close

Вход

Забыли?

вход по аккаунту

?

High-Resolution Solid-State 13C-NMR Spectroscopy of Polymers [New Analytical Methods (37)].

код для вставкиСкачать
High-Resolution Solid-state 13C-NMR Spectroscopy of Polymers
By Rudiger Voelkel*
New Analytical
Methods (37)
Dedicated to Professor Helmut Dorfel on the occasion of his 60th birthday
Until a few years ago, solid-state nuclear resonance yielded spectra containing broad lines
only. Meanwhile, CP/MAS-NMR spectroscopy has provided a method which gives narrow
nuclear resonance lines from a solid-state specimen as well. Using this technique, it is now
possible to produce spectra of “rare” nuclei (I3C, 29Si, I5N etc.) which are resolved in terms
of chemical structure. The analytical capabilities of NMR spectroscopy can be applied to
the solid state: it may be that it is necessary to identify compounds in the solid state because, for example, a solvent would alter the coordination sphere, or that it is desired to
monitor chemical reactions in the solid state, for example the baking of an enamel. Where a
substance in the solid state is concerned, high-resolution I3C-NMR spectroscopy provides
not only information about the chemical structure, but also about the solid state itself. To
mention just a few examples, information on the conformation, crystal structure and molecular dynamics, as well as molecular miscibility is given. This opens u p a broad spectrum of
applications, from a statement concerning the crystal modification of an active substance in
ready-to-use pharmaceutical preparations, e.g. tablets, to the question of whether two polymers are miscible with one another at a molecular level.
1. Introduction
Since the introduction of the CP/MAS (cross polarization, magic angle spinning) I3C-NMR experiment in its
present form[’] as a combination of dipolar decoupiing,[’I
sample rotation at the magic angler3]and cross polarization?] a number of articlesr51and monographs[61presenting
an introduction to the method have been published, as
well as articles providing an overview of individual areas
of application, e.g. polymers,[71 carbohydrates[’] and biopolymers.lg1
CP/MAS-NMR spectroscopy opens u p the possibility
of obtaining, even from a substance in the solid state, I3CNMR spectra, the signals of which can be correlated with
chemical structures. The resolution of individual signals
can be useful in two ways on the one hand, solid-state
NMR spectroscopy can be used in direct analogy with solution NMR spectroscopy as an analytical tool for the investigation of compounds which have low solubility or are
insoluble. On the other hand, it is also possible to make a
number of statements regarding the physics of the solid
substance. In both cases, the correlation of individual signals with chemical structural elements is an indispensible
prerequisite. Accordingly, Section 2 deals initially with the
question of the line width, and, in connection with the generation of the signals, quantifiability is also discussed. In
Section 3, relating to analytical applications, consideration
is given only to insoluble polymers. Attention centers on
reaction resins, which otherwise are analytically accessible
only with difficulty on account of their cross-linked nature.
They are however of great technical importance, for example as matrix materials in composites.
[*] Dr. R. Voelkel
Plastics Laboratory
BASF Aktiengesellschaft
D-6700 Ludwigshafen (FRG)
1468
0 VCH Verlagsgesellschafi mbH, 0-6940 Weinheim. 1988
In the second part (Sections 4 and 5 ) a number of physical applications are presented, which are less familiar to
chemists accustomed to NMR spectroscopy in solution.
One of the strengths of solid-state NMR spectroscopy is
that it provides information on questions of conformation,
crystal modification, molecular dynamics, and molecular
adjacency. An understanding of the physical behavior of
the sample is essential in order to take advantage of the
unique possibilities offered by solid-state NMR spectroscopy: for example, selectivity with respect to certain components is also useful in the case of analytical applications.
The present article is limited to the one-dimensional
methods which have already become well-established in
practice. Very promising additional possibilities offered by
two-dimensional NMR experiments will be dealt with separately by B. Blumich and H . W. Spiess.””
2. Line Width, Quantifiability
When reference is made to “magic angle spinning’’
NMR spectroscopy, what is almost always meant is the
combination of cross polarization (CP), the actual rotation
of the sample at the magic angle of 54.7” (MAS) and dipolar decoupling (DD) of the ‘H nuclei from the observed
I3C nucleus. With rigid solid substances only the combination of all three experimental “tricks” makes it possible
and practicable to record spectra of rare nuclei (most commonly I3C) resolved in terms of chemical structure.
Magic angle spinning, however, imposes the most stringent requirements on the experimental procedure, and
makes the most striking contribution to the narrowing of
the lines. The effect of MAS can be seen particularly well
in the case of pure polyacetylene with only one chemical
environment of the “C nuclei (Fig. 1). When the sample is
0570-0833/88/1111-1468 $ 02.50/0
Angew. Chem. Inr. Ed. Engl. 27 (1988) 1468-1483
ppm), and the question of the quantifiability of the spectra.
