SEARCH

SEARCH BY CITATION

Keywords:

  • films;
  • fluorescence;
  • glass transition;
  • interface;
  • luminescence;
  • nanolayers;
  • photophysics;
  • polystyrene;
  • thermal properties;
  • thin films

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. Acknowledgements
  7. REFERENCES AND NOTES

The effect of nanoscale confinement on the glass transition temperature, Tg, of freely standing polystyrene (PS) films was determined using the temperature dependence of a fluorescence intensity ratio associated with pyrene dye labeled to the polymer. The ratio of the intensity of the third fluorescence peak to that of the first fluorescence peak in 1-pyrenylmethyl methacrylate-labeled PS (MApyrene-labeled PS) decreased with decreasing temperature, and the intersection of the linear temperature dependences in the rubbery and glassy states yielded the measurement of Tg. The sensitivity of this method to Tg was also shown in bulk, supported PS and poly(isobutyl methacrylate) films. With free-standing PS films, a strong effect of confinement on Tg was evident at thicknesses less than 80–90 nm. For MApyrene-labeled PS with Mn = 701 kg mol−1, a 41-nm-thick film exhibited a 47 K reduction in Tg relative to bulk PS. A strong molecular weight dependence of the Tg-confinement effect was also observed, with a 65-nm-thick free-standing film exhibiting a reduction in Tg relative to bulk PS of 19 K with Mn = 701 kg mol−1 and 31 K with Mn = 1460 kg mol−1. The data are in reasonable agreement with results of Forrest, Dalnoki-Veress, and Dutcher who performed the seminal studies on Tg-confinement effects in free-standing PS films. The utility of self-referencing fluorescence for novel studies of confinement effects in free-standing films is discussed. © 2008 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 46: 2754–2764, 2008


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. Acknowledgements
  7. REFERENCES AND NOTES

In 1994, Keddie et al.1 reported the seminal study of the effect of nanoscale confinement on the glass transition temperature (Tg) of amorphous polymers. Using ellipsometry to characterize Tg via measurement of film thickness as a function of temperature, they observed major reductions in Tg relative to its bulk value when the thickness of a polystyrene (PS) film supported on silica was less than ∼ 40 nm. For example, a 17-nm-thick PS film exhibited a 21 K reduction in Tg relative to bulk PS. They hypothesized that the relaxation of constraints to cooperative segmental motion at the free surface could lead to enhanced mobility and thus a reduction in Tg with nanoscale confinement that originates at the free surface. Experimental support for this hypothesis was provided in 2003 by Ellison and Torkelson2 who demonstrated via a novel, multilayer-fluorescence method that a reduction in Tg can be measured several tens of nanometers from the free surface of a PS film and that the gradient in Tg becomes stronger with a reduction in distance from the free surface. Since the study by Keddie et al.,1 many dozens of experimental and theoretical investigations have focused on how nanoscale confinement alters the glass transition and related properties of thin polymer films2–45 and polymer nanocomposites.45–50 The majority of the experimental studies have involved polymer films supported on silica slides. Such a system provides the combined advantages of ease in quantitatively tuning the extent of confinement by spin coating51 and a geometry that is well designed for simple application with a range of experimental methods for determining Tg.

