Article pubs.acs.org/Macromolecules
Dynamics of Di(propylene glycol) Dibenzoate‑d10 in Poly(vinyl acetate) by Solid-State Deuterium NMR Boonta Hetayothin,† Roy A. Cabaniss,‡ and Frank D. Blum*,§ †
Department of Chemistry and Materials Research Center, Missouri University of Science and Technology, Rolla, Missouri 65409, United States ‡ Department of Computer Science, Missouri University of Science and Technology, Rolla, Missouri 65409, United States § Department of Chemistry, Oklahoma State University, Stillwater, Oklahoma 74078, United States S Supporting Information *
ABSTRACT: Deuterium solid-state NMR and temperaturemodulated differential scanning calorimetry were used to probe the dynamics of the plasticizer di(propylene glycol) dibenzoate (DPGDB-d10) in mixtures with poly(vinyl acetate) (PVAc). The plasticizer, deuterated in the phenyl rings, was synthesized, and 2H NMR spectra were obtained from PVAc samples with 10, 22, 27, and 37% deuterated plasticizer content as a function of temperature. The dynamics of the plasticizer in the plasticized polymer system were found to be heterogeneous with respect to different plasticizer molecules undergoing different motions. The experimental 2H NMR line shapes were fitted using a set of simulated spectra obtained from the MXQET program. The simulations were based on the superposition of two types of motion: a two-site jump motion, i.e., 180° ring flips, plus isotropic motions. The presence of the polymer allowed more plasticizer molecules to undergo 180° ring flips than in the bulk plasticizer. For the average of the log of the rate constants for the ring flips (⟨log k⟩) versus 1/temperature was linear with an apparent energy of activation of 75 kJ/ mol for ring flips. From both NMR and TMDSC, the reduction in Tg was proportional to the amount of plasticizer added. In addition, the Tgs of DPGDB-d10/PVAc as a function of plasticizer content were found to be similar to those of PVAc-d3/DPGDB as determined by NMR The NMR data for both the polymer and plasticizer and TMDSC data may be fit to a plasticization model of Jenkel and Heusch with an interaction parameter b = −0.53, suggesting that both species were sensitive to the same local environment.
■
INTRODUCTION Plasticized polymers are often binary systems whose physical properties can be manipulated fairly easily with the addition of an external plasticizer. Polymeric materials are often found to contain plasticizers to improve their flexibility, softness, pliability, and processability.1−4 For instance, poly(vinyl chloride) (PVC) can be used as a plastic, flexible film, or tubing depending on the addition of plasticizers. Plasticizers reduce the glass transition temperatures (Tg) of polymers via a process called plasticization, which can be described by free volume, for example, or other theories.4 Small molecule plasticizers usually occupy the space between polymeric chains creating additional free volume in the polymer matrix, thereby causing the longer range segmental motions to occur more easily.5,6 Understanding the dynamic behavior of binary glassy systems is important because the dynamics are related to many properties found in polymeric materials, including processability, material deformation, and the diffusion of plasticizer. The dynamics of plasticized polymers are more complex than those of (nonplasticized) homopolymers since they generally involve the physical characteristics of at least two components, which have © 2012 American Chemical Society
different Tgs. The molecular details of these systems can be difficult to uncover, yet this detail determines the macroscopic properties. Poly(vinyl acetate) (PVAc) is a polymer, which has been largely used in dispersants, adhesives, paints, and coatings. Plasticization of PVAc has extended the range of applications of this polymer. Di(propylene glycol) dibenzoate (DPGDB) is an effective plasticizer for this polymer.2 The PVAC/DPGDB system has been studied with both NMR (PVAc-d3) and temperature-modulated differential scanning calorimetry (TMDSC). It was found that the behaviors measured by NMR and TMDSC were consistent with measuring the same phenomena, but in different frequency ranges.2 Follow-up studies on adsorbed PVAc-d3 on silica showed that the dynamics of the polymer was not affected by the plasticizer at adsorbed amounts less than 0.81 mg/m2.7 This result raised the obvious questions about the relationship between the behavior of the Received: July 16, 2012 Revised: September 29, 2012 Published: November 9, 2012 9128
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
Article
plasticizer and the polymer. As a first step, we explore this relationship in the bulk polymer−plasticizer system. In this work, we report the synthesis of a deuterated plasticizer (DPGDB-d10) and used 2H NMR to probe the dynamics of this deuterated plasticizer in a PVAc matrix using the phenyl deuterons on the plasticizer. The Tgs of DPGDB-d10/PVAc determined by 2H NMR and temperature-modulated differential scanning calorimetry (TMDSC) were compared with those of PVAc-d3/DPGDB where the deuterated polymer was used as the probe.2 The 2H spectra of DPGDB-d10 were also simulated in an effort to understand the dynamics based on combinations of jump and isotropic motions. A strong correlation between the motion of the plasticizer and the polymer was found, especially as the system goes through the glass transition region.
