NEWS & VIEWS

nature materials | VOL 4 | SEPTEMBER 2005 | www.nature.com/naturematerials 651

ANNE M. MAYESis at the Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA. e-mail: amayes@mit.edu

Polymer products are oft en composite materials that contain inorganic fi llers such as calcium carbonate, talc or carbon black as a means to

bolster the mechanical, thermal or barrier properties, or to decrease their overall cost. In recent years, the recognition that small fractions of nanometre-sized particles can impart performance enhancements above what is achieved with conventional micrometre-sized fi llers has sparked a fl urry of research activity on polymer nanocomposites1. Depending on polymer–particle interactions, polymer nanocomposites may exhibit an increase or reduction in glass-transition temperature (Tg) — above this temperature a polymer can undergo plastic deformation whereas below it’s a rigid and brittle material — relative to the pure polymer. As such systems oft en form complex morphologies with various levels of particle agglomeration, quantifying the eff ect of nanoparticle content on Tg would seem to be beyond predictive reach.

In a surprising fi nding described on page 693 of this issue, Sanat Kumar, Linda Schadler and co-workers2 demonstrate that mixing nanosized silica particles into a polystyrene matrix suppresses Tg in a way that quantitatively mimics Tg shift s that are observed when polystyrene is confi ned in a thin-fi lm geometry, if the spacing between particles is equated to the fi lm thickness. Most intriguingly, the nanocomposite–thin fi lm analogy holds despite the fact that the particle distributions in their samples are quite heterogeneous, with variably sized aggregates dispersed amongst individual particles. Th e results off er a framework for understanding and designing new nanocomposite materials, and lend insight on the glass-transition phenomenon for polymers in restricted geometries.

Th e glass-transition of polymer thin-fi lms has been a controversial subject, receiving much attention in the past decade since the fi rst observations that substrate-supported polystyrene thin-fi lms exhibit a Tg substantially below that of bulk polystyrene, and strongly dependent on fi lm thickness3. Th ese

fi ndings, reproduced in numerous experiments, have been interpreted in the context of a ‘two-layer model’, in which a near-surface layer a few nanometres thick and having a reduced Tg resides atop a bulk Tg region (Fig. 1). Elegant fl uorescence experiments reported by Ellison and Torkelson4 showed instead that a Tg gradient is present near the surface of a polymer fi lm that extends tens of nanometres into the fi lm. Th ese authors also found that the near surface region displayed a maximum in Tg when total fi lm thickness was comparable to a coil diameter (~25 nm for the polystyrene system studied), whereas near the substrate, Tg decreased with decreasing fi lm thickness. Th eir fi ndings suggest that the two interfaces of the fi lm are in dynamic ‘communication’ when suffi ciently close together, complicating the story from the two-layer model.

Depending on polymer-particle interactions, similar eff ects might be anticipated when nanoparticles are mixed in a polymer. Indeed, molecular dynamics simulations by Glotzer and co-workers5 showed that embedding non-interacting nanoparticles in a polymer matrix causes segments in the particles’ vicinity to relax faster, implying a local Tg reduction (Fig. 1). Reminiscent of Ellison and Torkelson’s thin-fi lm fi ndings, the segment mobility was found to decay to its bulk value at a distance from the nanoparticle surface of roughly the bulk coil diameter. Th e similarity to thin-fi lm results suggests that the ‘free’ aspect of the surface (that is, lack of a solid boundary) is not a prerequisite for the Tg reductions seen in thin polymer fi lms.

Inspired by Glotzer’s simulations, Kumar, Schadler and co-workers measured the Tg values of polystyrene/silica nanocomposites at various nanoparticle loadings, and compared them to Tg values reported in the literature for polystyrene

A comparative study of silica-fi lled polystyrene and thin polystyrene fi lms offers a new framework both for quantifying the typical shift in the glass-transition temperature of such systems, and prospects for the design of new nanocomposites.

NANOCOMPOSITES

Softer at the boundary

Region of reduced Tg

Substrate

Polymer

Figure 1 Regions of altered mobility. Schematic cross-section of a polymer thin-fi lm on a substrate (left) and a polymer nanocomposite (right) having equivalent surface-to-volume ratios. Shaded regions depict material displaying a glass transition below the bulk value. The work of Kumar et al.2 reveals a strikingly similar suppression of Tg in the confi ned geometries of polymer–particle nanocomposites and thin fi lms.

NV-print0905.indd 651NV-print0905.indd 651 12/8/05 10:20:26 am12/8/05 10:20:26 am

Nature Publishing Group© 2005

NEWS & VIEWS

652 nature materials | VOL 4 | SEPTEMBER 2005 | www.nature.com/naturematerials

fi lms2. In their system the nanoparticles are poorly wetted by the polymer in the absence of any surface treatment. Drawing on the two-layer model for thin fi lms, one might conjecture that the reduction in Tg observed in these systems would scale with the ratio of the particle surface area to the polymer volume. By considering particle aggregates as single particles of larger diameter, the authors calculated this ratio from electron micrograph images and plotted Tg reduction as a function of the inverse ratio (volume/surface area). Th ey fi nd that, comparing nanocomposite systems to thin fi lms of equal surface-to-volume ratio (such as the two systems depicted schematically in Fig. 1), the nanocomposites exhibit a greater Tg suppression.