In contrast to the many investigations of the Tg-confinement effect of polymer films supported on silica, relatively few studies have focused on the effects of confinement on the glass transition behavior and related properties of single-layer free-standing films,26–42 that is, films with two polymer–air interfaces or free surfaces. Such a system is of great technological importance in applications ranging from gas separation membranes to microelectronics. Using Brillouin scattering and ellipsometry, Dutcher and coworkers26–31 were the first group to characterize the Tg-confinement effect in free-standing polymer films, doing studies on both PS and poly(methyl methacrylate). The two most striking features associated with their PS free-standing films are that the Tg reductions can be much greater than those of PS films supported on silica (Tg reductions of as much 70–80 K were reported in freely standing films relative to bulk Tg26–29) and that at high molecular weight (MW) there is a strong MW dependence of the Tg reduction. Such a MW-dependent Tg reduction is not evident in supported PS films.10 Beyond the studies by Dutcher and coworkers, three other studies have been reported that provide direct Tg determinations of single-layer free-standing polymer films.34–36 Mattsson et al.34 used Brillouin scattering to demonstrate that the Tg-confinement effect in free-standing PS films is independent of MW for MWs less than ∼ 350 kg mol−1. Liem et al.35 used Raman spectroscopy to determine Tg reductions for free-standing PS films (MW = 600 kg mol−1); in a 50-nm-thick PS film, they observed that Tg was decreased by slightly more than 30 K relative to bulk PS, in approximate agreement with data by Dutcher and coworkers.28, 29 Miyazaki et al.36 used x-ray reflectivity to determine Tg values for two MWs (303 kg mol−1 and 2890 kg mol−1) of PS free-standing films, obtaining rough agreement with the MW dependence reported by Dutcher and coworkers.28, 29

Quantitative inferences have been made about the Tg-confinement effect in PS and poly(vinyl acetate) free-standing or suspended films from data obtained by ingenious studies on the thickness and temperature dependences of relaxation times obtained via dielectric spectroscopy37 and of creep compliance.38, 39 In the case of PS films, the data are in approximate accord with the results of Dutcher and coworkers.29 In addition, two studies have reported Tg characterization via differential scanning calorimetry (DSC) in multilayer films of polymer that have ultrathin individual layers that are not well consolidated.52, 53 (By well consolidated, we mean a multilayer film that has been exposed to temperatures above Tg in such a manner that the interfaces have healed, leading to an overall film structure that has properties the same or similar to that of a single layer film of the same overall thickness as the originally multilayer film.) Both studies report some reduction in Tg relative to bulk polymer with decreasing nanoscale layer thickness. However, in the case of the PS,53 the Tg reductions are smaller in magnitude than even those that are commonly reported for supported PS films1, 2, 9, 10; the limited Tg reductions may be related to the extent of partial consolidation of the layers that can occur during DSC measurement, among other effects.

The effect of confinement on Tg and related properties has also been the subject of recent study in complex geometries related to single-layer free-standing films,54–60 including polymeric nanowires and nanofibers54, 55 and patterned polymer nanostructures with high levels of free surface,58, 60 for example, nanostructed films of the type that may be generated during lithographic processing in the microelectronics industry. In each case, reductions in Tg were observed with confinement.

Here, we report the first determinations of Tg in free-standing polymer films via a fluorescence-based method. Although the current study is limited to single-layer films, one of our long-term goals is to perform studies on well-consolidated, multilayer free-standing films with only one layer containing a fluorescence-dye-labeled polymer. Such studies will ultimately provide us the ability to measure distributions of Tg across freely standing films, similar to those reported by Ellison and Torkelson2 and Priestley et al.11, 15, 61 in supported polymer films.

In most of our previous fluorescence studies of Tg-confinement effects, we measured fluorescence intensity as a function of temperature, the Tg value being obtained from the intersection of the lines associated with the temperature dependences of the rubbery-state and glassy-state fluorescence intensities.2, 9–15, 45, 46, 58, 61 We have found that such measurements do not provide data of sufficient quality to yield accurate and precise Tg determinations in free-standing films; this is presumably due to film rippling on cooling, resulting in changes in the surface area of the film exposed to the fluorescence excitation light. In seeking to eliminate this effect, we have discovered that the fluorescence spectral shape of a particular pyrenyl dye covalently attached to polymer changes as a function of temperature and that the use of an intensity ratio to characterize these changes in spectral shape provides an accurate and precise determination of Tg. By employing an intensity ratio, we obtain self-referencing fluorescence data that are independent of film rippling effects because the intensity ratio is independent of the surface area of the film exposed to the excitation light.

