Chemical Engineering

UC Berkeley

Teaching

Overview

Our research program is based on soft microstructured materials such as block copolymer melts, polymer blends, and microemulsions. Principles of solid state physics provide a recipe for creating soft materials. The modulus of an ordered phase is inversely proportional to the thermally driven mean square displacement of the ordered units and the lattice spacing. The soft materials that we study are thus characterized by large lattice spacing ranging from nanometers to microns, and large fluctuations wherein the magnitude of the mean square displacement approaches, and sometimes exceeds, the lattice spacing. These characteristics often lead to slow relaxation processes, which, in turn, enable fundamental investigations of non-equilibrium phenomena such as nucleation.

We specialize in the study of soft structures that self-assemble from the liquid state. The self-assembled nature of these structures has important consequences-they form spontaneously and do not require machining or lithography. The structure need not be permanently fixed; it can be altered by changing external variables such as temperature, pressure, and electric fields. Subtle changes in the external variables could lead to profound changes in the mechanical, optical and electrical properties of the material if they are accompanied by a microstructural transition. These changes could be accomplished repeatedly and reversibly if the structures are at equilibrium. These materials thus have the potential of performing complex functions, if we can understand the physical origins of their complex responses and control them to produce useful results.

Our program is concerned with both synthesis and characterization of soft polymer materials. We measure their response to changes in external variables by a variety of in-situ probes such as small angle neutron scattering and depolarized light scattering. Our objective is to develop a fundamental understanding of the relationship between the measured response and the molecular structure of the components that comprise our system.

Interdisciplinary Laboratory for Soft Condensed Matter Studies

In July 2000, my students and I took upon the task of building an inter-disciplinary laboratory for synthesis and characterization of soft condensed matter. Making these facilities available to other groups at Berkeley and neighboring institutions, and establishing long-term collaborations was an important goal of this endeavor. I have co-advised students and post-docs with the following College of Chemistry Faculty: Alex Bell, Arup Chakraborty, Jean Frechet, Alex Liddle, John Newman, and Clay Radke. Students from the following groups have made use of our facilities: Alex Bell (Chemical Engineering), Alex Katz (Chemical Engineering), Susan Muller (Chemical Engineering), Carolyn Bertozzi (Chemistry), Jean Frechet (Chemistry), Don Tilley (Chemistry), Ronald Gronsky (Materials), Kevin Healy (Bioengineering), John Kerr (LBNL), Alice Gast (MIT), and Bob Waymoth (Chemistry, Stanford).

Recent Research Accomplishments

1. Microstructured Morphologies and Bulk Properties in Block Copolymer Membranes for Fuel Cell Applications

The fact that wet membranes dry up when heated in air appears to be an inescapable fact of life. The purpose of our research is to establish a new systematic methodology for controlling the water retention of polymer electrolyte membranes. Block copolymer membranes comprising hydrophilic phases with widths ranging from 2 to 5 nm become wetter as the temperature of the surrounding air is increased at constant relative humidity. The widths of the moist hydrophilic phases were measured by cryogenic electron microscopy experiments performed on humid membranes (Figure 1). Simple calculations suggest that capillary condensation is important at these length scales. The correlation between moisture content and proton conductivity of the membranes is demonstrated and these membranes have the potential to increase the operating temperature of polymer electrolyte fuel cells (PEFCs).

