Soft Matter Physics.   Polymers.   Experiments.

Soft Matter Physics is the study of physical systems that are strongly influenced by temperature and mechanical forces.  These typically involve materials with complex many-body interactions such as polymers, liquids, liquid crystals, colloids, granular materials, and many biological systems.  How molecular packing and conformations change with temperature arises from an interplay of entropic and enthalpic considerations defining available thermodynamic free energy minima.  However, the dynamics of these complex systems can often be very slow such that the system spends most of its time kinetically trapped in non-equilibrium states.

Polymers are long chain molecules that frequently exhibit unique properties because of their macromolecular nature.  Entanglements between these large polymer molecules leads to slow and complex dynamics, which means that polymers can easily become trapped in non-equilibrium states during processing and fabrication of materials.  Polymers are ubiquitous in our lives.  Synthetic polymers make up much of the objects we use in daily life (nylon, polyester, styrofoam, plexiglass), while natural polymers make up much of who we are (DNA, proteins, polysaccharides) and much of our natural world (cellulose, cotton, wood, paper).  Understanding how polymer molecules behave in different environments, especially at nanoscale dimensions near interfaces, are fundamental to the behavior of many materials and to the development of new technologies.

Experiments are used to test our understanding of how systems behave.  We want to understand the effects of forces, interfaces, and other perturbing influences on the properties of polymer materials.

Our lab’s approach to research is to develop and use unique experimental methods to study previously inaccessible phenomena in order to provide molecular-level insight into new and existing polymer materials.  We frequently tackle long-standing issues in the field by developing a simplified sample geometry to address the phenomenon in question.

Over the years our research has focused on problems associated with polymer glasses, interfacial phenomena, polymer blends and miscibility.  Below is a list of research topics our lab has worked on.  Click on each link to learn more about our lab’s contributions to these topics.

• Understanding coupled glass transition dynamics across polymer interfaces – Recent work from our lab has uncovered broad profiles in local Tg(z) across glassy-rubbery polymer interfaces demonstrating glass transition dynamics can become coupled between neighboring polymer domains over long distances.  Current efforts are aimed towards understanding the cause of this unexpected phenomenon.

• Impact of surface bound chains – We are working on understanding the unexpected large increases in local Tg(z) that occur next to surfaces with tethered chains.  Grafted and adsorbed polymer chains are widely used to modify interfacial interactions in polymer materials and nanocomposites, but the underlying mechanisms by which this causes property changes to the neighboring polymer matrix are not well understood.

• Developing nanoscale measures of local polymer properties – An ongoing focus of our lab is the development of new experimental methods to measure localized polymer properties at nanoscale dimensions in an effort to understand how their perturbed behaviors near interfaces are interrelated.

• Nanoscale property changes in thin polymer films – Ongoing research in our lab is working to understand how various different material properties (glass transition temperature Tg, physical aging, refractive index and density, …) are altered in thin films, in particular at the local nanoscale level as a function of distance from interfaces.

• Tuning miscibility of polymer blends with electric fields – Experimental results demonstrating electric fields enhance the miscibility of PS/PVME blends.  By simply turning the electric field on and off, the blend can be repeatedly jumped from the one-phase to two-phase region at constant temperature.

• Stress during vitrification impacts stability of polymer glasses – Using ellipsometry we have characterized how the physical aging rate β of glassy polymer films can be changed by the application of stress to the film during vitrification (thermal cooling) of the glass.

• Polymers in art conservation – In collaboration with Chief Conservator Renée Stein at Emory’s Carlos Museum, our lab is characterizing polymers used in the preservation and restoration of artworks in an effort to understand how their preparation conditions impact their properties.

Historically the treatment of polymer blends expounded on in textbooks treated individual domains within blends as retaining their bulk component properties such that to a first approximation the blend’s macroscopic properties can be considered to be a weighted average of the individual component properties based on the fraction of each component.  However, it has been recognized for some time that the morphology of the blend plays a large role in determining the blend’s macroscopic properties and material performance (a central premise behind structure-property relations).  Yet, even in this treatment the material differences are frequently assumed to be localized at the polymer-polymer interfaces between domains with the domains themselves still thought to exhibit bulk properties.

It is common to think that material properties follow the local composition, which transitions sharply within a few nanometers at the interface between immiscible polymers.  However, this traditional view is not consistent with studies on polymers in nanoconfined systems such as thin films which over the past two decades have demonstrated large property differences caused by interfacial perturbations when film thicknesses or domain sizes become comparable to approximately 100 nm (see Nanoscale property changes in thin polymer films).  Given the prevalence of new synthesis and processing technologies that can create so-called nanostructured blends with domain sizes less than 100 nm, there is a pressing need to understand how local properties change across and near polymer-polymer interfaces.

Recent work from our lab has uncovered broad profiles in local Tg(z) across glassy-rubbery polymer interfaces demonstrating glass transition dynamics can become coupled between neighboring polymer domains over long distances.  Current efforts in our lab are aimed towards understanding the cause of this unexpected phenomenon.

The glass transition temperature (Tg) is a fundamental property that defines a material’s transition on cooling from an equilibrium rubbery melt (typical rubbery modulus ~MPa) to a nonequilibrium glassy solid (typical glassy modulus ~GPa).  We are interested in understanding how the local value of Tg and other related properties change across a polymer-polymer interface from one domain to another.

