<?xml-model href='http://www.tei-c.org/release/xml/tei/custom/schema/relaxng/tei_all.rng' schematypens='http://relaxng.org/ns/structure/1.0'?><TEI xmlns="http://www.tei-c.org/ns/1.0">
	<teiHeader>
		<fileDesc>
			<titleStmt><title level='a'>Segmental Diffusion in Attractive Polymer Nanocomposites: A Quasi-Elastic Neutron Scattering Study</title></titleStmt>
			<publicationStmt>
				<publisher></publisher>
				<date>01/04/2019</date>
			</publicationStmt>
			<sourceDesc>
				<bibl> 
					<idno type="par_id">10108121</idno>
					<idno type="doi">10.1021/acs.macromol.8b01716</idno>
					<title level='j'>Macromolecules</title>
<idno>0024-9297</idno>
<biblScope unit="volume">52</biblScope>
<biblScope unit="issue">2</biblScope>					

					<author>Eric J. Bailey</author><author>Philip J. Griffin</author><author>Madhusudan Tyagi</author><author>Karen I. Winey</author>
				</bibl>
			</sourceDesc>
		</fileDesc>
		<profileDesc>
			<abstract><ab><![CDATA[We present a systematic study of segmental dynamics in model attractive polymer nanocomposites comprising poly(2-vinyl pyridine) (P2VP) and 26-nm diameter colloidal silica nanoparticles (NPs) using quasi-elastic neutron scattering (QENS). Unlike most dynamic measurements, QENS provides both spatial and temporal information on small length scales (~1 nm), fast time scales (~1 ns), and therefore at temperatures far above the glass transition. We find that on these length and time scales, P2VP segmental motion is well-described by classic translational diffusion even under extreme confinement, where the average interparticle spacing is on the order of the Kuhn length. The average segmental diffusion coefficient decreases monotonically with increasing NP concentration by up to a factor of ~5 at the highest NP concentrations (50 vol%). Interestingly, this reduction in segmental dynamics is very weakly dependent on P2VP molecular weight spanning the unentangled (10 kg/mol) to the highly entangled regimes (190 kg/mol). This stands in contrast to the well-documented molecular weight effect on segmental dynamics in attractive polymer nanocomposites at lower temperatures.]]></ab></abstract>
		</profileDesc>
	</teiHeader>
	<text><body xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink">
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Introduction:</head><p>The addition of nanoparticles (NPs) to a polymer matrix, forming a polymer nanocomposite (PNC), can significantly enhance the thermal, mechanical, and functional properties of the host matrix. <ref type="bibr">1,</ref><ref type="bibr">2</ref> Furthermore, PNC materials have wide-ranging tunable properties that can be dominated by the polymer, the nanostructured filler, or the interfacial region. As such, they are appealing materials for a variety of fields. Several questions still exist regarding the dynamic properties of free and interfacial chains. Polymer dynamics in polymer nanocomposites and polymer melts significantly influence or dictate their processability, applications, glass transition temperature (Tg), and various macroscopic properties (such as creep, toughness, and transport). In addition, due to the large surface area to volume ratio of NPs, PNCs are a model system to study the perturbation to polymer dynamics caused by a solid interface.</p><p>At the largest length scale, elastic recoil detection measurements have been conducted to probe center-of-mass polymer chain diffusion in PNCs with NPs that are athermal <ref type="bibr">3</ref> , attractive <ref type="bibr">4</ref> , grafted <ref type="bibr">5</ref> and anisotropic <ref type="bibr">[6]</ref><ref type="bibr">[7]</ref><ref type="bibr">[8]</ref><ref type="bibr">[9]</ref> . This work has recently been reviewed. <ref type="bibr">10</ref> Similar dynamics were probed using nuclearmagnetic-resonance techniques as well. <ref type="bibr">11</ref> On a smaller length scale, segmental dynamics in PNCs have received more attention due to their relevance toward ion transport, small molecule separation, and the glass transition. This topic has also recently been reviewed from different perspectives. <ref type="bibr">[12]</ref><ref type="bibr">[13]</ref><ref type="bibr">[14]</ref><ref type="bibr">[15]</ref><ref type="bibr">[16]</ref> However, additional fundamental studies are needed to explore the complex parameter space and understand the underlying physics of interfacial and confined polymer dynamics.