2. Order-Disorder Transition in Copolymers with Quenched Sequence Disorder (Eitouni, Rappl, Gomez)

Amphiphilic molecules have proven to be excellent materials for studying the physics that underlie the phase behavior of complex systems. For example, studying the phase behavior of block copolymers with ordered sequences and well-defined molecular architectures has yielded deep insights into the spontaneous formation of ordered microphases including lamellae, hexagonally packed cylinders decorated with rings, etc. Random copolymers, wherein two or more kinds of monomers are connected together in a disordered sequence, are a class of amphiphilic molecules with quenched disorder. A large number of high volume and commercially important materials such as styrene-butadiene rubber (SBR), acrylonitrile-butadiene-styrene copolymers (ABS), Nafion, and polyurethanes, belong to this category. While evidence for the presence of order has been obtained in some random copolymers, experimental data that probe the thermodynamic properties of copolymers with quenched disorder do not exist. We are not aware of any random copolymer system wherein the order-disorder transition (ODT) has been determined experimentally. We have recently obtained the first experimental data set that establishes the existence and nature of order-disorder transitions in copolymers with quenched disorder.
The order-disorder transition is announced by subtle and continuous changes in both birefringence and SAXS signals. The characteristic length scale of the concentration fluctuations in disordered RGCs decreases as the ordering transition is approached, and is a universal function of reduced temperature. This is a surprising result because the characteristic length scale of the concentration fluctuations of most systems increases as the phase transition is approached (e.g. critical polymer blends and block copolymers). This data is shown in Figure 1a. The experimentally determined location of the order-disorder transition is in reasonable agreement with field theoretic predictions made with no adjustable parameters. Kinetic barriers arising due to the presence of quenched disorder and architectural complexity limit the extent of long-range order in these systems. This is vividly seen in transmission electron micrographs obtained from weakly ordered RGCs, as shown in Figure 1b. These signatures of the ODT of copolymers with sequence disorder are, however, dramatically different from those seen in copolymers with ordered sequences.

Figure 1a. Decrease of the characteristic length of the disordered concentration fluctuations in a randomly grafted copolymer as the ordering transition is approached. Scattering vector of the SAXS peak, qpeak, as a function of reduced temperature, TODT/T, for a polybutadiene-graft-polystyrene copolymer in the neat state (circles) and with added solvent (7 and 22 wt. % represented by squares and plusses, respectively). Characteristic length of the disordered concentration fluctuations is 2p/qpeak, T is the sample temperature and TODT is the order-disorder transition temperature of each system. The curve represents the least-squares quadratic fit through the data. The inset shows qpeak versus T.

Figure 1b. Typical transmission electron micrograph of BgS[0] at room temperature showing a highly defective lamellar phase. Dark regions are polybutadiene-rich. Lamellae are more coherent in some localized regions (e.g. within the oval).

3. Effect of Chemical Oxidation on the Self-Assembly of Organometallic Block Copolymers (Eitouni)

Controlling the spontaneous formation of ordered domains in soft materials such as block copolymers is a subject of considerable current interest. These systems may enable the development of stimuli-responsive materials for applications such as actuators and photonics due to the reversible nature of order formation. However, the stimuli that are typically used to control the morphology of block copolymers, e.g. temperature, pressure, solvent concentration, etc., are not well-suited for applications that require repeated switching between morphologies. In this paper we demonstrate that the oxidation state of redox-active species incorporated in block copolymer chains can be used to control order formation. Since the state of redox active species can be controlled electrochemically, this work lays the foundation for using electrochemical potential to control the self-assembly of block copolymers.
Poly(isoprene-block-ferrocenyldimethylsilane) (IF) and poly(styrene-block-ferrocenyldimethylsilane) (SF) copolymers were synthesized following the procedure of Ian Manners and coworkers. Oxidation of the ferrocene moiety to the ferrocenium cation was accomplished by reaction with silver nitrate. The chemical structure of a partially oxidized IF copolymer is presented in Figure 2a. The order-disorder transition temperatures of a series of SF and IF samples with oxidation amounts ranging from 0 to 8% were determined by small angle X-ray scattering. The results of our analysis are summarized in Figure 2b where the order-disorder transition temperature (TODT) is plotted versus percent oxidation for both the SF and IF copolymers. It is evident that TODT decreases systematically with increasing levels of oxidation in both cases.
We have demonstrated that chemical oxidation of SF and IF copolymers leads to stabilization of the disordered state. Changing the redox state of 8% of the ferrocene moieties results in a reduction of the order-disorder transition temperature by as much as 40 oC. We suggest that this stabilization is due to the entropy of a very small fraction of dissociated counter ions that are introduced during the oxidation step. An interesting property of ferrocene is the fact that its oxidation state can be altered reversibly by the application of small electric fields (e.g. ± 2V/cm). Pioneering work by Abbott and coworkers used this to control the self-assembly of small molecule amphiphiles containing ferrocene. Our results suggest that the self-assembly of ferrocene-containing block copolymers may be controlled by the application of electrochemical potential in a suitable electrolyte.