 
Communication:   Experimentally Determined Profile of Local Glass Transition Temperature Across a Glassy-Rubbery Polymer Interface with a Tg Difference of 80 K
Roman R. Baglay and Connie B. Roth,
Journal of Chemical Physics   2015, 143, 111101.

Within a glassy-rubbery model system, we have measured the dynamical profile of the local glass transition temperature Tg(z) as a function of position z across a polymer-polymer interface using a localized fluorescence method.  A canonical glassy polymer, polystyrene (PS) with nominal Tg = 100 °C, is placed in contact with a canonical rubbery polymer, poly(n-butyl methacrylate) (PnBMA), with nominal Tg = 20 °C.  Both domains are made large enough to avoid other interfacial interactions associated with the sample boundary.  The local Tg(z) of 10-15 nm thick pyrene-labeled layers are then measured at different positions z from the polymer-polymer interface (z = 0) within both domains on either side of the interface.  Clearly the local Tg(z) value will transition from one extreme to the other:   from 100 °C far from the interface on the glassy PS side to 20 °C far from the interface on the rubbery PnBMA side.  What is surprising is that the dynamical Tg(z) profile is very broad, spanning 350-400 nm from one bulk Tg value to another, not at all matching the composition profile as would be expected based on traditional textbook teachings.  In addition, the dynamical Tg(z) profile is highly asymmetric, spanning much further into the glassy PS side than the rubbery PnBMA side.  Theoretical considerations suggest this dynamical length scale should be related to that associated with cooperative motion in glasses.

 
Local Glass Transition Temperature Tg(z) of Polystyrene Next to Different Polymers:   Hard vs. Soft Confinement
Roman R. Baglay and Connie B. Roth,
Journal of Chemical Physics   2017, 146, 203307.

Local Glass Transition Temperature Tg(z) Profile in Polystyrene next to Polybutadiene With and Without Plasticization Effects
Benjamin L. Kasavan, Roman R. Baglay, and Connie B. Roth,
Macromolecular Chemistry and Physics   2018, 219, 1700328.

We have demonstrated that this broad and asymmetric Tg(z) profile is common to a range of polymer systems.  For polystyrene (PS) domains placed next to a variety of different weakly immiscible polymers with Tgs both lower and higher than that of PS, we find that the local Tg(z) profile exhibits a common behavior consistently recovering bulk Tg at a distance of z = 100-125 nm from the interface when PS is next to a “hard” versus z = 225-250 nm when PS is next to a “soft” interface.

One of the key findings from our 2017 Journal of Chemical Physics paper is that the broad, asymmetric Tg(z) profile only develops when the polymer-polymer interface is annealed to equilibrium.  If the measurement is done with minimal of annealing of the interface, the Tg(z) profile is much sharper.  This indicates that some factor(s) during polymer-polymer interface formation causes this broad coupling of dynamics across the interface.  We pursued a series of experiments to decipher which of these factors was important:   broadening of the interface, chain connectivity across the interface, and roughening of the interface.  These were recently reviewed in:
Polymers Under Nanoconfinement:   Where Are We Now in Understanding Local Property Changes?
Connie B. Roth, Chemical Society Reviews   2021, 50, 8050-8066.

 
Local Glass Transition Temperature Tg(z) Within Polystyrene Is Strongly Impacted by the Modulus of the Neighboring PDMS Domain
Yannic J. Gagnon and Connie B. Roth,
ACS Macro Letters   2020, 9, 1625-1631.

The differences between hard versus soft interfaces, along with theoretical efforts in the literature, suggests that the modulus of the neighboring domain may also play an important role in determining this behavior.  As such we have collected Tg(z) profiles in polystyrene (PS) next to polydimethylsiloxane (PDMS) with varying crosslink density.  The local Tg(z) in PS at z = 50 nm from the PS/PDMS interface varies by more than 40 K with varying PDMS modulus from 0.9 to 2.6 MPa.  However, the Tg(z) profiles also persist out to a much shorter distance (z = 65-90 nm), which we attribute to the much narrower polymer-polymer interface formed between this strongly immiscible system, as the χ parameter for PS/PDMS is an order of magnitude larger that the weakly immiscible systems shown above.

 
Experimental Study of the Influence of Periodic Boundary Conditions:   Effects of Finite Size and Faster Cooling Rates on Dissimilar Polymer-Polymer Interfaces
Roman R. Baglay and Connie B. Roth,
ACS Macro Letters   2017, 6, 887-891.

Another key factor we are pursuing with ongoing work is the role of finite domain size.  Our first look at this found that the presence of a second PS/PnBMA interface already began perturbing the measured Tg(z) values in PS when the PS domain became less than approximately 400 nm.  The Tg(z) profile across a 300 nm PS layer did not correspond simply to the additive effects of two PS/PnBMA interfaces suggesting that finite size effects also modify the behavior.  We believe understanding these effects is essential to bridging the gap between our earlier results on single interfaces with semi-infinite domains and block copolymer systems.

Surface bound chains such as grafted and adsorbed polymer chains are widely used for modifying interfacial interactions in polymer materials and nanocomposites, but the mechanisms by which the changes to material properties are conferred to the system are not well understood.

 
Optimizing the Grafting Density of Tethered Chains to Alter the Local Glass Transition Temperature of Polystyrene near Silica Substrates:   The Advantage of Mushrooms over Brushes
Xinru Huang and Connie B. Roth,
ACS Macro Letters   2018, 7, 269-274.