2.1. Line Width
CIS
trans
__7
250
bl
200
u,.~= 136.3
127.L
SSB
558
- "r
cis
A
n
200
100
d
- 6
Fig. 1. Solid-state "C-NMR spectra of CIS- and trans-polyacetylene. a) Stationary powder sample. The spectrum scans the entire range of the chemical
shift. b) With sample rotation (v,=4500 Hz) at the magic angle. A narrow
line is obtained at the isotropic value of the chemical shift, and spinning
sidebands (SSB) are spaced at intervals of v, (all spectra cross polarized with
3 ms, dipolar decoupled with v I H= S O kHz, each 20000 recordings at intervals of 2 s).
stationary, a powder spectrum is measured, which sweeps
over the entire range of the chemical shift. The principal
values a,,
of the chemical shift tensor can be obtained by
matching of the theoretical spectrum.[6a."a1 However, if the
sample were rotated at speeds of rotation which are large
in relation to the width of the powder spectrum (in Hz),
the result obtained would be only one sharp line at the isotropic value of the chemical shift, the position of which is
different for the cis and trans isomers of polyacetylene.
When the sample is rotated more slowly, spinning sidebands (SSB) occur, in addition to the isotropic signal, at a
spacing corresponding to the frequency of rotation (in Fig.
1,4500 Hz). The chemical shift tensor can be reconstructed
from the intensity of these
If the recording
is synchronized stepwise with the rotor position in a twodimensional measurement, the sideband intensity gives the
orientation distribution of anisotropic samples.f'0,'21In the
event o f interfering superpositions of the spinning sidebands with other signals, it is possible to suppress the sidebands by appropriate sequences of pulses.['31 When such
suppression is applied, the signal intensities are often distorted to a great extent. Except in the case of extremely
broad signals, for example when examining coal o r
charged, electrically conductive polymers, another measurement at a different speed of rotation frequently proves to
be more advantageous.
With regard to the use of MAS-l3C-NMR spectroscopy
as an analytical tool, especially for compounds which have
low solubility or are insoluble, two questions are of prime
importance: the question of the attainable resolution, i.e.
the line widths which can be obtained in comparison with
the range of the chemical shift ("C: approximately 200
Angew. Chem. Int. Ed. Engl. 27 (1988) 1468-1483
In solid-state N M R spectra the line widths are usually
10-100 times greater than those of solution N M R spectra.
The attainable line width of 10-100 Hz at 7T for rigid,
crystalline solid substances is determined, on the one
hand, by susceptibility variations and by dipolar interactions which are not entirely averaged out. On the other
hand, effects which are typical of solid substances frequently limit the resolution which can be
Errors in the setting of the magic angle play virtually no part,
provided that adjustment is carried out using KBI-"~]and
double-bearing rotors are used.
When several molecules are situated at non-equivalent
positions within a crystal, or if the symmetry of the crystal
is less than that of the molecule, molecular positions which
are equivalent to one another in solution can lead to several signals in the solid state. In the case of small differences in shift, a broadening of the signal may result"61 (Fig.
2). Similarly, the freezing-in of conformations in the solid
substance leads to the non-equivalence of signals. Such a
conformational effect on the solid-state spectrum has been
described both in the case of peptides["] and also in the
case of synthetic polymer^.['^.^^^
9
L
2
1
1&Y
G-
~
-- i o
100
3
'
4 5,67
6'0 - io . 5-
-6
Fig. 2. Effect of the crystal structure on the solid-state N M R spectrum (for
the example of a urea derivative). Two molecules in non-equivalent positions
lead to a doubling of the signals (each 1000 recordings at intervals of 6 s, CP
2 ms, v I H = v I C = 5 0kHz, v,=3650 Hz).
Conformation and, to a lesser extent, packing, are variables which contribute to the determination of the line position. This frequently results in complications when interpreting the solid-state N M R spectra of crystalline samples
with respect to chemical structures; on the other hand, the
influence of conformation and packing enables the spectroscopist to distinguish various modifications of chemi1469
dipolar coupling becomes inefficient; the natural line
width of I3C-NMR signals ( T Z )due to relaxation via the
C-H dipolar interaction then limits the resolution attainable.[241Figure 4 shows an appropriate example: in this
case, a spin-lock frequency in the median kHz range was
employed. The maximum line width is therefore to be expected at approximately 30 K above the glass-transition
temperature (in the case of poly(buty1 acrylate)
TG= - 46°C) (cf. Section 5.3). The agreement between the
decoupling frequency m I H and the jump frequency of the
motion leads to line broadening. There is also a broadening due to modulation of the resonance frequency on account of the anisotropy of the chemical shift, if the mobility is comparable with the rotor frequency (in most cases,
approximately 4000 H Z ) . [ ' ~ ~ ~
C)
1s o
100
7
.
50
.
150
100
150
100
I
.
.
.
,
50
261
-6
Fig 3. Effect of crystal modification and crystallinity on the solid-state N M R
spectrum of an amine hydrochloride. a) Crystallization from 2-propanol/diisopropyl ether. b) Recrystallized from ethyl acetate. Attention is drawn to the
non-equivalence within the positions a, b, g and h and the modification effect (compare (a) with (b)). c) Amorphous sample (quenched from the melt).
