SAXS patterns of five different diblocks [63]: samples with volume fractions and values of XN predicted to have morphologies which were disordered, cylindrical, and lamellar range showed SAXS patterns consistent with theoretical predictions but one sample close to the PDMS double gyroid-lamellar boundary had a scattering pattern typical of cylinders. This study also examined a diblock which was expected on the basis of volume fraction to form BCC spheres of PDMS in a PS matrix, but the scattering pattern did not have sufficient peaks to be interpretable. The discovery of the gyroid phase in PS-PDMS further supports the idea that the volume fraction boundaries between phases are shifted towards larger volume fractions of PDMS in samples containing a PDMS minority. In [65], the double gyroid was found to exist at strong segregations in the region which lamellae occur in theoretical predictions. The protocol used in this work was similar to that in the study performed by the Register group: the polymers were cast from toluene and annealed at 130 0 C for one week before examination using TEM and SAXS. Four samples, all in the expected lamellar region, 37 were discussed: one which had a morphology consisting of PDMS cylinders, two samples contained a PDMS double gyroid, and one sample was composed of lamellae. One other study on block polymers containing PS and PDMS showed even stranger phase behavior [66]; a PDMS-PS-PDMS triblock with a 50:50 PS:PDMS volume composition displayed a morphology consisting of large PDMS spheres in a PS matrix. While this sample was obtained in an unknown way from a film of unknown thickness (not from a bulk sample) and cast from toluene and annealed at 100 0C for 1.5 hours, making it probable that its bulk equilibrium structure was not obtained, the formation of a morphology consisting of spheres where lamellae are predicted is highly unexpected! If the predicted phase boundaries are accepted, this would require the PS phase to be swollen at least seven times as much as the PDMS phase at the glass transition of PS (which itself will not happen at high solvent concentrations) and then exist in a deep potential well so that it is not erased by annealing. While this is possible, it is more probable that there is some equilibrium phase that is not lamellae which is present at this composition. Further support for the conclusion that this morphology is not an artifact of preparation methods is its formation when cast from both toluene and bromobenzene [66]. Work has also been done on the hydrogenated analogs of PS-PDMS-PS triblocks (polycyclohexane-PDMS-polycyclohexane) [62]. Here, a selected set of triblocks were studied using SAXS, TEM, and DMA in order to locate the ODTs and morphologies of these samples. All were within the lamellar volume fraction region and all displayed lamellae. Prior work on the equilibrium bulk morphology of PI-PDMS diblocks consists of an examination of the ODTs of two PI-PDMS diblocks using SAXS and dynamic mechanical analysis (DMA) [67]. In this study, the presence of a double gyroid morphology where lamellar morphology was expected was also found. While both samples investigated had morphologies which were expected to be within the lamellar region (with volume fractions of 0.46 and 0.63), the DMA data was consistent with lamellae for the sample closer to 50:50 PI:PDMS and the SAXS data was consistent with the double gyroid morphology for a sample close to the lamellae-double gyroid phase boundary. Similarly to PS-PDMS and PI-PDMS, Poly(ethylenepropylene) (PEP)-PDMS has received attention due to the high x parameter between the component blocks and associated ability to form ordered structures on small length scales. These morphologies have been studied using SAXS, small angle neutron scattering (SANS), and DMA. The typical experiment that has been performed is the location of the ODT using SAXS, SANS, and/or DMA, determination of the scattering form factor, and the assignment of a morphology based upon this data. The morphologies found agreed with theoretical predictions in some cases: in [68] BCC cylinders of PDMS were found in the expected volume fraction range and in [67] when two samples which were expected to display lamellae at room temperature were examined by SAXS, the resulting data were consistent with this hypothesis. One of the samples which had a volume fraction closer to those in the double gyroid range, however, went through two OOTs upon heating; first into a layered structure and then into a double gyroid structure before ultimately disordering at the ODT. Another study [69] also found an anomalous pressure dependence of the ODT for a dilock with almost symmetric composition; instead of having a monotonic change in 38 ODT as pressure increased, the ODT had a minimum at one of the intermediate pressures examined. The combined evidence from these three studies shows that the well-known PS-PI phase diagram cannot fully describe the behavior of PEP-PDMS diblocks. The morphologies of PEO-PDMS block copolymers have also received much interest due to the small scale highly segregated structures that can be found and the formation of ordered liquid structures. This system has been extensively characterized above the PEO melting point using optical microscopy and SAXS in [70]. While many of the volume fractions studied in this work were formed by creating blends of diblocks (which changes the thermodynamics of mixing by altering the chain stretching that occurs; however, the volume fraction boundaries between morphologies should be close to those found in pure diblock systems), an extensive range of blend compositions at varying temperatures was covered; the authors were able to present a phase diagram as a function of composition and the ratio between molecular weight and temperature, which scales the same way as xN. In this case, the phase behavior was quite different from that found in most other diblock systems: only lamellae, cylinders, and a phase with cubic symmetry were found. In this case, the phase diagram is again remarkably asymmetric with respect to volume fraction; while on the PDMS-rich side of the phase diagram the boundaries between morphologies are in the expected locations, only lamellae are present in the PDMS-poor region and the region in which they are found extends into much higher volume fractions than for PS-PI. Most other studies have focused on the PDMS-poor side of the phase diagram, so this thesis work provides a nice complement. Polybutadiene (PB) is another polymer that has been combined with PDMS to form a diblock in order to take advantage of the phase separation at small length scales offered by PDMS-containing diblocks; in this case, it is the crystallization of short and spatially-confined PDMS chains that is of interest. One study [71] examined nine different PB-PDMS diblocks to answer this question. Unfortunately, in this case little can be said except that the presence of PDMS spheres and some combination of PDMS spheres, PDMS cylinders, and lamellae were present at the varying volume fractions; the preparation procedure was not fully described, only one descriptive micrograph was provided for each image, and no other technique was employed to confirm the assigned phases. The morphologies were assigned based upon examination of a sample cast from a THF solution (temperature, time of casting, and whether or not thin film or bulk samples were investigated was not discussed); crystallization behavior was investigated using a DSC on bulk powders. However, if the morphology claims in the paper are accepted at face value, it appears that PDMS spheres were present well into the region where PDMS cylinders are expected to be found. One other paper [72] has also examined PB-PDMS morphology. Here, one diblock with a volume fraction where cylinders are expected was synthesized and analyzed using SAXS and DMA. As it was heated, it underwent a transition from a cylindrical morphology through BCC packed spheres into a disordered structure, as would be expected from theory. PDMS block copolymers containing polyamides have been synthesized to probe the effect of crystallization of one of the component blocks on overall morphology and to understand how structure varies with volume composition and processing [73]. In this work, it was found that the observed morphologies for two sphere-forming diblocks (one which 39 has a volume fraction just on the disorder side of the disordered-spherical boundary and the other which has a volume fraction within the cylinder region near the cylinder/double gyroid boundary) depend strongly upon both the volume fraction and processing treatment. It was found that the diblock with a very low percentage of PDMS would develop different structures when solvent cast than when cooled from a melt; in the former case, the PDMS separated from the solvent before crystallization of the nylon block and so spheres of PDMS are dispersed in spherulites while in the latter, a morphology with a much more regular arrangement of PDMS spheres within a nylon 6 matrix with less crystallinity was obtained by quenching in the structure present above the crystallization temperature of nylon 6. For the diblock with a larger amount of PDMS, more regularly sized and placed spheres of PDMS in a matrix of nylon 6 were obtained by both solvent casting and melt casting; there was also no significant difference in the nylon crystallinities obtained. 40 Sample morphology #PDMS 0.04 0.05 0.08 0.10 0.12 PDMS Spheres 0.12 0.17 0.18 0.20 0.20 0.25 0.29 0.18 0.23 0.23 0.26 0.27 0.29 PDMS Cylinders 0.32 0.33 0.33 0.34 0.37 0.39 0.39 0.44 41 XN 15000 93 93 35 35 20 78 24 43 33 97 960 27 50 80 28 380 28 24 380 39 95 45 420 460 110 Reference Argon, [73], 1991 Berg, [33], 2003 Rueda, [63], 2010 Lodge, [68], 2002 Lodge, [68], 2002 Lodge, [68], 2002 Berg, [33], 2003 Valles, [72], 2008 Kunieda, [70], 2003 Kunieda, [70], 2003 Huang, [71], 1992 Argon, [73], 1991 Valles, [72], 2008 Rueda, [63], 2010 Berg, [33], 2003 Ndoni, [35], 2009 Uragami, [53], 1999 Ndoni, [36], 2011 Ndoni, [34], 2004 Uragami, [53], 1999 Huang, [71], 1992 Ndoni, [34], 2004 Rueda, [63], 2010 Uragami, [53], 1999 Hill, [65], 2009 Register, [64], 1995 Sample morphology #PDMS PDMS DG 0.37 0.39 0.39 0.41 0.42 PDMS HPL Lamellae 0.39 0.39 0.36 0.36 0.37 0.39 0.39 0.43 0.44 0.45 0.47 0.48 0.48 0.50 0.50 0.53 0.53 0.56 0.57 0.58 0.60 0.65 0.65 0.65 0.71 Other block HPL 0.65 0.65 0.64 Other block DG 0.64 0.65 0.65 0.60 0.60 0.67 0.67 0.82 0.84 Other block Cylinders 42 XN 22 23 21 290 170 20 11 50 38 41 26 25 120 62 210 46 10 9 44 44 67 52 92 260 57 260 120 16 14 230 13 12 19 14 12 10 68 52 78 60 47 200 Reference Berg, [33], 2003 Ndoni, [36], 2011 Ndoni, [36], 2011 Hill, [65], 2009 Hill, [65], 2009 Ndoni, [36], 2011 Ndoni, [36], 2011 Kunieda, [70], 2003 Kunieda, [70], 2003 Huang, [71], 1992 Ndoni, [36], 2011 Ndoni, [36], 2011 Huang, [71], 1992 Huang, [71], 1992 Hill, [65], 2009 Huang, [71], 1992 Bates, [67], 1996 Bates, [67], 1996 Kunieda, [70], 2003 Kunieda, [70], 2003 Deloche, [59], 2003 Huang, [71], 1992 Rueda, [63], 2010 Uragami, [53], 1999 Deloche, [59], 2003 Uragami, [53], 1999 Register, [64], 1995 Bates, [67], 1996 Bates, [67], 1996 Uragami, [53], 1999 Bates, [67], 1996 Bates, [67], 1996 Bates, [67], 1996 Bates, [67], 1996 Bates, [67], 1996 Bates, [67], 1996 Kunieda, [70], 2003 Kunieda, [70], 2003 Kunieda, [70], 2003 Kunieda, [70], 2003 Richter, [57], 2010 Register, [64], 1995 Sample morphology Other block Spheres Disordered qPDMS 0.89 0.76 0.76 0.80 0.80 0.92 0.02 0.12 0.18 0.39 0.46 0.48 0.64 0.65 0.76 0.80 0.86 0.94 XN 51 33 28 31 29 190 39 20 22 11 11 9 28 10 27 28 29 25 Reference Richter, [57], 2010 Kunieda, [70], 2003 Kunieda, [70], 2003 Kunieda, [70], 2003 Kunieda, [70], 2003 Register, [64], 1995 Rueda, [63], 2010 Lodge, [68], 2002 Valles, [72], 2008 Ndoni, [36], 2011 Bates, [67], 1996 Bates, [67], 1996 Bates, [67], 1996 Bates, [67], 1996 Kunieda, [70], 2003 Kunieda, [70], 2003 Kunieda, [70], 2003 Kunieda, [70], 2003 Table 1.1: A chart detailing the morphologies found in the literature for PDMS-containing diblocks (it does not include triblocks or other multiblocks also described in the text); it contains the data displayed in figures 1-7 and 1-8. Some data found in the literature discussed was purposefully excluded if there were significant reasons to believe that the data presented was either not sufficient to establish the morphology claimed, e.g due to large amounts of obvious homopolymer contamination in the TEM images, etc. Note that because much of the relevant information was not available in the papers, in some cases calculations were made to arrive at the volume fraction and XN values. If necessary, volume fractions were computed from mole or weight fractions using commonly available densities. x values were taken from [74] if available or calculated from solubility parameters found in [75]; in the latter case, the reference volume was assumed to be 100O. The temperature used was m01 that at which the last thermal treatment was done (e.g. annealing, crosslinking, etc.) or room temperature if the sample was not thermally treated after casting. For works in which several samples were examined at a large variety of temperatures, the chart displays multiple entries for each transition temperature ±50 C: the lower temperature for the phase present below the transition temperature and the upper temperature for the phase present above the transition temperature. 1.6 Scope of Thesis Project This thesis seeks to expand upon previous work by experimentally elucidating the determinants of PDMS-containing block polymer structure by examining the effect of volume 43 fraction, temperature, and processing history on the morphologies of selected PS-PDMS and PI-PDMS diblocks of varying volume fractions and molecular weights. The majority of the PS-PDMS diblocks investigated have a XN of ~ 100 at 150 0 C (the temperature at which they were annealed and thus where final structure formation occurred) and the majority of the PI-PDMS diblocks examined have a xN of ~ 100 at 25 0 C (due to the rubbery nature of both blocks, structural changes can occur at room temperature). The structures of three low molecular weight diblocks (two of PS-PDMS and one of PIPDMS) were examined with several minute resolution while undergoing differing thermal programs by using SAXS at Brookhaven National Laboratory in order to determine the effect of processing on final morphology and to explore the temperature dependence of the phase diagram at varying values of <pPDMS- 44 Chapter 2 Strongly Segregated Morphologies of PS-PDMS and PI-PDMS Diblocks 2.1 Introduction and motivation This chapter describes the determination of the bulk morphologies of selected PS-PDMS and PI-PDMS diblocks using SAXS and TEM. The diblocks were designed to provide insight into the strongly segregated morphologies of these materials (all samples had values of XN close to 100 at the final processing temperature); diblocks in this regime are of interest for application development because they are relatively stable with temperature (the boundaries between different morphologies in 4>PDMs become nearly independent of composition as XN increases) and because the stronger segregation of material means there is less molecular-scale mixing so the different properties associated with each block (e.g. etch resistance) are more pronounced. The samples span a range of volume fractions (12%-65% PDMS; see Table 2.1) and three different morphologies (PDMS spheres, PDMS cylinders, and lamellae) are experimentally observed. 2.2 Materials synthesis The polymers used in this thesis were synthesized in the laboratory of Professor Apostolos Avgeropoulos at the University of Ioannina in Ioannina, Greece by his graduate students. The information described below was kindly provided by Professor Avgeropoulos. The total process consisted of three steps: purification of the reagents and solvents, polymerization, and molecular characterization. Because of the high reactivity of species present during the polymerization process, ultrapure reagents, solvents, and polymerization environments are necessary for the elimination of premature products and side reactions while allowing for production of monodisperse products. Therefore, extensive procedures were undertaken to prevent contaminants from being present during the reaction. As the polymerization procedures were performed in both benzene and tetrahydrofuran (THF) in different phases of the reaction, both of these solvents needed to be purified. 