pattern into a narrow line shape is observed. A narrow resonance is a result of an isotropic reorientation of the nuclei. Phenyl-ring flips are motions that are expected to occur in chemical species with aromatic rings. These have been observed in materials such as synthetic polymers31−34 or biopolymers.35,36 In the absence of molecular motion, the 2H NMR line shapes for aromatic rings can result in a Pake pattern with a splitting between the two horns of about Δ = 120−135 kHz. For phenyl rings, continuous rotational diffusion or 180° jumps about the symmetry axis reduce the quadrupole splittings by factors of 8 and 4, respectively, and result in changes in the line shapes. An 180° ring flip results in a 120° change in the tensor component that is parallel to the C−D bond, as shown in Figure 1.35−38
■
NMR BACKGROUND Deuterium (2H) NMR is a powerful technique for probing molecular motion in polymeric systems via changes in line shapes or measurement of the spin−lattice relaxation times.8−16 The 2H NMR line shapes are affected not only by the time scales of molecular motion of about 103−107 s−1 but also by the type of motion.16−18 NMR has been particularly useful for the determination of behavior in plasticized polymer systems2,19−24 and briefly outlined in our previous work.2 The deuterium NMR spectrum of a static X−D bond is dominated by the electrostatic interaction of the nuclear quadrupole moment (eQ) with the electric field gradient (eq) at the quadrupolar nucleus. The deuterium NMR transition frequencies are determined by (i) the Larmor frequency, ω0, (ii) the quadrupole coupling constant, e2qQ/h, (iii) the asymmetry parameter, η, and (iv) the orientation of the electric-field gradient (EFG) relative to the applied magnetic field or25 ω = ω0 ± δ(3 cos 2 θ − 1 − η sin 2 θ cos2 φ)
Figure 1. Positional changes of aromatic deuterons undergoing 180° ring flips.
■
(1)
with δ = 3e qQ/8h. The polar angle, θ, is the angle made between the principal axis of the electric field gradient tensor (which in organic compounds is typically along the C−D bond) and the magnetic field. The spherical polar angle, φ, also specifies the azimuthal orientation of the principal axis system of the electric field gradient tensor with respect to the external magnetic field. The asymmetry parameter, η, defines the off-axis contribution of the EFG and is usually zero for aliphatic C−D bonds, indicating that the electric field gradient tensor is axially symmetric.16,26−28 However, for aromatic C−D bonds, the motionally averaged field gradient tensor is not particularly axially symmetric.29 The quadrupole coupling constant (qcc) determines the width of the line shape which can be obtained from the superposition of resonance frequencies for two sets of transitions for mz of +1 → 0 and 0 → −1 as a function of the angle θ. The splitting between the two transitions, Δv, responsible for the line shape can be obtained from26,29 2
Δv =
3 2 (e qQ /h)(3 cos2 θ − 1) 4
EXPERIMENTAL SECTION
Synthesis of Deuterated Benzoyl Chloride-d5.39 Benzoyl chloride-d5 was prepared from a nucleophilic substitution of benzoic acid-d5 (Cambridge Isotope laboratories, Inc., Andover, MA) and thionyl chloride (Aldrich, Milwaukee, WI). Benzoic acid-d5 (5.017 g, 0.0395 mol) was added to a 50 mL two-necked round-bottom flask equipped with a drying tube. Thionyl chloride (4.070 g, 0.0342 mol) was then added, dropwise, over 30−40 min. The mixture was stirred under warm conditions about (35−40 °C) and refluxed for 1−1.5 h until hydrogen chloride gas was no longer evolved (tested with pH paper). The crude product was distilled twice until it was colorless (or clear pale yellow). The yield of the purified deuterated benzoyl chloride was 83%. Synthesis of Deuterated Plasticizer, Di(propylene glycol) Dibenzoate-d10.39 A mixture of 