</p><p>There are several methods that can be used to analyze segmental dynamics including nuclear magnetic resonance (NMR), dynamic mechanical measurements (DMA), broadband dielectric spectroscopy (BDS), temperature-modulated differential scanning calorimetry (TMDSC), neutron spin echo (NSE) and quasi-elastic neutron scattering (QENS). The approximate timescales associated with many of these measurement techniques, as they pertain to segmental dynamics, are schematically represented in Figure <ref type="figure">1</ref>. Also included in Figure <ref type="figure">1</ref> is a characteristic polymer segmental relaxation (&#61537;-process) curve following Vogel-Fulcher-Tammann (VFT) temperature dependence,</p><p>where &#61556;0, B and T0 are fitting parameters related to high temperature relaxation time, fragility, and Vogel temperature, respectively. To probe the &#61537;-process with a particular technique, the measurement temperature must be chosen such that the &#61537;-process falls within the accessible temporal window of the technique. For example, TMDSC measurements are particularly useful for measurements near Tg (~0.5 -50 s) and BDS is useful for its coverage of over 6 decades in relaxation times at temperatures above Tg. <ref type="bibr">17</ref> QENS, the focus of this paper, has the advantage of spanning even shorter time scales while providing simultaneous temporal and spatial information to capture the timescales and geometries of measured motions. QENS measurements have been performed on filled rubber <ref type="bibr">18,</ref><ref type="bibr">19</ref> , polymer/layered silicate nanocomposites <ref type="bibr">20,</ref><ref type="bibr">21</ref> , and other PNC systems <ref type="bibr">[22]</ref><ref type="bibr">[23]</ref><ref type="bibr">[24]</ref><ref type="bibr">[25]</ref><ref type="bibr">[26]</ref><ref type="bibr">[27]</ref><ref type="bibr">[28]</ref> . For example, in a PNC where free chains were removed by extraction, Roh et al. found slower relaxations and increased dynamic heterogeneity for 1,4polybutadiene segments near aggregated carbon black NPs compared to segments in bulk polymer. <ref type="bibr">19</ref> For crystalline poly(dimethyl siloxane)/SiO2 PNCs, multiple polymer processes were analyzed and generally, bulk-like dynamics were measured below the polymer glass transition while slow dynamics attributed to interfacial polymer were identified at higher temperatures. <ref type="bibr">23</ref> Additionally, QENS and NSE experiments on low molecular weight polyethylene glycol and SiO2 (an attractive interaction) shows physically adsorbed chains are dynamically active with pico-second segmental dynamics at 413 K, but slowed relative to bulk. <ref type="bibr">26</ref> Recently, polymer nanocomposites comprising poly(2-vinyl pyridine) (P2VP) and colloidal silica (SiO2) have emerged as model systems for studying the properties of attractive PNCs. <ref type="bibr">[29]</ref><ref type="bibr">[30]</ref><ref type="bibr">[31]</ref> In these systems, it has been experimentally shown that a physically adsorbed bound polymer layer of thickness ~Rg spontaneously forms on the NP surface in solution and persists in the melt state. <ref type="bibr">30</ref> On the atomic level, hydrogen bonding between the nitrogen in P2VP and native hydroxyl groups on the surface of SiO2 was directly observed by sum frequency generation and X-ray photoelectron spectroscopic techniques. <ref type="bibr">31</ref> The segmental dynamics of P2VP/SiO2 PNCs have also been studied by dielectric spectroscopy as a function of NP concentration <ref type="bibr">29</ref> , NP size <ref type="bibr">32</ref> , polymer molecular weight <ref type="bibr">33</ref> , and interfacial bonding strength <ref type="bibr">34</ref> . An interfacial layer on the order of a few nanometers with suppressed segmental dynamics has been identified in these experiments, and the degree to which the dynamics are suppressed increases with decreasing polymer molecular weight, among other characteristics. <ref type="bibr">15,</ref><ref type="bibr">31,</ref><ref type="bibr">33</ref> Analysis of small-angle X-ray scattering <ref type="bibr">33</ref> , pycnometry data <ref type="bibr">31</ref> , and various spectroscopic techniques <ref type="bibr">31</ref> suggest that shorter chains pack more efficiently at the NP surface, leading to a greater fraction of physisorbed segments, slower interfacial dynamics, and a thicker interfacial layer. Despite this progress, a recent review by Sokolov et al. highlighted that the impact of temperature on the structure and dynamics of interfacial polymer, and whether the molecular weight effect is a kinetically trapped phenomena or an equilibrium state, remains an open question.</p><p>In this article, we present a systematic study of segmental dynamics in attractive polymer nanocomposites comprising P2VP and colloidal SiO2 NPs using TMDSC (T~Tg) and QENS (T~Tg+150K).</p><p>We show that motions of P2VP segments on ~1 nm length scales and ~1 ns time scales are well described by classic translational diffusion, even at NP concentrations of ~50 vol% where the average interparticle spacing is ~2 nm. The average segmental diffusion coefficient decreases with increasing NP concentration by up to a factor of ~5 and is mostly independent of temperature over the studied temperature range. In contrast to the well-documented molecular weight dependence of segmental diffusion in the deeply supercooled regime, our measurements of the same dynamic process at higher temperatures show reduced segmental dynamics that are largely independent of matrix molecular weight. Finally, by comparing TMDSC, BDS, and QENS, our results suggest that temperature has a significant impact on the NP-induced perturbation to segmental dynamics in PNCs and highlights the unique and complementary insights that can be provided by QENS.