45 This was accomplished by using ground, degassed calcium hydride as a drying agent and performing a distillation. In the case of THF, two additional distillations were performed: the first using a sodium mirror to indicate purity and the second containing a potassium: sodium alloy in a 3:1 ratio. The three monomers used in synthesis: styrene, isoprene, and hexamethylcyclotrisiloxane (D3 ) were also purified. These reagents were similarly dried over ground calcium hydride before further preparation. At the conclusion of the preparation procedures, the monomers were distilled into and stored in an apparatus containing ampoules and break seals. Commercially synthesized styrene was also reacted with dibutylmagnesium in an ice bath to remove impurities. After drying, commercially synthesized isoprene was distilled into a flask containing a small amount of n-butyllithium and stirred for thirty minutes at 00C twice. As D3 is a solid at room temperature, it was dissolved in benzene during its purification procedure and before introduction to the reaction. After the solution was formed, it was distilled and degassed. Then, living PSLi anions were added to the flask and benzene and D3 was distilled again. The resulting product was stored at -20 0 C. The final purifications that must occur before synthesis occurs are those for the initiator which begins the reaction and the termination agent which terminates it. sec-BuLi, the initiator, was received in a 1.4M hexane solution which was first introduced into a vacuum line and the hexane was removed. Then, benzene was added and the solution was stored in an ampoule before usage during polymerization. Methanol, which was used to terminate the reaction, was first dried using calcium hydride and then degassed and distilled into ampoules. After the necessary materials have been prepared, polymerization was performed. The polymerization contains of four steps: preparing the polymerization vessels, performing the PS or PI polymerization, performing the PDMS polymerization, and termination. The reaction vessel consists of a round-bottomed flask connected to a vacuum line and purge section and to which all of the necessary reagents are connected by break seals. During polymerization, pure benzene was used as a solvent. Before performing the reaction, this benzene was first mixed in the purge section with n-BuLi (which reacts with any impurities that may be present), frozen with liquid nitrogen and degassed, and then distilled into the main flask. After this step, the purge section was sealed off from the main flask and the polymerization can begin. Polymerization of both the PS-PDMS and the PI-PDMS diblocks was accomplished by first polymerizing the styrene or isoprene monomers in benzene at room temperature using sec-BuLi as an initiator. In both cases, the monomer and initiator were introduced into the flask and polymerization was allowed to occur (for 18 hours in the case of PS and 24 in the case of PI). After polymerization of the first block was complete, the polymerization of the PDMS block was performed. The procedure in both cases was identical, except that in the case of the PI-PDMS diblocks THF was first added to the reaction and mixed for 20 minutes before performing the second polymerization. D3 was polymerized to form PDMS in one of two multi-step procedures. The first step in both cases was to add D3 to the reaction vessel and allowed to react for 18 hours. Then, enough THF was added to the mixture to form a solution with equal amounts of benzene and THF. Following these two steps, one of two approaches were taken. The first was to allow a four hour 46 polymerization of D3 so that half of the monomers could react. The second was to lower the reaction temperature to -20 0 C for seven days to allow complete polymerization of the D3 . At the conclusion of the polymerization, approximately one milliliter of methanol was added to flask to terminate the reaction. Upon the conclusion of the polymerization, each polymer was analyzed using membrane osmometry (MO), gel permeation chromatography (GPC) and nuclear magnetic resonance (NMR) to determine the molecular weight and composition of the product. For GPC, a PL-GPC-120 (Polymer Laboratories) equipped with three PLgel 5m MIXED-C columns and an RI detector light source upgrade (PL-GPC 120/220) was used. The experiment was performed in THF at 35 0 C with a flow rate of 1 milliliter per minute. NMR studies were performed using a Bruker AC-250 spectrometer at 250 megahertz. The polymers were dissolved in CDCl3 and examined at 25 0 C. The following chemical shifts in the 1H-NMR spectrawere used for calculation of the volume fraction: 6.7ppm-7.7ppm for the five aromatic protons of polystyrene, 4.9ppm-5.1ppm for the one proton of the double bond in PI(1,4), 4.6ppm-4.7ppm for the two protons of the double bond in PI(3,4) and 0.3ppm-0.6ppm for the six protons of PDMS. 2.3 Experimental procedures The diblocks studied in this chapter were all prepared in a uniform way in order to ensure consistency in final morphology assignment (see Chapter 3 for more information about how processing conditions can affect observed morphology). The procedure used here is similar to those described in [64] and [65], making comparison of results with these works the most clear-cut. In each case, approximately 0.1 gram of each diblock was dissolved in toluene to form a 5 wt% solution. These solutions were poured into ceramic crucibles and between one and ten crucibles were placed symmetrically around a central crucible containing toluene inside a hood. This setup was covered with a glass dish and the solvent was allowed to slowly evaporate; total drying time was approximately seven days. Then, these samples were placed inside a vacuum oven and annealed for seven days at 150 0 C under vacuum. The samples were then slowly cooled to room temperature under vacuum by turning down the oven temperature to 230C. This final cooling step required approximately 12 hours (in some cases noted on the chart structure formation was not observable using SAXS but was if another solvent was used or annealing was not performed and data from those trials was used to make the phase diagram.) Samples were prepared for TEM examination by using a cryo-microtome to cut very thin slices which were then picked up and placed on copper grids using a sucrose solution. First, the samples were trimmed using a glass knife and then they were cut on the diamond knife (final slices taken were ~ 50 nm thick.) The sample, air, and knife temperatures were all set at -140 0 C during cutting (well below the ~ -120 0 C glass transition of PDMS, the ~ -70 0 C glass transition of PI, and the ~ 100 0 C glass transition of PS) to ensure limited sample deformation during cutting. The sample speeds were ~ 1" during trimming and during cutting. The grids onto which the slices were deposited were examined ~ 0.2" using an optical microscope to look for regions where thin pieces of sample had deposited. These grids were then imaged using a JEOL 2000FX TEM operating at 120 kV. No 47 staining was necessary due to the inherent scattering contrast between the silicon and oxygen in the PDMS backbone and the carbon in the PS and PI backbones (scattering cross-section increases with increasing Z of atomic nuclei.) Samples were prepared for SAXS examination by placing the annealed materials between two pieces of Kapton tape (which is nearly x-ray transparent at the Cu-Ka wavelength). The materials were then studied using a Rigaku SAXS instrument using a rotating anode and producing Cu-Ka x-rays with a wavelength of 1.54 A . The sample chamber was evacuated before each one hour data collection session began. Image plates were used to record the raw scattering data and Polar software (available at http://precisionworksny.homestead.com/EN/Technical-software-polar.html) was used for analysis. Silver behenate was used to calibrate the distance between the samples and the image plate each time the sample holder was removed and replaced (silver behenate has a characteristic peak at q = 1.076nm- 1 .) 2.4 Results The morphology assignments and characteristics of samples investigated are displayed in Table 2.1. These results are also shown graphically on their own and in comparison to the diagrams shown in Chapter 1 in Figure 2-1. It is clear from both of these figures that the samples investigated here have morphologies which occur at volume fractions within the ranges of those found in the literature and for the most part in the ranges predicted by SCFT. The glaring exception to this is S-64LMW. This may be explained by the fact that at its lower molecular weight (and associated lower value of XN at both room and the annealing temperatures) means that it exists in the region of the phase diagram where the phase boundaries all curve together. In this region, the phases are closer together in volume fraction and the domains are less strongly segregated; therefore, it takes less perturbation in volume fraction to cause the preferred phase to shift and other factors such as chain packing length, etc. may have more of an influence. 2.5 TEM results The TEM images for the samples investigated using this technique are shown in Figures 2-2 (PS-containing samples) and 2-3 (PI containing samples.) It is clear that the PScontaining samples studied here contain PDMS cylinders and lamellae and that the PIcontaining samples display PDMS spheres and cylinders. Each image other than that from S-64LMW contains multiple grains at different orientations to the incoming electron beam; therefore different projections of the individual morphologies can be observed. For instance, samples S-17, S-22, and S-37 contain both end-on and in-plane cylinders and S43 and S-53 both show lamellar-lamellar grain boundaries. These multiple views confirm that the morphological assignment is valid and not an artifact of projection along a special axis. 48 Identity of non-PDMS block PI Total molecular weight (kg/mol) 44.9 41.2 43.8 53.5 43.2 36.1 66.1 31.1 45.9 46.1 7.0 4.3 50.9 73.4 60.7 43.8 42.9 10.7 0.16 0.17(GPC) 0.21 0.22 0.30 0.34 0.37 0.43 (GPC) 0.53 0.56 0.64 0.45 0.12 0.19 xN at 150 0 C 91 84 95 110 95 76 130 70 110 110 14 10 70 100 S-16 S-17 S-21 S-22 S-30 S-34 S-37 S-43 S-53 S-56 S-64LMW S-45LMW 1-12 1-19 0.23 0.25 0.44 0.39 82 59 57 14 1-23 1-25 1-44 I-39LMW #PDMS Sample name SAXS d-spacing (nm) 23* 34 40 N/P 39** 37 66 39 48 47 13 10 36 N/P 38 53* 65 16 SAXS morphology TEM morphology S* C C N/P C** C N/P L L L N/P N/P N/P N/P C C* L N/P N/A PDMS C PDMS C PDMS C PDMS C L L PDMS C PDMS S PDMS S PDMS C ble 2.1: The list of samples analyzed in this thesis and their relevant characterization data. These materials were synthesiz the laboratory of Professor Apostolos Avgeropoulos at the University of Ioannina in Ionannina, Greece. The molecu eights and volume fractions of the samples were also determined there; gel permeation chromatography (GPC) was u establish the former and nuclear magnetic resonance (NMR) the latter (except where noted that the characterizati s performed using GPC.) The x-ray measurements and TEM measurements were performed at the Institute for Sold anotechnologies at the Massachusetts Institute of Technology in Cambridge, Massachusetts. - means that the releva easurement was not performed. N/P means that there were not sufficient peaks in the SAXS data to assign the value w rtainty. * Indicates that toluene was the casting solvent and no annealing treatment was performed. ** Indicates th clohexane was used as the casting solvent and the annealing treatment was that described in the text. 140 120 100 Z 80 U U 60 40 8 PDMS spheres 20 a PDMS cylinders a 0 0 a 0.2 0.8 0.6 0.4 1 PDMS DG PDMS perforated lamellae 8 Lamellae PDMS volume fraction * Other block perforated larrellae 140 Other block DG Other block cylinders - 120 = 100 U Other block spheres - - Disordered 60 40 20 0 0 0.2 0.4 0.8 0.6 1 PDMS volume fraction Figure 2-1: Updated phase diagrams (compare to figures 1-8 and 1-7) of PDMS-containing diblocks reflecting the work done in this thesis. Extended data for the samples are available in Table 2.1. The SAXS and TEM data used for this table was that taken after the sample underwent a one week cast from a 5 wt% solution and then a one week anneal at 1500C under vacuum. (a) shows only the samples characterized in this thesis; the samples that have PS as the complementary block are indicated by square markers while those that have PI as the complementary block are indicated by circular markers. (b) shows the samples characterized in this thesis plus those with values of XN between 0 and 150 investigated in other works. The samples which were studied here have larger markers. 50 igure 2-2: TEM images of the PS-PDMS samples which were characterized using this technique. The PS cylinder morphol ccurred when #PDMS < 0.37 for the strongly segregated samples and for the only low molecular weight sample. Lame ccurred when 4PDMS> 0.43. See Table 2.1 for more information. ontinued Figure 2-2: TEM images of the PS-PDMS samples which were characterized using this technique. The PD phere morphology occurred when #PDMS < 0.19. PDMS cylinders were present when 'PDMS = 0.23. Table 2.1 for m formation. . . ........................ .. gure 2-3: TEM images of the PI-PDMS samples which were characterized using this technique. See Table 2.1 for m formation. 2.6 2.6.1 SAXS results Morphology determination using SAXS As discussed in Chapter 1, SAXS is a powerful tool for determining sample morphology because it can be used to determine the characteristic spacings at which there are correlations in electron density and thus can be used to determine the space group of materials. In diblock polymers, the four ordered morphologies belong to different space groups and so SAXS can be used to distinguish between them. As stated before, scattering occurs when x-ray beams which are scattered constructively interfere (and thus is intimately dependent upon material structure) and the spacing between the planes in the sample containing the scattering centers is given by d = sr. = The ratios of d-spacings at which scattering occurs for a particular morphology can be analytically determined based upon the location of scattering centers in that morphology; the d-spacing between planes is affected by the lengths of the axes of the unit cell and the distance between planes for which constructive interference occurs depends upon the locations of scattering centers which are not at the corners of the unit cells. Therefore, the ratios between the q values at which peaks are present in the SAXS data can be used to assign morphologies to materials. The lamellar structure is the simplest ordered morphology which occurs in diblock polymers and it has periodicity along one axis. The constructive interference that results in appreciable intensity of scattered x-rays must therefore occur for q values which are perpendicular to this direction. The distance between scattering planes for any given scattering peak is thus to a first approximation the wavelength of the scattered light divided by the number of wavelengths present between lamellae. This means that the ratio of d-spacings where scattering is present should be 1:j:j:1 etc and thus the ratio of q-values should be 1:2:3:4 etc. In practice, due to the influence of the form factor, which modulates scattering intensity based upon the interference of waves emanating from different parts of one scattering center, certain reflections can be absent or strongly suppressed. Determination of the d-spacings and scattering vectors for which constructive interference occurs for more complex morphologies is aided by the use of the concept of the reciprocal lattice (much of the following discussion is adapted from course notes from Bruce Clemens' MSE 205 class at Stanford University.) The reciprocal lattice is made up of the points defined by the translation vectors a*, a*, and a* which are related to the vectors a1 , a 2 , and a3 of the real space lattice for a material by: a, a 2 xa 3 ai - (a 2 x a3 ) (2.1) a a 3 xa a 2 - (a 3 x a1) (2.2) a* = , x(2.3) a 3 - (a1 54 x a2) Each vector of the reciprocal lattice is perpendicular to the two vectors of the real lattice which do not have the same index (as can be easily seen by observing that the numerator is the cross product of these two vectors) and has a length that is equal to the reciprocal of that of the real lattice vector which shares its index (as can be seen from the cancellation of the magnitude of the cross product terms in the numerator and denominator and the presence of the real space vector in the denominator). The spacing where Ghkl is a vector which G between planes which have intercepts -k1 al a2 a3 is given by IJhk1I is perpendicular to these planes and is defined by ha*+ka*+la*. The former can be verified by observing that the diffracting planes can be defined by any two vectors connecting any of the three intercepts and noting that the dot product of these vectors with Ghkl is always zero (e.g. 1 - 2 - Ghk1 =1 - ha* - 1 - ka* = 0 (any other terms drop out due to the perpendicularity of reciprocal lattice vectors to two of the real lattice vectors.) The latter can be seen by noting that the distance between the scattering planes is given by the dot 1 ha*1 -- [GlkiI G1 I ( h product of gh and the unit normal to the planes or h -aIIGhkLI = IGhkL With this knowledge, determination of the expected ratios of d-spacings of planes from which scattering occurs for cylindrical and spherical morphologies can be performed easily. Diblock polymers with cylindrical morphology have periodicity along two axes and their morphologies can be conceived of as two-dimensional plane lattices with plane group of p6mm. The primitive unit cell for this lattice consists of a parallelogram with two 600 angles and two 1200 angles. The planes from which scattering occurs can be thought of as lines within this plane lattice. In the reciprocal lattice for this system, where i and j are the unit vectors in the x and -+ _3) and a* = 2 (ii a* = y directions for a Cartesian plane lattice. Thus, IGhk I is given by d-spacings of these planes are planes are therefore 1: ~: : : . 2___2 2 2VrhT+k -+hk 2,h/ k2 +hk and so the The ratios between the d-spacings of scattering etc. and the ratios of the q values of these spacings are 1:x/5:2:F7 etc. Diblocks which have a morphology consisting of spheres occupying a BCC lattice have a reciprocal lattice given by a* =t2, ai = j, and a* = f. Therefore, |GhktI is given +1 In this case, scattering does by rh2 + k 2 + 12 and d-spacings are given by not occur from any arbitrary plane because unlike in the previously discussed cases the real lattice for this morphology is non-primitive (it contains more than one lattice point) and scattering from an object (PDMS sphere) on this additional lattice point can cause destructive interference of x-rays. Determination of the planes from which constructively interfered waves are scattered requires a more advanced treatment, the results of which indicate that allowed planes are those for which h + k + 1 is even. Therefore, the ratio of d-spacings at which scattering occurs in this lattice is given by 1: : : etc at q values of 1:V2:V25:2 etc. Expected scattering from samples with a double gyroid morphology can be determined in a similar way but is not discussed here due to the lack of such samples in this study. With this knowledge, the morphology of a sample containing more than one peak in its SAXS pattern can be easily determined by comparing its scattering with that of the known diblock morphologies. It most likely has the morphology whose predicted scattering intensity it most closely matches. 55 It is important to note here that disordered morphologies can also give rise to correlation hole scattering due to the relatively higher frequency of intra-chain contacts as opposed to interchain contacts which results in variations in electron density at a characteristic distance. 2.6.2 Data Figures 2-4 and 2-5 show small angle scattering from the PDMS-containing diblocks studied here. It is encouraging to see that SAXS patterns with more than one peak predict the same morphology as that displayed in their TEM images. 