Characterization of Molecular Motion in Solid Polymers by Variable

Characterization of Molecular Motion in Solid Polymers by Variable ...https://pubs.acs.org/doi/pdfplus/10.1021/bk-1984-0247.ch006implications for insi...
1 downloads 0 Views 1MB Size
6 Characterization of Molecular Motion in Solid Polymers by Variable Temperature Magic Angle Spinning CNMR Downloaded via UNIV OF CALIFORNIA SANTA BARBARA on July 10, 2018 at 22:20:08 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

13

W. W. FLEMING, J. R. LYERLA, and C. S. YANNONI IBM Research Laboratory, San Jose, CA 95193 The inclusion of a variable temperature magic-angle spinning capability for solid state C NMR spectroscopy makes feasible the investigation by C relaxation parameters of structural and motional features of polymers above and below Tg and in temperature regions of secondary relaxations. Herein, we report variable temperature (50K to 323K) spectral data on semicrystalline poly(propylene) and glassy PMMA. Illustrative of the data are the T and T results for isotactic poly(propylene) over the temperature range 50K to 300K. All carbons in the repeat unit show minima in T and T which reflect methyl group reorientation motion at the appropriate measuring frequencies (15 MHz and 57 kHz). The T data for CH and CH carbons indicate the importance of spin-spin as well as spin-lattice pathways in their rotating frame relaxation over much of the temperature interval studied. An interesting spectral observation is the strong motional broadening of the methyl group in the temperature region of the T minimum. These and other facets of the poly(propylene) data as well as similar data for PMMA are discussed with respect to their implications for insight into polymer chain dynamics in the solid state. 13

13

1



1





2



One of the principal advantages of C P M A S experiments is that resolution in the solid state allows individual-carbon relaxation experiments to be performed. If a sufficient number of unique resonances exist, the results can be interpreted in terms of rigid-body and local motions (e.g., methyl rotation, segmental modes in polymers, etc.) (1,2). This presents a distinct advantage over the more common proton relaxation measurements, in which efficient spin diffusion usually results in averaging of relaxation behavior over the ensemble of protons to yield a single relaxation time for all protons. This makes interpretation of the data in terms of unique motions difficult.

0097-6156/ 84/ 0247^0083506.00/ 0 © 1984 American Chemical Society Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

84

N M R A N D

MACROMOLECULES

Relaxation parameters of interest for the study of polymers include 1) C and * H spin-lattice relaxation times ( T and T ) , 2) the spin-spin relaxation time T , 3) the nuclear Overhauser enhancement ( N O E ) , 4) the proton and carbon rotating-frame relaxation times ( T ^ and T ^ ) , 5) the C - H cross-relaxation time T , and 6) the proton relaxation time in the dipolar state, T (2). Not all of these parameters provide information in a direct manner; nonetheless, the inferred information is important in characterizing motional frequencies and amplitudes in solids. The measurement of data over a range of temperatures is fundamental to this characterization. The initial studies of carbon relaxation in polymers have emphasized T j and T measurements, which provide information on molecular motions in the M H z and kHz frequency ranges, respectively. Schaefer and Stejskal have carried out the pioneering work in their investigations of glassy polymers (1). In particular, they stress the utility of T measurements for probing the dynamic heterogeneity of the glassy state and as a potential source of insight into the mechanical and other physical properties of polymers at the molecular level. Garroway and co-workers (3) reported the first variable-temperature ( V T - M A S ) T j results in their study of epoxy resins, and together with VanderHart, (4) have detailed the complications in extracting information on molecular motion from T experiments. In this paper, we report the first extensive sub-ambient V T - M A S C Tj and T data on macromolecules. The emphasis of the study was placed on isotactic poly (propylene) (PP) and atactic poly (methylmethacrylate) ( P M M A ) as they represent semi-crystalline and glassy polymers, respectively. Specifics of the investigation were directed to the issue of elucidating sidechain and backbone motions from the high frequency relaxation experiments. 1

3

1 C

1 H

2

p

C

1

l

p

H

D

p

l

l

p

p

1

l

3

p

Experimental 1

3

The C data at 15.1 M H z were acquired on a modified Nicolet TT-14 N M R system. The features of this spectrometer and of the spinning assembly have been reported previously (5). Samples were machined into the shape of Andrew-type rotors and used directly for the various studies. Temperature variation was achieved by cooling or heating the helium gas used for driving the rotor. The temperature was controlled to ± 2 ° C with a home built temperature sensing and heater/feedback network. Spin-lattice relaxation times, Τ j were collected using a pulse sequence developed by Torchia (6) which allows cross-polarization enhancement of the signals. The T data were determined at 57 kHz using T methodology of Schaefer et al. (1). The PP examined was a 90% isotactic, 70% crystalline sample. The P M M A was an atactic commercial polymer. l

l

p

p

Results and Discussion 1

3

Figure 1 shows the C P M A S C spectra of PP as a function of temperature. The interesting feature is the progressive broadening of the methyl resonance

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

6.

FLEMING ETAL.