Figure 2a. Schematic representation of the chemical structure of a partially oxidized IF diblock copolymer.	Figure 2b. SAXS results. Phase diagram of SF and IF polymers as a function of percent oxidation.

4. Interfacial Kinetics of Polymer Blends (Reynolds co-advised by Radke)

In most binary polymer blends, phase separation of the constituent polymers is the equilibrium state and unfavorably affects the properties of the blend. The properties of the blend are determined not only by the polymers themselves but by the morphology and length-scale of the dispersion and by the properties of the interface. By adding interfacially active polymers such as diblock copolymers to the blend, the interfacial tension can be reduced and in some cases may be sufficiently lowered to allow the formation of a microemulsion or ordered structure. Understanding the dynamics of both the phase separated dispersions and the thermodynamically stable microemulsions, is largely a question of understanding the dynamics of the interfaces.
We use an A / A-B / B ternary polymer blend to study interfacial dynamics, where A is saturated polybutadiene with 63% 1,2 addition and B is saturated polybutadiene with 90% 1,2 addition. The equilibrium properties of the interface were investigated using 450 nm thick homopolymer trilayers (A | B | A), created by spincoating. The top layer was in fact a blend of homopolymer A and the diblock copolymer A-B, however after several days at room temperature the diblock copolymer had diffused throughout the trilayer and reached an equilibrium concentration profile. The diblock copolymer was labeled with deuterium so that the concentration profile through the film could be measured using dynamic secondary-ion mass spectroscopy. By making various films with different amounts of the diblock copolymer, we constructed an adsorption isotherm for the system. Self-consistent field theory was also used to produce a theoretical adsorption isotherm, using parameters calculated from independent small-angle neutron scattering measurements and no adjustable constants. The agreement between experiment and theory was found to be quantitative (Figure 3a).
Once the equilibrium properties had been characterized we moved on to the dynamic characteristics of the interface, specifically the rate of adsorption and desorption of the diblock copolymer from the A/B interfaces. The films were prepared in the manner described earlier, however rather than letting the films come to equilibrium the secondary-ion mass spectroscopy measurements were made a set amount of time after the sample was prepared, ranging from half an hour to five days. In this way the time dependence of the concentration profile was found (Figure 3b). At short times the diblock copolymer was observed to deplete from the first layer and adsorb at the first A/B interface, while at longer times a slower exchange between the first and second A/B interfaces was seen. This slow equilibration gave us an estimate of the adsorption time constant of 1 day.

Figure 3a. Adsorption isotherm measured experimentally using dynamic secondary-ion mass spectroscopy and predicted theoretically using self-consistent field theory.	Figure 2b. SAXS results. Phase diagram of SF and IF polymers as a function of percent oxidation.