In this first study on grafted chains by our group, we demonstrated that end-tethered polystyrene (PS) chains with a molecular weight Mw = 100 kg/mol increased the local glass transition temperature Tg(z) of PS matrices by several decades and persisted out to distances of z = 100-125 nm away from the substrate interface.  Surprisingly, the largest local Tg(z=0) increase next to the substrate interface of +50 K above the bulk Tg of PS occurred for a very low grafting density of σ = 0.011 chains/nm^2, within the mushroom-to-brush transition regime.  We believe the good interpenetration that can be obtained between the grafted and matrix chains at these low grafting densities is central to creating the large Tg increases observed.  These results are extremely puzzling because typically the glass transition in polymers is not associated with chain connectivity as packing frustration occurs at the segmental level.

 
End-Tethered Chains Increase the Local Glass Transition Temperature of Matrix Chains by 45 K Next to Solid Substrates Independent of Chain Length
James H. Merrill, Ruoyu Li, and Connie B. Roth, ACS Macro Letters   2023, 12, 1-7

Following up on this work, we have shown that low grafting densities of end-grafted PS chains can substantially increase the local glass transition temperature Tg(z=0) of PS matrices next to silica susbstrates by +45 K, regardless of end-group grafting chemistry.  Surprisingly, the magnitude of the Tg increase is independent of grafted chain length from 9 kg/mol to 200 kg/mol and grafting density σ from 0.003 to 0.33 chains/nm^2 spanning the mushroom-to-brush transition regime.  The large tens-of-degree increase in local Tg(z=0) observed in this athermal system by only the immobilization of chain ends from covalent bonding suggests a chain connectivity mechanism that substantially increases the local activation energy required for cooperative rearrangements.

Ongoing work in our lab is investigating this unexpected phenomenon further by studying the local glass transition temperature Tg(z) next to the substrate interface as a function of tethered chain length and grafting density to identify the underlying factors controlling this behavior.

 
Review and Reproducibility of Forming Adsorbed Layers From Solvent Washing of Melt Annealed Films
Michael F. Thees, Jennifer A. McGuire, and Connie B. Roth,
Soft Matter   2020, 16, 5366-5387.

It is well known that in solution, polymer chains naturally segregate and adsorb to interfaces.  In polymer melts, however, it is less clear the extent to which chains can similarly adsorb to interfaces.  The study of such adsorbed chains are complicated by the difficulty of accessing and identifying the surface bound chains.  Efforts to investigate adsorbed layers in melt films and polymer nanocomposites frequently rely on solvent washing to expose such near-surface, “bound layer” chains. 

In this publication, we review recent literature efforts to quantify the residual adsorbed layer thickness h_ads(t) measured after a given solvent washing procedure as a function of annealing time t of the film at an elevated temperature prior to this solvent rinse, a procedure frequently called the “Guiselin’s experiment”.  We identify and experimentally test a common protocol for forming adsorbed layers h_ads(t) from solvent washing melt films, and find them to be far less reproducible and reliable than implied in the literature, strongly dependent on solvent washing and substrate cleaning conditions.  This review also summarizes literature understanding developed over several decades of study on polymer adsorption in solution, which experimentally demonstrated that polymer chains in solution are highly mobile, diffusing and exchanging on the surface even in the limit of strong adsorption, contradicting Guiselin’s assumption.  A number of open questions and implications are discussed related to thin films and polymer nanocomposites.

Ongoing work in our lab is studying how these different populations of surface bound chains alter the local glass transition temperature of neighboring polymer chains in bulk films, using these local Tg measurements to infer the adsorbed layer structure formed during melt annealing.

 
Experimental Study of Substrate Roughness on the Local Glass Transition of Polystyrene
Xinru Huang, Michael F. Thees, William B. Size, and Connie B. Roth,
Journal of Chemical Physics   2020, 152, 244901.

Another form of surface treatment is to vary the roughness of the interface.  For decades, computer simulations have shown that increased interface roughness leads to slower dynamics.  In this study, we investigated this experimentally by creating silica substrates with increasing roughness using a hydrogen fluoride vapor treatment.  We found that the local glass transition temperature Tg near silica substrates increased by 10 K with increasing roughness, but only for extremely large roughness scales not normally associated with the glass transition, leaving the mechanism for this observed behavior uncertain.

As part of our research into the nanoscale properties of polymer materials, we develop experimental methods that give us access to different types of material properties such as the glass transition temperature Tg, physical aging, modulus, density, and phase separation.  In some cases we are adapting existing methods to give us improved sensitivity at nanoscale dimensions, in other cases we are developing new methods to provide information about how localized polymer properties are perturbed near interfaces.  Overall, we are interested in understanding how different material properties that are related to each other in bulk systems are affected and related to each other in nanoscale materials.

 
Temperature Dependent Perylene Fluorescence as a Probe of Local Polymer Glass Transition Dynamics Yixuan Han and Connie B. Roth,
Soft Matter   2022, 18, 6094-6104.

We’ve investigated the temperature dependence of perylene’s fluorescence emission spectrum doped in bulk polymer matrices of poly(methyl methacrylate) (PMMA), polystyrene (PS), poly(2-vinyl pyridine) (P2VP), and polycarbonate (PC) films.  Indentifying a temperature-dependent self referencing region (SRR) in the spectra, we show how an intensity ratio can be defined to provide a local measure of the dynamic glass transition temperature Tg.  The temperature dependence of the intensity ratio reflects intermolecular collisions of the perylene dye with the surrounding polymers segments.  We find it follows a non-Arrhenius temperature depenedence above the glass transition temperature Tg, well described by the bulk Williams–Landel–Ferry (WLF) equation, while transitioning to an Arrhenius temperature dependence below Tg.
In collaboration with Simon Blakey’s group in Chemistry at Emory and grad student Mark Maust, our current efforts are focused on developing this perlyene dye into a new localized measure of the glass transition reflecting a dynamic Tg.