2.7 2.7.1 Discussion PS-containing diblocks Using a combination of TEM and SAXS, it was possible to assign morphologies to eleven samples, the results of which are summarized in Table 2.1. There was one sample which was found to display the PDMS sphere morphology (S-16), which was assigned on the basis of its volume fraction (16% PDMS) and SAXS pattern. The SAXS pattern for this sample shows three distinct peaks with a ratio in values of q of approximately i:/2:v'/5, which is consistent with a spherical morphology and not a cylindrical morphology. The d-spacing was determined to be 23 nm using the SAXS data. Six high molecular weight samples samples show the PDMS cylinder morphology (S17, S-21,S-22, S-30, S-34, and S-37). with volume fractions ranging between 17% and 37% PDMS. TEM images were used to establish the morphologies of S-17, S-22, S-34, and S37; for all of these samples except for S-22, SAXS is able to provide corroboration for the assignment. The TEM images show either end on cylinders (which appear as circles) or cylinders in the plane of the section (which appear as lines.) The morphologies S-21 and S-30 were determined solely using SAXS; both sample, however, have three peaks which rules out the possibility of a lamellar structure and makes the presence other structures highly unlikely. For most samples, strong {110} and {210} peaks are observed in addition to the primary peak while the {200} peak is strongly suppressed. S-17 is an exception where the {200} peak is displayed but not the {210} peak. For sample S-37, only the {210} peak is observed (explained below.) D-spacings for these samples with the exception of S-37 were determined by assuming that the first peak in the diffraction pattern was due to the {100} peak and using the relationship d = '.q Since S-37 has a large molecular weight (66.1 ±9)) it is not unreasonable to assume that the first peak would be inside the beam stop. For this to be the case, the peak that is displayed could be no lower index than {210}. The d-spacings for all of the other samples cluster together, which makes sense because their molecular weights and segregation strengths at the annealing temperature are similar. If we assume that the domain length scales with the molecular weight of the sample to the two thirds power, it is not unreasonable to estimate that the peak present in S-37 is from the {210} planes: we can look at the ratio of (')3 = 1.5 and estimate that the d-spacing for S-37 should be approximately 55 nm. This is not too 56 PS-containing samples with cylindrical morphology PS-containing sample with spherical morphology I I II II 210 100 S-37 100 y.v 110 210 S-34 100 110 111 110 210 S-16 Sj S-30 i ii. , 1 ; 4P",Y( ,,iN,14V1 *1 i I. S-22 100 I c 110 210 S-21 100 110 200 /1 A S-17 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 Figure 2-4: SAXS data from the PS-PDMS samples which are studied in this thesis. See Table 2.1 for more information PS-containing samples with lamellar morphology f I I I I I 100 200 300 S-56 100 300 i I. S-53 100 V i 300 S-43 ', Y 0 0.1 0.2 7 V 0.3 0.4 q (1/nm) 0.5 0.6 0.7 ontinued Figure 2-4: SAXS data from the PS-PDMS samples which are studied in this thesis. See Table 2.1 for m formation. Pi-containing samples with spherical morphology Pl-containing samples with cylindrical morphology 100 110 210 1-19 1-25 100 100 i 110 I. E C 1-12 210 300 .2 1-23 1 h~ v V A.1 4 'If 0 [1 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 I i I 0.1 0.2 0.3 I 0.4 q (1/nm) I I 0.5 0.6 I 0.7 Figure 2-5: SAXS data from the PI-PDMS samples which are studied in this thesis. See Table 2.1 for more information PI-containing sample with lamellar morphology I 0 0.1 i i 0.2 0.3 i 0.4 q (1/nm) | | 0.5 0.6 0.7 ontinued Figure 2-5: SAXS data from the PI-PDMS samples which are studied in this thesis. See Table 2.1 for m formation. Room temperature SAXS for processed LMW samples 1-39LMW S-45LMi I S-64LMV\ '~ IV A 4t 0 0.2 0.4 0.6 A A 0.8 1 q (1/nm) gure 2-6: SAXS data from all of the low molecular weight samples investigated in this thesis. See chapter 4 for m formation. far from the value of 66 which is found when the diffraction peak present arises from the {210} planes. One low molecular weight sample, S-64LMW, also displays the PDMS cylinder morphology, as can be seen by an inspection of its TEM images. The results from SAXS, which displayed only one peak, were inconclusive. Note that the cylinders for this sample are much smaller than those present in the larger molecular weight samples, as would be expected. In order to observe the cylinders at all, a higher magnification is needed than was typically used for the other samples; to see them well, a large amount of additional power is needed. An inspection of the image with higher magnification shows that the cylinders have a large volume fraction compared to the matrix, making the morphology not unreasonable despite its odd placement on the phase diagram. S-43, S-53, and S-56 all display the lamellar morphology. TEM images were obtained of S-43 and S-45, which both display dark stripes (PDMS) on a lighter background. All three samples have SAXS patterns which are consistent with cylinders: all samples display both the {100} and {300} peaks while S-56 also shows a {200} peak. As the molecular weight decreases from 46 -kmot of the for S-56 to 31.1 ISmot for S-43, the q vector {100} peak moves outwards, as expected. This trend is also consistent with the expected scaling as a function of molecular weight: the weight ratio (}6j)2 = 1.3 while the d-spacing ratio -= 1.2. 39 2.7.2 PI containing diblocks The morphologies for six PI-containing diblocks: 1-12, 1-19, 1-23, 1-25, and 1-44 were also established using the procedures described above. These samples spanned a range of volume fraction from 12% to 44% PDMS and displayed morphologies of PDMS spheres and cylinders, and lamellae. 1-12 and 1-19 are identified as displaying a morphology consisting of PDMS spheres in a PI matrix; their TEM images show dark spheres on a lighter background. 1-12 displayed one peak in the SAXS pattern, while 1-19 displayed zero. This is probably due to a lack of long range order in the samples. 1-23 and 1-25 both show a PDMS cylinder morphology. For 1-23, this is established both by TEM image which shows both end-on and in plane dark cylinders in a lighter matrix and the SAXS data which consists of four well-resolved peaks corresponding to the {100}, {110}. {210}, and {300} diffracting planes. 1-25 can also be assigned a cylindrical morphology because of its SAXS data: the three {100}, {110}, and {210} peaks are characteristic of hexagonally packed cylinders. It is surprising that 1-25 has the larger d-spacing despite it's lower molecular weight (53 nm vs 58 nm); however, this may be because the sample was unannealed and thus a solvent-swollen state may have been preserved. No order was displayed in the annealed sample, but perhaps it would have had a smaller d-spacing if extended annealing were performed. 1-44 is the only PI-containing sample examined which displays a lamellar microstructure. This can be seen by examination of the SAXS pattern, which shows the {100}, {200}, and {300} peaks. The d-spacing for this sample is 65 nm. 62 Chapter 3 Effect of Processing on Morphologies of PS-PDMS and PI-PDMS Diblocks 3.1 Introduction Chapter 2 discussed the structures of samples which had undergone a very specific processing treatment. However, knowledge of the true equilibrium morphologies of these materials is impossible to obtain because it is impossible to determine when samples have achieved equilibrium. To provide some insight into whether or not the morphologies observed above occur after many different treatments (providing strong evidence for their equilibrium nature or at least strong metastability) and to investigate the possibility of using different processing treatments to tune ultimate structure, different casting procedures were studied with and without annealing treatments. In this chapter, only the higher molecular weight diblocks were examined; the effect of temperature treatments and chain mobility of the lower molecular weight diblocks is discussed in 4. SAXS was used to investigate the morphologies observed in each case. Examination of Table 3.1 suggests that the solvent used during casting has the potential to have a profound effect on the final morphology. It is clear that PDMS has a lower solubility parameter than either PS or PI and so solvents which are good for PDMS are less for PS and PI and vice versa. When diblock polymers are dissolved in a solvent, their equilibrium structure depends on the solvent concentration: solvation dilutes contacts between unlike species and so lowers the effective x parameter while also increasing the size of the chains as they pervade more volume. Of course, at low block copolymer concentration the system is a homogeneous solution. With decreasing solvent concentration, a disorder to order transition occurs where the block copolymer forms a swollen periodic microdomain structure. If a particular solvent preferentially segregates to the domains of one of the blocks, then that block pervades a larger volume than it does in the neat block copolymer solid. Additionally, the presence of a solvent introduces mobility into the structure and thus lowers the glass transition of the blocks. Combining these two effects, it is expected that as the solvent evaporates the structure observed after drying is complete would be that which was present when one of the blocks experienced its glass transition (i.e. the glass transition temperature reaches room temperature) such that chain mobility 63 Material PDMS PS PI Toluene Cyclohexane Methyl cyclohexane Octane Solubility parameter ( 15.7 18.3 16.7 18.2 16.8 16.0 15.6 ) Block preference PS/PI neutral/PI PDMS PDMS Table 3.1: Solubility parameters for the polymers and solvents studied in this thesis. All values are taken from [75]. becomes appreciably slowed. If the solvent is preferentially swelling one of the blocks during this "freezing in" of the structure, it is possible that its effective volume fraction may be larger; for materials with volume fractions near morphology boundaries, there exists the possibility of locking in a non-equilibrium phase during casting. Annealing raises the temperature to above the glass transition of all polymers studied here and increases chain mobility, making structural rearrangement possible and thus increasing the probability of achieving the morphology which is at equilibrium for the conditions present during this treatment. However, for strongly segregated and well entangled polymers, the kinetics of rearrangement may be slow enough to require very long annealing times. 3.2 Sample preparation To examine the influence of casting solvent (and thus relative swelling of the PDMS block during casting) and subsequent annealing treatment on PDMS-containing diblocks, each sample was cast from three (for PS-containing materials) or four (for PI-containing materials) solvents. These samples were then cut in half and half was examined directly using SAXS and the other half was annealed. Sample casting, annealing, and SAXS preparation procedures were identical to those described in Chapter 2 except for the solvent used during casting. Similarly, SAXS measurements were conducted in an identical way for consistency. 3.3 Results Tables 3.2 and 3.3 describe the morphologies observed after casting from different solvents before and after annealing. Figures 3-2 and 3-1 show the SAXS data. Data is only shown for samples which were cast from multiple solvents. In many cases there is no effect on morphology (which is encouraging when assigning an appropriate equilibrium structure); the absence of higher order peaks indicate that the morphology is less ordered after some treatments compared to others but does not indicate structural change. Larger effects 64 are found in the d-spacings of the materials studied; in most cases it appears that the as cast d-spacing can very greatly but that the final d-spacing after annealing is relatively consistent independent of initial casting results. The latter shows that the annealing treatment employed is appropriate for allowing structural relaxation and rearrangement to a morphology which is more stable than any initial one. The former is possibly due to nonequilibrium swelling which results in domains or matrices under tension or compression after the solvent has been removed but which can rearrange when mobility but not solvent is reintroduced. Thus, it is fair to say that the protocol used in Chapter 2 is appropriate for establishing the equilibrium bulk morphology of the PDMS-containing diblocks studied and that using casting solvent to tune morphology may require different samples (possibly those having longer chains, which would order at higher solvent concentrations magnifying differential swelling effects and have reduced mobility both above and below the glass transition temperature.) 3.4 3.4.1 Discussion PS-containing samples Nine PS-containing samples are investigated in this chapter, three of which show morphologies which can be tuned by changing the processing treatment. These three samples do not display many similarities; they all have molecular weights which are in the middle of the range studied in this thesis (so they may be large enough to have sufficient entanglements to promote stable morphology formation after solvent evaporation but short enough to rearrange in kinetically accessible time frames during processing.). All of the samples investigated here do have two things in common: they only show long range order under certain conditions (which are unique to each sample), and that each sample mostly displays consistent d-spacings after annealing regardless of the initial solvent treatment. S-16 has two scattering patterns which contain enough information to assign for a morphology to be assigned to the relevant processing treatment: casting without annealing from toluene and casting without annealing from cyclohexane. In the first case, three peaks which have ratios of the scattering vector of 1:v12:/ are shown and in the second two peaks which have ratios of the scattering vector of 1:/5 are displayed. The scattering that occurs from the toluene cast sample fairly unambiguously indicates a spherical morphology; this is the only morphology that exists in this volume fraction range which has a ratio of the first two peaks of 1:V/ and the next peak q value is V3 times larger than the initial value, which is also consistent with a spherical morphology. The S-16 sample which was cast from cyclohexane only displays two peaks which have a ratio of 1:v5. In this case, it is possible that the v5 peak is suppressed or that the sample displays a cylindrical morphology; either case indicates a change in morphology, as the intensity of the higher order scattering peaks depends on the volume fraction of the minority block. Therefore, the PS matrix becomes less swollen as a more neutral solvent is used and either smaller spheres or cylinders are formed. There is a large change (~ 25%) in the d-spacing when cyclohexane is used during casting instead of toluene, indicating that a significant morphology shift occurs. 65 Sample name S-16 S-17 S-22 S-30 S-34 S-37 S-43 S-53 S-56 1-12 1-19 1-23 1-25 Molecular weight (kg/mol) 44.9 41.2 53.5 43.2 36.1 66.1 31.1 45.9 46.1 50.9 73.4 60.7 43.8 4PDMS Toluene as cast (annealed) 0.16 0.17 0.22 0.3 0.34 0.37 0.43 0.53 0.56 0.12 0.19 0.23 0.25 S (N/P) C (C) N/P (N/P) N/P (N/P) C (C) C (N/P) N/P (L) C (L) L (L) N/P (N/P) N/P (N/P) N/P (C) C (N/P) Morphologies as cast from: Methyl cyclohexane Cyclohexane as cast (annealed) as cast (annealed) S/C (N/P) C (C) N/P (N/P) N/P (C) L (C) C (C) N/P (L) N/P (L) N/P (L) N/P (N/P) N/P (N/P) N/P (N/P) N/P (C) N/P (N/P) N/P(C) C (C) N/P (N/P) N/P (C) N/P (N/P) N/P (L) N/P (L) N/P (L) N/P (N/P) N/P (N/P) N/P (N/P) C (C) Octane as cast (annea - (-) (-) (-) (-) (-) - (-) - (-) - (-) - (-) N/P (N/P S (S) C (C) C (-) able 3.2: Summary of sample morphologies obtained by SAXS after different casting and annealing treatments for mples whose morphology was discussed in Chapter 2. The diblocks were cast from the relevant solvent for one week at ro mperature; half of each sample was then annealed for one week at 150 0 C. Morphologies were determined by the relativ lues of the positions of the peaks in the scattering spectrum, which are characteristic of different morphologies. S indica spherical morphology, C indicates a cylindrical morphology, and L indicates a lamellar morphology. - indicates that periment was not performed and N/P indicates that there were insufficient peaks in the SAXS data to assign a morpholo #PDMS S-16 S-17 S-22 S-30 S-34 S-37 Molecular weight (kg/mol) 44.9 41.2 53.5 43.2 36.1 66.1 0.16 0.17 0.22 0.3 0.34 0.37 Toluene as cast (annealed) 23 (N/P) 31 (34) 28 (N/P) N/P (N/P) 35 (37) 43 (66) S-43 31.1 0.43 N/P (39) S-53 S-56 1-12 1-19 1-23 1-25 45.9 46.1 50.9 73.4 60.7 43.8 0.53 0.56 0.12 0.19 0.23 0.25 41 (48) 44 (47) 31 (36) 38 (N/P) N/P (38) 53 (56) ample name d spacings (nm) as cast from: Methyl cyclohexane Cyclohexane as cast (annealed) as cast (annealed) 29 (N/P) 28 (34) N/P (30) N/P (39) 30 (38) 31 (47) N/P (N/P) N/P(35) 36 (36) 31 (39) N/P (38) N/P (50) 31 (38) 30 (39) 42 38 37 N/P N/P 51 (46) (44) (34) (N/P) (N/P) (55) 34 36 31 N/P N/P 52 (45) (49) (36) (N/P) (N/P) (55) Octane as cast (annea - (-) (-) (-) (-) (-) (-) - (-) - (-) - (-) 35 (35) N/P (N/P 47 (47) 52 (-) ble 3.3: Summary of d-spacing values obtained by SAXS after different casting and annealing treatments for the samp ose morphology was discussed in Chapter 2. The diblocks were cast from the relevant solvent for one week at ro mperature; half of each sample was then annealed for one week at 150 0 C. D-spacings are obtained from the relationsh = 2. In some cases, the ratios of q at which the main peaks appear indicate a morphology change, so the changes in pe acings are attributable to changes in both the sizes of the domains and the distances between the first planes from wh attering appears. See Table 3.2 for more information about the latter. - indicates that the experiment was not perform d N/P indicates that there were insufficient peaks in the SAXS data to assign a morphology. S-17 scattering as a function of processing conditions S-16 scattering as a function of processing conditions 100 Methyl cyclohexane; annealed 110 Methyl cyclohexane; annealed Methyl cyclohexane; as cast Methyl cyclohexane; as cast 100 Cyclohexane; annealed 110 Cyclohexane; annealed c 100 I. Cyclohexane; as cast 110 100 110 Cyclohexane; as cast c 100 110 200 Toluene; annealed Toluene; annealed 100 100 110 110 111 Toluene; as cast Toluene; as cast 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 igure 3-1: Scattering from PS-containing samples studied in this thesis after different processing treatments. The diblo ere cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one wee 50 0 C. See Tables 3.2 and 3.3 for more information. .... S-22 scattering as a function of processing conditions I I I S-30 scattering as a function of processing conditions I I 100 Methyl cyclohexane; annealed 110 200 Methyl cyclohexane; annealed Methyl cyclohexane; as cast 100 100 110 110 Methyl cyclohexane; as cast 210 Cyclohexane; annealed i $ I Cyclohexene; annealed Cyclohexane; as cast C Cyclohexane; as cast Toluene; annealed Toluene; annealed Toluene; as cast Toluene; as cast - 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 A V'K 0.6 0.7 ontinued Figure 3-1: Scattering from PS-containing samples studied in this thesis after different processing treatments. blocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for eek at 150 0 C. See Tables 3.2 and 3.3 for more information. S-37 scattering as a function of processing conditions S-34 scattering as a function of processing conditions 100 110 210 Methyl cyclohexane; annealed Methyl cyclohexane; annealed Methyl cyclohexane; as cast Methyl cyclohexane; as cast 100 100 110 110 ii 210 210 *1 1 Cyclohexane; annealed Cyclohexane; annealed I 100 200 100 110 Cyclohexane; as cast Cyclohexane; as cast 100 210 110 210 Toluene; annealed Toluene; annealed 100 100 110 110 \1 Toluene; as cast Toluene; as cast ..Ab III 0 N1 - I 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 ontinued Figure 3-1: Scattering from PS-containing samples studied in this thesis after different processing treatments. iblocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for eek at 150 0 C. See Tables 3.2 and 3.3 for more information. S-53 scattering as a function of processing conditions S-43 scattering as a function of processing conditions 100 100 300 300 Methyl cyclohexane; annealed Methyl cyclohexane; annealed 100 Methyl cyclohexane; as cast Methyl cyclohexane; as cast 100 300 i 300 Cyclohexane; annealed Cyclohexane; annealed I I I I Cyclohexane; as cast Cyclohexane; as cast 100 100 300 300 Toluene; annealed Toluene; annealed 100 110 200 210 Toluene; as cast Toluene; as cast 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 ontinued Figure 3-1: Scattering from PS-containing samples studied in this thesis after different processing treatments. T blocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for eek at 150 0 C. See Tables 3.2 and 3.3 for more information. S-56 scattering as a function of processing conditions 100 400 200 300 Methyl cyclohexane; annealed Methyl cyclohexane; as cast 100 200 i 300 I. I Cyclohexane; annealed Cyclohexans; as cast 100 300 200 Toluene; annealed 100 200 Toluene; as cast 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 ontinued Figure 3-1: Scattering from PS-containing samples studied in this thesis after different processing treatments. blocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for eek at 150 0 C. See Tables 3.2 and 3.3 for more information. 