Conformational and packing disorder leads to broad signals ((a) and (b) each
2000 recordings plus 3000 recordings added with inverted 1st order SSB
[13a], in order to suppress the sidebands, vr=4SO0 Hz, C P 2ms.
V , , = V , ~ = S O kHz; c) 4000 recordings at intervals of 2 s, C P 2 ms, without
SSB suppression).
cally identical compounds (Fig. 3). Under conditions of increasing temperature, e.g. by interconversion of conformers["] o r tautomerism,[201there may be a dynamic exchange
between signals which are distinguishable in the solid-state
NMR spectrum. Whenever the atoms undergoing exchange with one another differ in chemical shift by a few
ppm at the most, slow dynamic processes with a lifetime in
the region of
s can be investigated by line shape analysis.
Below the glass-transition temperature the conformational and packing disorder leads to relatively broad lines
for amorphous samples. Line widths between 100 and 500
Hz (Fig. 3c) are typical at 75 MHz ("C). Accordingly, differing tacticities in polymers can only be measured in the
solid state if they result in differing conformations in the
or if the formation of hydrogen bonds is restricted to a specific tacticity.[ZZ1
In the case of elastomers having a glass-transition point
below room temperature, the modulation of the C-H dipole coupling due to motion plays a part in determining the line
Indeed, if the frequency of the
jump process and that of the decoupling field (usually
w ,H / 2n. 50 kHz) are comparable, the suppression of the
5
1470
-
200
150
100
50
0
- 6
Fig. 4. Effect of the molecular dynamics on the CP/MAS-NMR spectrum of
polybutyl acrylate. If the frequency of the molecular dynamics is comparable
with the strength of the decoupling field ( 0 , ~ / ( 2 ~ 1 ) = 5 0kHz), a line broadening results (T=261 K). In contrast to T=296 K, there are spinning sidebands at T=223 K. Attention is drawn to the absence of the signal from the
chain (arrow in the spectrum at T=223 K, 6x40) at T = 296 K.
When measurements are taken far above the glass transition, which is the case with rubbers even at room temperature, the dipolar interaction is to a large extent averaged
out by the high molecular mobility, so that very narrow
lines are measured even without dipolar decoupling.Wc.24.251 on account of the small line widths of rubbers, it i s possible to determine the microstructure of elastomers and elastomeric copolymers even in the solid state.
Even the scalar "C-'H coupling was resolved in a twodimensional experiment in the case of polyisoprene.'z61
2.2. Quantifiability
The I3C-NMR signal in solid-state NMR spectroscopy is
enhanced by cross polarization/' a transfer of magnetization from ' H to I3C in a double resonance experiment. In
order to achieve the transfer of magnetization, a Hartmann-Hahn contact is at present almost always used, i.e.,
the simultaneous application of two radio-frequency fields
Angew. Chem. Int. Ed. Engl. 27 (1988) 1468-1483
in resonance with I3C and ’H respectively with the same
frequency of the nuclei in the rotating coordinate system.
Other cross polarization technique^[^^.^'] are not considered in the text which follows. The reason for using cross
polarization for enhancing the I3C signal resides, on the
one hand, in the ratio of the gyromagnetic ratios yH/
yc = 4. Additionally the maximum permissible repetition
rate with cross polarization is determined by the very much
faster relaxation of the ‘H and not by the I3C; in favorable
circumstances it is possible to achieve an enhancement factor of u p to 20 in the signal-to-noise ratio. This is why
cross polarization is frequently a necessary prerequisite for
the practical feasibility of MAS-I3C-NMR spectroscopy.
Among polymers, the elastomers are an exception-measurements are made far above the glass-transition temperature and excitation takes place directly by a 90” pulse with
the resonant frequency of the I3C nuclei.
Signal enhancement by cross polarization implies that
the signal intensities are influenced by the kinetics of this
transfer of magnetization[6a.281
and are not a priori quantitative. Figure 5 shows the growth of the magnetization for
I 3l
1nA
the signal intensity M ( t ) passes through a maximum and
then falls off exponentially with increasing contact time t .
The time constant Tlp(’H)of the decay is common to all C
atoms [Eq. (a)-(d)].[6“,7h.281
M(t)=M,L-’f(t)
where
and
In order to treat the protons as a spin reservoir with a single spin temperature, the derivation of equation (a) is
based on a strong homonuclear H-H dipole interaction,
i.e. a short proton relaxation time T,(’H). Equation (a) describes the rise of the I3C magnetization with a time constant A - ’ TCH,during the contact of the I3C nuclei with the
’ H reservoir which in the ‘‘locked’’ condition is cold with
regard to the spin temperature. Superimposed on the rise
of the I3C magnetization, there is a contrary process: the
“heating” of the proton reservoir, i.e. reduction of the proton magnetization, due to the relaxation of the protons in
the rotating coordinate system with the time constant
TI,( I H). As a result of the reduction of the proton magnetization during the double resonance experiment, the maximum I3C magnetization attainable in the cross polarization
experiment also falls to Mo/l - exp( - t/T,,( ’H)). For large
radio-frequency field strengths, M o = y H / y c Me,, i.e. the
theoretically attainable I3C magnetization M o amounts to
approximately four times the equilibrium magnetization
Meq in the external field Bo. As already mentioned, an even
more significant gain in the signal-to-noise ratio is observed in cross polarization, because the repetition rate of
the experiment is not determined by the longitudinal relaxation of the I3C nuclei but by the very much more rapid
relaxation of the protons.