4.141 g (0.0285 mol) of benzoyl chloride-d5 and 1.925 g (0.0143 mol) of di(propylene glycol) (Aldrich, Milwaukee, WI) was added to a 50 mL two-necked round-bottom flask equipped with a drying tube. The reaction (Scheme 1) was refluxed at 90−110 °C until the hydrogen chloride gas was expelled. The reaction mixture was allowed to cool to room temperature, and then 20−25 mL of diethyl ether (Aldrich, Milwaukee, WI) was added. The solution was poured into a separatory funnel and washed with 20 mL of water. The aqueous layer was discarded. The resulting organic layer was washed with a 20% w/w Na2CO3 (Aldrich) solution until it became neutral (tested with pH paper). The Na2CO3 solution layer was discarded. The diethyl ether in the organic layer was evaporated, yielding a brownish viscous liquid. The crude product was further purified by redissolving it in diethyl ether, and activated charcoal (neutral, Aldrich) was added for decolorizing. The solution mixture was stirred with activated charcoal for 15 min, then the liquid portion was decanted, and the diethyl ether was evaporated. The product obtained was a clear viscous liquid with a pale yellow color. 1H NMR (400 MHz, CDCl3) δ: 1.34 (m, 6H), 3.66 (m, 4H), 4.27 (m, 1H), 5.30 (m, 1H). There was no visible intensity of the aromatic protons even at very high magnification, suggesting a high degree of deuterium incorporation. Preparation of Plasticized Polymers. A 25% w/w solution of PVAc, Mw = 170 kDa (Scientific Polymer Products, Inc., Ontario, NY), was prepared in a capped test tube using toluene as a solvent. Various amounts of deuterated plasticizer DPGDB-d10 were added to make 10, 22, 27, and 37% w/w plasticized polymers. The solutions were shaken using a mechanical shaker for 24 h. The polymer solutions were then cast on glass slides and allowed to dry at ambient temperature for 24 h.
(2)
A typical qcc for a C−D bond is about 160−180 kHz; therefore, the splitting, D, for the most intense peaks (i.e., the “horns” for θ = 90) for a static C−D is three-fourths of 160−180 or 120−135 kHz for an axially symmetric C−D. Both the motional rate (time scale of motion) and the type of motion affect the line shapes.17,30 The changes in NMR line shapes are sensitive to the exchange rates of typically about 103− 107 Hz. As the motional rates increase, a collapse of a powder 9129
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
Article
Scheme 1. Synthesis of Di(propylene glycol) Dibenzoate-d10
Figure 2. Selected simulated NMR line shapes at various jump rates (k, in kHz) that were made using MXQET for two-site, 180° ring flips. The resulting films were colorless and transparent, suggesting there was no apparent phase segregation between the plasticizer and polymers. Characterization of Plasticized Polymers by TGA and TMDSC. The compositions of the polymer/plasticizer mixtures were analyzed by
high-resolution thermogravimetric analysis (Hi-Res TGA 2950) (TA Instruments, New Castle, DE). The scans were run with a heating rate of 20 °C/min using high-resolution mode from room temperature to 750 °C in air. The Tg’s were measured using temperature-modulated 9130
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
■
differential scanning calorimetry (TMDSC) (TA Instruments, model 2920, New Castle, DE). The sample pans were referenced against empty pans, and the cell was purged with nitrogen gas at a flow rate of 50 mL/ min. The samples were held at −40 °C for 5 min, heated to 150 °C at a rate of 2.5 °C/min with a modulation amplitude of ±0.5 °C and a period of 60 s, then held for 3 min, cooled to −40 °C at the same rate, and held for another 3 min. After the first heating and cooling scan, the second heating scan was applied with the same conditions as the first. In order to standardize the effects of previous thermal history, the