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Experimental Section:</head><p>Materials: All poly(2-vinylpyridine) (P2VP) was purchased from Scientific Polymer Products, Inc. and used as received. To study the molecular weight dependence of interfacial dynamics, unentangled (10 kg/mol), lightly-entangled (40 kg/mol), and well-entangled (190 kg/mol, M/Me &#8776; 11) P2VP were also studied. <ref type="bibr">30</ref> Throughout this paper, these samples will be referred to as 10, 40, and 190  Representative TEM micrographs illustrate that NPs remain well-dispersed in the as-dried PNCs (Figure <ref type="figure">S1</ref>). The presence of a physically adsorbed bound layer in P2VP/SiO2 PNCs is known to promote good NP dispersion and prevent NP-NP aggregation. <ref type="bibr">30,</ref><ref type="bibr">34,</ref><ref type="bibr">37</ref> Thermogravimetric Analysis (TGA): Polymer degradation behavior and NP concentrations were measured via TGA using a TA instruments SDT Q600. For each measurement, a sample of 5-10 mg was placed in a platinum pan and heated from 300 K to ~1100 K at a rate of 5 K/min under air purge. NP concentrations listed in Table <ref type="table">1</ref> were calculated with the TGA results and densities of 1.2 and 2.3 g/cm 3 for P2VP and SiO2, respectively. 33</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Temperature-Modulated Differential Scanning Calorimetry (TMDSC):</head><p>The calorimetric glass transition was measured via TMDSC with a TA Instruments Q2000. All measurements were made upon cooling a sample of &#8805; 5 mg of polymer at a rate of 5 K/min with a modulation time of 30 sec and amplitude of &#177;0.5 K over a temperature range of Tg &#177; 60 K. Tg was defined as the inflection point of the heat flow thermograms and all results were reproduced.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Broadband Dielectric Spectroscopy (BDS): Segmental dynamics of bulk P2VP was measured</head><p>with BDS using a Solartron ModuLab XM MTS with the femto-ammeter accessory. Polymer films were placed between steel electrodes and separated with 50 &#61549;m silica spacers. Samples were annealed in the cryostat at 420 K until the imaginary permittivity spectra stopped changing. Isothermal frequency sweeps from 10 -1 -10 6 Hz were measured every 5 K between 380 and 450 K on cooling. Measurements were made after heating again to ensure reproducibility.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Quasi-Elastic Neutron Scattering (QENS):</head><p>Inelastic neutron scattering measures the double differential scattering cross-section (d 2 &#61555;&#61487;d&#61527;d&#61559;), which is related to the probability that a given incident neutron is scattered into a solid angle d&#61527; with an energy transfer d&#61559;. The double differential scattering cross-section has incoherent and coherent contributions, each of which can be expanded and related to the dynamic structure factors:</p><p>where k0 and k1 are the magnitude of incident and final wave vectors, respectively, N is the number of nuclei, &#61555; is the incoherent and coherent scattering cross sections of the nuclei, and S(q, &#61559;) is the incoherent and coherent dynamic structure factors. Because &#120590; &#119894;&#119899;&#119888; &#119867; (~80 barns) is much larger than all other atoms in this system (&#61555;other &lt; ~6 barns), we can generally assume that the signal is dominated by the incoherent contribution to Equation <ref type="formula">2</ref>. <ref type="bibr">38</ref> For example, using Equation <ref type="formula">2</ref>, ~90% of the signal in P2VP is incoherent and even at our maximum SiO2 concentration of 50 vol%, the polymer accounts for ~63% of the total scattering. <ref type="bibr">38</ref> Sinc(q, &#61559;) is the time and space Fourier transform of the self-part of the van Hove correlation function and combines spatial (q=k1-k0) and temporal (&#61559;) information for correlations between the single nuclei. <ref type="bibr">39</ref> Quasi-elastic neutron scattering measures Sinc(q, &#61559;) centered at &#61559;=&#61488;&#61484; and is typically used to probe diffusive motions on molecular length scales&#61486; QENS measurements were made at the High-Flux Backscattering Spectrometer (HFBS, NG2) at the NIST center of neutron research in Gaithersburg, MD, USA. <ref type="bibr">40</ref> Samples containing at least 200 mg of polymer were folded and sandwiched into aluminum foil and placed in cylindrical aluminum cans for measurements. Each sample was approximately 50 &#61549;m thick.