Molecular Motion in Solid

Figure 1. C P M A S

1

3

Polymers

C spectra of poly (propylene) as a function o f

temperature.

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

85

86

N M R A N D

MACROMOLECULES

as the temperature is lowered. At « 1 1 0 K , the resonance is broadened to the point of disappearing from the spectrum. However, at temperatures below 77K, the methyl resonance narrows and reappears in the spectrum. This broadening phenomenon arises as the reorientation rate of the methyl group about the C3 axis becomes insufficient to average (stochastically) dipolar interaction with the methyl protons. A t the onset of broadening, the methyl motion has a correlation time comparable to the inverse of the strength (in frequency units) of the proton decoupling field. This reduces the efficiency of the rf decoupling and leads to a maximum linewidth of the carbon when the motions occur at the frequency corresponding to the amplitude of the proton decoupling field. Rothwell and Waugh (7) have developed the theory for T (the inverse of the C linewidth) for an interplay between stochastic and coherent motions. For such a system, the profile of linewidth vs. temperature shows a maximum when the correlation time for molecular motion, T , is equal to the modulation period of the decoupling, (Ι/ω^). In the "short correlation time" limit ( ω τ < < 1 ) (high temperature), the linewidth is reduced by the rapid motional averaging, while in the "long correlation time" limit ( ω τ , > > 1 ) (low temperature), the linewidth is reduced by efficient decoupling of C - H dipolar interactions. The spectra of PP in Figure 1 are consistent with the progression of the methyl resonance through the linewidth regions as the temperature is lowered. The reappearance (narrowing) of the methyl resonance at 77K indicates that the "long correlation time" regime has been reached (7). Further proof of the progressive changes in correlation time for methyl rotation as the temperature is lowered is provided by the decoupling-field dependence of the linewidth. A t about 160K, the methyl linewidth is independent of decoupling field, while at 77K the linewidth varies as the inverse square of the decoupling field. This is the expected dependence for the transition between the extreme narrowing and long correlation time regimes (7). Finally, from the expression for the C linewidth derived by Rothwell and Waugh [Eq. (1)] and the correlation time and temperature of the T minimum observed by McBrierty et al. (8), 2

1

3

c

1

0

1

1

(

3

l

2

_ l . yhl*

(

T

c

\

p

( 1 )

we calculate, for ω^=5Ί kHz (the value of the decoupling field used to obtain the spectra in Figure 1), that the maximum broadening for the methyl resonance would occur at 109K, in excellent agreement with our observations. This broadening of the methyl resonance observed in PP is also found in polycarbonate, P M M A , and epoxy polymers. It should be a general phenomenon for rapidly reorienting side groups or main-chain carbons in polymers. For semicrystalline systems, where the local molecular structure is relatively homogeneous, severe broadening should result in the "disappearance" of resonance lines from the spectra. For glassy systems, where there is more heterogeneity in the local molecular environment, the

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

6.

F L E M I N G ET A L .

87

Molecular Motion in Solid Polymers

effect may result in significant changes in resonance lineshape as a function of temperature as the carbons in differing environments undergo severe broadening. Of course, the phenomenon may be used to determine T for the group undergoing the motion (7); however, the severe broadening does limit the ability to measure high-frequency relaxation times in such temperature intervals. The C spin-lattice relaxation times for isotactic PP are shown in Figure 2. Primarily, the data represent that of the crystalline component. The semilog plots of intensity vs. time were nearly exponential for each of the carbons at all temperatures. Over the temperature range, each carbon in the repeat unit displays an individual relaxation time. The methyl relaxation appears to be dominated by methyl C reorientation. If it is assumed that a C - H heteronuclear relaxation mechanism is operative, a calculation of the methyl carbon relaxation time based on a Bloembergen-Purcell-Pound (BPP) formalism and the correlation time at the proton T j minimum (8) at - 1 1 0 ° C gives a value of 10 ms at - 1 1 0 ° C , in good agreement with the observed value of 17 ms. In addition, the methyl motion also seems to dominate the backbone relaxation. This is evidenced by the shorter T j observed for the methine carbon relative to methylene (despite there being two direct C - H interactions for the methylene carbon). Apparently, backbone motions are characterized by such small amplitudes and low frequencies that contributions from the direct C - H interactions to spectral density in the M H z region of the frequency spectrum are minor relative to those from side groups. The 1 / r distance dependence of dipolar relaxation thus accounts for both the long T j values of C H and C H carbons (one to two orders of magnitude) relative to the methyl carbon and the shorter T j values for methine carbons relative to methylene carbon. The fact that the observed T j minimum for C H and C H carbons is close to that reported for a proton T j minimum (at 30 M H z ) (8) in PP that was assigned to methyl reorientation provides unequivocal support for the dominance of the T j relaxation by methyl protons. c