Tg reported is that determined from the second heating scan. The results are shown as differential reversing heat flow rate (dQrev/dT) vs temperature. A 10 °C smoothing was applied to the thermograms to reduce the highfrequency noise and highlight the transition without significantly distorting the thermogram. The reported Tg was taken at the peak of the derivative of the reversing heat flow vs temperature. The Tg for the pure plasticizer was measured using a Q2000 TA Instruments DSC. Characterization of Plasticized Polymers by Solid-State Deuterium (2H) NMR. The 2H NMR spectra were obtained using a Tecmag Discovery 400 MHz NMR spectrometer equipped with a highpower amplifier, a fast digitizer, and an Oxford AS-400 wide bore magnet. A fixed-frequency wide-line probe (Doty Scientific, Columbia, SC) with an 8 mm (diameter) sample coil was used. The quadrupoleecho pulse sequence (delay−90y−τ−90x−τ−acquisition) was used with a 2H frequency of 61.48 MHz. The 90° pulse width was 2.8 μs, and an echo time (τ) of 30 μs was used. Prior to collecting the data, the probe was tuned at each temperature. The raw data were left shifted so that the Fourier transform was started from the top of the echo. Approximately 4000−70 000 scans were collected, depending on the operating temperature, and no line broadening was used. The spectra were taken at intervals of 5 °C from −30 to 80 °C, depending upon the plasticizer content. The temperature was controlled with an accuracy of ±1 °C. The spectra were processed using the Mestrec software package (Santiago de Compostela University, Spain), and stacked spectra were scaled to the same heights for comparison. NMR Simulations. The experimental NMR line shapes were simulated using a FORTRAN-based program, MXQET. 16,40,41 Simulations of the deuterium NMR line shapes have been used to model the motional mechanism (types of motion) and motional rates in different systems in which the change of NMR line shapes are sensitive in a frequency range between 104 and 108 s−1 (Hz). The simulations were based on the phenyl ring motions modeled as two-site jumps with 180° ring flips, with an asymmetry parameter, η = 0.04. Several authors have reported the asymmetry parameters for phenyl rings in the range 0.03−0.06.37,40,42−45 The 180° jump about the C2 symmetric axis does not result in an isotropic motion, even at fast jump rates, k > 108.40,41 Therefore, the addition of an axial of motion was needed to further average the qcc to obtain an isotropic pattern. For this work, it was sufficient to add a Gaussian, a mixed Gaussian/Lorentzian, or a Lorentzian single central component to mimic the isotropic resonance. The liquid-like behavior of the plasticizer, as evidenced by the relatively narrow resonance above the Tg, was not the focus of this study. The experimental line shapes were then fitted to a series of simulated spectra using a mathematical routine (MATLAB, the Mathworks, Inc., Natick, MA). The weights of each of the simulated spectra were found by minimizing the squares of the differences between the experimental spectra and the sum of the simulated ones, i.e., a constrained leastsquares fit. Some examples of the simulated spectra for different ring-flip jump rates are shown in the Figure 2. In the absence of motion, the 2H NMR line shapes for aromatic rings can result in a rigid lattice spectrum with a splitting between the two horns of about D = 120−135 kHz. Continuous rotational diffusion (D/8 = 15−17 kHz) or 180° jumps (D/4 = 30−34 kHz) reduce the quadrupole splitting and change the line shapes. For our 2H NMR spectra, the simulation was based on a two-site jump model for two different orientations of a phenyl deuteron that executes 180° ring flips. There was little or no evidence for continuous rotation of the phenyl rings in our system.