5. Designing Balanced Surfactants for the Organization of Immiscible Homopolymers (Ruegg)

Mixtures of oil and water organize into a variety of structures, such as vesicles, bilayers, lamellae, and microemulsions, upon addition of a surfactant. In a mixture of oil, water, and a non-ionic surfactant, microemulsions have been shown to form with as little as 6% of the non-ionic surfactant in that mixture. This amount of surfactant can be reduced to 3% with the addition of small amounts of large molecule surfactants. We would like to minimize the amount of surfactant required in polymeric systems for the formation of microemulsions through the careful design of diblock copolymer surfactants.
Traditionally, A-B diblock copolymers have been used as surfactants for A and B homopolymers. However, experimentally this approach has only been effective with homopolymers A and B that are either miscible or weakly segregated. Also, the minimum amount of an A-B diblock that will organize A and B homopolymers has experimentally been around 10% (this has also been shown to be true using mean-field theories). We would like to design our polymeric surfactant based upon the non-ionic surfactants for oil and water toward the goal of organizing the immiscible polymers A and B with small amounts of this effective surfactant.
Our surfactant design consists of an A-C diblock copolymer, in which the C-block has attractive interactions with the B homopolymer and repulsive interactions with the A homopolymer. Thus the binary interactions in this polymeric system (A/B, A/C, and B/C) are analogous to that in the oil/water/non-ionic surfactant system.
We have studied the phase behavior of an A/B/A-C system with the following components: component A is a saturated polybutadiene with 90% 1,2-addition (sPB90), component B is polyisobutylene (PIB), and the diblock copolymer consists of block A (also sPB90) and block C which is a saturated polybutadiene with 63% 1,2-addition (sPB63). We synthesized the polybutadiene polymers through anionic polymerization and saturated the double bonds with either hydrogen or deuterium gas, thus giving our system a neutron contrast. All of the binary chi parameters that govern the thermodynamics of this A/B/A-C system have been measured and are summarized in Figure 4a. The binary interactions (A/B, A/C, and B/C) are analogous to that in the oil/water/non-ionic surfactant system (in the aqueous system, component A is oil, component B is water, and component A-C is an alkyl polyglycol ether surfactant). The A-C diblock copolymer will thus have "balanced" interactions with the A and B homopolymers due to the combination of attractive and repulsive interactions.
We have conducted a study utilizing homopolymers that are weakly segregated (chi*N ~ 2). Through neutron and light scattering studies we have observed the formation of lamellae and spherical or bicontinuous microemulsions. A phase diagram summarizing these results is shown in Figure 4a. This phase diagram looks very similar to the upper half of the fish phase diagram that is observed in oil/water/non-ionic surfactant systems. However, it is the gray region in Figure 4b that we are really interested in. Can we organize these weakly segregated polymers with small amounts of diblock copolymer? Preliminary experiments and theoretical predictions suggest that we can. We have recently observed a microemulsion peak from a blend with the same components in Figure 2 and only 5% diblock copolymer. Also, these experiments are in agreement with theoretical predictions from self-consistent field theory. While this study of minimizing the amounts of surfactant required to organize weakly segregated or immiscible polymers is far from complete, the initial findings look promising.
In addition to the polymers utilized in the above study, we have synthesized a series of A, B and A-C polymers with a variety of molecular weights. We thus will study surfactancy as a function of segregation of the homopolymers as well as that of the block copolymer. We will systematically make a variety of blends varying the size of each of the components as well as the amount of diblock and will use neutron scattering to determine the structure of each of the blends at a variety of temperatures. Thus we not only wish to minimize the amount of surfactant required for weakly segregated blends, but also for blends with all degrees of segregation.

Figure 4a. Binary chi parameters	Figure 4b: A/B/A-C Phase Diagram

6. Initial Stages of Nucleation in liquid-liquid Polymer Blends (Rappl)

Nucleation and growth is the universal process that underlies phase transformations from a metastable non-equilibrium state to a stable equilibrium state. Classical nucleation theory, which has been used to describe diverse phenomena such as crystallization, boiling, and liquid-liquid phase separation, is based on the assumption that these phase transitions are triggered by the formation of microscopic nuclei. Growth of the new phase requires spontaneous formation of nuclei that are larger than a critical size. We refer to this critical length scale as Rc. Classical theory assumes that the nuclei are small regions in space that have all of the characteristics of the stable state. The relevant characteristics, which include concentration, density, crystal structure, etc., depend on the particular phase transition under consideration. Nuclei with sizes smaller than Rc or with characteristics that are different from that of the new equilibrium phase are non-viable and are predicted to decay. The evolution of a metastable state into a stable one is thus characterized by a complex collection of structures with differing sizes and compositions.

We have endeavored to study the early stages of nucleation for liquid-liquid phase separation of a binary polymer blend. The large molecular weights and associated chain entanglement effects lead to slow dynamics such that the evolution of the phase separating structure can be observed by time-resolved small angle neutron scattering (SANS). When an off-critical homogeneous blend is quenched (by either a reduction in temperature or an increase in pressure) into the two-phase region of the phase diagram, the scattered intensity increases with time at small scattering vectors but remains time invariant at large scattering vectors; the demarcation between these two regimes is labeled the critical scattering vector, qc, Fig. 5a. As the size of the emerging structures is inversely related to the scattering vectors at which they scatter, the existence of qc is evidence for a critical size below which phase separated structures do not exist, Rc = 1/qc.