</p><p>First, in a fixed window scan (FWS), the elastic scattering intensity Sinc(q, &#61559;=&#61488;) was measured as a function of temperature, starting at 50 K with a heating rate of 1 K/min. Second, Sinc(q,&#61559;) was measured at select temperatures over a q-range of 0.25-1.75 &#197; -1 and an energy range spanning -17-17 &#956;eV (with a resolution of 0.8 &#956;eV as defined by the elastic scattering of vanadium at 50 K). These q-and energy ranges correspond to molecular dynamic processes with length and time scales of approximately 3-25 &#197; and 40 ps-2 ns, respectively. The measurement temperatures were guided by extrapolating dielectric relaxation times (similar to Figure <ref type="figure">1</ref>) and further refined by choosing a temperature where mean squared displacements measured via FWS were ~7 &#197; 2 or at least 3 &#197; 2 for the lowest temperature measurements. QENS spectra were collected for 12 hours under vacuum after a 30-minute equilibration at the measurement temperature.</p><p>Analysis was primarily conducted in DAVE software. <ref type="bibr">41</ref> In this PNC system, relatively high temperatures are necessary to observe the segmental diffusion process in the dynamic window of QENS and are mostly unexplored. The thermal degradation temperatures of bulk P2VP, as defined by the temperature at which 5 wt% polymer is lost in TGA are &gt;600 K (Table <ref type="table">1</ref>).</p><p>(Full thermograms are presented in Figure <ref type="figure">S2</ref>.) The maximum temperature of FWSs and QENS measurements are sufficiently below the onset of thermal degradation, by at least 75 K (Table <ref type="table">1</ref>).</p><p>Nevertheless, a thorough analysis of molecular weight (GPC), Tg (TMDSC) and thermal degradation (TGA) of samples after QENS measurements are presented in Section 9 of Supporting Information. Although the thermal degradation behavior did not change after measurements (Figure <ref type="figure">S10</ref>), the molecular weight and glass transition temperature decreased slightly (Table <ref type="table">S1</ref> and Figure <ref type="figure">S9</ref>, respectively). It is important to note that molecular weights measured after QENS measurements are still categorically different and span the unentangled to well-entangled regimes. These changes in chain length and Tg are expected from slight polymer degradation, but do not affect the reported measurements of segmental dynamics or main conclusions of this work. <ref type="bibr">42</ref> To confirm this, Section 9 of Supporting Information also includes a timedependent analysis of QENS, showing the sample measurement was the same at the beginning and end of the experiment. This result demonstrates the reliability of these QENS measurements.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Results:</head><p>The glass transition temperature (Tg) of P2VP/SiO2 PNCs as measured by TMDSC are shown in Figure <ref type="figure">2</ref> and listed in Table <ref type="table">1</ref>. The absolute Tg shown in Figure <ref type="figure">2</ref> increases with P2VP molecular weight (MW) and NP concentration (Figure <ref type="figure">2</ref>, 40 kg/mol). It is well-established that the addition of highly attractive NPs causes an increase in Tg resulting from the slowing down of the primary structural relaxation (&#945;-process) at the NP-polymer interface. <ref type="bibr">15</ref> As such, it is expected that increased NP concentration causes an increase in Tg due to the larger volume fraction of 'interfacial' polymer affected by the NP surface.</p><p>Furthermore, the impact of the same concentration of NPs is much more pronounced for unentangled polymer (~4 K for 10 kg/mol P2VP) than in well-entangled polymer (&lt;1 K for 190 kg/mol P2VP). This increased perturbation for lower molecular weight PNCs has recently been reported and described by differences in interfacial packing. <ref type="bibr">31,</ref><ref type="bibr">33,</ref><ref type="bibr">34</ref> Importantly, our measurements agree with calorimetric measurements of Tg on similar systems.   P2VP segments in bulk and PNCs show similar low mobility for T &lt; Tg, consistent with thermal harmonic vibrations in the glassy state. <ref type="bibr">43</ref> For T &gt; Tg, polymer segments exhibit a dramatic increase in &#8249;x 2 &#8250; as they become more mobile and are able to relax in the melt state. The bulk polymer and both PNCs show the change in slope occurring at similar T-T g bulk , as expected from the small increase in calorimetric Tg with the addition of NPs. However, at T&gt;Tg, polymer segments in P2VP/SiO2 PNCs show significantly reduced mobility with increasing NP concentration. Also shown in Figure <ref type="figure">3</ref>   <ref type="bibr">38</ref> The MSD of nuclei in the dried SiO2 NPs show no change of slope and minimal mobility over all temperatures, as expected from thermal vibrations. <ref type="bibr">44,</ref><ref type="bibr">45</ref> Figure <ref type="figure">3</ref> shows the overall mobility of hydrogens in the sample but it is difficult to separate various