1-19scattering as a function of processing conditions 1-12 scattering as a function of processing conditions I II Octane; annealed Octane; annealed Octane; as cast Octane; as cast Methyl cyclohexane; annealed Methyl cyclohexane; annealed Methyl cyclohexane; as cast I Cyclohexane; annealed IL - 0 I aI II 0.1 0.2 0.3 ii. Methyl cyclohexane; as cast I. I Cyclohexane; annealed Cyclohexane; as cast Cyclohexane; as cast Toluene; annealed Toluene; annealed Toluene; as cast Toluene; as cast I 0.4 q (1/nm) I I I 0.5 - - ai 0.6 0.7 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 igure 3-2: Scattering from PI-containing samples studied in this thesis after different processing treatments. The diblo ere cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for one wee 50 0 C. See Tables 3.2 and 3.3 for more information. 1-25 scattering as a function of processing conditions 1-23 scattering as a function of processing conditions 100 100 110 210 Octane; annealed 100 Octane; as cast 100 110 210 Octane; as cast Methyl cyclohexane; annealed 100 Methyl cyclohexane; annealed 210 Methyl cyclohexane; as cast 100 Methyl cyclohexane; as cast *1 Cyclohexane; annealed I. I 210 Cyclohexane; annealed Cyclohexane; as cast Cyclohexane; as cast 100 110 Toluene; annealed 210 Toluene; annealed 100 110 210 Toluene; as cast Toluene; as cast t I 0 0.1 0.2 0.3 0.4 q (1/nm) 0.5 0.6 0.7 0 I 0.1 I 0.2 j 0.3 I 0.4 I 0.5 I I 0.6 0.7 q (1/nm) ontinued Figure 3-2: Scattering from PI-containing samples studied in this thesis after different processing treatments. blocks were cast from the relevant solvent for one week at room temperature; half of each sample was then annealed for eek at 150 0 C. See Tables 3.2 and 3.3 for more information. S-17 displayed sufficient SAXS data to assign a morphology in five different processing cases: casting from toluene with and without annealing, casting from cyclohexane with and without annealing, and casting from methyl cyclohexane with annealing. In all of these cases, there is remarkable similarity between the SAXS data obtained: all samples have two peaks with a scattering vector ratio of 1:vr5, and the sample cast from toluene and then annealed displays an additional peak at a q value of twice the principal peak. This is strong evidence for the microdomain morphology of PDMS cylinders inside of a PS matrix. This is undoubtedly the case for the sample which was cast from toluene and then followed by an anneal because of the additional peak. The other scattering patterns could arise from a cylindrical morphology displaying only two peaks, but this is unlikely because the q values at which the peaks occur are similar across samples and because a spherical microdomain morphology would be most favored after a toluene casting (where the scattering data unambiguously points to a cylindrical structure.) The d-spacings present after annealing are almost identical (34 nm in two cases and 35 nm in the third) and fairly similar before annealing (31 nm and 28 nm); however, there is a significant increase in the d-spacing after annealing. This indicates that one of the domains (probably PDMS) is swollen after the casting treatment but that this artifact can be eliminated by appropriate annealing. S-22 only exhibits ordering after casting from methyl cyclohexane; the morphology that forms in this case is stable with respect to annealing. In both cases, the first two peaks show a ratio of 1:v/5, providing strong evidence for a cylindrical morphology at this volume fraction and segregation strength. The sample which had also been annealed shows a third peak which is twice as large as the principal scattering vector, providing even stronger evidence for this morphology assignment. With the exception of the third peak, both scattering patterns look qualitatively similar which is confirmed by their identical d spacings (36 nm.) The other two samples which display a prominent peak are those which have been cast from cyclohexane and annealed and cast from toluene with no further treatment. Both of these samples have SAXS patterns which look qualitatively similar except that the large broad peak is shifted to a lower q value for the toluene cast sample. It is difficult to conclusively assign d-spacings to these patterns, as assuming that these peaks arise from the {100} peak gives relatively small d-spacings (30 nm and 28 nm) while assuming that they are from the {110} peak of a cylindrical sample gives the large d-spacings of 52 nm and 48 nm (it is unlikely that this would display a different morphology based upon its volume fraction and the morphologies of the nearby samples on the phase diagram.) The former seems more reasonable because the values are closer to those from the methyl cyclohexane casting conditions, but in this case it cannot be determined unambiguously. S-30 has one sample with a SAXS pattern which contains enough peaks to definitely assign it a morphology. This is the sample which was cast from cyclohexane and then annealed. It contains three peaks with q values which have a ratio of 1:V35:v/7, indicating a cylindrical morphology with a d-spacing of 39 nm. After casting from methyl cyclohexane and then performing an annealing treatment, only one prominent peak is displayed, but it does have the same d-spacing as the prior sample, indicating that the morphologies are probably similar (although the latter has much less long range order.) The sample which 75 was cast from methyl cyclohexane but left unnanealed also contained only one prominent peak, the d-spacing in this case is the substantially smaller 31 nm. Again, it appears that methyl cyclohexane swells one of the domains (probably PDMS) during casting but that this effect is erased upon annealing. S-34 is one of the three samples that does exhibit a change in morphology with processing treatment. While in four cases the morphology displayed is unambiguously cylinders, S-34 has a lamellar morphology when cast from cyclohexane and left unannealed. This can be seen by the ratio in the q values of the two scattering peaks present; it is 1:2. While it is possible that this is indicative of a cylindrical morphology with the v5 peak suppressed, the shape and locations of the peaks that are present are dramatically different from the other cases displaying cylinders (including those from this same sample after annealing.) The other four cases (annealed and not after casting from toluene, annealed after casting from cyclohexane, annealed after casting from methyl cyclohexane) all have the ratio of 1:v/5 for the first two scattering peaks, indicative of a cylindrical morphology. All of these except for the unannealed sample cast from toluene also have a v? peak, strengthening the validity of this assignment. Once again the d-spacings are all quite similar after annealing (37-38 nm) and the unannealed toluene cast sample has a d-spacing which is only slightly smaller (35 nm). The sample which was cast from cyclohexane and not annealed has a much smaller d-spacing of 30 nm, indicating that the morphology present here is substantially different from that observed in the other cases. It is unexpected that the neutral solvents induces a different morphology formation from the two solvents which are preferential for opposite blocks; to induce a transition from PDMS cylinders to lamellae, the PDMS would have to be swollen which would be more likely to occur in methyl cyclohexane than in cyclohexane. This may be due to differences in volatility between the solvents: if the methyl cyclohexane evaporated more slowly, there would be have been more time for the chains to rearrange while there was still solvent in the system and thus a metastable structure would be more challenging to lock in. S-37 is another sample whose SAXS patterns change greatly when the processing treatment is varied. Although all of the indexable patterns are consistent with a cylindrical morphology, even a quick glance at the data shows that the peak positions and shapes are inconsistent across differing preparation procedures. Four morphologies can be assigned without ambiguity: three from the SAXS data and one from the TEM data (see chapter 2 for more information about indexing the SAXS data from S-37 which has been cast from toluene and subsequently annealed). The unannealed toluene cast sample and the unannealed cyclohexane cast sample both have two peaks in their scattering patterns with a q value ratio of 1:v/5. When the cyclohexane cast sample is annealed, the V7_ peak develops, providing confirmation of the cylindrical morphology assignment. The two samples which were cast from methyl cyclohexane have insufficient peaks for morphology determination. The d-spacings for this sample vary widely (more than 100%!), showing that the morphology is not very robust to processing treatment. However, across all cases the d-spacing does increase with annealing. In contrast to S-37, S-43 displays remarkably similar SAXS data across the processing conditions examined here. In all cases, the unannealed samples are insufficiently ordered to provide useful SAXS data for morphology determination but the annealed samples 76 are indicative of a lamellar morphology with two peaks with a q value ratio of 1:3 and a d-spacing of 38-39 nm. The samples which were cast from cyclohexane and methyl cyclohexane but were not annealed also are remarkably similar: they have a single peak in their diffraction pattern and d-spacings of 31 and 30 nm, respectively. Here, again, annealing promotes the formation of larger periodicities. S-53, however, is another sample which displays differing morphologies depending upon the processing treatment. Lamellae are present after annealing for every sample, but cylinders are present if the sample is cast from toluene but not annealed. The lamellae display only the 1 and the 3 peak and have d-spacings ranging between 45 and 48 nm. The sample with the cylindrical morphology has four peaks whose scattering vectors have the ratio 1:v/5:2:,F and a d-spacing of 41 nm. The sample cast from cyclohexane but not annealed has a similar d-spacing of 42 nm (making it possible that it also shows a cylindrical morphology), while that cast from methyl cyclohexane and left unannealed has the substantially smaller d-spacing of 34 nm. It is likely that the cylinders which are present are made up of PDMS because the cylinders are present when the solvent is neutral or PS-preferential, indicating that PS is being swollen. As S-53 is heavier than S-43, it is possible that it is easier to trap S-53 in a metastable state and explaining the presence of a trapped morphology for S-53 but not S-43. S-56 unambiguously shows a lamellar morphology for all processing treatments except casting from cyclohexane without subsequent annealing and casting from methyl cyclohexane without subsequent annealing. This is given by the ratios of the scattering peaks, which are 1:2 in the toluene cast but annealed sample, 1:2:3 in the toluene cast and annealed sample and the cyclohexane cast and annealed sample, and 1:2:3:4 in the methyl cyclohexane cast and annealed sample. In all cases, annealing increases the long range order in S-56, which is reflected in the increased presence of higher order peaks. The d-spacings after annealing are also similar, ranging from 44 nm to 49 nm and increasing from the unannealed value in all cases. The unannealed samples have d-spacings of 44 nm, 38 nm, and 36 nm for the samples cast from toluene, cyclohexane, and methyl cyclohexane, respectively. The decreasing d-spacing with increasing preference of the solvent for the PDMS block means that the smaller d-spacings arise when the PS undergoes a glass transition with more solvent in the PDMS block. The PDMS block can then contract farther as the solvent evaporates, as it is rubbery throughout the casting and annealing processes. 3.4.2 PI-containing samples Four PI-containing samples are discussed here. It is surprising that any of these samples should display sensitivity to casting solvent and annealing: both PI and PDMS have glass transitions substantially below room temperature so it is expected that any rearrangement from a metastable to a stable morphology should occur at room temperature. However, while no morphology transitions are observed, it is clear that certain processing conditions do a much better job at promoting long range order than others. These are discussed below. For no processing case studied here does 1-12 exhibit sufficient long range order to have 77 a SAXS pattern from which a morphology can be inferred. In all cases, the d-spacings observed from the one scattering peak are relatively similar (ranging between 31 and 37 nm), with an even closer correspondence between the peaks coming from samples which have been annealed (ranging between 34 and 36 nm.) In the case of 1-19, processing using octane promotes the formation of spheres. The ratio between the peak spacings (1.75) is not consistent with spherical spacing, but is consistent with scattering due to poorly ordered monodisperse spherical domains [76]. In this case, scattering is expected when qR=5.76 and qR=9.10, which have a ratio of 1.58. This information can also be used to compute the sphere radius, which is 28.8 nm. Measuring the spheres in chapter 2, a value of - 40 nm is obtained for the diameter. This means that either substantial deformation occurred during microtoming while the toluene cast and annealed sample was prepared for TEM examination or that the casting solvent has a profound effect on ultimate d spacing. 1-23 is another sample which only displays long range order under certain processing conditions. When cast from toluene and subsequently annealed, it displays good long range order with four peaks in its SAXS pattern with a q ratio of 1:V'5:V7:3, indicating that the sample has a cylindrical microdomain structure. It also displays appreciable long range order after casting from octane (with and without subsequent annealing), showing peaks with a q ratio of 1:V35. Since the cylindrical morphology appears both when the solvent is preferential to PI and when it is preferential to PDMS, it is clear that cylindrical microdomains comprise the equilibrium morphology for this material. The d-spacings are quite different depending on the solvent used (38 nm for toluene vs 47 for octane); octane should swell the cylinders but since both blocks should be mobile it is unclear why there is this large discrepancy. 1-25 also displays a cylindrical morphology for several different processing conditions. It has sufficient peaks in its SAXS pattern to be interpretable for all cases except for that cast from toluene and subsequently annealed and that cast from cyclohexane and not subsequently annealed. The other six cases contain scattering patterns with both the {100} and {210} peaks, in addition to the {110} peak for the toluene cast but not annealed sample. These patterns all indicate a cylindrical morphology, meaning that is the equilibrium microdomain structure. The d-spacings here follow the same trends observed above: the d-spacings after annealing are relatively constant (55-56 nm) and slightly larger than the unannealed d-spacings (51-53 nm.) In this case, the unannealed s-spacings are also relatively close together, indicating that casting solvent has little effect on ultimate morphology in this case. 78 Chapter 4 Morphologies of low molecular weight PS-PDMS and PI-PDMS diblocks as a function of temperature 4.1 Overview The study of the morphologies of weakly segregated PDMS-containing diblocks provides a complement to investigations into the strongly segregated PDMS-containing diblock morphologies studied in chapters 2 and 3 because it allows for both insight into the morphologies that are formed in a different segregation regime and an increased understanding of factors that affect morphology development from either the disordered or metastable equilibrium structures present during thermal processing. Therefore, the structures of weakly segregated PDMS-containing diblocks at varying temperatures were investigated using SAXS at Brookhaven National Laboratory's X27C beamline. Due to the high flux of synchrotron x-ray sources, it was possible to record sufficient scattering intensity for forming an image of the sample's diffraction in less than one minute. This capability was coupled with the depression of the glass transition of the PS blocks and enhanced mobility of the PI and PDMS blocks due to the short chain lengths allowed for structural rearrangements on readily observable laboratory timescales. The following sections describe this work and the results that were obtained. 4.2 4.2.1 Experimental procedures Sample descriptions Two PS-PDMS diblocks and one PI-PDMS diblock were investigated. These materials were synthesized at the laboratory of Professor Apostolos Avgeropoulos at the University of Ioannina in Ioannina, Greece. Their chain structural characteristics are described in 4.1 and the physical properties of the component polymers are described in 4.2. The morphological characterization data for S-64LMW is given in chapters 2 and 3. The anticipated values of XN as a function of temperature are shown in 4-1. All of these 79 XN as a function of temperature 30 Annealing temperature 25 20 z Anticipated ODT 15 S-64LMW 10 1-39LMW Anticipated ODT Anticipated ODT S-45LMW 5 0 0 50 100 150 Temperature (Celsius) 200 250 300 Figure 4-1: Calculated values of XN(T) at different temperatures for the three samples studied in this section. Values of x were computed based upon the work in [74]. The annealing temperature and anticipated ODTs for each sample are also shown; anticipated ODTs are based upon the figure in [15]. materials have a composition where the volume fraction of PDMS is relatively close to 0.5 and the value of xN is close to the value at the ODT (~-10-15) at room temperature, indicating that these samples feature some of the smallest molecular weights that can be present in phase-separated PDMS-containing materials. 