1

3

3

6

2

2

The T data (Figure 3) for the C H and C H carbons also give an indication of methyl group rotational frequencies. As the temperature is lowered below 163K, the T j for these carbons increases and the T j decreases by roughly an order of magnitude between 163K and 95K, suggesting that the contribution of methyl proton motion to M H z spectral density is decreasing, while increasing in the kHz regime. The C H and C H T do not change greatly over the temperature interval from 163K to ambient, and, in contrast to the T j behavior, the C H carbon has the shorter T . The interpretation of the carbon T data is complicated by the fact that spin-spin (cross-relaxation) processes, as well as rotating-frame spin-lattice processes, contribute to the relaxation (4). Only the latter provide direct information on molecular motion. Although both processes show a dependence on the number of nearest-neighbor protons, the relative insensitivity of T to temperature and the approximate 2:1 ratio of C H / C H T values also suggest that spin-spin processes dominate the relaxation above 163K. (If spin-spin effects dominate the rotating-frame relaxation and the carbon cross-relaxation to the proton l

p

2

2

2

l

l

p

l p

p

l

2

C

p

H

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

N M R A N D

1

MACROMOLECULES

3

Figure 2. The C spin-lattice relaxation times at 15 M H z for isotactic poly(propylene) methylene ( · ) , methine (O), and methyl (A) carbons.

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

Molecular Motion in Solid

FLEMING ETAL.

Polymers

11 I I I 1—J I 1 1—ι—ι—ι—ϊ 2 4 6 8 10 12 14 1000/Τ(°Κ) Figure 3. The T data at 57 kHz for C H (O), C H ( · ) , and C H carbons in poly (propylene). l

p

2

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

3

(•)

90

N M R A N D

MACROMOLECULES

dipolar reservoir is less efficient than the corresponding proton spin-lattice relaxation T , the observed T for the C H carbon will be about 1 . 7 - 2 . 0 χ that of the C H carbon, based on the approximate twofold difference in the second moments (due to protons) for the two types of carbons. This is roughly the result observed in the data displayed in Figure 3.) Below 163K, the T of both carbons shorten and tend toward equality, indicating that spin-lattice processes derived from methyl reorientation are becoming competitive with the spin-spin process in relaxing the backbone carbon magnetization. McBrierty et al (8) report a proton T minimum at 97K, which reflects methyl reorientation at kHz frequencies. No clear minimum is observed in the C data, perhaps due to the interplay of the spin-spin and spin-lattice processes. Nonetheless, it is apparent that the methyl protons are responsible for the spin-lattice contributions to the C H and C H T values. 1

D

l

p

2

l

p

l

1

p

3

2

l

p

Further evidence for the effect of spin-spin processes on T in PP is given in Figure 4, which shows T plotted against the reciprocal of the temperature as a function of the rotating frame field. As indicated in Table 1, in the case of motion dominating T , there is a square dependence of T on field. For spin-spin domination, there is an exponential dependence. The results at room temperature clearly display a dependence greater than the 4 χ suggested for motion and the field variation. Only at temperatures less than 150K with large rotating-frame fields are strong motional effects observed. As previously discussed, these arise from methyl rotation. The domination of both spin-lattice relaxation times for C H and C H carbons in PP by methyl reorientation is clearly disappointing, since the potential for information on backbone motion due to the high resolution of the C P M A S experiments is not realized. The implication is that it may not be possible to observe backbone motion in crystalline materials having rapidly reorienting side groups without resorting to deuterium substitution of these side groups. The T j data for various carbons in P M M A are given in Figure 5. Cleai deviations from nonexponential behavior of the magnetization were often observed. Behavior different from that observed for PP presumably arises because the high degree of stereoregularity and high crystallinity of the PP provide a more homogeneous local environment than in glassy P M M A , where distributions of relaxation times are commonly observed, owing to site heterogeneity. For P M M A , the reported relaxation times represent the long-time portion of the magnetization decay curves. The results for P M M A tend to cluster over the temperature range studied, except for the α-methyl carbon. The rapid relaxation for this carbon in the temperature range from 2 0 ° C to - 7 0 ° C is consistent with the proton T j minimum at about -23 ° C assigned to α-methyl rotation at M H z frequencies (9). The T data for P M M A are summarized in Table 2. As in the case of PP, the a - C H undergoes motional broadening and disappears from the spectrum near the minimum in T . In P M M A , severe broadening occurs in the temperature range between 140K and 200K. At lower temperatures, the l

l

p

p

l p

l

p

2

l

p

3

l p

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

6.

FLEMING ETAL.

Molecular Motion in Solid Polymers

Ο CH Δ CH ©A 33 KHz 2

1001



ΟΔ45 •A 63

KHz

KHz

50 r •

A 6

J 8

L 10

(1000/T)

Figure 4. T h e T j for methine (circles) and methylene (triangles) carbons poly(propylene) as a function of the rotating-frame field and temperature; (a) 33 k H z , (b) 45 k H z , and (c) 63 k H z .

Randall; NMR and Macromolecules ACS Symposium Series; American Chemical Society: Washington, DC, 1984.

92

N M R A N D MACROMOLECULES

T A B L E 1.

Motion:

1

3

CT

l

p

formalism.

BPP dipolar

_ l . = ΞίφΙ f(r ) = { c

2

«

ω

ι

)

J

(