Article
RESULTS The amount of the plasticizer in the samples was analyzed using thermogravimetric analysis (TGA). The decomposition thermograms for bulk PVAc and plasticized samples are shown in Figure 3. The plasticizer decomposed in the range 290−325 °C, while
Figure 3. Thermogravimetric (TGA) analysis of (A) 0%, (B) 10%, (C) 22%, (D) 27%, and (E) 37% DPGDP-d10 in PVAc.
decomposition of the polymer occurred at higher temperatures, beyond 325 °C.2 The amount of plasticizer estimated from the TGA experiments was within 5% of those estimated from the compositions of the original mixtures. The quadrupole echo 2H NMR spectra for the pure plasticizer, DPGDB-d10, were taken at various temperatures ranging from −60 to −20 °C and are shown in Figure 4. The sample became liquid-like at temperatures above −20 °C. Unfortunately, it was not possible to obtain a good spectrum at −25 °C due to severely poor signal-to-noise (S/N) ratio. The spectra from −60 to −30 °C were consistent with a rigid powder pattern although the intensity of the spectra was reduced at higher temperatures
Figure 4. 2H NMR spectra for the pure plasticizer DPGDB-d10. 9131
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
Article
Figure 5. Experimental (black) and simulated (gray) 2H NMR spectra for a 10% DPGDB-d10/PVAc sample.
Accordingly, the Tg was defined at the temperature where the signal height of the central resonance of the spectrum was about 3 times (roughly equal intensities) that of the height of the horns. The onset of the glass transition was taken as the temperature at which there was a discernible isotropic resonance. Uncertainties in the Tg assignments are estimated to be less than ±5 °C. The 2H NMR spectra for the 22% DPGDB-d10/PVAc sample are shown in Figure 6 as a function of temperature. The signal-tonoise increases were due to increased DPGDB-d10 concentration. The NMR glass transition takes place roughly between 34 and 45 °C and is centered at Tg (NMR) of about 39 °C. The line shape of 22% DPGDB-d10/PVAc spectra did not seem to change much at temperatures below the Tg although the contribution from the 180° ring flips increased somewhat as the temperature increased. The spectra for the 27% and 37% plasticized samples are shown in Figures 7 and 8, respectively. These spectra were also similar to those of the 22% DPGDB-d10/PVAc. The collapse of the powder pattern occurred within a range of 29−34 °C or Tg (NMR) of about 32 °C for the 27% plasticized sample. For the 37% DPGDB-d10/PVAc, the powder pattern collapsed between 10 and 19 °C, and the Tg (NMR) was assigned as 15 °C. In general, a shift to higher mobility at the same temperatures was also observed as the plasticizer content increased. This effect was consistent with a lowering of the Tg expected with increased amounts of plasticizer. The summary of Tg (NMR) for DPGDBd10/PVAc at different plasticizer content is shown in Table 1 along with the Tonset and the resulting (roughly) half-width for the transition. The simulations of the experimental 2H NMR spectra are shown superimposed as lighter lines in Figures 5−8. The fits were quite good for the rigid and liquid-like ranges but not as good in the glass transition region where the signal-to-noise
(reduced S/N). The splitting of about 128 kHz between the horns was as expected with a full splitting in the wings of about 256 kHz (difficult to see). As the temperature was increased to −20 °C, more scans were necessary. As shown in Figure 4, there was only a very weak intensity for some phenyl rings undergoing 180° flips (inner horns at 29 kHz separation) in this low temperature range. The Tg(NMR) of plasticizer DPGDB-d10 was estimated to be about −25 °C. For DPGDB-d10 with PVAc, the quadrupole echo 2H NMR spectra were taken as a function of temperature and plasticizer amount. The resulting experimental and simulated spectra are shown in Figures 5−8. As previously noted, the phenyl deuterons appeared to have two motionally different states. Rigid C−D bonds were observed with a splitting of about 120−125 kHz between the two outer horns and phenyl rings executing 180° flips with splitting of about 30−34 kHz between the two inner horns. At low temperatures, the spectra for the 10% DPGDBd10/PVAc sample (shown in Figure 5) showed powder patterns from both static rings and those executing 180° ring flips. The rig flips were not observed for the bulk plasticizer at low temperatures. In the low temperature range of about 24−39 °C, the line shapes did not seem to change much with temperature. As the temperature was increased to as much as 80 °C, the powder pattern collapsed with the rise of a central resonance. The Tg based on the plasticizer motion was estimated to be about 59 °C. The central component increased in intensity and became a smooth single narrow resonance by 80 °C. Although it was easy to observe the spectra going from a Pake pattern (glassy state) to a relatively narrow isotropic resonance (rubbery state) with increasing temperature, assigning a Tg was not necessarily that straightforward. In this work, we have assigned T g s that are consistent with previous work. 7 9132
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
Article
Figure 6. Experimental (black) and simulated (gray) 2H NMR spectra for a 22% DPGDB-d10/PVAc sample.