4.2.2 DSC procedures DSC experiments were performed on diblocks with the molecular weight and volume fractions as listed in Table 4.1. These samples were cast from toluene and annealed at 150 0 C using the protocol for sample preparation that was described in Chapter 2. During the experiments, 4-6 mg of sample were heated from room temperature to 180 0 C; held there for three minutes; cooled to -80 0 C; held there for three minutes; and then heated to room temperature. In all cases, the rate of temperature change was 50 C per minute. A reference pan also underwent this same procedure and the heat flow to it was subtracted from the heat flow to the pans containing the samples. These parameters were chosen in order to both mimic the experimental conditions present during the first set of SAXS studies 80 Sample name S-64LMW S-45LMW I-39LMW Molecular weight of non-PDMS block ( ) 4.3 2.3 6.6 Molecular weight of PDMS block (-)k 2.1 1.9 4.2 PDMS 0.64 0.45 0.39 Estimated xN at ODT 13 11 12 Estimated ODT (EC) 180 85 220 able 4.1: The composition and anticipated ODT's of the low molecular weight diblocks studied in this section. The molecu eights were determined by GPC and the volume fractions were determined by NMR. Estimated values of XN at the OD ere estimated based on the figures in [15] and the value of X was determined from the equations presented in [74]. Polymer Entanglement molecular PS PI PDMS weight (kg/mol) 13.3 5.4 12.3 tran Radius of gyration normalized by the 0Glass square root of sample molecular weight A) 0.659 (PS-PDMS1 = 43 A) (PS-PDMS2 = 32 A) 0.971 (PI-PDMS1 = 64 A) 0.676 (PS-PDMS1 = 31 A) (PS-PDMS2 = 30 A) (PI-PDMS1 = 44 A) temperatur 100 -70 -123 able 4.2: Physical properties of the chains that form the diblocks investigated in this thesis. The entanglement molecu eights and radii of gyration are taken from [77]; note that these assume infinitely long chains and so may overestimate dii of gyration. The first number given for radius of gyration is the ratio of the radius of gyration to the square root of sam olecular weight; the latter are the calculated values for the relevant samples. The glass transition data is taken from [ e measured values were performed on longer chains and so may overestimate the values of Tg. described below (heating up to 180 0 C), to provide information about transitions during both heating and cooling, and to provide data on possible low temperature transitions. 4.2.3 SAXS procedures Three different trips were made to Brookhaven to study the scattering of these diblocks at elevated temperatures and different experimental procedures were used based upon an evolving understanding of the morphology changes the samples underwent when subject to x-ray irradiation and temperature changes. In total, ten different experiments were performed. The different procedures are summarized below and by sample in table form in Table 4.4. The first procedure was used for S-64LMW-A, S-64LMW-B, and S-64LMW-C; S45LMW-A, S-45LMW-B, and S-45-LMW-C; and I-39LMW-A and I-39LMW-B. The second was used for S-45LMW-D and I-39LMW-C. Both procedures involved initially preparing the samples, sandwiching them between Kapton tape, and placing them inside an Instec HCS600V hotstage. The experiments were performed in air, x-ray wavelength was set at 1.371 A and silver behenate powder was used to calibrate the sample to detector distance. The first procedure involved continuously performing three minute exposures for each sample as they completed the thermal programs described in Figures 4-2, 4-3 and 4-4. Several different strategies were employed to prepare the samples before performing the dynamic temperature experiments. S-64LMW-A and S-45LMW-A were first cast from toluene, annealed at 150 0 C for one week, and then annealed at temperatures slightly above the temperature where they began to macroscopically soften but did not flow (at approximately 80 0 C and 50 0 C, respectively.) S-64LMW-B and S-45LMW-B were also cast from toluene and then annealed for one week at 150 0 C but did not receive the last pretreatment. S-64LMW-C and S-45LMW-C were taken from uncast material. I-LMW39-A was cast from toluene and annealed for one week at 150 0 C; I-LMW39-B consisted of uncast material. In all cases, the cooling from the annealing temperature to room temperature occurred in air. As can be seen in a later section, the substantially mobile short chains at room temperature allow for structural rearrangements on the time scale of hours; the highest molecular weight diblock with a glassy block (S-64LMW samples) attained identical equilibrium morphologies after different cooling programs. Possible differences in peak sharpness at the initially observed temperature may be due to beam alignment with particularly organized or disorganized regions of sample; enhanced ordering after cooling may be due to preferential alignment on the Kapton tape surface. I-39LMW-A, which has two blocks with glass transition temperature below room temperature, showed structural rearrangement on the time scale of minutes during cooling. The second procedure, described in Table 4.3, consisted of a multistep process described below for temperatures anticipated to be below the ODT. In this case, the high temperature was selected to be above the anticipated ODT based upon prior work (800C for Sample II-D and 50 0 C for Sample III-C). For temperatures above this anticipated ODT, only steps 6-9 were performed at each temperature. Each exposure was only performed for 40 seconds for these samples in order to minimize sample 83 Temperature programs for S-64LMW diblocks 200 180 S-64LMW-B S-64LMW- S-64LMW-C- 160 140 0 a) ---- 120 100 80 60 40 I 20 0 100 200 300 400 Time (minutes) 500 600 700 Figure 4-2: The temperature programs for the S-64LMW samples described in Table 4.4. The temperature profiles used for S-64LMW-A and B were selected in order to observe scattering at a wide and complementary range of temperatures and investigate whether or not the scattering was dependent upon thermal history. The temperature profile for S-64LMW-C was selected in order to observe scattering at many different and closely spaced temperatures and to detect where morphological transitions in this range occur with fine precision. 84 Temperature program for S-45LMW diblocks 120 S-45LMW-B 110 100 90 80 F 70 S-45LMW-A (D a0 E 6 6 S-45LMW-C O 50 40 30 20 0 100 200 300 400 500 Time (minutes) 600 700 800 900 Figure 4-3: The temperature programs for the S-45LMW samples described in Table 4.4. These temperature profiles were chosen in order to sample a wide range of temperatures at which morphology transitions could occur and to provide comparative data for morphology transitions upon heating and cooling. 85 Temperature program for I-39LMW diblocks 70 1-39LMW-A 60 50 -39LMW-B I40 0 - E 0 - -- 20 10 0 0 100 200 300 400 500 600 Time (minutes) 700 800 900 1000 Figure 4-4: The temperature programs for the I-39LMW samples described in Table 4.4. The two temperature profiles were selected to investigate structure in I-39LMW at temperatures both above and below room temperature and to investigate morphology transitions as a function of temperature, thermal history (i.e. heating vs. cooling), and time. 86 Step Step Step Step Step Step Step Step Step Step 1 2 3 4 5 6 7 8 9 10 Perform initial exposure at 25 0 C Heat to highest temperature(TH) Perform exposure Anneal at TH for 40 minutes Perform exposure Cool to temperature of interest Perform exposure Anneal at temperature of interest for 40 minutes Perform exposure Repeat steps 2-9 for all temperatures of interest. Table 4.3: The second procedure used in SAXS studies at BNL. exposure to the x-ray beam. 4.3 DSC results The raw DSC data is shown in Figure 4-5. By convention, endotherms are shown as negative peaks. It is apparent that both S-64LMW and S-45LMW display endothermic peaks (from approximately 500 C - 70 0 C and 30 0 C - 50 0 C, respectively) during the heating runs, indicating a phase transition. Because PS and PI should all be amorphous both above and below these temperatures, and PDMS should be amorphous in this range (it has a melting temperature of appromixately -50 0 C) it is likely that this is related to the ordering of phase separated domains or due to depressed glass transitions in the PS block due to short chain lengths which allows for increased mixing and diffusion. According to [78], the ODT should be a first-order phase transition, so it is not unexpected that it may be associated with a release of heat. The lack of a measured transition for I-39LMW may be due to the decreased interfacial energy between PI and PDMS chains compared to PS and PDMS chains so that less heat is required to break the ordered structure; the amount required may be too small to be observable. Or, it may be due to the presence of an ODT at a higher temperature than those investigated here or the rubbery PI and PDMS may be able to mix and diffuse sufficiently at room temperature so that there is no one temperature (i.e. the glass transition) above which mixing can occur but below which the structure may be relatively immobile. While endothermic peaks signifying a first-order phase transition are present, two expected features of the DSC curves are absent. These include the crystallization of PDMS at approximately -50 0 C and the glass transition of polystyrene in between room temperature and the ODT. The glass transition of polystyrene may be suppressed due to the combination of its short chain length and the ODT. Its short chain length allows for rearrangements below its nominal glass transition at any given chain length (see Section 4.4.3) and thus does not prevent the occurrance of an ODT due to limited mobility while mixing with PDMS above the ODT plasticizes the PS so that it is no longer glassy. The endothermic peak may also arise due to the glass transition as described above and 87 Prior preparation procedure S-64LMW-A Continuous x-ray exposure during x-ray data collection? Yes S-64LMW-B Yes 1) Cast from toluene for one week at 25 0 C 2) Anneal at 150 0 C for one week S-64LMW-C S-45LMW-A Yes Yes Precipitated from reactor 1) Cast from toluene for one week at 250 C 2) Anneal at 150 0 C for one week 3) Anneal at 50 0 C for one day S-45LMW-B Yes 1) Cast from toluene for one week at 25 0 C 2) Anneal at 150 0 C for one week S-45LMW-C Yes Precipitated from reactor S-45LMW-D I-39LMW-A No Yes Precipitated from reactor 1) Cast from toluene for one week at 25 0 C 2) Anneal at 150 0 C for one week I-39LMW-B Yes Precipitated from reactor I-39LMW-C No Precipitated from reactor Sample Name 1) Cast from toluene for one week at 25 0C 2) Anneal at 150 0 C for one week 3) Anneal at 80 0 C for one day Table 4.4: Summary of different sample preparation procedures used during this experiment. The initial scattering pattern displayed at 25 0 C was consistent across prior preparation methods. The thermal programs for samples receiving continuous x-ray exposure can be found in Figures 4-2, 4-3 and 4-4 while those for the samples not receiving continuous x-ray exposure can be found in Table 4.3. 88 thus mask the smaller change that would be present due to a glass transition. The crystallization of PDMS may be suppressed due to the small domain sizes (amorphous PDMS may be able to pack better into smaller areas or may have a lower surface tension with PS) or possible intermixing of PS into the domains due to weak segregation which disrupts packing. 4.4 SAXS results This section contains results obtained by following the two different protocols outlined above. The results from the second protocol will be discussed first, as they are cleaner and a better representation of the equilibrium morphologies of the diblocks investigated as a function of temperature. First, a little-known fact is demonstrated in Figure 4-6: the PDMS-containing samples studied here are quite beam-sensitive and have structures which are strongly affected by x-ray dose. In the case shown below, a sample was heated from room temperature to 90 0 C, quenched to 60 0 C, and then quenched to 250 C while being continuously exposed to x-rays (except when the shutter was closed briefly in between images.) The total time exposed to the x-ray beam was approximately one hour. Then, the sample was translated slightly such that the x-ray beam illuminated a new area of the sample and re-exposed; a completely different (and representative of the sample's expected morphology based upon initial data) scattering image was revealed. This means that the scattering measured using the first protocol was dependent on the temperature of the material when the images were taken, the thermal history of the sample, and the cumulative x-ray dose the sample had received before that point in time. Because the latter was not controlled and is less technologically relevant (because samples are almost never exposed to high intensity x-rays while undergoing processing), the data obtained using this procedure is very interesting but less useful to other scientists. It is also more difficult to compare to other experiments in this work. Clues to determining the factors responsible for the producing the differences in scattering between regions which have and have not been exposed to the x-ray beam for an extended period of time can be found in [79]; here, the structures of blends of PEOPDMS diblocks and D4 (the monomer from which PDMS is polymerized) were studied using SAXS. A figure from this paper which shows scattering data from blends of D4 and PEO-PDMS is reproduced below. A comparison with scattering data from samples S-64LMW-A and S-64LMW-B in this work (see Figure 4-8) shows a very interesting correspondence. It appears that the x-ray beam may depolymerize some of the PDMS to form D 4 ; this may be accompanied by some cross-linking which makes the old structure unrecoverable upon cooling, or the new structure formed by a D4 and diblock blend may be stable across a range of elevated and room temperature. The presence of what appears to be two superimposed scattering patterns for these samples examined using the initial protocol (see later sections for more details) might be due to the coexistence of two mixtures in thermodynamic equilibrium or partial conversion of the prior equilibrium morphology (which was locked in due to x-ray crosslinking) to the new equilibrium morphology. 89 4.4.1 Minimal radiation exposure results The SAXS results obtained for samples S-45LMW-D and I-39LMW-C are shown in Figure 4-9. It is clear that for the temperatures examined, there is little change in the structure as observed by scattering for S-45LMW-D and loss of a higher-order peak between 90 0 C and 130 0 C for I-39LMW-C. There is no indication that either sample goes through the ODT; at the ODT, the scattering peak should broaden appreciably and the intensity should dramatically decay. The decreasing intensity of the secondary peak in I-39LMW-C could indicate a transition between a strongly ordered morphology and one that is weakly ordered, meaning that the ODT is not too much higher than the temperatures investigated here. The subtle changes in peak location for S-45LMW-D and changing intensity of the secondary peak in I-39LMW-C indicate that structural rearrangements are not limited by kinetic factors in this case and that the scattering patterns are representative of the actual structures. The representative character of the scattering from samples after relatively short annealing times at the temperatures of interest is important from a technological perspective. This is because samples are often processed through several steps and allowed to rest for undetermined amounts of time before their morphology is examined. Rapid kinetic change means that treatment steps may not be responsible for the final effect that is measured and slow kinetic change means that structures that are determined cannot be reliably predicted to exist for arbitrary lengths of time. Therefore, technological applications which rely on block copolymer structures require knowledge of the predictability of morphological change at room temperature. Any structural rearrangement at laboratory temperatures, means that the sample has substantial mobility despite the fact that polystyrene at this molecular weight should be a glass at room temperature. The increased mobility could be due to the plasticization of PS by PDMS while the morphology is disordered and/or the fact that the PS chains are much shorter than their entanglement molecular weight of 13.3 1 and the PDMS chains are shorter than their entanglement molecular weight of 12.3 6, both determined in [77]. The lack of entanglements means that structural rearrangements occur much more quickly because the chains can relax due to Rouse dynamics instead of diffusing through a network of entanglements. This means that any processing steps which purposefully change the displayed morphologies of these materials must be preserved through some means other than the PS glass transition if it is desired for these non-equilibrium structures to persist over laboratory or product time scales. 