In the quantitative evaluation of the spectra,[7b1magnetizations of the various positions in the molecule which are
generated with a fixed contact time I are compared. The
scaling of Ma is different for the individual positions;
the relaxation time in the rotating coordinate system,
T,,(I3C), and the cross polarization constant TCHare both
individual quantities for the various positions in the molecule [see equations (a) and (c)]. Ttp(’H)is determined from
the gradient l/Tlp(’H) of the straight line at long contact
times in a logarithmic plot of the signal amplitudes, assuming Tl,(’H)>> TcH (Fig. 5). The initial part of the curve is
analyzed according to equation (d) in order to obtain
l/TcH l/TlP(l3C)(Table 1). TlP(l3C)can be measured in
a separate experiment (cf. Section 5.2). Tlp(’H), T,, and
TlP(l3C)are used to compute the scaling factors A [Eq. (c)]
and f(t) [Eq. (d)]; finally, A/f(t) gives the factor by which
the measured intensity according to equation (b) must be
multiplied in order to guarantee a “correct” count of the
spins (Table 2).
’
\:
0
1
2
3
tlrnsl
4
5
O\
6
7
Fig. 5. Rise olthe ‘ V N M K algndl durlng cross polarization, for the example
of an epoxy resin of bisphenol-A diglycidyl ether and diaminodiphenylmethane (0.172 epoxide units/bisphenol-A unit). The individual positions in the
molecule have an individual time o f build-up (cf. Table I). In the case of
long contact times, the signal amplitude declines with the common time constant T,,,(’H) (4000 recordings in each instance, at intervals of Zs, vr=3850
Hz, v,<.=v,,,=50 kHz). AQ=acquisition time of the FID, 0 Cl, 0 C2,
C3, 0 C4, x C5, V C6.
+
the various positions in a hardened epoxy resin. The mechanism which couples the I3C nuclei to the ‘H “reservoir”
is the dipole-dipole interaction. O n account of its l / r 6
dependence upon the spacing r between I3C and ’H,
the signal build-up for the various resonance lines takes
place at differing speeds: quaternary C nuclei without
‘H nuclei within bonding distance cross-polarize in the
Hartmann-Hahn contact[41more slowly than H-substituted
C nuclei. Where methyl groups are present, the dipolar interaction is reduced by the rotation about the ternary axis,
so that a CH3 group behaves in a similar way to a C H
group (cf. Fig. 5). In the case of single-phase systems with
adequate dipole-dipole interaction between the protons
Angew Chem. Inr Ed. Engl. 27 (I988)1468-1483
+
147 1
Table I . Dynamics of the cross polarization in the case of an epoxy resin [cf. Fig. 5 and Eq. (a)-(d)]. The time constant for the transfer of magnetization, T,,,, can
be determined from the rise of the signal amplitudes A at the commencement of the C-H contact in Figure 5 (time constant in column (2)). TtP(''C) must be
obtained from a separate experiment. T,,,('H) is obtained from the graph in Figure 5 for long contact times. The scaling I for the various signals is computed from
T, ,,, T,,,("C) and T18,('H)
in accordance with equation (c).
~
Cl
cz. c5
c3, c 4
C6
3.18
1.80
3.71
2.85
0. I9
0.10
0.35
0.28
2.99
1.70
3.36
2.57
~~~
0.064
0.059
0.100
0.1 10
0.111
0.196
0.099
0.130
0.953
0.863
1.001
0.980
[a] With v , c = v , ~ = 6 2 . 5kHz, v,=3850 Hz. [bl With vtC=3O kHz (separate measurement). [c] T,,('H)=3 ms.
Table 2. Falsification of the signal intensities by cross polarization. The magnetization Mocorresponding to a true "spin counting" is falsified by a factor of f(r)/A
on account of the cross polarization dynamics [Eq. (b)]. The smaller the difference in f ( ( l ) / I for the individual positions in the molecule, the better is the
reproduction of the quantitative conditions in the spectrum. f(l) was computed for two contact times f ( I = 1 ms and 1=2 ms respectively) using equation (d) and
the experimental times derived from Table 1. I was taken from Table 1. The values in column 7 and the renormalized values (column 8) show that quantitativity is
satisfied to an accuracy of ? 5% for I = 2 ms (0.95 5 f @ ) / Ai1.05).