The log of the average jump rates (log ⟨k⟩) as a function of temperature for different plasticizer contents (10, 22, 27, and 37%) are shown in the Supporting Information. The average correlation times, ⟨τc⟩, are proportional to the inverse of these jump rates. The plot of log ⟨k⟩ values as a function of inverse temperature for different plasticizer contents seemed to be linear for small amounts of plasticizer (10−22%). As the amount of plasticizer increased, the deviation from linearity was more pronounced. The activation energies for the 180° ring flips from linear portions for both the 10 and 22% plasticizer content were calculated to be 25 kJ/mol. However, with a broad distribution of jump rates (or correlation times), ⟨τc⟩ was sometimes dominated by longer correlation times. Therefore, the average of the logarithm of the jump rate, ⟨log k⟩ was also calculated for comparison and shown in Figure 9. The ⟨log k⟩ values, as a function of inverse temperature for different plasticizer contents, seemed to fall within the same curve and satisfactorily fit the Arrhenius equation. The calculated activation energy from this ⟨log k⟩ ranged from about 75 ± 2 kJ/mol for all of the samples taken together. Thermograms for samples containing different amounts of plasticizer were obtained using temperature-modulated differential scanning calorimetry (TMDSC) and are shown in Figure 10. The derivatives of the reversible heat flow rates were plotted as a function of temperature for the different DPGDB-d10 bulk PVAc samples. The Tg (TMDSC) was chosen as the maximum of the derivative of the reversible heat flow curve. The Tgs (TMDSC) observed for DPGDB-d10 bulk PVAc for a 10, 22, 27, and 37% plasticizer content were 31, 15, 7, and −7 °C, respectively, as shown in Table 3. These Tgs (TMDSC)
ratios of the spectra are poorer. The MXQET program was used to obtain a basic set of spectra with various jump rates ranging from 1 to 1 × 109 Hz. The simulations were based on a two-site jump model for a phenyl deuteron that executed 180° ring flips with some examples shown in Figure 2. MATLAB was used to fit the superpositions of simulated spectra to experimental spectra. The slow jump rates are effectively rigid in the NMR simulations. Good fits required a broad distribution of spectra with different jump rates covering the range above. The distributions of (log) jump rates as a function of temperature and plasticizer content are presented using bar graphs and described in the Supporting Information. For simplicity, the phenyl ring motions were categorized into three groups or regimes as slow, intermediate, and fast with respect to their jump rates, which resulted in different powder pattern intensities. It is known that a 180° jump about the C2 symmetric axis does not yield an isotropic motion, even with a fast exchange rate, k > 108.36,40,41 Fast isotropic motions, which result in a single narrow peak, occurred with overall motions and ring flips in the fast exchange limit, where the spectra were no longer sensitive to the jump rates. In the slow regime (1−104 Hz), there was no significant intensity loss during the quadrupole echo. However, when the exchange rate became faster, at about 104−107 Hz (classified as the intermediate regime), a significant loss of intensity was observed.25 Only a small amount of intensity loss occurred in the fast regime, k > 108 Hz. The motional components fractions used in the fittings at different temperatures are given in Table 2. Because of the signalto-noise ratios of the spectra, the fits showed changes with some local bumpiness, but the changes over a wider temperature ranges were clearly apparent. 9133
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
Article
Figure 7. Experimental (black) and simulated (gray) 2H NMR spectra for a 27% DPGDB-d10/PVAc sample.