Self-referencing fluorescence measurements in polymer systems have been previously obtained using wavelength-shifting probes as well as probes or polymers that exhibit changes in intensity ratio via excimer fluorescence or nonradiative energy transfer (also called fluorescence resonance energy transfer).13, 62–76 Self-referencing fluorescence approaches have been employed in various studies of polymer properties and behavior, including bulk Tg,13, 63, 64 sorption and drying,65 mixing,66 conversion during polymerization or curing,62, 66, 72 micelle or aggregate formation,67–69, 73 phase separation,70 and characterization of conformation in solution71 and local polarity in heterogeneous polymeric systems.75, 76 In the current study, we show that Tg values can be obtained in supported polymer films and, for the first time, in free-standing polymer films using a self-referencing fluorescence intensity ratio. Additionally, we provide the first demonstration of the applicability of this approach for determination of Tg as a function of confinement in free-standing PS films and show that our data are in approximate agreement with data obtained by Forrest and coworkers.26–29

EXPERIMENTAL

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. Acknowledgements
  7. REFERENCES AND NOTES

Materials Synthesis and Characterization

1-Pyrenylmethyl methacrylate-labeled PS [MApyrene-labeled PS77; Mn = 701,000 g mol−1, Mw/Mn = 1.31,78 by gel permeation chromatography (GPC) relative to PS standards using tetrahydrofuran (THF) as eluent] and 1-pyrenylmethyl methacrylate-labeled poly(isobutyl methacrylate) (MApyrene-labeled PiBMA; Mn = 170,000 g mol−1, Mw/Mn = 1.82, by GPC using a universal calibration method) were synthesized by bulk free radical polymerization of styrene (Aldrich) and isobutyl methacrylate (Aldrich), respectively, in the presence of trace amounts of 1-pyrenylmethyl methacrylate (MApyrene77) (Polysciences). 1-Pyrenylbutyl methacrylate-labeled PS (BApyrene-labeled PS; Mn = 464,000 g mol−1, Mw/Mn = 1.57, by GPC) was synthesized by bulk free radical polymerization of styrene (Aldrich) in the presence of trace levels of 1-pyrenylbutyl methacrylate (BApyrene). The BApyrene was synthesized by esterification of methacryloyl chloride (Aldrich) and 1-pyrenyl butanol (Aldrich) following synthesis procedures described in ref.79 for a related monomer. All polymers were precipitated into excess methanol and washed by dissolving in THF and precipitating in methanol seven times to remove any residual dye-labeled monomer before drying under vacuum. Bulk Tg values were measured by differential scanning calorimetry (DSC) (Mettler-Toledo 822e, second heat, onset method, 10 K min−1) and fluorescence methods (temperature dependence of intensity ratios or integrated intensities) and were found to agree within experimental error: MApyrene-labeled PS Tg,bulk = 375 K by DSC and fluorescence; MApyrene-labeled PiBMA Tg,bulk = 338 K by DSC and fluorescence. In the case of BApyrene-labeled PS, Tg,bulk = 375 K by DSC and fluorescence. The pyrenyl-dye label contents were measured by UV-Vis absorbance (Perkin-Elmer Lambda 35): MApyrene-labeled PS is 1.1 mol % MApyrene and 98.9 mol % styrene; MApyrene-labeled PiBMA is 0.8 mol % MApyrene and 99.2 mol % styrene; and BApyrene-labeled PS is 0.3 mol % BApyrene and 99.7 mol % styrene.

Film Preparation

Single-layer supported films of MApyrene-labeled PS and PIBMA and BApyrene-labeled PS were prepared by spin coating51 dilute polymer solutions onto quartz slides. Depending on desired film thickness and polymer molecular weight, spin coating was done from toluene solutions at concentrations of 1.0–4.5 wt % and spin speeds of 800–2000 rpm. The films were dried in vacuum at Tg,bulk + 20 K for 12 h. Free-standing films were prepared by first spin coating dilute polymer solutions onto freshly cleaved mica. These mica-supported films were annealed under vacuum at Tg,bulk + 20 K for 12 h to remove residual solvent. Free-standing films were fabricated by using a water transfer technique31 to transfer the films from the mica substrates onto nylon sample holders (washers) with 10 mm diameter holes. Film thickness was verified by spin coating a second film at the same time from the same solution onto a silicon slide with native silicon oxide layer and measuring its thickness via spectroscopic ellipsometry (J. A. Woollam Co. M-2000D). (Within experimental error, identical film thicknesses as measured by ellipsometry are obtained when, using the same solution and rotational speed during spin coating, films are spun on silicon and measured directly or are spun on mica and then floated onto a silicon wafer and measured. Thus, our determination of film thickness for freely standing films spun on mica via analysis of supported films spun on silicon is appropriate.)