~~
t = l ms
C nucleus
(Fig. 5)
c1
c2, c 5
c3, c 4
C6
1 = 2 ms
1 [a1
0.953
0.863
1.001
0.980
0.68
0.55
0.69
0.66
0.71
0.64
0.69
0.67
1.00
0.89
0.97
0.94
0.512
0.486
0.513
0.510
[a] From Table 1. [b] Calculated according to equation (d) with the values from Table 1 for
(Cl)
AS TCH/Tip(13C)10.1 and TCH%T,,,('H), the scaling
factor ;1 differs for the individual positions in the epoxy
resin only to a slight extent (cf. Table 1, 0 . 8 5 4 % 1.0). If
the value of TCHis short in comparison with Tio(l3C)and
TI,( 'H), at least approximately quantitative cross polarization spectra are recorded, even without measurement of
the cross polarization dynamics, as already described in
the literat~re."~]
However, it is necessary to ensure adequately long contact times, t > 3 T C H . The second term in
f(t) [Eq. (d)], which is dependent upon the C position, then
approximates to the value zero and thus f(t) becomes almost unitary (Table 2, t = 2 ms 0.485f(t)10.52). As the
value of T,,('H) is unitary, too, as a result of spin diffusion, the first term in equation (d) is a constant for all positions in the molecule.
The condition of a contact time t which is long in comparison with the cross polarization constant TCH,t2 3 TCH,
clearly means that, for the sake of quantifiability, a measurement on the rising branch of the I3C magnetization
should be avoided (cf. Fig. 5). The price which is paid is a
deterioration in the signal-to-noise ratio (cf. f(t)/;1 in Table
2 for t = 1 and 2 ms; the magnetization of CH, (CI), for
example falls from 0.71 M o to 0.54M0, i.e. to 2/3). On the
other hand, the distortion of the signal intensities f(t)/L as
compared with the true intensity ratios is very small at
f = 2 ms (Table 2, column 8, t 5%).
To be able to evaluate quantitatively, measurements
close to the T , , minima should be avoided. It is also true
that the line widths increase close to [T,,,(i3C)]m,,(cf. Sections 2.1 and 5.3) and that close to [T,t>(iH)],,,tnthe signal
amplitudes decline strongly [Eq. (a)].
1472
I=
0.54
0.56
0.5 I
0.52
1.00
I .05
0.95
0.97
I a n d 2 ms. [c] For easier comparison scaled to f(r)/L= 1 for CH,
To sum up, then, sufficiently long contact times ( 22 ms)
give spectra of homogeneous samples which are, as a rule,
at least approximately quantitative, even without laborious
normalization. It is not necessary to measure the cross polarization dynamics or to use calibration samples. An exception is systems which are strongly heterogeneous in
terms of dynamics, e.g. polymeric glasses modified for
high impact strength. In this case a quantitative comparison of the rubber and matrix signals is not permissible. The
problem of quantifiability of partially crystalline samples
has been investigated in detail with polyethylene as a
model compound.[*']
3. Reaction Plastics
The application of high-resolution solid-state I3C-NMR
spectroscopy to insoluble polymer network^"^' is, in particular, of industrial interest, since such cross-linked polymers participate in a multiplicity of processes which are important at the industrial level. Examples are, to mention
just a few, the baking of enamels, controlled thermal decomposition in the production of carbon fibers, adhesion
and coating processes, and especially the curing of resins.
Investigations, in addition to those on resins based on
furfuryl alcohol13o1and p~lydiacetylene'~']
as well as endcapped p ~ l y i r n i d e s , 'have
~ ~ ~ principally been carried out on
methacrylate-based resins,"31 urea/formaldehyde condens a t e ~ [as~ ~well
] as phenolic1351
and epoxy resin^.^'^.''.'^^
In the case of the dimethacrylates, the investigations
were aimed at the determination of the residual double
Angew. Chem. Inr. Ed. Engl. 27j1988) 1468-1483
bond^'"^.‘'^ and at the part played by the melaminelformaldehyde adduct, which is added in most cases. The
is essentially determined by the reaction of
the hydroxymethyl groups of the melamine/formaldehyde.
The extent to which it is possible to make statements concerning the number of hydroxymethyl radicals by means of
solid-state N M R spectroscopy is shown by Figure 6. A
tetrafunctional hardener. The network structure present in
the finished articles (and thus also the mechanical properties) is critically dependent upon the final conversion obtained during curing. In the MAS-I3C-NMR spectrum
(Fig. 7) measured for the cured resin, the individual signals
can be correlated with the different chemical structure^.^^^^'
The possibilities described in the literature with regard to
solid-state N M R spectroscopy in the investigation of
epoxy resins have recently been
L
-
100
50
0
100
50
0
h
n
A
I
.
50
60
. .
150
1
-6
2
a
3
b,c
.
5
5
d
I
Fig. 6. Heat treatment of a melamine/forrnaldehyde condensate. a) After 2 h
at 200°C. a large number of hydroxymethyl groups is still present (6=70). b)
After 24 h at 240°C, no CH,OH groups can be detected any longer, but there
1000 recordings in each inare signs of degradation (CH3 and >N-CO-);
stance, waiting time 2 s, v,= 3900 Hz).
pure melamine/formaldehyde condensate (a foam) was
used. At the end of the foaming process, there are initially
considerable proportions of N-CH,OH functions, which
can, however, be virtually entirely eliminated by subsequent tempering. A similar behavior is shown by urea/formaldehyde condensates.[341It is important that tempering
should not be excessive, since this would not only lead to
oxidative decomposition, but also to the formation of methyl groups because of undesired reactions (Fig. 6 ) . Both
effects have also been observed in the case of phenolic resins, in spite of the different starting materials
In
order to monitor the hardening of phenolic resins and the
formation of methylene bridges from the hydroxymethyl
radicals, the contact time for the transfer of magnetization
from ‘ H to I3C was varied. With the aid of model substances, suitable conditions of measurement for a quantitative determination of the ratio of methylene bridges to aromatic rings were
(cf. Section 2.2 with regard to
quantitative evaluation).