types of polymer motion by monitoring only the elastic scattering intensity. For example, any protons mobile on the experimental length and time scale will contribute to &#8249;x 2 &#8250;, regardless of their motion being diffusion, reorientations, rotations, or other motions. To better understand and characterize the segmental mobility and dynamic processes, isothermal QENS measurements of SS(q,&#61559;) were made. According to Figure <ref type="figure">3</ref>, measurement temperatures of at least ~Tg+100 K will place segmental dynamics in the experimental length and time scale. Figure <ref type="figure">4a</ref> shows a representative QENS spectrum of bulk 40 kg/mol P2VP at 550 K and q=1.21 &#197; -1 and is compared to P2VP with NP concentration of 25 and 50 vol% in Figure <ref type="figure">S4</ref>. With this relatively simple single Lorentzian model, the data show no significant or systematic residuals (Figure <ref type="figure">S4</ref>) and therefore more complex models, such as the addition of another Lorentzian, are unwarranted.</p><p>Figure <ref type="figure">4b</ref> shows the full width at half max (FWHM) of the Lorentzian component for bulk 40 kg/mol P2VP plotted as a function of q 2 for several temperatures. The quasi-elastic broadening is found to increase linearly with q 2 , indicative of translational diffusive motions where the slope is related to the diffusion coefficient (FWHM~Dq 2 ). <ref type="bibr">46</ref> One may expect signatures of Rouse dynamics at low q (where FWHM~q 4 ) <ref type="bibr">25,</ref><ref type="bibr">46</ref> , but this is not apparent in our data. With a Kuhn segment length of ~1-2 nm for P2VP, the length scales probed by QENS are likely smaller than those associated with Rouse dynamics. <ref type="bibr">47</ref> The apparent non-zero y-intercept in Figure <ref type="figure">4b</ref> is expected from the presence of q-independent reorientational motions (such as pyridine ring fluctuations or &#61538;-relaxations) as well as potential contributions from multiple scattering events, which are expected to be minimal for the present sample dimensions. More complex models, such as jump diffusion <ref type="bibr">46</ref> , are often applied to polymeric systems but do not appreciably improve the fits as compared to the translational diffusion model for both bulk P2VP and PNCs. Importantly, the observed quasi-elastic broadening is substantially larger than the energy resolution, especially for T &#8805; 515 K. As expected, at higher temperatures, the observed FWHM increases as segmental mobility increases.</p><p>Using 535 K as an example, Figure <ref type="figure">5a</ref> shows the quasi-elastic broadening is reduced with increasing SiO2 NP concentration. The q-dependence of the FWHM for all systems and all measurement temperatures are included in Figure <ref type="figure">4b</ref> and S5. Using the translational diffusion model, the segmental diffusion coefficients (D&#61537;) were extracted from the slope of Figure <ref type="figure">5a</ref> and are plotted as a function of inverse temperature in Figure <ref type="figure">5b</ref>. These extracted diffusion coefficients can be directly compared to TMDSC and BDS through &#61556; ~ (D&#61537;q 2 ) -1 where a q of 0.63 &#197; -1 was chosen. As shown in Figure <ref type="figure">1</ref>, the measured QENS relaxations times for bulk 40 kg/mol P2VP at T &gt; 515 K are consistent with BDS and TMDSC measurements, suggesting that the observed dynamics are related to the primary structural relaxation process. A detailed discussion of analysis for BDS measurements and a comparison to literature is provided in Section 7 of Supporting Information.</p><p>At the high measurement temperatures (T&gt;Tg+100 K) and over the narrow temperature range studied by QENS, D&#61537; shows Arrhenius behavior for bulk and PNC materials (Figure <ref type="figure">5b</ref>). Although D&#61537;&#61484;PNC &lt; D&#61537;&#61484;Bulk, all materials exhibit similar activation energies. The effect of NP concentration is further highlighted by normalizing D&#61537; relative to bulk measurements at the same temperature (Figure <ref type="figure">5c</ref>) showing a monotonic decrease in the average polymer segmental diffusion coefficient with increasing NP concentration. Specifically, D&#61537; drops by ~40% with 25 vol% NP and ~80% with 50 vol% NP concentration when P2VP is lightly entangled. Furthermore, over the narrow temperature range measured, the reduction in diffusion coefficient is independent of temperature within experimental error. This slowing of segmental dynamics with increased NP concentration is consistent with the increase in Tg (Figure <ref type="figure">2</ref>) and decrease in MSD for all T&gt;Tg (Figure <ref type="figure">3</ref>), despite the isolation of diffusive motions of protons. Assuming monodisperse NPs randomly dispersed in the polymer matrix, the average interparticle distance (ID) is given as</p><p>where dNP and &#981;NP are