4.4.2 Initial protocol heating results The last scattering patterns measured during the heating cycles of the first protocol are discussed in this section . For these experiments, all seven samples studied displayed a predictable change in peak character as the temperature increased: an initially sharp peak at room temperature broadened asymmetrically (only towards lower q with a sharp dropoff in intensity at higher q) as the temperature increased resulting in a final high-temperature broad peak at a significantly lower q than the original peak. The asymmetric broadening is characterized by peaks which are negatively skewed (the maximum in intensity occurs 90 at a higher q than the mean q of the peak); in some cases, it appears as if two peaks can be resolved in one scattering pattern (one at lower q which is growing and one at higher q which is shrinking). The low q peak may be due to a blend between depolymerized D 3 and the diblock; the D3 may swell the PDMS, increasing its effective volume fraction and the domain spacing and causing scattering at a lower value of q. The low temperature, high q peak occurs due to scattering from ordered structures in the phase separated state as described in section 1.3. Figure 4-10 shows the integrated intensity obtained from the last scattering patterns taken at selected temperatures during heating runs of samples S-64LMW-A, S-64LMW-B, and S-64LMW-C. From the data shown in this figure, it is clear that while the basic trend described above is consistent across all three samples, the actual scattering patterns vary greatly depending upon thermal history. Overall, it is possible to generalize that from approximately 80 0 C to 165 0 C, the high q scattering peak is replaced with a low q. Figure 4-8 (described initially above) shows several scattering profiles as recorded on the CCD camera for data whose character is not fully described by integrated intensity curves. All of the patterns shown display increasing texture. For S-64LMW-A, preferred orientation begins to develop at 80 0 C, increases at 105 0 C and 120 0 C, and transitions into many large 0 grains (indicated by the speckled diffraction peaks in the image) at 150 C. In all of these images, the sharp high q boundary of the high q peak and the gradual development of a low q peak as temperature is raised is evident. For sample S-64LMW-B, four-fold symmetry 0 0 begins to develop at 90 0 C which transitions to six-fold symmetry at 110 C and 125 C. At 120 0 C and 165 0 C, speckles begin to develop in sample S-64LMW-B, again indicating the presence of large and highly-ordered grains with random alignment around the xray beam. The development of a low q peak and decrease in intensity of the high q peak with increasing temperature can also be seen in these samples. S-64LMW-C, all S-45LMW samples, and all I-39LMW samples do not display the textures shown in figure 4-8 (instead displaying cylindrically symmetric rings around the beam), and so this data is only shown for S-64LMW-A and S-64LMW-B. The development of orientation and speckles may be due to the broad nature of the transition due to a gradual cumulative effect of x-ray beam dose: ordered grains transform piecewise from an ordered morphology to a disordered one; scattering vectors disappear at one orientation from the higher q peak to reappear at the lower q peak (see (d) and (f) in 4-8). The combination in S-64LMW of longer chains than those in S-45LMW and more hindered motion than I-39LMW causes slower morphology rearrangements and thus slower transitions between morphologies, allowing for observation of this phenomenon. The lack of thermal pretreatment for S-64LMW-C means that fewer large grains were formed to begin with, making this effect essentially impossible to detect. The scattering patterns of S-45LMW-A, B, and C show a similar trend with temperature as those of S-64LMW-A, B, and C. The scattering intensities displayed as a function of q and temperature are shown in figure 4-11 In this case, the lower total molecular weight reduces XN and thus the transition temperature for the decay in ordered structure scattering to modified scattering due to the combination of heating and x-ray dose to between 45 0 C and 80 0 C. The scattering data from I-39LMW-A and B are shown in figure 4-12. These samples 91 display an increase in low q scattering occurring at approximately 40 0 C to 50 0 C followed by almost total suppression of scattering above 500 C. One interesting thing to note about the data obtained from the samples as they were heated is that the transition temperatures are similar to those detected by DSC in the case of S-64LMW (transition temperatures are just above that observed using DSC) and S-45LMW (transition temperature range overlaps with that observed by DSC at the low end.) If the DSC is measuring the macroscopic softening of the sample, this makes sense (i.e. below this temperature, any physical interactions between the x-ray and the sample may not yield a change in morphology due to the glassiness of the PS block.) This indicates that there appears to be both a temperature and a dose effect during the heating run. Perhaps when the x-ray dose is small, the effect of the sample and x-ray interaction is not seen until the sample begins to rearrange on its own due to metastability of the lower temperature ordered structure; at this point, its rearrangement occurs piecewise until the entire sample takes on the new equilibrium diblock-D 3 blend morphology. Then, when the sample is cooled (described below), the initial morphology is not recovered due to the high dose of x-ray changing the equilibrium morphology and limiting the diffusivity and mobility of the chains which fixes the morphology. 4.4.3 Morphology change through the transition The transition from the initial morphology to the final morphology can be seen most clearly by examining the scattering from a transitioning sample as a function of time. Continuous three minute exposures over a total time period of 30 minutes were taken at each temperature for which scattering data is measured; selected series of these at the lowest temperatures during which morphology transitions occurred are shown in Figure 4-13. The scattering from S-64LMW-A at 80 0 C is shown in (a); here, the relatively sharp initial peak moves inwards, broadens, and flattens as time increases. At this temperature and dose, two peaks cannot be discerned; instead, the scattering data displays a fairly flat top. The movement to low q of the entirety of the peak is due to the replacement of the initial morphology with a new morphology containing a larger d-spacing. This could be due to the depolymerization of parts of the PDMS domains to form D3 which then swells the remaining PDMS chains and increases the PDMS domain size and spacing. This trend is present across all three samples. The piecewise nature of the transition could be due to the grains which are less strongly segregated forming the new structure first, leaving behind the sections of sample which scatter at lower q values. The broadening can easily be attributed to decreasing perfection of the ordered structure as it transitions to a disordered morphology and the flattening could be due to the superposition of a high dose peak and a lower dose peak. Similar trends are apparent in S-45LMW-A at 50 0 C (shown in (b) of this figure): an initially relatively narrow and intense peak broadens, shifts to lower q, and decreases in intensity as time increases for similar reasons. In contrast, in I-39LMW-B at 40 0 C, it is apparent that there is the development of a lower q peak while the higher q peak decreases in intensity, broadens, and shifts inwards. In this case, it is clear that the x-ray dose dependent peak grows at the expense of the initial peak 92 and that the initial phase peak shifts slightly to lower q due to some limited swelling of the structure. The temperatures investigated here were all slightly above the transition temperatures detected by DSC: thus their equilibrium phases should either be different from their low temperature phases or more kinetically achievable; it is curious that this is where the effects of x-ray dose begin to be seen. It could be due to a weakening of the driving force for ordering of the neat sample combined with an increasing concentration of dose-affected material. 4.4.4 Initial protocol cooling results In Figure 4-14, the last scattering patterns obtained at each temperature during the cooling runs performed are presented; S-64LMW-A and B are shown in (a), S-45LMW-A and B in (b), and I-39LMW-A and B in (c). In this case, the effect of x-ray dose is readily apparent (compare to Figures 4-10, 4-11, and 4-12 and 4-9. For the most part, during cooling the dose-dependent peak decreases in intensity without any compensatory growth in the initial peak: the initial structure is never recovered due to irreversible changes in the molecules and hence the morphology upon prolonged x-ray exposure. The curves become fairly featureless as increasing exposure destroys more and more of the PDMS block and the sample becomes composed of PS or PI hooked to ever shorter PDMS in a blend with D 3 and show almost no evidence of any consistent periodic variations in electron density. I-39LMW-B does display some ordering at lower temperatures, but the annealing time at these temperatures is much longer than that used in the other samples studied (two hours instead of thirty minutes) and both blocks are in the rubbery regime at all temperatures investigated and so have higher mobilities. Note also that the peaks are present at a lower q value than expected based upon the trend in peak movement during heating: the maxima are present at q- 3.8 at moderate temperatures during heating runs and at q- 3.2 lower temperatures during cooling runs. These peaks are actually closer in q to the high temperature peaks that occur at - 3.3 - indicative of irreversible structural change. Perhaps if these samples were exposed to x-rays for longer periods of time they would experience similar changes in scattering data as those shown for the PS-containing samples. Or, the difference in the chemistries of PS and PI may be responsible for the differing behavior in the x-ray beam of PS-PDMS and PI-PDMS diblocks. 4.4.5 Conclusions While DSC experiments show a range of temperatures at which some thermodynamic transition is occuring for the two PS-containing samples, only inconclusive data is available from SAXS to confirm this or to establish an ODT for the PI-containing diblock. Because of this, it is impossible to extract a x parameter for PS-PDMS or PI-PDMS based upon the work done here or to conclusively establish the ODT for these samples. It is possible, however, to say that even low molecular weight diblocks which can rearrange in dramatically brief timescales (on the order of minutes) are strongly segregated enough to produce scattering indicative of ordered structures at a variety of temperatures. This 93 means that the X parameter is very high in both of these systems, with promising applications for pattering due to the small length scale and stability of the structures obtained independent of processing treatment. There has also been evidence presented that these samples are highly sensitive to x-rays when exposed at elevated temperatures, providing important information for experimentalists who wish to do further work on the structures of these materials and the kinetics of molecular rearrangements. It also provides the possibility for new processing treatments; novel dose-dependent structures are formed which might be of interest for applications. 94 DSC plot for S-64LMW DSC plot for S-45LMW 0.5 0.8 0.6 0.4 0 0.2 0 -0.5 -0.2 -0.4 -0.6 -0.8 -1.5 -1.2 -2 -100 0 100 Temperature(Celsius) (a) 150 200 -1.4 -100 -50 0 50 Temperature(Celsius) 100 1 200 (b) igure 4-5: Data from DSC scans of S-64LMW (a), S-45LMW (b), and I-39LMW (c). The scans were performed at 5 mi nd began at room temperature. The samples were heated to 180 0 C, held for three minutes, cooled to -80 0 C, held fo inutes, and then heated back to room temperature. The morphology transitions observed in SAXS can be seen to vary egrees in the DSC data. For S-64LMW, the endothermic peak at approximately 50 0 C to 70 0 C corresponds somewhat w e morphology transition observed in SAXS between 80 0C and 165 0 C. For S-45LMW, the endothermic peak at approxima 00 C to 500 C is at the low end of the 40 0 C to 80 0 C range where the morphology transition was observed to occur with SA here do not appear to be any thermal transitions in I-39LMW. The thermal transitions are only observed during the hea ycles, consistent with the faster kinetics of structural rearrangement observed during SAXS. The large vertical lines wh ccur at 25 0 C are due to the rapid change in temperature of the DSC pans to the initial experimental temperature. DSC plotfor 1-39LMW 0.5 -0.5 -1.5 -2 -100 -50 0 50 Temperature(Celsius) 100 150 200 (c) ontinued Figure 4-5: Data from DSC scans of S-64LMW (a), S-45LMW (b), and I-39LMW (c). The scans were performed 0 0 mOt and began at room temperature. The samples were heated to 180 C, held for three minutes, cooled to -80 C, held minutes, and then heated back to room temperature. The morphology transitions observed in SAXS can be seen to vary egrees in the DSC data. For S-64LMW, the endothermic peak at approximately 50 0 C to 70 0 C corresponds somewhat w e morphology transition observed in SAXS between 80 0 C and 165 0 C. For S-45LMW, the endothermic peak at approximat 0 C to 500 C is at the low end of the 40 0 C to 80 0 C range where the morphology transition was observed to occur with SA here do not appear to be any thermal transitions in I-39LMW. The thermal transitions are only observed during the hea ycles, consistent with the faster kinetics of structural rearrangement observed during SAXS. The large vertical lines wh ccur at 250 C are due to the rapid change in temperature of the DSC pans to the initial experimental temperature. .......... ....... .. ...... ...... ...... ..... 0 a 0 b c o d 0 e o 0 f Figure 4-6: Images recorded on the marCCD detector for a sample with the same characteristics as sample II-D where 40 second exposures were continuously taken throughout the duration of the experiment; this sequence of images obviates the need for the revised protocol and suggests that during x-ray irradiation some event such as cross-linking which promotes the formation of locked-in non-equilibrium morphologies or some event such as depolymerization of the PDMS occurs which irreversibly changes the equilibrium morphology occurs. (a) shows the image taken at 25 0 C before any temperature treatment has been performed. (b) shows the scattering from this sample at 90 0 C after raising the temperature. (c) shows data taken after holding the sample at 90 0 C for 40 minutes. (d) shows the image taken after a quench to 60 0 C. (e) shows the data obtained after holding at 60 0 C for 18 minutes. (f) shows data taken after the sample was then cooled to 25 0 C and held there for one hour and 15 minutes. (g) shows data taken after this sample was removed, translated a small distance, and re-exposed to the x-ray beam. This sequence shows that sample exposure to the x-ray beam strongly affects its structure and the resulting data obtained. 97 (a) (a) (b) I Oi# I., (~) (C) t (d) Figure 8.2 2D-patterns in the q,, q, range 0.24. Row (b), m 25: 5 samples around the L,- H transition with f 0, 0,05, 0 10, - 0,71 nm ' (20 range - 1 ) of oriented morphologies in the system Si,,C,Ek m/D, 58"C, with the central structure featuring 0.15, and 0.20. Row (c), m = 25: 5 samples at around the H- parasitic scattering (hmited by the 3' sIt), I,k O, transition with beam-stop and attenuated dirert beam image. Row (a), m - 14: 5 typical L,, 4 = 0.25, 0.35, 0,40, 0 45, and 0,50 Row (d). m = 51.6: 5 samples around the H-1,-0, transition with dek 0, 0.075, samples with A, - 0, 0.04, 0.12, 0.20, and 0.17. 0.20, and 0.50. - Figure 4-7: Scattering data and caption taken from [79] showing scattering from a PEOPDMS and D3 blend. Compare this to the scattering shown in Figure 4-8. It is clear that the images obtained after x-ray irradiation look qualitatively similar to those shown above, providing evidence for the hypothesis that the x-ray beam causes depolymerization of the PDMS to form D3 . 98 A B A B A 00 b B f c d A A e B h igure 4-8: Images of the scattering patterns recorded by the CCD camera for selected temperatures for samples S-64LMW nd S-64LMW-B. In these images, the integrated intensity presented in Figure 4-10 does not fully capture the x-ray scatter at occurred due to the development of orientation during the experiment. S-64LMW-A at 80 0 C is shown in a; S-64LMW 90 0 C in b; S-64LMW-A at 105 0 C in c; S-64LMW-B at 110 0C in d; S-64LMW-A at 120 0C in e; S-64LMW-B at 125 0 C i 64LMW-A at 150 0 C after 30 minutes in g and after approximately 70 minutes in h; S-64LMW-B at 165 0 C in i. .... ... ................ ... of temperature for1-39LMW-C Scattering asa function Scatteding as a function of temperature for S-45LMW-D 70C I 800 60C I 50C 65C I I. I S 55C 40C I' 30C I 45C 2SC I 35C 20C I 10C 25C 0 0.2 0.6 0.4 0.8 1 0 0.2 0.4 0.6 q (1/nm) q (1/nm) (a) (b) 0.8 1 igure 4-9: Scattering data from samples S-45LMW-D and I-39LMW-C following the revised protocol described. (a) sh cattering from sample S-45LMW-D and (b) - (d) show scattering from I-39LMW-C. It appears that there is very little cha ith temperature in the scattering data for S-45LMW-D for the temperature range investigated (25-80 0 C.) For I-39LMW ere appear to be one prominent and one secondary peak from 10-80 0 C, which becomes one peak which steadily loses inten rough 200 0 C. ............ ......... - .. .... ..... Scattering as a function oftemperature forI-39LMW-C Scattering as a function of temperature for 1-39tMW-C 150C 200C 140C 130C 19C c cc 120C I II 80C 110C 100C 170C 90C 160C 80C 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 q (1/nm) q (1/nm) (c) (d) 0.8 1 ontinued Figure 4-9: Scattering data from samples S-45LMW-D and I-39LMW-C following the revised protocol describ ) shows scattering from sample S-45LMW-D and (b) - (d) show scattering from I-39LMW-C. It appears that there is v tle change with temperature in the scattering data for S-45LMW-D for the temperature range investigated (25-80 0 C.) 39LMW-C, there appear to be one prominent and one secondary peak from 10-80 0 C, which becomes one peak which stead ses intensity through 200 0 C. ....... .. ............. Scattering as a function of temperature forS-64LMW Scattering as a function of temnperature for S-64LMW 95C-C 65C-C I I 9C-C 65C-8 90C-B 6C-C soc-A 800-C 8CC5CC-CC 0 0.2 0.4 0.6 q (1/nm) 0.8 5CC-A S0C-A 25c-B 75C-C 25-A 70C-C 1 0 0.2 0.4 0.6 0.8 q (1/nm) igure 4-10: Scattering from S-64LMW-A, B, and C as a function of temperature. In all cases, the data from the last expos ken at a particular temperature during the heating cycle is plotted. S-64LMW-A data is shown in red; S-64LMW-B een; S-64LMW-C in green. For all samples, it is clear that in the region of 70 0 C-165 0 C the low temperature high q p suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in e mple, indicating that there is both a time and temperature effect in the morphological transition that is occurring; ehavior, however, is the same in all cases. Scattering as a function oftemperature forS-64LMW 120C-C - - 0 0.2 - 0.4 0.6 q (1/nm) Scattering as a function of temperature for S-64LMW 0.8 150C-C - 120C-A 150C-A 115C-C 145C-C 1100-C 140C-C 110C-8 135C-C 105C-C 130C-C 105C-A 130C-B 1O0C-C 125C-C 1 0 0.2 0.4 0.6 0.8 q (1/nm) ontinued Figure 4-10: Scattering from S-64LMW-A, B, and C as a function of temperature. In all cases, the data fr e last exposure taken at a particular temperature during the heating cycle is plotted. S-64LMW-A data is shown in -64LMW-B in green; S-64LMW-C in green. For all samples, it is clear that in the region of 70 0 C-1650 C the low temperat gh q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatu each sample, indicating that there is both a time and temperature effect in the morphological transition that is occurri e behavior, however, is the same in all cases. Scattering as a function oftemperature forS-64LMW 180C-8 180C-A 165C-B 0 0.2 0.6 0.4 0.8 1 q (1/nm) ontinued Figure 4-10: Scattering from S-64LMW-A, B, and C as a function of temperature. In all cases, the data f e last exposure taken at a particular temperature during the heating cycle is plotted. S-64LMW-A data is shown in -64LMW-B in green; S-64LMW-C in green. For all samples, it is clear that in the region of 70 0 C-165 0 C the low temperat gh q peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatu each sample, indicating that there is both a time and temperature effect in the morphological transition that is occurri e behavior, however, is the same in all cases. Scattering asa function of temperature forS-45LMW Scattering as a fuinction of temperature for S-45LMW 150C-C I 52C- 36C-C - - - -50C-C 34C 50CA 32C-C 48C-C 30C- 44C-C 30CAC 44C-C 0 0.2 0.4 0.6 q (1/nm) 0.8 25C-B 42C-C 25C-A 40C-C 1 0 0.2 0.4 0.6 0.8 1 q (1/nm) igure 4-11: Scattering from S-45LMW-A, B, and C as a function of temperature. In all cases, the data from the last expos ken at a particular temperature during the heating cycle is plotted. S-45LMW-A data is shown in red; S-45LMW-B een; S-45LMW-C in green. For all samples, it is clear that in the region of 40 0 C-85 0 C the low temperature high q p suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperatures in e mple, possibly indicating that longer annealing times are necessary in this regime to achieve equilibrium morphologies; ehavior, however, is the same in all cases. ............. forS-45LMW of temperature asa function Scattering Scattering as a function of temnperature for S-45LMW 95C-B I110C-85C-B 8OC-B 75C-B 75C-A 100C-B 65C-A 0 0.2 0.6 0.4 q (1/nm) 0.8 1 0 0.2 0.6 0.4 0.8 1 q (1/nm) ontinued Figure 4-11: Scattering from S-45LMW-A, B, and C as a function of temperature. In all cases, the data from st exposure taken at a particular temperature during the heating cycle is plotted. S-45LMW-A data is shown in red 5LMW-B in green; S-45LMW-C in green. For all samples, it is clear that in the region of 40 0 C-85 0 C the low temperature h peak is suppressed and a high temperature low q peak is formed. This transition occurs at slightly different temperature ach sample, possibly indicating that longer annealing times are necessary in this regime to achieve equilibrium morpholog he behavior, however, is the same in all cases. ..... .....