 
Physically Intuitive Continuum Mechanics Model for QCM:   Viscoelasticity of Rubbery Polymers at MHz Frequencies
Yannic J. Gagnon, Justin C. Burton, and Connie B. Roth,
Journal of Polymer Science   2022, 60, 244-257.

In our first collaboration between the Burton Lab and the Roth Lab, we have developed a continuum physics model that describes the propagation of shear waves through a polymer film caused by the resonance oscillation of a quartz crystal microbalance (QCM) sensor.  By solving the resulting set of coupled equations numerically, we can fit the experimental data for the resonance frequency Δfn and dissipation ΔΓn shifts as a function of harmonic number n, over an extended harmonic range approaching film resonance, with zero approximations.  This allows us to determine the frequency-dependent modulus of polymer films at MHz frequencies, modeled as linear on a log–log scale.  This first paper demonstrates application of this continuum physics model on rubbery polymer films of polybutadiene (PB) and polydimethylsiloxane (PDMS) films, showing excellent agreement with time–temperature shifted rheometry data from the literature.  We are now working on applying this method to glassy-rubbery bilayer films.
See:  Understanding coupled glass transition dynamics across polymer interfaces

 
Changes in the Temperature-Dependent Specific Volume of Supported Polystyrene Films with Film Thickness
Xinru Huang and Connie B. Roth,
Journal of Chemical Physics   2016, 144, 234903.

Ellipsometry measures the change in polarization of light as it reflects off of a sample.  By modeling the sample with an optical layer model using Fresnel equations, the film thickness h and refractive index n of the film can be determined to great precision.  Here we use ellipsometry to measure the temperature-dependent index of refraction of polystyrene (PS) films supported on silicon and investigate the validity of the commonly used Lorentz-Lorenz equation for inferring changes in density or specific volume from very thin films.  We conclude that the derivation of the Lorentz-Lorenz equation becomes invalid for very thin films as the film thickness approaches ∼20 nm, and that reports of large density changes greater than ±1% of bulk for films thinner than this likely suffer from breakdown in the validity of this equation or in the difficulties associated with accurately measuring the index of refraction of such thin films.  We do observe small shifts of 0.4 ± 0.2%, outside of our experimental error, in the density (inverse of specific volume) for thin films that occur simultaneously in both the liquid and glassy regimes uniformly together.  We have since gone on to investigate these effects in a number of different polymers.
See:  Nanoscale property changes in thin polymer films

 
Effect of Adjacent Rubbery Layers on the Physical Aging of Glassy Polymers
Phillip M. Rauscher, Justin E. Pye, Roman R. Baglay, and Connie B. Roth,
Macromolecules   2013, 46, 9806-9817.

We have also developed ellipsometry methods to measure the physical aging rate of polymer films (see below).  In this work, we developed fitting methods that would allow us to extract the physical aging rate of thin glassy layers within rubbery-glassy bilayers.  In addition, we refined an existing fluorescence method developed by the Torkelson Lab to measure the local glass transition temperature Tg using pyrene dyes covalently bonded to polymer chains.  This has allowed us to map profiles in the local glass transition temperature Tg(z) within a number of different systems.
See:  Understanding coupled glass transition dynamics across polymer interfaces
and Impact of surface bound chains

 
Characterization of Phase Separation of Polystyrene / Poly(vinyl methyl ether) Blends Using Fluorescence
Annika Kriisa, Sung S. Park, and Connie B. Roth,
Journal of Polymer Science, Part B: Polymer Physics   2012, 50, 250-256.

We have characterized for the first time changes in the fluorescence emission spectra of pyrene and anthracene dyes covalently bonded to polystyrene (PS) upon phase separation from poly(vinyl methyl ether) (PVME).  The specific chemical structure of the fluorescent labels is found to affect the measured phase separation temperature T_s, with fluorophores covalently attached in closer proximity to the PS backbone identifying phase separation a few degrees earlier.  The sharp increase in fluorescence intensity upon phase separation that occurs for all fluorophores with little change in spectral shape is consistent with a mechanism of static fluorescence quenching resulting from the specific interaction with a nearby quenching molecular unit.  Based on recent work that has identified a weak hydrogen bond occurring between the aromatic hydrogens of PS and the ether oxygen of PVME, we believe a similar weak hydrogen bond is likely occurring between the PVME oxygen and the aromatic dyes providing a local (few nanometer) sensitivity to phase separation.  Our observations indicate that fluorescence is sensitive to the early initiation of phase separation.  In comparing similar dyes, fluorophores in which the dye is covalently bonded closer to the PS backbone denote phase separation on average 3 °C earlier than those fluorophores in which the dye is located further away on a covalent tether.  As the PVME component moves further away from the PS component upon phase separation, local breaking of these weak hydrogen bonds connecting the PVME oxygen with the aromatic rings of PS and the aromatic dyes, leads to the observed increase in fluorescence intensity.  This reliable and reproducible method of measuring the phase separation temperature T_s allowed us to quantify how the presence of electric fields shifted the miscibility of PS/PVME blends. 
See:  Tuning miscibility of polymer blends with electric fields

 
Streamlined Ellipsometry Procedure for Characterizing Physical Aging Rates of Thin Polymer Films
Elizabeth A. Baker, Perla Rittigstein, John M. Torkelson, and Connie B. Roth,
Journal of Polymer Science, Part B: Polymer Physics   2009, 47, 2509-2519.