Most strikingly, by equating the harmonic average (the reciprocal of the arithmetic mean of reciprocals) of the interparticle spacing to fi lm thickness, the authors fi nd that the nanocomposite and thin-fi lm data completely superpose2. Th e remarkable agreement between the two geometries seems surprising, given the 3D nature of the confi nement in nanocomposites, the presence of particle curvature, the broad range of particle sizes and so on. Th e equivalence between interparticle spacing and fi lm thickness implies that a dynamic cross-talk exists between the interfaces of polymer thin-fi lms, consistent with the results of Ellison and Torkelson4. Th ese fi ndings further beg the question: for a fi xed interparticle spacing, what is the role of particle size? Is there a particle size below which no Tg suppression is expected, or does the eff ect hold down to molecular dimensions, connecting reductions in Tg arising from geometric constraints to those achieved by the addition of common plasticizers?

Th e precise origin of the Tg suppression in constrained geometries remains a subject of inquiry3,5,6 along with the nature of the glass transition itself. Nevertheless, the intriguing work by Kumar, Schadler and co-workers reported in this issue off ers new light in which to view the properties of polymer nanocomposites. Th eir results may explain, for instance, why silica nanoparticles added to a glassy polymer membrane imparts substantially enhanced permeability to organic molecules7. Th eir fi ndings might also be used in designing new materials — the addition of weakly wetting particles to a conventional polymer electrolyte, for example, should suppress Tg and thereby raise ionic conductivity above that of the unfi lled polymer8. Th us, these results promise further advancements in both fundamental and applied aspects of the thermomechanics of polymer nanocomposites.

REFERENCES1. Jordan, J., Jacob, K. I., Tannenbaum, R., Sharaf, M. A. & Jasiuk, I. Mater. Sci.

Eng. A 393, 1–11 (2005).2. Bansal, A. et al. Nature Mater. 4, 693–698 (2005).3. Forrest, J. A. Eur. Phys. J. E 8, 261–266 (2002).4. Ellison, C. J. & Torkelson, J. M. Nature Mater. 2, 695–700 (2003).5. Starr, F. W., Schroeder, T. B. & Glotzer, S. C. Phys. Rev. E 64, 021802 (2001).6. Ellison, C. J., Mundra, M. K. & Torkelson, J. M. Macromolecules 38, 1767–

1778 (2005).7. Merkel, T. C. et al. Science 296 519–522 (2002).8. Croce, F., Appetecchi, G. B., Persi, L. & Scrosati, B. Nature 394, 456–458 (1998).

MATERIAL WITNESS

Taking lessons from the bookWhen the bicycle was voted by Britons last year as the greatest invention, some technology experts were dismayed. As a keen cyclist, I was reluctant to complain myself. More frustrating, however, was the time span considered: the past 250 years, perhaps on the basis that older inventions would be mere historical curiosities by now. On the contrary, this eliminated some of the most worthy candidates, of which the foremost is surely the book.

Just as literary critics love to discuss the ‘death of the novel’, so publishers obsess about the ‘death of the book’, if only to debunk the notion. In part, this is a question about the future of 100,000-word theses in the era of the word-bite and web page; but also at issue is the fate of the book as physical object, a series of printed paper sheets between covers.

This invention has its drawbacks. Libraries consume space. Books are heavy, combustible, edible (to some species), and they fall apart. Electronic media would eliminate these problems at a stroke.

And yet they have not done so. The impact of electronics on reading and writing has been hugely asymmetric. These words will not encounter paper until they are ready for the pages of Nature Materials. Many people barely pick up a pen now except to sign their name; but some newspapers can put virtually their entire content online without fear of losing paper sales.

Some commentators are convinced that this is only a matter of time; and they may be right. Electronic paper has improved remarkably, and will surely hit the mass market within a decade. Power sources for such devices are getting lighter and longer-lived; manufacturing of high-resolution screens becomes ever cheaper, largely as a result of inventive materials solutions.

In his new book The Singularity is Near (Viking, 2005), inventor Ray Kurzweil suggests that technologies evolve in series of S-shaped curves — slow initial growth followed by rapid expansion that eventually levels off — with exponentially decreasing ‘cycle’ times. He suggests that the false starts and fundamental shortcomings of early ‘electronic books’ are mirrored by those of electronic pianos (a technology Kurzweil pioneered), and that once these problems are overcome, books will be seen as no less susceptible to ‘digital’ replacement than the acoustic piano.

But even if this is true, the extraordinary resilience of the book has something to teach us about the nature of technology. The reasons for the book’s current dominance over electronic alternatives are not purely technical but are bound up with the human interface. What seems like imperceptible flicker on a standard computer screen confuses the eye and slows reading speed, so that we still prefer to print out long texts. Electronic books are wonderful for text searches, but don’t yet have a browse facility that compares with flicking through paper pages. And even if books are biodegradable, do we still trust that words are as safe in an electronic memory as they are in paper and ink?

Philip Ball

NV-print0905.indd 652NV-print0905.indd 652 12/8/05 10:20:27 am12/8/05 10:20:27 am

Nature Publishing Group© 2005