- for1-39LMW Scattering as a function oftemperature 500-A I 48C-B 46C-8 44C-8 42C-B 1~ II 40-B II 4*C-A 30C-A 250-A 0 0.2 0.4 0.6 0.8 1 q (1/nm) igure 4-12: Scattering from I-39LMW-A and B as a function of temperature. In all cases, the data from the last expos ken at a particular temperature during the heating cycle is plotted. I-39LMW-A data is shown in red; I-39LMW-B in gre or both samples, in the range of 30 0C-50 0C the low temperature high q peak is suppressed and a high temperature lo eak is formed. ......... 0.3 Scattering at constant temperature for S-45LMW-A at 50C Scattering at constant temperature for S-64LMW-A at 80C Increasing time Increasing time . 0.35 0.4 0.45 q (1/nm) 0.5 0.55 0.6 0.45 0.5 (a) igure 4-13: SAXS data taken from diblocks undergoing structural evolution at inute exposures from samples transitioning between ordered and disordered gh temperature correlation hole peak at the expense of the low temperature -64LMW-A at 80 0 C; (b)shows S-45LMW-A at 500 C; (c) shows I-39LMW-B at 0.55 0.6 q (1/nm) 0.65 0.7 (b) a constant temperature. Ten sequential th structures are shown and the growth of ordered structure peak is evident. (a) sh 40 0 C. ........... Scattering at constant temperature for 1-39LMW-B at 40C Increasing time 0.2? 0.3 0.35 q (1/nm) 0.4 0.45 (c) ontinued Figure 4-13: SAXS data taken from diblocks undergoing structural evolution at a constant temperature. equential three minute exposures from samples transitioning between ordered and disordered structures are shown and rowth of the high temperature correlation hole peak at the expense of the low temperature ordered structure peak is evid a) shows S-64LMW-A at 80 0 C; (b)shows S-45LMW-A at 50 0 C; (c) shows I-39LMW-B at 40 0 C. cooling forS-45LMW during oftemperature as a function Scattering Scattering as a function of temperature durin cooling for S-64LMW 105C-B 1650- I I 1150-4 95C-B - 130C-I 9OC4B 120C-A 65C-B 80C-B 75C-B 105C4 I I J 7C-A 90C-B BOC-A 50C-A 30C-A 0 0.2 0.6 0.4 0.8 1 0 0.2 0.5 0.4 q (1/nm) q (1/nm) (a) (b) 0.8 i igure 4-14: SAXS data taken from all samples during their cooling runs. Note the lack of structure development for any of amples which received only a 30 minute anneal at the colder temperatures. I-39LMW-B shows morphology developmen onger temperatures due to the longer two hour anneal at each temperature that it received. (a) shows data from S-64LMW nd B; (b) shows data from S-45LMW-A and B; (c) shows data from I-39LMW-A and B. during coolngforI-39LMW Scattering asa frctIlon oftemperature 4CC-A 35C-8 3CC-B 250-8 20C-8 IS-1 15C-B 1CC-A C 0.2 0.4 0.6 0.8 q (1/nm) (c) ontinued Figure 4-14: SAXS data taken from all samples during their cooling runs. Note the lack of structure developm r any of the samples which received only a 30 minute anneal at the colder temperatures. I-39LMW-B shows morpholo evelopment at longer temperatures due to the longer two hour anneal at each temperature that it received. (a) shows d om S-64LMW-A and B; (b) shows data from S-45LMW-A and B; (c) shows data from I-39LMW-A and B. .. .... ............. ... ........... ............. ..... .... ... Chapter 5 Conclusions and future work In this thesis, several different topics related to determining the morphologies of PDMScontaining diblocks were investigated. These included work on the structures of highly segregated materials both at equilibrium and as a function of processing treatment (casting solvent and annealing) as well as attempted determination of the ODT for several low molecular weight diblocks. Overall, the results were mixed: while the volume fractions at which bulk morphologies of the larger molecular weight diblocks were observed were reasonable given prior work and theory and they were relatively stable independent of the sample preparation procedure, determination of an ODT or interpretation of the scattering data obtained at elevated temperatures was inconclusive. The most promising future directions for this work involve further study of the elevated temperature SAXS data and determination of what is causing the dose-dependent scattering effects and the application of bulk morphology knowledge to the creation of devices for which it is suited and to fundamental studies investigating the perturbation of known and well-characterized morphologies under confinement. The former is most probably due to the depolymerization of PDMS to form D4 and cross-links; the former causes new structures to become the equilibrium morphologies and the latter arrests structural change by impeding the chain mobility required for morphological transitions. 5.1 Bulk morphologies The bulk morphologies for strongly segregated diblocks found in this thesis match fairly well with predictions based upon prior experimental and theoretical work. Figure 2-1 shows this quite nicely. In part a, it appears as if the lamellar volume fraction range is smaller than would be expected: there are samples with a cylindrical morphology in PI- and PS-containing samples with volume fractions at the low end of the expected lamellar region. Similarly, there is a PI-containing sample with a cylindrical morphology at a volume fraction at the high end of the expected lamellar region. The combination of these two effects means that the lamellar range may be narrower than theoretically expected; the phase diagram may be compressed inwards from both ends. Other work, most notably [64] suggests that the phase diagram may be asymmetrical; however, it appears that the samples investigated with a PDMS majority composition fell well within 112 the anticipated ranges for the morphologies (i.e. not close to morphology boundaries) and in two of the three cases were in fact closer to the boundary with larger PDMS content than that with lower so that the narrowing of the central region would not be observed. Examination of part b of Figure 2-1 further shows that the morphologies obtained in this work fall in similar regions of volume fraction and segregation strength as others previously observed in other PDMS-containing systems. In many of these cases, it appears that the boundaries between morphologies are closer to 50% PDMS on both the PDMS-rich and PDMS-poor sides of the phase diagram. These results are encouraging because they verify prior work and indicate that PDMS-containing diblocks display consistent morphologies which match theoretical predictions in many different experiments. The one exception to this is the one low molecular weight sample on the phase diagram, which although having a majority of PDMS by volume displays (64 % ) a PDMS cylinders morphology. This may be explained by the weakly segregating nature of the system; at the bottom of the phase diagram, all the theoretical boundaries curve together and the predicted morphologies are in less of a energy well than at stronger segregations. Here, other effects such as block flexibility, local chain chemistry, etc. may affect the morphology that is displayed. Another encouraging result from the work on bulk morphologies is the correspondance between SAXS and TEM results. In cases where both tools were used to investigate morphology and there were enough peaks in the SAXS data to assign a morphology, the morphology assignment was consistent across TEM and SAXS for samples which were prepared in the same way. SAXS morphology assignment can be less than clear-cut due to possible ambiguity in peak location (especially for broad or low intensity peaks.) TEM morphology assignment can be misleading if anomalous grains are investigated or they do not occur at sufficient angles with respect to the electron beam for enough different projections to be observed. However, when measurements made by the two different techniques indicate the same morphology it means that there is a high probability that the assigned morphology is correct. Future work in this area could expand upon the results presented here in several ways. These could include pairing a larger diversity of other chemical blocks (see Chapter 1 for other blocks which have previously been investigated; the diversity is endless) with PDMS and determining whether the boundaries between morphologies as a function of chiN and volume fraction continue to be consistent. In a similar vein, PDMS-containing triblocks, starblocks, and other architectures could be studied by these same methods to determine the influence of archeology on morphology. Other opportunities include synthesizing more samples with volume fractions inside the observed cylinder-lamellar transition region (around 40% PDMS by volume) to better pin down how the morphology varies here. Finally, advanced theoretical modelling which explicitly accounts for the chemistries and chain characteristics of PDMS, PS, and PI could be performed to see if it can account for the data presented here. 5.2 Processing treatments Additional support for the assignment of accurate equilibrium morphologies to the diblocks studied based upon the TEM and SAXS work done on samples cast from toluene 113 and then annealed is provided by the robustness of these morphologies to processing treatments. It was shown here that the morphologies obtained did not vary significantly with the preference of the solvent for PDMS or the attached block and with whether or not an annealing treatment took place; this means that even if one of the blocks was swollen and an alternate morphology was present at some point during casting, this effect was erased by the time that the final structure was formed. This could be due to the strong segregation between the blocks and thus strong driving force for the formation of an ordered morphology and/or of substantial mobility being present in the material throughout the casting process. In any case, since the morphologies found were for the most part independent of initial processing, the final morphology can be said to be robust and easily attainable. The d-spacings for some samples did vary somewhat with casting solvent prior to annealing, but since they all obtained the same value after annealing it is fairly clear that the annealing treatment was sufficient to relax the morphology to its equilibrium structure. Further work in this area could be performed to investigate the affects of different annealing times and temperatures, the use of solvent annealing or combined solvent and thermal annealing. Faster casting procedures could also be used to attempt to lock in swollen structures by evaporation of solvent before the polymers have time to rearrange into their preferred morphologies. Since PDMS-containing diblocks are increasingly used in academia and industry for patterning and other technologies where the ultimate morphology is extremely relevant to the ultimate application, especially in thin films, the importance of processing steps in determining final structure is of great importance. 5.3 Morphologies as a function of temperature Overall, the results from this study are somewhat less clear-cut than those discussed previously. It is fair to say that both low molecular weight PS-PDMS samples had a transition which was detectable by DSC which matched temperatures in which there were changes in the scattering patterns of continuously x-ray exposed diblocks. The PI-PDMS sample did not display a transition in either the DSC experiment or the SAXS experiment with minimal dosing across a larger range than that measured by DSC; however, some decrease in order is apparent upon increasing temperature because a secondary peak in the scattering pattern is lost as the temperature is raised. In all cases, it appeared that exposure to the x-ray beam had substantial effects on the SAXS data: experiments where the samples were continuously exposed to x-rays exhibited peaks which broadened asymmetrically towards low q (in some cases showing increased orientation) and appeared similar to SAXS data obtained from blends of PEO-PDMS and D4 while diblocks with very limited exposure to x-rays displayed peaks which were fairly stationary and only changed in intensity (if at all) in the same temperature ranges. It is also of note that the kinetics for structural arrangement were very fast in the latter case: while the samples were held at each temperature for approximately 40 minutes before taking the final data, this data differed imperceptibly from the data taken after initially quenching to the target temperature. This is probably due to the short chain lengths of all the blocks which both depress the glass transitions of all of the blocks thus providing good chain mobility and 114 result in a lack of entanglements above the glass transitions. In the former case, however, the transition occurred relatively slowly (on the order of at least 30 minutes during heating cycles and sometimes with no change at all in 30 minutes during cooling cycles), indicating some sort of crosslinking and/or change of equilibrium state and/or other effect occurring due to the x-ray flux which made the structure more resistant to change. There is much more work that could be done to provide further insight into all the phenomena that are present here. It could be interesting to use other methods (such as DMA) to locate the ODT to provide confirmation of the ranges found. Additionally, more experiments could be done using the revised protocol to both pin down ranges over which the transition occurs and then precisely target the ODT. Experiments could be performed significantly above the ODT in order to determine the mean-field and correlation hole scattering, from which X and its possible composition dependence can be extracted. Other work could build upon the studies discussed above and investigate the effect of solvent concentration on structure by obtaining scattering from solvated diblocks with different concentrations of solute. This could determine whether or not there are other morphologies present during casting which could be locked in by fast solvent evaporation, cross-linking, or other means. Other studies could focus on better determining what is happening during x-ray exposure; these could take the form of exposing large regions of sample to similarly intense x-rays (or precisely determining which region of sample is in the beam path) and then examining the properties of the resulting material using GPC, FTIR, NMR, DSC, and TEM to determine any chemical or structural changes in the material. Deep UV illumination could also be investigated as a route to promoting the formation of these morphologies. The x-ray induced morphologies might be well-suited for certain technological applications and in any case would expand the repertoire of diblock morphologies which are currently available to engineers; thus, determining precisely how they are induced and what the resultant structures are would be quite interesting. 