We have developed a new method of characterizing the physical aging rate of thin polymer films in an efficient manner using ellipsometry.  Ellipsometry measures the change in polarization of light as it is reflected from the sample. By using Fresnel’s equations, it is possible to determine the film thickness and index of refraction of a very thin layer supported on a known substrate (typically silicon).  Using visible light from 400-1000 nm, we are able to measure the film thickness to within a fraction of a nanometer.  In comparison to other methods available for characterizing the structural relaxation of thin films, this method has the advantage that it can be easily applied to different polymers with varying chemical structure.  The structural relaxation of the sample is characterized by measuring the decrease in film thickness and increase in index of refraction resulting from the slow increase in density of the material as a function of time.  We quantify the physical aging rate β from the time-dependent decrease in the film thickness as a function of logarithmic time: β = − d(h/h0) / d(log t).  This streamlined ellipsometry method for characterizing physical aging in polymer films has enabled a number of studies investigating changes in polymer glass stability in our lab and others.
See:  Stress during vitrification impacts stability of polymer glasses
and Nanoscale property changes in thin polymer films

The study of the glass transition in confined geometries is an active field of research in polymers, small molecules, colloids, and computer simulations.  Confinement associated with decreasing system size (e.g. decreasing film thickness) is seen as a method of perturbing dynamic heterogeneity in glasses in order to gain insight into the length scales associated with cooperative motion.  Polymer systems have been studied to the greatest extent because of the ease with which ultrathin films of known thickness can be made, and for their technological importance in applications such as microelectronics and gas separation membranes.  Large shifts in the glass transition temperature Tg have been observed with decreasing film thickness for over two decades for film thicknesses less than ~100 nm, which are believed to be correlated to a host of other property changes.  Interfaces dominate the behavior causing in large perturbing influences to the local structure and dynamics of the material as a function of distance from the interface.  It is believed that similar mechanisms are behind the success of polymer nanocomposites.

Polymers Under Nanoconfinement:   Where Are We Now in Understanding Local Property Changes?
Connie B. Roth, Chemical Society Reviews   2021, 50, 8050-8066.
This review provides a summary of local glass transition temperature Tg changes near interfaces, comparing across different types of interfaces:   free surface, substrate, liquid, and polymer–polymer.  Local versus film-average properties in thin films are discussed, making connections to other related property changes, while highlighting several historically important studies.  Emphasis is made to identify observations and open questions that have yet to be fully understood such as the evidence of long-ranged interfacial effects, finite domain size, interfacial breadth, and chain connectivity.

Refractive index and density changes in thin films

We have recently used ellipsometry to study how the refractive index (related to mass density) changes in thin films.  By developing a linear gradient model for the refractive index with depth n(z), we have demonstrated that a gradient in refractive index emerges in thin polymer films for thicknesses less than ~50 nm, addressing recent reports of physically unrealistic density increases.  Counter to common expectations of a simple free volume correlation between density and dynamics, we find that the direction of refractive index (density) gradient indicates a higher density near the free surface, which we rationalize as the enhanced dynamics near the free surface enabling more optimized denser molecular packings.

Gradient in Refractive Index Reveals Denser Near Free Surface Region in Thin Polymer Films
Yixuan Han and Connie B. Roth,
Journal of Chemical Physics   2021, 155, 144901.

Comparing Refractive Index and Density Changes with Decreasing Film Thickness in Thin Supported Films Across Different Polymers
Yixuan Han, Xinru Huang, Alan C. W. Rohrbach, and Connie B. Roth,
Journal of Chemical Physics   2020, 153, 044902.

Glass transition temperature Tg changes in ultrathin free-standing polystyrene films

Using transmission ellipsometry, we have measured the thermal expansion of ultrathin, high molecular weight (MW), free-standing polystyrene (PS) films over an extended temperature range.  For two different MWs, we observed two distinct reduced glass transition temperatures (Tgs), separated by up to 60 K, within single films with thicknesses h less than 70 nm.  The lower transition follows the previously seen MW dependent, linear Tg(h) behavior, while we also observed the presence of a much stronger upper transition that is MW independent and exhibits the same Tg(h) dependence as supported and low MW free-standing films.  This represents the first experimental evidence indicating that two separate mechanisms can act simultaneously on thin free-standing polymer films to propagate enhanced mobility from the free surface into the material.  The change in thermal expansion through the transitions indicate that ~90% of the film (matrix) solidifies at the upper transition with only ~10% of the material remaining mobile, freezing in at the lower transition.  Surprisingly, when we compare our results to the existing literature, and especially the low MW free-standing film data, we conclude that the upper transition encompasses the free surface region and associated gradient in dynamics.  This leaves open the question about where the small (~10%) fraction of material that has ultrafast, MW dependent dynamics resides within the film.  Subsequent measurements demonstrated that physical aging occurs below this stronger upper transition, including above the MW-dependent lower transition, confirming that the majority of the film solidifies at the upper MW-independent transition.