Fluorescence Measurements

Fluorescence was measured using a Photon Technology International fluorimeter with 1–4 nm bandpass excitation and 1 nm bandpass emission slits. Film temperature was controlled by a microprocessor controller (Minco Products) with a Kapton ribbon heater attached to a flat aluminum plate that was also used as a clamping device to hold the sample. For supported films, quartz slides on which the dye-labeled polymer film was spin coated were placed directly atop the thin aluminum plate with the film side facing away from the plate. A quartz cover slide was placed on top of the film and a clamping device held all the pieces together. For free-standing films, the nylon washers holding the polymer films were placed and clamped between two aluminum plates that had 15 mm-diameter-holes to accommodate the excitation and emission of fluorescence from the freely standing films. A quartz cover was placed on top of the aluminum plates and a clamping device held all the pieces together. The excitation wavelength was 324 nm, and the fluorescence emission intensity was monitored at 370–415 nm.

All but two of the fluorescence-based Tg measurements were obtained on cooling from the equilibrium rubbery or liquid state. (The other two measurements were obtained upon heating films from the glassy state to the rubbery state. See Discussion.) Films were heated to at least Tg + 20 K and held for a minimum of 10 min before measuring the fluorescence emission spectrum. Then the temperature was decreased by 5 K at a time, and at each temperature the sample was held for 5 min to ensure thermal equilibrium before again measuring the fluorescence emission spectrum. To determine Tg, the temperature dependence of the ratio of the intensity of the third peak to the intensity of the first peak (I3/I1) of the pyrenyl dye label emission spectra was used. The intensity ratio was obtained by dividing the intensity of the third peak located at 386.0 ± 3.5 nm by the intensity of the first peak located at 375.0 ± 3.5 nm, and intensities were averaged over a 1–2 nm range. In fitting the temperature dependences of I3/I1 in the rubbery and glassy states of the polymer film, data points well outside Tg were used for the linear fits. To initiate the fitting procedure, data points were added to the rubbery- and glassy-state linear regressions at the extrema in the temperature range of the data. Related details on determination of Tg values from fits to temperature-dependent fluorescence data are provided in refs.2,9, and10.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. Acknowledgements
  7. REFERENCES AND NOTES

Figure 1 illustrates the strong effect of temperature on the spectral shape of the fluorescence of MApyrene-labeled PS, MApyrene-labeled PiBMA, and BApyrene-labeled PS films. The films are supported on silica (quartz) slides and are sufficiently thick to exhibit bulk Tg response. The MApyrene-labeled PiBMA film provides the clearest indication of four peaks in the 370–415 nm wavelength range, which are characteristic of the fluorescence from the excited-state of a single pyrene molecule or pyrenyl chromophore.75–77, 80–82 In contrast to the fluorescence from the PiBMA film, the second fluorescence peak is virtually absent in the PS films.

thumbnail image

Figure 1. Temperature dependence of the fluorescence emission spectra of pyrene dye-labeled polymers supported on silica (quartz) slides: (a) 270-nm-thick MApyrene-labeled PS film, (b) 405-nm-thick MApyrene-labeled PiBMA film, and (c) 345-nm-thick BApyrene-labeled PS film. Spectra have been normalized to one by the peak intensity in the low temperature emission spectrum. Downward pointing arrows indicate the approximate wavelengths at which I1 and I3 intensities are measured. Inset: Molecular structure of MApyrene (a) and (b), and molecular structure of BApyrene (c).