5.4 Outlook for the future The primary reason that PDMS-containing diblock polymers are interesting and relevant to study is that their chemical properties are ideal for nanoscale patterning applications. The combination of a large x paramter which causes ordered morphologies to exist even in very low molecular weight materials with a large differential etch resistance of PDMS and other polymers to oxygen plasma, HF, and other etchants which allows for removal of one block allows for the creation of highly ordered and topological patterns. For the most part, PDMS-containing diblocks have been used extensively for thin film patterning. However, there is a large variety of geometries into which these materials could be patterned. A better understanding of the factors that determine PDMS-containing diblock morphology in the bulk allows for better designed and interpretable studies about how these morphologies are perturbed if confinement is introduced in any of the three dimensions. While the majority of experimental and theoretical work on block copolymer structure has focused on determining bulk structures, understanding structures that are present inside of confined geometries allows for the further development of block polymers as 115 technologically relevant materials. Technologies that make use of bulk block polymers are limited by the types and qualities of morphologies that are thermodynamically and kinetically accessible as well as by the range of available chemistries that can be used in synthesis. The presence of non-equilibrium defects such as grain boundaries and dislocations are also problematic and can be difficult to eliminate. Recently, research has been pursued to address these issues by using a variety of techniques to change the thermodynamic boundary conditions that are present when the block polymer microstructure is formed. Elimination of randomly located defects and control of the orientation and spatial registration of the domains have been pursued using "directed self assembly" [80]. Approaches have included affecting the energetics of candidate morphologies by chemical means, designing casting conditions in which the block polymer undergoes a disorder to order transition while geometrically confined to dimensions at which the energies of the confining surfaces have a significant effect on the overall free energy of the material, and the combination of these two strategies. For all of these cases, block polymers have mainly been confined between surfaces of constant curvature (flat, cylindrical or spherical surfaces), and so are subject to the same geometrical constraint at all locations within the material. In general, the symmetries of the confining geometry and of the block polymer's equilibrium morphology interact to produce the final morphology; the symmetries of the confining surfaces should be satisfied if possible in the equilibrium morphologies, so the domain spacing and IMDS shapes are usually perturbed from their bulk values and forms. Novel morphologies found under confinement include the perturbation of cylinders to helices or network structures, the formation of structures which are concentrically nested inside of each other, and many others including the necessary introduction of "equilibrium defects." In most cases, this is due to the surface energetics, maintenance of the repeat period, or maintenance of the minority domain's relevant length while conserving the volume fraction of material throughout the confined morphology. The understanding of morphologies that form as a function of imposed boundary symmetries and boundary surface preferences is thus a powerful tool for the prediction of new structures; it also allows the exclusion of morphologies which strongly violate the required symmetries, and surface boundary conditions, and provides insight into how the bulk morphologies may be manipulated. PDMS-containing diblocks are ideal for investigating the formation of such structures for several reasons: their morphologies are well-characterized, they have ideal properties for adapting the new structures into relevant technologies, the mass thickness contrast between PDMS and many other polymers makes imaging easy, and the strong segregation at small length scales means that a variety of chain lengths can be experimentally accessed so that large ranges of confinement ratios can be studied. For more information about prior work in this area and its technological potential, see [81]. 116 Bibliography [1] Y. Jung, W. Jung, and C. Ross, "Nanofabricated concentric ring structures by templated self-assembly of a diblock copolymer," Nano Lett, vol. 8, pp. 2975-81, Sept. 2008. [2] Y. Kang, J. J. Walish, T. Gorishnyy, and E. L. Thomas, "Broad-wavelengthrange chemically tunable block-copolymer photonic gels.," Nature Materials, vol. 6, pp. 957-60, Dec. 2007. [3] J. Yoon, R. T. Mathers, G. W. Coates, and E. L. Thomas, "Optically Transparent and High Molecular Weight Polyolefin Block Copolymers toward Self-Assembled Photonic Band Gap Materials," Macromolecules, vol. 39, pp. 1913-1919, Mar. 2006. [4] J. Yoon, W. Lee, and E. L. Thomas, "Self-Assembly of Block Copolymers for Photonic- Bandgap Materials," MRS Bulletin, vol. 30, no. October, pp. 721-726, 2005. [5] T. Deng, C. Chen, C. Honeker, and E. L. Thomas, "Two-dimensional block copolymer photonic crystals," Polymer, vol. 44, pp. 6549-6553, Oct. 2003. [6] A. Urbas, M. Maldovan, P. DeRege, and E. Thomas, "Bicontinuous cubic block copolymer photonic crystals," Advanced Materials,vol. 14, pp. 1850-1853, Dec. 2002. [7] A. C. Edrington, A. M. Urbas, P. DeRege, C. X. Chen, T. M. Swager, N. Hadjichristidis, M. Xenidou, L. J. Fetters, J. D. Joannopoulos, Y. Fink, and E. L. Thomas, "Polymer-Based Photonic Crystals," Advanced Materials, vol. 13, pp. 421-425, Mar. 2001. [8] Y. Fink, A. Urbas, M. Bawendi, J. Joannopoulos, and E. Thomas, "Block copolymers as photonic bandgap materials," Journal of Lightwave Technology, vol. 17, no. 11, pp. 1963-1969, 1999. [9] L. Zhu, S. Cheng, B. Calhoun, Q. Ge, R. Quirk, E. Thomas, B. Hsiao, F. Yeh, and B. Lotz, "Crystallization temperature-dependent crystal orientations within nanoscale confined lamellae of a self-assembled crystalline-amorphous diblock copolymer," Journal of the American Chemical Society, vol. 122, pp. 5957-5967, June 2000. [10] Y.-L. Loo, R. A. Register, and A. J. Ryan, "Modes of Crystallization in Block Copolymer Microdomains: Breakout, Templated, and Confined," Macromolecules, vol. 35, pp. 2365-2374, Mar. 2002. 117 [11] L. Zhu, S. Cheng, B. Calhoun, Q. Ge, R. Quirk, E. Thomas, B. Hsiao, F. Yeh, and B. Lotz, "Phase structures and morphologies determined by self-organization, vitrification, and crystallization: confined crystallization in an ordered lamellar phase of PEO-b-PS diblock copolymer," Polymer, vol. 42, pp. 5829-5839, June 2001. [12] P. Weimann, D. Hajduk, C. Chu, K. Chaffin, J. Brodil, and F. Bates, "Crystallization of tethered polyethylene in confined geometries," Journal of Polymer Science Part B: Polymer Physics, vol. 37, pp. 2053-2068, Aug. 1999. [13] L. Zhu, B. R. Mimnaugh, Q. Ge, R. P. Quirk, S. Z. Cheng, E. L. Thomas, B. Lotz, B. S. Hsiao, F. Yeh, and L. Liu, "Hard and soft confinement effects on polymer crystallization in microphase separated cylinder-forming PEO-b-PS / PS blends," Polymer, vol. 42, pp. 9121-9131, Oct. 2001. [14] F. S. Bates and G. H. Fredrickson, "Block copolymer thermodynamics: theory and experiment.," Annual Review of Physical Chemistry, vol. 41, pp. 525-57, Jan. 1990. [15] M. Matsen and F. Bates, "Unifying Weak- and Strong-Segregation Block Copolymer Theories," Macromolecules, vol. 29, pp. 1091-1098, Jan. 1996. [16] A. Khandpur, S. Foerster, F. Bates, I. Hamley, A. Ryan, W. Bras, K. Almdal, and K. Mortensen, "Polyisoprene-polystyrene diblock copolymer phase diagram near the order-disorder transition," Macromolecules, vol. 28, no. 26, pp. 8796-8806, 1995. [17] M. Matsen and F. Bates, "Block copolymer microstructures in the intermediatesegregation regime," Journal of Chemical Physics, vol. 106, no. 6, pp. 2436-2448, 1997. [18] M. Matsen and M. Schick, "Stable and unstable phases of a diblock copolymer melt," Physical Review Letters, vol. 72, no. 16, pp. 2660-2663, 1994. [19] M. Matsen, "The standard Gaussian model for block copolymer melts," Journal of Physics: Condensed Matter, vol. 14, no. 2, pp. R21-R47, 2002. [20] M. Matsen and F. Bates, "Origins of complex self-assembly in block copolymers," Macromolecules, vol. 29, pp. 7641-7644, Jan. 1996. [21] R. Segalman, "Patterning with block copolymer thin films," Materials Science and Engineering: R: Reports, vol. 48, pp. 191-226, Mar. 2005. [22] H. Xiang, K. Shin, T. Kim, S. Moon, T. J. McCarthy, and T. P. Russell, "The influence of confinement and curvature on the morphology of block copolymers," Journal of Polymer Science Part B: Polymer Physics, vol. 43, pp. 3377-3383, Dec. 2005. [23] M. Ma, E. L. Thomas, G. C. Rutledge, B. Yu, B. Li, Q. Jin, D. Ding, and A.-C. Shi, "Gyroid-Forming Diblock Copolymers Confined in Cylindrical Geometry: A Case of Extreme Makeover for Domain Morphology," Macromolecules, vol. 43, pp. 30613071, Mar. 2010. 118 [24] T. Higuchi, A. Tajima, K. Motoyoshi, H. Yabu, and M. Shimomura, "Frustrated phases of block copolymers in nanoparticles.," Angewandte Chemie (International ed. in English), vol. 47, pp. 8044-6, Jan. 2008. [25] P. Chen, H. Liang, and A.-C. Shi, "Microstructures of a Cylinder-Forming Diblock Copolymer under Spherical Confinement," Macromolecules, vol. 41, pp. 8938-8943, Nov. 2008. [26] K. I. Winey, E. L. Thomas, and L. J. Fetters, "Swelling of lamellar diblock copolymer by homopolymer: influences of homopolymer concentration and molecular weight," Macromolecules, vol. 24, pp. 6182-6188, Nov. 1991. [27] F. Bates and G. Fredrickson, "Block copolymers-designer soft materials," Physics Today, vol. 52, no. February, p. 32, 1999. [28] E. W. Cochran, C. J. Garcia-Cervera, and G. H. Fredrickson, "Stability of the Gyroid Phase in Diblock Copolymers at Strong Segregation," Macromolecules, vol. 39, pp. 4264-4264, June 2006. [29] M. Szwarc, "'Living' Polymers," Nature, vol. 178, no. 4543, pp. 1168-1169, 1956. [30] B. Alberts, A. Johnson, and J Lewis, Molecular Biology of the Cell. New York: Garland Science, 4th ed., 2002. [31] Http://www.jeol.com, "JEM-2100F Transmission Electron Microscope," 2011. [32] Http://www.rigaku.com, "S-MAX3000: High-resolution anisotropic measurements with maximum flexibility," 2011. [33] S. Ndoni, M. E. Vigild, and R. H. Berg, "Nanoporous materials with spherical and gyroid cavities created by quantitative etching of polydimethylsiloxane in polystyrenepolydimethylsiloxane block copolymers.," Journal of the American Chemical Society, vol. 125, pp. 13366-7, Nov. 2003. [34] M. Hansen, M. Vigild, R. Berg, and S. Ndoni, "Nanoporous Crosslinked Polyisoprene from PolyisoprenePolydimethylsiloxane Block Copolymer," Polymer Bulletin, vol. 51, no. 5, pp. 403-409, 2004. [35] P. P. Szewczykowski, K. Andersen, L. Schulte, K. Mortensen, M. E. Vigild, and S. Ndoni, "Elastomers with Reversible Nanoporosity," Macromolecules, vol. 42, pp. 5636-5641, Aug. 2009. [36] L. Schulte, A. Grydgaard, M. R. Jakobsen, P. P. Szewczykowski, F. Guo, M. E. Vigild, R. H. Berg, and S. Ndoni, "Nanoporous materials from stable and metastable structures of 1,2-PB-b-PDMS block copolymers," Polymer, vol. 52, pp. 422-429, Jan. 2011. 119 [37] C.-C. Chao, T.-C. Wang, R.-M. Ho, P. Georgopanos, A. Avgeropoulos, and E. L. Thomas, "Robust block copolymer mask for nanopatterning polymer films.," ACS Nano, vol. 4, pp. 2088-94, Apr. 2010. [38] M. Rodwogin, C. Spanjers, C. Leighton, and M. Hillmyer, "Polylactide- Poly (dimethylsiloxane)- Polylactide Triblock Copolymers as Multifunctional Materials for Nanolithographic Applications," ACS Nano, vol. 4, no. 2, pp. 725-732, 2010. [39] Y. S. Jung and C. A. Ross, "Orientation-controlled self-assembled nanolithography using a polystyrene-polydimethylsiloxane block copolymer.," Nano Letters, vol. 7, pp. 2046-50, July 2007. [40] C.-C. Chao, R.-M. Ho, P. Georgopanos, A. Avgeropoulos, and E. L. Thomas, "Silicon oxy carbide nanorings from polystyrene-b-polydimethylsiloxane diblock copolymer thin films," Soft Matter, vol. 6, no. 15, pp. 3582-3587, 2010. [41] E.-Q. Chen, Y. Xia, M. J. Graham, M. D. Foster, Y. Mi, W.-L. Wu, and S. Z. D. Cheng, "Glass Transition Behavior of Polystyrene Blocks in the Cores of Collapsed Dry Micelles Tethered by Poly(Dimethylsiloxane) Coronae in a PS- b -PDMS Diblock Copolymer," Chemistry of Materials, vol. 15, pp. 2129-2135, June 2003. [42] J. H. Waller, D. G. Bucknall, R. A. Register, H. W. Beckham, J. Leisen, and K. Campbell, "C60 fullerene inclusions in low-molecular-weight polystyrenepoly(dimethylsiloxane) diblock copolymers," Polymer, vol. 50, pp. 41994204, Aug. 2009. [43] H. Cho, S. Kim, and S. Park, "Fabrication of gold nanoparticles and silicon oxide corpuscles from block copolymers," Journal of Materials Chemistry, vol. 20, no. 6, pp. 1156-1160, 2010. [44] W. Hu, J. T. Koberstein, J. P. Lingelser, and Y. Gallot, "Interfacial Tension Reduction in Polystyrene/Poly(dimethylsiloxane) Blends by the Addition of Poly(styreneb-dimethylsiloxane)," Macromolecules, vol. 28, pp. 5209-5214, July 1995. [45] M. Maric and C. W. Macosko, "Block copolymer compatibilizers for polystyrene/poly(dimethylsiloxane) blends," Journal of Polymer Science Part B: Polymer Physics, vol. 40, pp. 346-357, Feb. 2002. [46] D. Cho, W. Hu, J. T. Koberstein, J. P. Lingelser, and Y. Gallot, "Segregation Dynamics of Block Copolymers to Immiscible Polymer Blend Interfaces," Macromolecules, vol. 33, pp. 5245-5251, July 2000. [47] K. Chang, C. W. Macosko, and D. C. Morse, "Ultralow Interfacial Tensions of Polymer/Polymer Interfaces with Diblock Copolymer Surfactants," Macromolecules, vol. 40, pp. 3819-3830, May 2007. [48] M. Wagner and B. Wolf, "Effect of block copolymers on the interfacial tension between two immiscible homopolymers," Polymer, vol. 34, no. 7, pp. 1460-1464, 1993. 120 [49] P. Choi, P.-F. Fu, and L. Guo, "Siloxane Copolymers for Nanoimprint Lithography," Advanced FunctionalMaterials,vol. 17, pp. 65-70, Jan. 2007. [50] S. Ndoni, P. Jannasch, N. B. Larsen, and K. Almdal, "Lubricating Effect of Thin Films of Styrene-Dimethylsiloxane Block Copolymers," Langmuir, vol. 15, pp. 38593865, May 1999. [51] X. Chen and J. A. J. Gardella, "Surface Modification of Polymers by Blending Siloxane Block Copolymers," Macromolecules, vol. 27, pp. 3363-3369, June 1994. [52] W. Lee, H. Kim, and H. Lee, "Proton exchange membrane using partially sulfonated polystyrene-b-poly(dimethylsiloxane) for direct methanol fuel cell," Journal of Membrane Science, vol. 320, pp. 78-85, July 2008. [53] T. Miyata, S. Obata, and T. Uragami, "Morphological Effects of Microphase Separation on the Permselectivity for Aqueous Ethanol Solutions of Block and Graft Copolymer Membranes Containing Poly(dimethylsiloxane)," Macromolecules, vol. 32, pp. 3712-3720, June 1999. [54] N. Kizilcan, N. Oz, B. Ustamehmetoglu, and A. Akar, "High conductive copolymers of polypyrrole-aw-diamine polydimethylsiloxane," European Polymer Journal, vol. 42, pp. 2361-2368, Oct. 2006. [55] N. Wu, L. Huang, A. Zheng, and H. Xiao, "Surface properties of block and graft polystyrene-polydimethylsiloxane copolymers," Journal of Applied Polymer Science, vol. 99, pp. 2936-2942, Mar. 2006. [56] H. R. Fischer, K. Tempelaars, A. Kerpershoek, T. Dingemans, M. Iqbal, H. V. Lonkhuyzen, B. Iwanowsky, and C. Semprimoschnig, "Development of flexible LEOresistant PI films for space applications using a self-healing mechanism by surfacedirected phase separation of block copolymers.," A CS Applied Materials & Interfaces, vol. 2, pp. 2218-25, Aug. 2010. [57] L. Willner, R. Lund, M. Monkenbusch, 0. Holderer, J. Colmenero, and D. Richter, "Polymer dynamics under soft confinement in a self-assembled system," Soft Matter, vol. 6, no. 7, pp. 1559-1570, 2010. [58] C. Lorthioir, A. Alegria, J. Colmenero, and B. Deloche, "Heterogeneity of the Segmental Dynamics of Poly(dimethylsiloxane) in a Diblock Lamellar Mesophase: Dielectric Relaxation Investigations," Macromolecules, vol. 37, pp. 7808-7817, Oct. 2004. [59] C. Lorthioir, P. Auroy, and B. Deloche, "End-chain segment ordering in lamellar sublayers of a diblock copolymer.," The European Physical Journal. E, Soft Matter, vol. 11, pp. 3-6, May 2003. [60] M. S. Kent, "A quantitative study of tethered chains in various solution conditions using Langmuir diblock copolymer monolayers," Macromolecular Rapid Communications, vol. 21, pp. 243-270, Mar. 2000. 121 [61] K. Almdal, M. A. Hillmyer, and F. S. Bates, "Influence of Conformational Asymmetry on Polymer - Polymer Interactions: An Entropic or Enthalpic Effect?," Macromolecules, vol. 35, pp. 7685-7691, Sept. 2002. [62] S. Maheshwari, M. Tsapatsis, and F. S. Bates, "Synthesis and Thermodynamic Properties of Poly(cyclohexylethylene- b -dimethylsiloxane- b -cyclohexylethylene)," Macromolecules, vol. 40, pp. 6638-6646, Sept. 2007. [63] M. Ninago, A. Satti, A. Ciolino, E. Valles, M. Villar, D. Vega, A. Sanz, A. Nogales, and D. Rueda, "Synthesis and morphology of model PS-b-PDMS copolymers," Journal of Polymer Science Part A: Polymer Chemistry, vol. 48, no. 14, pp. 3119-3127, 2010. [64] J. H. Chu, P. Rangarajan, J. L. Adams, and R. A. Register, "Morphologies of strongly segregated polystyrene-poly(dimethylsiloxane) diblock copolymers," Polymer, vol. 36, no. 8, pp. 1569-1575, 1995. [65] N. Politakos, E. Ntoukas, A. Avgeropoulos, V. Krikorian, B. Pate, E. Thomas, and R. Hill, "Strongly segregated cubic microdomain morphology consistent with the double gyroid phase in high molecular weight diblock copolymers of polystyrene and poly (dimethylsiloxane)," Journal of Polymer Science Part B: Polymer Physics, vol. 47, no. 23, pp. 2419-2427, 2009. [66] J. R. Sargent and W. P. Weber, "Novel synthesis of triblock PDMS-PS-PDMS copolymers," Journal of Polymer Science Part A: Polymer Chemistry, vol. 38, p. 482, Feb. 2000. [67] K. Almdal, K. Mortensen, A. J. Ryan, and F. S. Bates, "Order, Disorder, and Composition Fluctuation Effects in Low Molar Mass HydrocarbonPoly(dimethylsiloxane) Diblock Copolymers," Macromolecules, vol. 29, pp. 5940-5947, Jan. 1996. [68] X. Wang, E. E. Dormidontova, and T. P. Lodge, "The Order-Disorder Transition and the Disordered Micelle Regime for Poly(ethylenepropylene- b -dimethylsiloxane) Spheres," Macromolecules, vol. 35, pp. 9687-9697, Dec. 2002. [69] D. Schwahn, H. Frielinghaus, K. Mortensen, and K. Almdal, "Temperature and Pressure Dependence of the Order Parameter Fluctuations, Conformational Compressibility, and the Phase Diagram of the PEP-PDMS Diblock Copolymer.," Physical Review Letters, vol. 77, pp. 3153-3156, Oct. 1996. [70] M. Uddin, C. Rodriguez, A. L6pez-Quintela, D. Leisner, C. Solans, J. Esquena, and H. Kunieda, "Phase behavior and microstructure of poly (oxyethylene)-poly (dimethylsiloxane) copolymer melt," Macromolecules, vol. 36, no. 4, pp. 1261-1271, 2003. [71] W. Li and B. Huang, "Relationship between crystallization of the PDMS block and morphology of poly(butadiene-b-dimethylsiloxane)," Journal of Polymer Science Part B: Polymer Physics, vol. 30, pp. 727-732, June 1992. 122 [72] A. Ciolino, L. Gomez, D. Vega, M. Villar, and E. Valles, "Synthesis and physicochemical characterization of a well-defined poly(butadiene 1,3)-blockpoly(dimethylsiloxane) copolymer," Polymer, vol. 49, pp. 5191-5194, Nov. 2008. [73] C. Veith, R. Cohen, and A. Argon, "Morphologies of poly (dimethylsiloxane)-nylon-6 diblock copolymers and blends," Polymer, vol. 32, no. 9, pp. 1545-1554, 1991. [74] E. W. Cochran, D. C. Morse, and F. S. Bates, "Design of ABC Triblock Copolymers near the ODT with the Random Phase Approximation," Macromolecules, vol. 36, pp. 782-792, Feb. 2003. [75] J. Brandrup, E. H. Immergut, E. A. Grulke, A. Abe, and D. R. Bloch, eds., Polymer Handbook, vol. 7. New York: John Wiley & Sons Inc, 4th ed., 2005. [76] W. Lee, J. Yoon, H. Lee, and E. Thomas, "Direct 3-D Imaging of the Evolution of Block Copolymer Microstructures Using Laser Scanning Confocal Microscopy," Macromolecules, vol. 40, pp. 6021-6024, Aug. 2007. [77] L. Fetters, D. Lohse, D. Richter, and T. Witten, "Connection between polymer molecular weight, density, chain dimensions, and melt viscoelastic properties," Macromolecules, vol. 27, no. 17, 1994. [78] L. Leibler, "Theory of microphase separation in block copolymers," Macromolecules, vol. 13, pp. 1602-1617, Nov. 1980. [79] D. Leisner, M. Uddin, A. Lopez-Quintela, T. Imae, and H. Kunieda, "Mesophase Morphologies of Silicone Block Copolymers in a Selective Solvent Studied by SAXS," in Self-Organized Surfactant Structures (T. Tadros, ed.), pp. 161-174, Weinheim: Wiley-VCH Verlag GmbH & Co. KgaA, 2010. [80] J. Cheng, C. Ross, H. Smith, and E. Thomas, "Templated Self-Assembly of Block Copolymers: Top-Down Helps Bottom-Up," Advanced Materials, vol. 18, pp. 25052521, Oct. 2006. [81] C. R. Stewart-Sloan and E. L. Thomas, "Interplay of symmetries of block polymers and confining geometries," European Polymer Journal, vol. 47, pp. 630-646, Apr. 2011. 123