Figure 8. Experimental (black) and simulated (gray) 2H NMR spectra for a 37% DPGDB-d10/PVAc sample.
decreased an average of 7 °C for every 5% plasticizer added. The reductions of the Tgs (TMDSC) as the amount of plasticizer increased were similar to those in the previous work, although there are some small differences because the polymer in the
previous study was deuterated and made in a separate synthesis.2 The width of the glass transition obtained from the width at halfheight depended upon the amount of plasticizer in the sample. The higher the plasticizer content was, the broader the transition 9134
dx.doi.org/10.1021/ma301463n | Macromolecules 2012, 45, 9128−9138
Macromolecules
Article
Table 1. 2H NMR Glass Transition Temperature Ranges for DPGDB-d10/PVAc at Various Plasticizer Contents7 Tg (NMR) % plasticizer
Tonset (°C)a
Tmid (°C)b
approx half-width (°C)
10 22 27 37 100
45 24 20 0 −30
59 39 32 15 −25
14 15 12 15 5
Where a narrowed component in the spectrum was first observed and Tmid is an estimate of the middle of the transition as defined in the text. b The reported Tg, see text, where the height of the central component is 3 times the height of the horns in the spectrum. a
Figure 9. Averages of the log jump rates values, ⟨log k⟩, for DPGDB-d10 as a function of temperature for the different plasticizer contents (10, 22, 27, and 37%).
was. A single, relatively narrow transition was observed from the TMDSC experiments, although at higher plastizer contents, there may be some additional thermal activity on the low temperature side of the main transition. There was no indication of any isolated bulk-like polymer or plasticizer in the TMDSC traces. This seemed to correlate well with observations of the physical appearance of the transparent films, indicating that phase separation had not occurred.
■
DISCUSSION The spectra for the pure plasticizer sample were noteworthy for what they did and also did not show. Those spectra changed from Pake powder patterns at −30 °C and below to a narrower central resonance at −20 °C. This change was indicative of rigid plasticizer (motion slow compared to about 250 kHz in this case) to faster isotropic overall (liquid-like motions). In other words, the system went from solid-like to liquid-like in this 10 °C range. In between these two regimes, there existed a region of intermediate motion, where no useable spectrum could be obtained. Compared to the motion of the plasticizer in the polymer-containing system (vide infra), the pure plasticizer
Figure 10. TMDSC derivative reversing heat flow curves for (A) 0%, (B) 10%, (C) 22%, (D) 27%, and (E) 37% plasticized-d10 PVAc. The maximum of the derivative curve is taken as the Tg (TMDSC).
motions were more homogeneous. A broader distribution of motions, like those in the polymer-containing system, would allow spectra to be obtained around −25 °C, for example. Spectra
Table 2. Motional Components Used To Simulate Experimental Line Shapesa,b 10% plast T (°C) −20 −10 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
S
35 33 22 23 16 c 14 3 c c 3
I
32 32 38 36 26 30 27 29 3 c c
F
33 36 40 41 51 57 29 10 3 12 1
22% plast L
0 0 0 0 7 13 30 58 94 88 96
S
60 31 23 26 22 30 41 16 27 c c 1
I
20 48 49 46 51 46 35 46 10 1 c c
F
20 21 28 28 27 24 24 10 c 4 5 1
27% plast L
0 0 0 0 0 0 0 28 63 95 95 98
S 46 52 24 47 41 32 52 16 c 2 1 c
I 41 29 59 39 40 48 36 46 1 c c 1
F 13 19 17 14 19 20 c 16 4 4 3 c
37% plast L 0 0 0 0 0 0 12 22 95 94 96 99
S
I
F
L
26 62 36 49 17 20 c c
64 28 57 38 47 11 c c
10 10 7 c 18 c 4 c
0 0 0 13 18 69 96 100
a Components: S, slow (k ≤ 1.0 × 104 Hz); I, intermediate (1.0 × 104 < k < 1.0 × 108 Hz); F, fast (k ≥ 1.0 × 108 Hz) where k is the jump rate, and L, liquid-like, which modeled as an isotropic component. bGiven as the percentages of each component in the simulated spectra. cA small (