the NP diameter and volume fraction, respectively. <ref type="bibr">48</ref> Thus, at &#981;NP = 25 and 50 vol%, ID is ~9.5 nm and ~2.2 nm, respectively. From dynamic and static measurements from various techniques, the length scale of the perturbed interfacial layer (from the perspective of segmental dynamics) is often reported as ~2-5 nm from the NP surface. <ref type="bibr">15</ref> As such, to a first approximation, the 50 vol% PNC can be considered an "all-interfacial" PNC wherein nearly all of the polymer segments are within the interfacial layer. Therefore, Figure <ref type="figure">5c</ref> suggests that the interfacial layer in this strongly attractive PNC system is dynamically active at these high temperatures and the measured segmental diffusion coefficient from QENS is slowed by nearly one order of magnitude.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Effect of chain length on interfacial dynamics:</head><p>To study the effect of molecular weight on interfacial dynamics, we studied PNCs with 25 vol% SiO2 dispersed in P2VP with molecular weights ranging from unentangled to well-entangled (10, 40, and 190 kg/mol), Table <ref type="table">1</ref>. The results from FWSs for each polymer and PNC are shown in Figure <ref type="figure">6</ref>. To account for the molecular weight dependence of T g bulk (Figure <ref type="figure">2</ref>), &#8249;x 2 &#8250; is normalized to T=T g bulk -100 K and temperature is presented relative to T g bulk . Data from all three MWs essentially collapse onto a master curve for bulk polymer and 25 vol% PNCs, showing that segmental mobility is largely independent of molecular weight (even in the case of PNCs). Isothermal QENS measurements were performed at 525, 535, and 535 K for PNCs and bulk polymers with MWs of 10, 40, and 190 kg/mol, respectively. At these temperatures, which are all ~Tg bulk +160 K, segments in bulk exhibit a similar average MSD (~7 &#197; 2 relative to &lt;x 2 &gt; at T=50 K), as shown in Figure <ref type="figure">S3</ref>. Given the weak temperature dependence of D&#61537;&#61484;PNC / D&#61537;&#61484;Bulk (Figure <ref type="figure">5c</ref>), we will compare these QENS measurements as isothermal.</p><p>All bulk and PNC materials exhibit classic characteristics of translational segmental diffusion (FWHM ~ D&#61537;q 2 ), as shown in Figure <ref type="figure">5a</ref> and Figure <ref type="figure">S7</ref>. Figure <ref type="figure">7a</ref> shows that D&#61537; for bulk P2VP is approximately independent of MW, differing by less than 30% from each other. Small variations in bulk D&#61537; are attributed to slight differences in measurement temperature relative to Tg (Table <ref type="table">1</ref>, Figure <ref type="figure">2</ref>) and fragility. The measured D&#61537; values for 25 vol% PNCs are also shown in Figure <ref type="figure">7a</ref> and the segmental diffusion coefficients are significantly suppressed upon the addition of attractive NPs.</p><p>To quantitatively compare the impact of molecular weight on segmental diffusion in PNCs, D&#61537;,PNC is normalized by D&#61537;,Bulk in Figure <ref type="figure">7b</ref>. The error bars in Figure <ref type="figure">7b</ref> represent the propagated error in fitting the q 2 dependence of the quasi-elastic broadening and do not include the errors associated with small differences in NP concentration, measurement temperatures, etc. For all molecular weights of P2VP, the addition of 25 vol% SiO2 NPs causes a substantial reduction in the average segmental diffusion coefficient.</p><p>For PNCs with 10 kg/mol and 190 kg/mol P2VP, D&#61537;,PNC/D&#61537;,bulk is 42&#177;5% (T=525 K) and 61&#177;8% (T=535 K), respectively. These conclusions are similar to those from the FWS presented in Figure <ref type="figure">6</ref>, that the addition of NPs significantly reduces the segmental dynamics in PNCs but the effect is weakly dependent on molecular weight. In contrast, a much stronger molecular weight dependence was observed in the difference between Tg in PNCs and bulk (&#61508;Tg): 4.2 K for 10 kg/mol and 0.5 for 190 kg/mol (Figure <ref type="figure">2</ref>). In addition, at the highest molecular weight of 190 kg/mol, although Tg approaches the bulk value, the dynamics measured by QENS at high temperature are still measurably reduced. contrast, our QENS analysis provides an average diffusion coefficient that is significantly reduced suggesting that this method is insensitive to the faster diffusion corresponding to bulk-like P2VP. Others have reported similar findings when comparing inelastic neutron scattering and other techniques, including NMR <ref type="bibr">26</ref> and ellipsometry <ref type="bibr">45</ref> . This has been explained by differences in dynamic sampling, wherein inelastic neutron scattering is biased to the slower processes. <ref type="bibr">45</ref> This is consistent with our data. We extract an average diffusion coefficient in PNC systems that is slower than that of