Even at a very early stage, epoxy resins were investigated in detail using MAS-NMR s p e c t r o ~ c o p y ; [ ~such
~”.~~~
resins have great technological importance as matrix resins
for composite materials consisting of polymer and carbon
fibers embedded therein for strengthening. The epoxy resins are made from two components, which may be, for example, a bifunctional resin (diepoxide) and a diamine as
Angew. Chem. In!.
Ed. Engl. 27(1988) 1468-1483
200
150
9
6 8
5’7’ 6‘ e
100
/
’
.
,
a
’
’
’
.
50
-6
Fig. 7. Solid-state ‘.’C-NMR spectrum of an epoxy resin based on the diglycidyl ether of bisphenol-A (DGEBA), cured with 4,4’-diaminodiphenylmethane (DDM). By evaluating the epoxide signal (arrow, signal 6,6= 53.7) the
epoxide content can be measured Lo an accuracy of approximately k0.04
epoxide units/bisphenol-A unit and the curing conversion can be determined. The figure shows the spectrum of the starting mixture containing 0.64
epoxide units/bisphenol-A unit and a partially cured sample containing 0.42
epoxide units/bisphenol-A unit. The signals I or 9 are suitable for internal
standardization. They are not affected by superposition of DDM (4000 recordings in each instance, 2 s, v, =4300 Hz, CP 2 rns).
In epoxy resins too, the line width of the signals is essentially based on an inhomogeneous broadening, which
arises as a result of a distribution of the values of the isotropic chemical shift on account of the differing conformat i o n ~ . ~ In
~ ”the
] case of the aromatic carbons ortho to the
ether bond in bisphenol-A, at a low temperature a signal
doubling was resolved. At room temperature and above,
these two signals coalesce as a result of “chemical” exchange of the two orrho carbons by 180” ring flips. The
correlation time of this movement was determined from investigations of the line shape and compared with mechanical relaxation data.l”]
Particular importance is ascribed to the question of the
quantifiability of the spectra when it is desired to measure
the curing conversion by means of N M R spectroscopy.
One possibility consists in investigating the dynamics of
the cross polarization[7b1and thus arriving at scaling factors for the areas of the individual signals (including sidebands) (cf. Section 2.2). Because of the necessity to evaluate areas, this process is, however, relatively inaccurate.
1473
A more reliable and simpler method is the evaluation of the
signal amplitudes of the solid-state NMR spectra of model
substances (e.g. oligomers of the diglycidyl ether of bisphenol-A). If the model substances are soluble, it is possible to
determine the epoxide content in solution and to plot calibration curves of the signal amplitude against the epoxide
content. For this purpose, the signal of the central carbon
(C6) in the glycidyl radical (6= 53.7, Fig. 7) is, for example,
normalized internally with reference to another signal of
bisphenol-A, e.g. to that of C9 (6= 35.2).
For a predetermined measurement condition, the residual epoxide content and thus an upper limit for the conversion can be determined by means of such a calibration line
from the signal amplitude of the epoxide. Residual epoxide contents of u p to approximately 1 epoxidel20 bisphenol-A are measurable and can be correlated with the curing conditions.
If cross-linkable thermoplastic materials are employed
as matrix materials, it is possible to utilize the advantage of
the thermoplastic processability and storage stability, without having to accept the sensitivity of non-crosslinking
thermoplastic materials, for example, to solvents. Figure 8
shows the spectrum of such a cross-linking thermoplastic
material based on aromatic thioether ether sulfones. After
tempering at 350°C, the homopolymer shows only a relatively small shift in peak intensity from 6> 130 to 6< 125,
which is in accord with a partial neutralization of the pure
para-substitution of the aromatic compounds by opening
and reforming the thioether linkages. On the other hand,
under the same temperature conditions a copolymer with
bisphenol-A units (Fig. 8b) shows a drastic change of the
spectrum (Fig. 8a). After 24 h at 350°C, the spectrum still
consists of the starting material spectrum only to the extent
of a proportion of 50% in terms of area. In 12 h, the signals
of the central propylidene unit in bisphenol-A (6= 29 and
41.6) fall to 60% of the initial intensity; after 24 h only 25%
is still present, which demonstrates the direct participation
of the comonomer in the cross-linking. In the case of a
thermal cross-linking under the very drastic conditions
chosen here, a multiplicity of in some cases very similar
structures are produced. Accordingly, the difference spectrum between product and starting material (Fig. 8c) also
shows only broad I3C-NMR signals. However, this is not a
weakness of MAS-NMR spectroscopy caused by the
method itself, but a consequence of the fact that an unspecific reaction cannot generate a specific spectrum.