We have studied the characteristics of the early stages of nucleation over a wide range of quench depths. Near the binodal, the activation barrier is too large for nucleation to be observed on experimental time scales. Nucleation is first observed for quench depths approximately midway between the binodal and spinodal. Spinodal decomposition, rather than nucleation and growth, is the predicted mechanism for phase separation below the spinodal, and extensions to classical nucleation theory near the spinodal are necessary to reconcile the differences between these two mechanisms (the structures created during the early stages of spinodal decomposition will not generally display equilibrium characteristics). Classical nucleation theory predicts that both the activation barrier and the critical nucleus size decrease to zero as the spinodal is approached, whereas the theory of spinodal decomposition predicts the divergence of the critical size at the spinodal. Extensions to classical nucleation theory predict that the activation barrier decreases continuously as the spinodal is approached, but that the critical nucleus size decreases with quench depth near the binodal, reaches a finite minimum, and then increases with quench depth ultimately diverging at the spinodal. Figure 5b shows our experimental results for the dependence of the critical nucleus size on quench depth. We see no evidence for a diverging length scale at the spinodal; note that the critical nucleus size is normalized Rc/Rg (Rg for the polymers used here is 16 nm) and that the abscissa is a normalized Flory-Huggins chi parameter, Quench Depth = chi /chis, defined such that Quench Depth = 1 is the spinodal. Included in this figure are the theoretical predictions based on the Cahn-Hilliard extension to classical nucleation theory near the spinodal.

Figure 5a. SANS profiles during nucleation	Figure 5b. Dependence of the critical nucleus size on quench depth

7. Order-Disorder Transitions in Block Copolymer Networks (Gomez)

The order-disorder transition in block copolymers has many potential useful applications, but because the disordered phase exhibits liquid-like behavior it limits the ability of the polymer to maintain structural integrity. Often, to capture the microstructure and macroscopic properties a polymer is cross-linked so that the resulting network inhibits macro-scale diffusion and resists deformation. We have found that by selectively cross-linking the polyisoprene chains of a polystyrene-polyisoprene block copolymer we can preserve the order-disorder transition, resulting in a diffusionless temperature-driven phase transition between an ordered soft solid and a softer, disordered solid.
Although the phase transition can only be preserved for a precise range of cross-linking densities, as seen in Figure 6a, it is surprising that we can cross-link the block copolymer in the disordered phase (160 degrees C) and still preserve the ordered phase at low temperatures (< 100 degrees C). Because of the symmetrical morphology of this polymer, this implies that we can create a random three-dimensional network and “squeeze” it into two dimensional planes, or lamellae. This constraint in dimensionality of the ordered-phase has a tremendous effect on the symmetry of the lamellae created at low temperatures, as seen in Figure 6b. Both depolarized light-scattering and transmission electron microscopy (TEM) results suggest that grains of the ordered phase are highly anisotropic with very little coherence parallel to the lamellae and considerable long range order perpendicular to the lamellae. The reason for the appearance of this unusual morphology is not clear.
We have thus created a responsive solid that undergoes a diffusionless, solid-solid phase transition that is robust and has many potential applications for nano-scale mechanical or optical switches. However, our interest remains in the rich novel thermodynamics that can be studied with cross-linked block copolymers.

Figure 6a. Phase Behavior of a polystyrene- Figure 6a. Phase Behavior of a polystyrene- has a 0.5 volume fraction.	Figure 6b. TEM image of a polystyrene-polyisoprene block copolymer network.

Acknowledgement

We gratefully acknowledge the following agencies for funding our work:

·This material is based upon work supported by the National Science Foundation under Grant No. 0305711 (Division of Materials Research, Division of Chemical and Transport Systems, Division of Electrical and Communications Systems)
·Department of Energy (Laboratory Directed Research and Development Program)
·American Chemical Society (Petroleum Research Fund)
·Tyco Electronics

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.