Two Simultaneous Mechanisms Causing Glass Transition Temperature Reductions in High Molecular Weight Free-Standing Polymer Films as Measured by Transmission Ellipsometry
Justin E. Pye and Connie B. Roth,
Physical Review Letters   2011, 107, 235701.

Above, Below, and In-Between the Two Glass Transitions of Ultrathin Free-Standing Polystyrene Films:   Thermal Expansion Coefficient and Physical Aging
Justin E. Pye and Connie B. Roth,
Journal of Polymer Science, Part B: Polymer Physics   2015, 53, 64-75.

Supported polystyrene films: Correlating physical aging and local Tg changes

We have observed changes in the physical aging for ultrathin supported polystyrene (PS) films that are correlated with the known local Tg reductions near a free surface.  Using our streamlined ellipsometry method for characterizing physical aging, an aging rate β is determined from the time-dependent decrease in the film thickness as a function of the logarithmic aging time.  We have measured the temperature dependence of the aging rate in ultrathin supported PS films and have found reduced physical aging rates at all temperatures for film thicknesses < 100 nm.  Our analysis of these results have demonstrated that the physical aging rates in these ultrathin films are not simply shifted in correspondence with the average glass transition temperature Tg(h) reductions in these films, but that the reduced physical aging rates result from a gradient in enhanced dynamics present near the free surface of these films.  The temperature dependence of the length scale characterizing the depth to which these enhanced dynamics penetrate into the film, from our analysis using a common two-layer model and a gradient model, are in very good agreement with similar length scales characterizing the gradient in Tg dynamics, strongly suggesting that these two phenomena are related.  Subsequent measurements focused at an aging temperature of 40 °C demonstrated an unexpected molecular weight (MW) dependence to the physical aging rate of ultrathin 31 nm thick PS films not present in bulk films, where samples made from ultra-high MWs ≥ 6500 kg/mol exhibit on average a 45% faster aging response compared with equivalent films made from (merely) high MWs ≤ 3500 kg/mol.  These results indicate the breadth of the gradient in dynamics originating from the free surface in these thin films is diminished for films of ultra-high MW PS.  These results contribute to growing literature reports signaling that chain connectivity and entropy play a subtle, but important role in how glassy dynamics are propagated from interfaces.

Physical Aging in Ultrathin Polystyrene Films:   Evidence of a Gradient in Dynamics at the Free Surface and Its Connection to the Glass Transition Temperature Reductions
Justin E. Pye, Kate A. Rohald, Elizabeth A. Baker, and Connie B. Roth,
Macromolecules   2010, 43, 8296-8303.

Unexpected Molecular Weight Dependence to the Physical Aging of Thin Polystyrene Films Present at Ultra-High Molecular Weights
Michael F. Thees and Connie B. Roth,
Journal of Polymer Science, Part B: Polymer Physics   2019, 57, 1224-1238.

Miscibility and phase separation of polymer blends has long been studied with the goal of understanding the specific interactions that affect blending.  This is especially important for polymers because large macromolecules limit the entropic contribution to the free energy of mixing allowing even weak interactions to play a significant role in miscibility.  Understanding how chemical structure affects specific interactions in polymer blends is important in developing blending models that can predict miscibility and blend properties from their individual constituents.  We seek a better understanding of blending interactions in polymers and methods to locally tune and control the enthalpic interaction of a blend.

Electric fields are one potential route to tuning the miscibility of polymer blends.  Several key energy technologies such as polymer solar cells and batteries involve polymer blends sandwiched between two electrodes such that the application of electric fields are part of the device’s application.  Organic photovoltaics have electrode spacings typically of order only 100 nm so that even a modest applied voltage of 1 V easily leads to electric field strengths of 10 MV/m.

How electric fields can modify the miscibility of blends has been an open question dating back to Debye in the 1960s.  There are conflicting reports on both the magnitude and direction that the phase separation temperature T_s can shift in the presence of electric fields.  Theoretical understanding of the phenomenon has been hampered by the lack of experimental data with unambiguously large shifts in T_s with increasing electric field strength E outside of experimental error.

Characterization of Phase Separation of Polystyrene / Poly(vinyl methyl ether) Blends Using Fluorescence
Annika Kriisa, Sung S. Park, and Connie B. Roth,
Journal of Polymer Science, Part B: Polymer Physics   2012, 50, 250-256.

Building on previous fluorescence studies, we developed a method to reliably and reproducibly measure the phase separation temperature T_s of polystyrene (PS) / poly(vinyl methyl ether) (PVME) blends.  Using this method, we have characterized for the first time changes in the fluorescence emission spectra of pyrene and anthracene dyes covalently bonded to PS chains upon phase separation from PVME.  The specific chemical structure of the fluorescent labels is found to affect the measured phase separation temperature T_s, with fluorophores covalently attached in closer proximity to the PS backbone identifying phase separation a few degrees earlier.  The sharp increase in fluorescence intensity upon phase separation that occurs for all fluorophores with little change in spectral shape is consistent with a mechanism of static fluorescence quenching resulting from the specific interaction with a nearby quenching molecular unit.  Based on work that has identified a weak hydrogen bond occurring between the aromatic hydrogens of PS and the ether oxygen of PVME, we believe a similar weak hydrogen bond is likely occurring between the PVME oxygen and the aromatic dyes providing a local (few nanometer) sensitivity to phase separation.  Our observations indicate that fluorescence is sensitive to the early initiation of phase separation.  In comparing similar dyes, fluorophores in which the dye is covalently bonded closer to the PS backbone denote phase separation on average 3 °C earlier than those fluorophores in which the dye is located further away on a covalent tether.  As the PVME component moves further away from the PS component upon phase separation, local breaking of these weak hydrogen bonds connecting the PVME oxygen with the aromatic rings of PS and the aromatic dyes, leads to the observed increase in fluorescence intensity.