bulk, even when the interparticle distance is nearly 10x larger than the measurement length scale. The discrepancy between neutron scattering and other techniques has also been described in terms of technique sensitivity and dynamic range, where the spectral shape is analyzed over only one order of magnitude in QENS (Figure <ref type="figure">1</ref>). This is also consistent with our PNC data being described by a single Lorentzian, despite the known heterogenous dynamic environment PNCs. Our direct comparison of TMDSC and QENS, along with similar measurements from BDS <ref type="bibr">29</ref> , demonstrate that considering differences in experimental probes and sensitivities is critical in future comparisons, especially in heterogeneous materials such as PNCs.</p><p>In QENS, when a segment relaxes slower than the experimental time scale (~2 ns), it appears immobile and therefore contributes to the elastic peak and is excluded from the QENS broadening analysis.</p><p>This effect can be directly quantified by the elastic incoherent structure factor (EISF), which represents the fraction of immobile nuclei and is calculated by the area of the elastic contribution relative to the sum of the elastic and quasielastic contributions. The EISF for each sample measured at ~Tg+160 K is shown as a function of q in Figure <ref type="figure">8a</ref>. For each bulk sample, nearly 80% of segments are mobile at q ~ 1 &#197; -1 but upon the addition of NPs, a smaller fraction of nuclei are mobile on this nanosecond time scale. Thorough analysis and fitting of the q-dependence of the EISF is beyond the scope of this study, but it is worth noting that our data follows a similar trend to comparable systems in literature. <ref type="bibr">20,</ref><ref type="bibr">46</ref> The addition of hydroxyl-terminated SiO2 NPs introduces additional scattering intensity that will contribute to the elastic fraction and therefore affect the EISF. To account for this contribution, we assume an upper estimate for the hydroxyl surface density of ~4.9 nm -2 and amorphous SiO2 and P2VP densities of 2.3 and 1.2 g/cm <ref type="bibr">3</ref> , respectively, and calculate the predicted incoherent and coherent scattering intensities of each sample using Equation <ref type="formula">2</ref>. <ref type="bibr">38,</ref><ref type="bibr">53</ref> For NP loadings of 25 and 50 vol%, the polymer scattering accounts for 83% and 63% of the total scattering contributions. It is important to note that the incoherent and coherent scattering from SiO2 does not affect the measured quasielastic broadening or D&#61537; because the nuclei are immobile on the experimental length and time scales (Figure <ref type="figure">3</ref>) and therefore contribute solely to the elastic scattering. This increase in elastic scattering was accounted for in the EISF, but in all samples, the reduction in mobile nuclei was found to be more than expected from just the addition of SiO2 NPs. This difference can be attributed to nuclei of the polymer that are slowed beyond temporal window of the experiment and therefore appear immobile, most likely belonging to segments closest to the attractive NP interface.</p><p>Using a simple three-phase model including SiO2, immobile polymer, and mobile polymer, the "interfacial width" can be extracted by correlating the measured fraction of immobile polymer to the increased NP-polymer interfacial volume resulting from increased NP concentration. The calculated interfacial width is shown in Figure <ref type="figure">8b</ref>, where the error bars represent the standard deviation of calculations for 0.55 &lt; q &lt; 1.6 &#197; -1 . The interfacial with is ~1 nm and is independent of NP loading (see 40 kg/mol) and molecular weight. This interfacial width does not delineate the slow segments from the bulk-like segments because the remaining segments are still slower than bulk (Figure <ref type="figure">5c</ref>). Instead, this interfacial width represents the estimated average distance from the NP surface after which segments relax within the window of QENS. Without accounting for scattering from SiO2 NPs, the interfacial width falsely appears to be ~2.5 nm but the MW dependence remains unchanged. The extracted interfacial widths via BDS <ref type="bibr">33</ref> and QENS (Figure <ref type="figure">8b</ref>) show surprisingly distinct molecular weight dependences. Whereas the interfacial thickness in BDS decreases from ~4 nm at low MW to ~2.5 nm at high MW (consistent with arguments of MW-dependent interfacial packing <ref type="bibr">31,</ref><ref type="bibr">33</ref> ), the interfacial thickness in QENS is consistently ~1 nm over the same molecular weight range. Although these are structural insights inferred from dynamic measurements, the same behavior is observed in direct measurements of segmental dynamics. In TMDSC (Figure <ref type="figure">2</ref>) and BDS <ref type="bibr">33</ref> measurements, the increase in glass transition temperature and decrease in segmental dynamics (respectively) are highly dependent on matrix MW, whereas we observed that D&#61537;,PNC/D&#61537;,Bulk in QENS is nearly independent of molecular weight.