Download figure to PowerPoint

It has been known for several decades that many pyrene-based dyes can exhibit strong sensitivity of fluorescence spectral shape to local, nanoscale environment. In a 1977 study, Kalyanasundaram and Thomas80 demonstrated that the vibrational structure of pyrene fluorescence depends on the polarity of the local dye environment. In a polar environment, there is an enhancement of the intensity of the 0–0 band (first peak starting at low wavelength in a highly structured spectrum) relative to the other peaks. In particular, the ratio of the fluorescence intensity of the third peak to the fluorescence intensity of the first peak, I3/I1, can provide a quantitative indication of the polarity nanoscale environment surrounding the pyrene dye, thereby making it useful in polarity-sensitive applications such as determination of critical micelle concentrations.82 Free pyrene and pyrene covalently attached to a polymer via a methylene linking group exhibit I3/I1 values that are strongly sensitive to polarity; however, a longer linking group leads to a loss of sensitivity.81

Figure 2 provides plots of I3/I1 as a function of temperature above and below the Tgs of the MApyrene-labeled PS, MApyrene-labeled PiBMA, and BApyrene-labeled PS films supported on silica (quartz) slides. (The downward-pointing arrows in Figure 1 indicate the approximate wavelengths at which the intensities of the first and third fluorescence peaks were measured.) In Figure 2(a,b), the lines are fits to the rubbery-state and glassy-state temperature data well above and below polymer Tg, which is determined from the intersection of the two lines. Thus, Figure 2(a,b) indicate that the temperature dependence of I3/I1 can provide good determinations of Tg in MApyrene-labeled PS (Tg = 375 K via this fluorescence ratio method and DSC) and MApyrene-labeled PiBMA (Tg = 338 K via this fluorescence ratio method and DSC). In contrast, Figure 2(c) indicates that the temperature dependence of I3/I1 does not provide sensitivity to Tg in BApyrene-labeled PS, which has a Tg of 375 K as measured by DSC and the temperature dependence of integrated fluorescence intensity. (The BApyrene-labeled PS yields a good determination of Tg via measurement of integrated intensity10 but not of intensity ratios of the type used in the current study.) We hypothesize that the loss of sensitivity of I3/I1 to Tg when the pyrene dye is connected to the polymer via a butyl linker is related to the loss of sensitivity to polarity with increasing length of the linking group.81 Further study of this issue is warranted.

thumbnail image

Figure 2. Temperature dependence of the ratio of fluorescence intensities at the third and first peaks of pyrene dye-labeled polymers supported on silica (quartz) slides: (a) 270-nm-thick MApyrene-labeled PS film, (b) 405-nm-thick MApyrene-labeled PiBMA film, and (c) 345-nm-thick BApyrene-labeled PS film. (Data were obtained upon cooling.)

Download figure to PowerPoint

Figure 3 shows that temperature also has a strong effect on the fluorescence spectral shape of MApyrene-labeled PS in free-standing films. Furthermore, the effect of temperature can be observed in free-standing films expected to exhibit bulk Tg behavior [the 190-nm-thick film in Fig. 3(a)] and a nanoconfinement-related reduction in Tg [the 50-nm-thick film in Fig. 3(b)]. Interestingly, the overall shape of the MApyrene fluorescence spectrum is very different in the free-standing and substrate-supported films. With free-standing films (Fig. 3), the maximum intensity is observed from the fourth peak (at 396–398 nm); in contrast, the maximum intensity is usually observed from the first peak in the silica-supported films (Fig. 1). This difference may be due to a polarity effect associated with the polymer–silica interface in supported films, for example, the hydroxyl groups naturally on the surface of the silica (quartz) slides can alter the local polarity sensed by pyrene dyes in PS films. Related differences in I3/I1 values have been reported on inclusion of 0.5 wt % well-dispersed silica nanoparticles (14-nm-diameter) in bulk PS films and in a comparison of 500-nm-thick and 24-nm-thick PS films supported on silica.83

thumbnail image

Figure 3. (a) Temperature dependence of the fluorescence emission spectra of free-standing films of MApyrene-labeled PS: (a) 190-nm-thick MApyrene-labeled PS and (b) 50-nm-thick MApyrene-labeled PS. Spectra have been normalized to one by the peak intensity in the low temperature emission spectra. Downward pointing arrows indicate the approximate wavelengths at which I1 and I3 intensities are measured. Inset: Molecular structure of MApyrene.