Electric Fields Enhance Miscibility of Polystyrene / Poly(vinyl methyl ether) Blends
Annika Kriisa and Connie B. Roth,
Journal of Chemical Physics   2014, 141, 134908.

Using the fluorescence method we developed containing trace levels of pyrene-labeled PS, we have reliably demonstrated how the miscibility of PS/PVME blends changes in the presence of uniform electric fields.  We have observed large shifts in the phase separation temperature T_s that is reliably and reproducibly well outside of experimental error.  For electric field strengths of 17 MV/m, we have measured shifts in T_s of up to 13.5 ± 1.4 K for a 50/50 PS/PVME mixture, some of the largest absolute shifts every reported.  We have also demonstrated that the effect is completely reversible with the blend returning to the same T_s value after the electric field has been removed.  Our shift in T_s with electric field strength (ΔT_s/E^2) showing enhanced mixing in the presence of electric fields is comparable to other experimental studies in the field on small molecule mixtures, but is still much larger than can be explained theoretically.  To understand the phenomenon theoretically, the free energy contribution of electric fields to blend miscibility must be identified.  Our experimental results demonstrating increased miscibility with increasing electric field strength E suggests that the term accounting for the dielectric contrast between different components dominates, suppressing concentration fluctuations parallel to the electric field direction and hence the formation of dielectric interfaces between domains during phase separation.  Although this reasoning qualitatively explains the direction of the T_s(E) shifts, quantitative agreement with the magnitude of the shift ΔT_s/E^2 is still lacking theoretically.

Jumping In and Out of the Phase Diagram Using Electric Fields: Time Scale for Remixing of Polystyrene/Poly(vinyl methyl ether) Blends
Annika Kriisa and Connie B. Roth,
ACS Macro Letters   2019, 8, 188-192.

We also demonstrate that electric field can be used as a control knob to repeatedly jump PS/PVME blends from the one-phase to two-phase region by simply turning on and off the electric field when the blend is held at a fixed temperature near the phase boundary.  This confirms that electric fields are shifting the miscibility transition.  The kinetics of the early stages of phase separation was also investigated, where the remixing time scale τ(T) with and without electric fields were well fit by a Vogel-Fulcher-Tammann expression consistent with a mobility-limited process.  Given the prevalence of advanced technologies such as photovoltaics, batteries, and sensors with built-in electrodes requiring idealized blend morphologies, electric field manipulation of phase behavior could have significant design benefits.

How stress imparts mobility to glasses is an ongoing subject of debate in soft matter.  Many experiments and simulations have shown that stress or strain applied to polymers, colloids, granular materials, etc. in the glassy state leads to enhanced mobility.  In some cases, such deformation can even appear to erase past physical aging as if “rejuvenating” the glass.  How deformation imparts mobility to glasses is a process often described by an potential energy landscape ’tilting’ mechanism, where the applied stress is thought to reduce energy barriers, allowing the system to transition to a higher energy state.

Upon vitrification (e.g., glass formation by a temperature quench), the material transitions from an equilibrium liquid to a non-equilibrium glass at the glass transition temperature T_g.  Structural relaxation of the non-equilibrium glassy state leads to a decrease in free volume of the material over logarithmic time scales.  The stability of the glassy state is quantified by measuring the physical aging rate β, characterizing this decrease in volume with logarithmic time. 

Although these volume contractions are minuscule (<1%), the resulting property changes termed physical aging (e.g., changes in mechanical properties and failure modes in particular) can be substantial and often adverse.  As an example, the permeability of gas-separation membranes is exceedingly sensitive to the free volume in the polymer gas-sieving layer.  Decreases in permeability of nearly 50% have been observed as a result of physical aging.

Importance of Quench Conditions on the Subsequent Physical Aging Rate of Glassy Polymer Films
Laura A. G. Gray, Suk W. Yoon, William A. Pahner, James E. Davidheiser, and Connie B. Roth, Macromolecules   2012, 45, 1701-1709.

Dating back to the mid-1990s, the gas separation membranes community have reported anomalous “accelerated” (faster) physical aging of free-standing polymer films with decreasing film thickness for a range of different polymers.  What was surprising was that these deviations in the aging rate from bulk behavior began at enormous film thicknesses of several microns, much larger than the ~100 nm thicknesses where most thin film interfacial effects occur.

Using a streamlined ellipsometry technique we developed to measure the physical aging rate of polymer films, we have systematically addressed various possible causes of this phenomenon from differences in molecular structure, quench depth below T_g, experimental technique, sample preparation, and stresses on the film.  We demonstrated that the physical aging rate of the material is strongly dependent on conditions during the formation of the glassy state.  Although supported films do not display any film thickness dependence to their aging rate at this large micron length scale, films quenched in a free-standing state exhibit a strong thickness dependence.  We suggested that differing quench conditions may impose unintended stresses trapping the glassy films into different states (potential energy minima), dictating the subsequent physical aging rate.  We then followed this up with studies that demonstrated this correlation conclusively by vitrifying films under different controlled stress conditions.