4. Physical influences
A number of physical properties of the solid substance
either influence the line position and line shape or determine the relaxation behavior of the nuclei (Table 3).
Table 3. Physical effects in CP/MAS-NMR spectra and their possible applications for the characterization of solid substances.
Effect
Information
on
on
by
Line position
conformation,
packing
coordination
chemical exchange
Relaxation
spin dynamics,
spin diffusion
crystal structure, polymorphism, par
tial crystallinity
catalysis
molecular dynamics, tautomerism,
conformational transformations
amorphous components, high-impact
modifier, block copolymers and graft
content from relaxation time-selective
spectra
adjacency
polymer blends, plasticizers, antioxidants, heterogeneous systems
molecular dynamics block copolymers, additives, polymer
mixtures, glass transitions, mechanical
properties
It is precisely the physical influence on the spectra
which broadens the range of application of CP/MASN M R spectroscopy considerably beyond the purely analytical application: as a result of the conformational effect
on the line position, it is possible to obtain information
concerning the conformation in the solid substance, and
modifications and crystalline or amorphous components
also become indirectly distinguishable (sections 4.1 and
4.2).
150
100
50
-6
Fig. 8. Solid-state "C-NMR spectra of a cross-linking thermoplastic material
consisting of a copolymer of poly(thioether ether) and poly(thioether ether
sulfone). a) Spectrum after 24 h at 350°C. b) Educt spectrum. c) Difference
spectrum a-b. The spectrum a consists only to the extent of 50% of the educt
signals. The signals (8 and 9) of the propylidene unit in bisphenol-A are reduced t o 25%, which demonstrates participation of this unit in the crosslinking mechanism. The newly formed structures are non-unitary, as is evident from the difference spectrum showing poor resolution (at the bottom in
the figure). (The sidebands have been elimtnated mathematically.)
1474
The combination of high resolution with relaxation rime
measurements gives, on account of the resolution of signals
of individual positions of a molecule, very detailed information, a t the local, molecular level, concerning the dynamics (Section 5 ) and the molecular adjacency relationships which are present (Section 4.4). However, the relaxation behavior may also be utilized with advantage for analytical applications, in order to achieve selectivity of the
spectrum with respect to specified components in the case
of heterogenous samples. Depending upon the selection of
the experimental parameters, a determination is made in
Angew. Chem. Inr. Ed. Engl. 2711988) 1468-1483
terms of crystalline or amorphous material (Section 4.2), or
in terms of the elastomer component or matrix (Section
4.3).
4.1. Conformation, Modification, Amorphous Components
The conformation plays an important part with regard to
the line position in solid-state NMR spectra: two methylene signals are observed at a spacing of 8.7 ppm (6=39.6
and 48.3 respectively) as a result of conformational nonas a conseequivalence in syndiotactic
quence of a y-gauche effe~t.[~'.~'I
In one half of the methylene groups, two y-carbon atoms are present in the gauche
(9) conformation, while in the other half both y-carbons
occupy the trans conformation. The conformation of isotactic polypropylene is distinct, in turn, from that of the
syndiotactic polymer and gives a methylene signal at
6= 44.5.12'.391
This is, besides polyvinyl alcohol,[zzlone of
the few examples where differing tacticities can be discerned in the solid-state NMR spectrum.
The methylene groups of the a- and y-modifications of
polypivolactone should be mentioned as a further example
of a signal doubling as a result of conformational nonequi~aIence.'~~]
Very much more frequent than the creation of two different conformations of otherwise equivalent positions of a
molecule in the same crystal is the formation of differing
conformations in polymorphism.
The modifications can then be detected because of the
conformational effect on the chemical shift (see Fig. 3).
Examples from the field of polymers are polybutylene terephthalate with a g t g conformation in the a form and ttt
conformation in the 0 form,1411a- and y-polypivolactone
(2, helix) in comparison with P-polypivolactone (planar
zig-zag conformation)[401 and peptides (a-helix vs. y~ h e e t ) . [ ' ' ~ .In
~ ~ the
I
case of isotactic poly( I-butene), the
various modifications differ only in the dihedral angle of
the g t conformation, which is, however, sufficient for their
identification by means of N M R s p e c t r o ~ c o p y . ~ ~ ~ . ~ ~ ]
While in the above-cited cases the conformation and
thus an intramolecular effect is responsible for the ability
of NMR spectroscopy to distinguish the modifications, in
other cases the conformations are the same and the modifications vary only in the nature of the packing of the polymer chains in the crystal. However, the intermolecular effect of the packing leads only to small differences in the
signal position: A6=0.8 ppm between the a - and the Bmodification of isotactic polypropylene[391and 1.4 ppm between orthorhombic and monoclinic p~lyethylene.'~~"]
In a-olefins of greater chain length (e.g. poly[(S)-3,7-dimethyl-I-octene]), the side chain is in rapid movement
even in the solid state, and the carbon atoms of the side
chain from C5 onwards are conformationally disordered.[441Thus, the situation is very similar to that in the case
of partially crystalline polymers, where the conformation,
packing and mobility of the crystalline component differ
from those of the amorphous component, which has appropriate consequences for the NMR spectrum: in addition to a sharper signal from the crystalline regions, there
is a broader signal, frequently shifted in its position, from
the chains in the amorphous region. This was first obAngew. Chem. I n i . Ed. Engl. 27(1988) 1468-1483
served in polyethylene;L45'there have meanwhile been further examples, e.g. polyethylene t e r e ~ h t h a l a t e , ~trans~~]
p ~ l y b u t a d i e n e , ' ~polymethylene
~~
oxide[4x1and polyethylene o ~ i d e . " ~ " . ~ ~ '
In polyethylene (see Fig. 9a), the signal of the amorphous regions is shifted by 2.3 ppm
The
reason for this is the proportion of gauche conformations,
which can be estimated from this difference in shift at approximately 1/3. By using cycloalkanes as model subs t a n c e ~ , ~ ' ~an~ .attempt
~']
was made to obtain information
on the conformation of the folds contained in the amorphous covering layers of polyethylene single crystals.['8h1
In this case, the problem resides principally in the low proportion by mass of the folds in the entire sample and thus
the poor signal-to-noise ratio of the corresponding signals.