Download figure to PowerPoint

Figure 4 provides plots of I3/I1 as a function of temperature for the 50-nm-thick and 190-nm-thick freely standing MApyrene-labeled PS films. As with the bulk, substrate-supported films in Figure 2, the plots in Figure 4 indicate that good determinations of Tg can be achieved in free-standing films. This is true regardless of whether the films exhibit bulk Tg response (Tg = 374 K or 101 °C for the 190-nm-thick film) or a nanoconfinement-related Tg reduction (Tg = 338 K or 65 °C for the 50-nm-thick film).

thumbnail image

Figure 4. Temperature and thickness dependences of the ratio of intensities at the third and first peaks of free-standing films of MApyrene-labeled PS: 190-nm-thick film (circles) and 50-nm-thick film (squares). (Data were obtained upon cooling.)

Download figure to PowerPoint

Figure 5(a) compares the experimental data on free-standing PS films obtained in the current study with data reported by Dalnoki-Veress et al.28, 29 for PS with Mn = 691 kg mol−1 and Mw = 767 kg mol−1. We have chosen to make this comparison because all but one free-standing film Tg data point taken in the current study are for PS with Mn = 701 kg mol−1, which matches well with the sample from Dalnoki-Veress et al.28, 29 The filled circles represent Tg values obtained by use of our fluorescence method on cooling the free-standing films from the rubbery state to the glassy state. We also provide two data points (open circles) that represent Tg values obtained by use of our fluorescence method upon heating the free-standing films from the glassy state to the rubbery state. We obtain good agreement between the measurements taken on cooling and on heating. This agreement may influence how future studies of the Tg-confinement effect are undertaken, because a large fraction of free-standing films undergo hole formation on cooling from the rubbery state, meaning that only a minority of such experiments lead to Tg determinations. Although not eliminated, hole formation effects may be reduced by taking measurements during a first heating cycle from the glassy state to the rubbery state.

thumbnail image

Figure 5. (a) Thickness dependence of Tg of free-standing films of MApyrene-labeled PS: Mn = 701 kg mol−1 on cooling (filled circles) and upon heating (open circles), and Mn = 1460 kg mol−1 on cooling (filled triangle). The open square symbols are Tg values of PS free-standing films with Mn = 691 kg mol−1 reported by Dalnoki-Veress et al.29 The horizontal dotted line corresponds to the bulk Tg value. (b) Magnified images of (a) in the 20–80 nm film thickness region. Diagonal lines are best fits to data reported by Dalnoki-Veress et al.29: Mn = 514 kg mol−1 (dash line), Mn = 691 kg mol−1 (solid line), Mn = 1180 kg mol−1 (dash-dot line), and Mn = 2070 kg mol−1 (dash-dot-dot line).

Download figure to PowerPoint

In agreement with the results of Dalnoki-Veress et al.28, 29 for free-standing PS films with Mn = 691 kg mol−1, we find that there is no effect of confinement on Tg down to a film thickness of ∼ 90 nm in our free-standing PS films with Mn = 701 kg mol−1. However, with decreasing thickness below 80–90 nm, there is a sharp reduction in Tg, with a 41-nm-thick film exhibiting a Tg of 326 K or 53 °C.

We also provide a single Tg data point for a higher molecular weight free-standing PS film with Mn = 1460 kg mol−1. This sample was made by combining 20 parts by weight of our MApyrene-labeled PS sample having Mn = 701 kg mol−1 with 80 parts by weight of an unlabeled, nearly monodisperse PS sample having Mn = 2000 kg mol−1. A 65-nm-thick free-standing film of this PS sample exhibits a Tg value of 342 K or 69 °C. This is 12 K below the Tg for a 65-nm-thick PS film made using only MApyrene-labeled PS with Mn = 701 kg mol−1. This comparison shows that although our fluorescence label is located on only the lower molecular chains in this bimodal mix, the Tg that is being sensed by the label is sensitive to the distribution of molecular weight.