Physical Aging of Polymer Films Quenched and Measured Free-Standing via Ellipsometry:   Controlling Stress Imparted by Thermal Expansion Mismatch between Film and Support
Justin E. Pye and Connie B. Roth,
Macromolecules   2013, 46, 9455-9463.

We have demonstrated that stress effects on cooling can lead to changes in physical aging rate β even when the stress imparted arises inherently from how the film is supported.  By cooling films on different holders, we conclusively demonstrated that stress imparted during vitrification caused by thermal expansion mismatch Δα between the film and sample support was controlling the subsequent physical aging rate β of the film.

The stress imparted to the free-standing polystyrene (PS) films on cooling when supported on rigid frames of different materials can be directly quantified by calculating the thermal expansion mismatch Δα between the film and rigid frame.  This stress is independent of film thickness, but depends on the thermal expansion mismatch between the film and rigid frame.  We demonstrated that the physical aging rate is independent of film thickness, but correlates with the difference in thermal expansion between the PS film and frame material, and hence the stress imparted to the film on cooling.

Stability of Polymer Glasses Vitrified Under Stress
Laura A. G. Gray and Connie B. Roth,
Soft Matter   2014, 10, 1572-1578.

In addition, we designed a simple jig that allowed us to apply a known stress to a free-standing polymer film during vitrification in order to quantify how the physical aging rate β changes with increasing stress.  This enabled us to investigate the impact of applying different stresses during vitrification, i.e., formation of a glass during thermal cooling, on the subsequent physical aging.  We found that the subsequent stability of the glassy system is affected by the stress applied on cooling, even after the stress has been removed.  The data showed an initial plateau in aging rate at low stresses that quickly transitions to a much higher aging rate at high stresses.  Above a minimum threshold in stress, the aging rate appeared to plateau at a higher value indicative of a less stable glass.  These effects could be understood in terms of the potential energy landscape (PEL) ’tilting’ mechanisms from theoretical efforts in the field.  We proposed that a glassy system formed under high stress is left trapped in a higher, shallower energy minimum with a corresponding faster physical aging rate.

The field of art conservation routinely uses polymer resins to repair and protect artworks that have been damaged.  The requirements for these polymer materials must meet various demands, the most important being that the resins can be removed or retreated without harming the original artwork, which may be thousands of years old.  The adhesive must also be stable for decades and ideally be nearly invisible by matching the refractive index of the repair to the material of the artwork.  In addition, although the repaired bonds must be strong, they should not be too strong such that any future fracture occurs in the bond at the existing fault lines and not a new locations in the art.

For historical reasons the field of conservation has long used a series of different acrylic resins originally developed by Rohm & Haas.  Although there is great interest within the field about the properties and best uses of these materials, there is limited funding available for direct fundamental research on these resins.

In collaboration with Chief Conservator Renée Stein at Emory’s Carlos Museum, the Roth lab is conducting research projects with undergraduate students to characterize these polymers used in the preservation and restoration of artworks.

As polymer properties are highly dependent on how they are prepared, we are hoping to provide some scientific insight into the various preparation steps commonly used in the field in an effort to understand which steps are important and why/how a given resin may function better in a given environment.

Sunmin Kim at the Carlos Museum Conservation Lab next to an ancient Egyptian coffin lid dating back to the 22nd Dynasty (900-700 BC).  Repair locations in white are visible, corresponding to B-72 bulked with glass miscrospheres that was injected into the cracks and around fragments as part of a triage treatment to stabilize the piece for transport to Atlanta.

Honors theses research by Benjamin Kasavan and Olivia Boyd have characterized the glass transition, refractive index, and fracture stress of various resins and blends.  Sunmin Kim is currently continuing these efforts.

Science.Art.Wonder, an Atlanta-based organization, pairs artists and scientists to design artwork that can communicate scientific research to the public.

Artist’s narrative by Nicole Cobb:

“Restoration” is a ceramic piece done in the style of Japanese kintsugi: the art of repairing broken pottery with gold. The gold holding the cracks in the piece together represents the various different polymer-based adhesives created for art restoration at the Carlos Museum. These adhesives are formed by dissolving Paraloid B-72 and B-48N acrylic pellets in acetone and must meet a Goldilocks standard. The adhesive strength should not be too weak such as to fail in holding the restored art together nor too strong such as to damage the restored art. Although these adhesives are typically imperceivable in restored artworks, they are represented in gold in this piece to highlight the importance in studying their material properties. This piece incorporates Kintsugi philosophy, which is about embracing flaws and reminds us that there is still utility and beauty to be found in broken things.

Scientific abstract by Sunmin Kim:

Since conservator Stephen Koob’s 1986 publication, the acrylic polymer Paraloid B-72 has been widely used as a reliably stable and reversible adhesive for restoration in the art conservation field. In recent years, conservators have begun to incorporate the less commonly used and stiffer polymer Paraloid B-48N into blends with B-72. Applied as a solution in acetone, these adhesives solidify due to their underlying polymers’ glass transition occurring by the evaporation of solvent over time. As these blended adhesives are applied to historically significant and irreplaceable artifacts, we must measure and record the strength of the adhesives prior to their incorporation. In our project, glass rods are broken and repaired with adhesives created out of varying B-72:B-48N ratios. The adhesives are evaluated for their tensile fracture strengths utilizing the Conservation Adhesive Tensile-to-Shear (CATS) Tester and additionally examined for their glass transitions using an optical instrument known as an ellipsometer.