</p><p>Although QENS, TMDSC, and BDS probe the same dynamic process (Figure <ref type="figure">1</ref>), a notable difference between these experimental methods is the measurement temperature, which has two important implications. First, it remains unclear how the structure, dynamics, and interactions of the interfacially bound polymer segments depend on temperature. Thus, TMDSC (T~Tg), BDS (T&lt;Tg+60 K) and QENS (T~Tg+160 K) may probe fundamentally different perturbations to the segmental dynamic process. Limited experimental data suggests a reduced interfacial width at elevated temperatures, which is in line with the observations in Figure <ref type="figure">8b</ref>. <ref type="bibr">15</ref> Second, the loops and trains of the adsorbed polymers may depend on processing conditions, even after long annealing. <ref type="bibr">15</ref> In fact, processing details including concentrations of polymer and NP solutions, solvent quality, and annealing conditions, may impact the adsorbed polymer conformations and subsequently the interfacial dynamics. As such, it is important for the field to consider and clearly report these details. At elevated temperatures, not only is the entropy of interfacial chains promoted and local free volume increased, but the relative strength of the hydrogen bond is decreased.</p><p>Samples measured for several hours at the high temperatures as required for QENS may provide enough time and thermal energy for segments to sample their local environment and energetic landscape and reach a more equilibrium conformation. Future studies of annealing time and temperature are needed to fully understand the influence of processing on interfacial dynamics.</p><p>Not only can QENS provide complimentary dynamic measurements to other techniques but it offers several unique capabilities, making it a useful complement to the field of segmental dynamics in PNCs. In this work, the use of spatial correlations in bulk and PNCs showed that although the observed segmental dynamics are temporally slowed, they are relatively unperturbed spatially. The fast time scales probed by QENS captures dynamics at elevated temperatures to elucidate the role of temperature on interfacial segmental dynamics. In addition, the ability to quantify the fraction of mobile and immobile species allowed the extraction of structural parameters from dynamic measurements. Finally, unique capability of neutron scattering that has yet to be fully exploited is selective H/D labeling to isolate and investigate different polymer dynamics and processes within the chain (through intrachain deuteration) or spatially in the PNC (through interchain deuteration).</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Conclusion:</head><p>Quasi-elastic neutron scattering was used to study segmental dynamics in highly attractive polymer nanocomposites (PNCs) at high temperatures (~Tg+150 K). We isolate the role of nanoparticle (NP) concentration, temperature, and matrix molecular weight on segmental dynamics in model PNCs made of P2VP and 26 nm colloidal SiO2. We monitor the elastic scattering as a function of temperature to reveal proton mobility over a wide temperature range and measure the dynamic structure factor under isothermal conditions to probe dynamics on length and time scales of ~ 1 nm and ~ 1 ns, respectively. We show segmental mobility is strongly reduced for all T&gt;Tg upon the addition of NPs. At the QENS length and time scales, we observe classic translational diffusion of segments in bulk and in PNCs, even when the average interparticle separation distance is ~2 nm (50 vol% SiO2). Simultaneously, the average segmental diffusion coefficient is reduced by ~80% (relative to bulk) at NP concentrations of 50 vol%, showing strong temporal suppression without spatial perturbations. Similar observations are made for PNCs with unentangled and well-entangled matrix polymers. The decrease in segmental mobility for all T&gt;Tg and a reduced diffusion coefficient are highly dependent on NP concentration, but nearly independent of matrix molecular weight.</p><p>Several dynamic probes have been used to study segmental dynamics in highly attractive PNC systems. By comparing our QENS results to solely dynamic measurements on slower time-scales (and therefore lower temperatures), we highlight categorically different observations on similar PNC systems.</p><p>Namely, calorimetric measurements (measured at T~Tg) show a much stronger molecular weight dependence than QENS (measured at T&gt;&gt;Tg). These discrepancies provide insights into the effect of temperature on the observed segmental dynamics in attractive PNCs. Furthermore, the unique ability of Krook for assistance with TEM, and Steve Szewczyk for assistance with various measurements.</p></div></body>
		</text>
</TEI>