To determine more rigorously the quality of the agreement between our data and the data by Dalnoki-Veress et al.,28, 29 we show in Figure 5(b) the Tg-confinement effect associated with free-standing PS films over the thickness range 20–80 nm. We also plot fit lines reported by Dalnoki-Veress et al.28, 29 for the Tg-confinement data for four MWs of PS. We observe that two of our three data points for the PS sample with Mn = 701 kg mol−1 (fluorescence data taken on cooling) touch or are on the fit line for the free-standing PS (Mn = 691 kg mol−1) data by Dalnoki-Veress et al.28, 29 Taken alone, our three data points for the PS sample with Mn = 701 kg mol−1 provide an acceptable fit to a straight line, albeit with a slope that is slightly smaller than that associated with the data by in refs.28 and29. It is noteworthy that our single data point for the PS sample with Mn = 1460 kg mol−1 falls between the Dalnoki-Veress et al. fit lines for free-standing films of PS with Mn = 1180 kg mol−1 and Mn = 2070 kg mol−1. Overall, our limited proof-of-principle data obtained using our self-referencing fluorescence method are in reasonable agreement with the data by Forrest, Dalnoki-Veress and Dutcher.

The positive results obtained in the current study suggest several important avenues for future investigation using our self-referencing fluorescence technique. First, given that free-standing films exhibit much stronger effects of confinement on Tg than substrate-supported films, it will be important to employ multilayer/self-referencing fluorescence methods to achieve the first characterization of the distribution of Tgs across free-standing films. This can provide critical data to address whether substantial Tg gradients are present across nanoconfined free-standing films and, if so, how the gradients compare with those in substrate-supported films. Second, as illustrated in Figure 2(b), our self-referencing fluorescence method works in polymers other than PS. Given that the vast majority of experiments involving free-standing films have been limited to PS, there is great need to undertake similar studies with other polymer systems. Third, using dye labels other than pyrene, our group has developed intensity-related fluorescence methods to investigate physical aging rates in polymer films,9, 84, 85 including the distribution of physical aging rates in substrate-supported polymer films.11, 61 If we can modify our approach to yield self-referencing fluorescence measurements that are sensitive to densification at constant temperature in the glassy state, we will have the opportunity to characterize the effect of confinement on physical aging in free-standing films and thereby determine whether the distribution of physical aging rates across a free-standing film is similar to or different from that associated with Tg.

Finally, it should be possible to determine whether the sensitivity of our self-referencing fluorescence method to the Tg-confinement effect in free-standing films is defined by a particular molecular weight average or is related to a molecular weight distribution in some complex manner. Such studies may also allow for a critical test of the mechanism suggested by de Gennes33 as a possible explanation for why the Tg-confinement effect is much stronger in free-standing films than in supported films. Binder86 has cogently explained de Gennes's mechanism as follows: “For thin enough films, one has to consider the possibility that the loops of (a) chain extend to the surface region of the thin film, where the mobility is greatly enhanced. These loops extending to the free surface then allow for a reptation-like dynamics (kinks may diffuse along the loop) and hence offer an alternative route towards structural relaxation… Obviously, this mechanism of faster structural relaxation can come into play only when (the film thickness is less than) the radius of a chain in the melt.” Such a mechanism provides a potential explanation for the important role of molecular weight on the Tg-confinement effect in free-standing PS films, even though the Tg-confinement exhibits no significant molecular weight dependence in supported PS films.10 (Related discussion on this and other potential mechanisms is given in ref.87.) This difference between free-standing and supported polymer films is one of the most intriguing results to arise from studies of the effects of nanoscale confinement on glass-formers. Any experimental progress that can be made in addressing the origin of this difference will be important.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. Acknowledgements
  7. REFERENCES AND NOTES

We acknowledge the support of the NSF-MRSEC program (Grant DMR-0520513) and Northwestern University, and we thank Robert Sandoval for assisting in the synthesis of the BApyrene-labeled PS.

REFERENCES AND NOTES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. Acknowledgements
  7. REFERENCES AND NOTES