On the other hand, with regard to solution-crystallized
trans-polybutadiene it is known that, in spite of the restriction due to the adjacent re-entry of the chains, the folds of
the chains adopt the same conformation as in entirely
amorphous
4.2. Partial Crystallinity
In order to separate the signals from the crystalline and
amorphous component respectively, use is frequently
made of the very much shorter longitudinal relaxation time
T ,( I3C) of the amorphous regions.[521In the case of direct
I3C-NMR signal generation using 90" pulses, short waiting
times preferentially emphasize the signal originating from
the amorphous regions (cf. Figs. 9a and 9b). The physical
cause of the shorter T7"(I3C) is the higher spectral density
in the MHz range due to fluctuating fields on account of
the higher molecular mobility. Since the transport of mag-
c
a1
CI
b'
JA
35
25
30
-6
20
35
30
25
20
-6
Fig. 9. Effect of condition?, 01 nieaaurrmenl o n r h r zpecrrum o l polyethylene
[56].The conformational difference between the crystalline (all-trans)and the
amorphous modification (trans/gauche) leads to the separated signals A and
B ( A s = 2.3 ppm). a) 90" excitation, 800 s waiting time reproduces the correct
intensities. b) 90" excitation, 4 s waiting time gives preference to the amorphous signal. c) CP contact (1 ms) gives preference to the crystalline signal.
d) "Spin locking'' of the protons (20 ms) before cross-polarization gives only
the signal of the crystalline components. Scaling to the same amplitudes of B
(a-c) and A (a, d).
1475
netization through spin diffusion for I3C nuclei at natural
"low power decoupling" (LPD) similar to that employed in
abundance is very slow as a result of the high degree of
NMR of solutions, since the dipolar interaction is to a
isotopic dilution, the Tfm('3C)is determined by the highlarge extent averaged out just by the high molecular mobilfrequency components of the local molecular dynamics:
ity and the sample rotation at the magic angle.[501
along the side chain of an ethylene-1-octene copolymer,
In this way, investigations were carried out not only on
the value of TI( I3C) rises on account of the increasing mothe homo polymer^[^*' but also, for example, on a polyprob i l i t ~ . [Thus,
~ ~ ] in the stricter sense, T , ( I 3 C )is not a phase
pylene grafted with butadiene, in order to determine quanproperty. Nevertheless, in appropriate systems a selectivity
titatively the proportion of grafts.[591Even in the case of
of the spectrum with respect to mobile ( = amorphous) and
segmented copolymers consisting of poly(tetramethy1ene
rigid ( = crystalline) components can be achieved as a reterephthalate) hard segments and poly(tetramethy1eneoxyterephthalate) soft segments
sult of the very much more rapid relaxation of the I3C nuclei in the mobile component.
fOfCHr)4-O-CO-C,
H,COt;;f(-O+C
H&}FO-CO-C,
H&CO-fm
Whilst the mobile components are preferentially recorded in a rapidly repeated 90" pulse experiment, con(HytreP'j
versely in the case of polarization transfer from 'H to I3C
only the soft phase is recorded at 90" I3C excitation and
with short contact times the signal of the crystalline comlow
power decoupling.[601The same applies to poly(acryponent (cf. Figs. 9a and 9c) is maintained at an intensified
lonitrile-butadiene-styrene)
block copolymers (ABS), in
level. The increased molecular mobility in the amorphous
which
case
the
spectrum
is
selectively
measured for the busubstance partially averages out the dipole-dipole interactadiene
blocks,
for
which
the
glass
transition
temperature
tion necessary for the transfer of magnetization from 'H to
is
far
below
room
In
the
case
of
polybutaI3C, and leads to slower transfer rates l/TcH for the modiene, the resolution is so good that a sequence analysis of
bile (amorphous) components. Both in the case of selecthe microstructure is possible as far as triads. In the case of
tion via the more rapid longitudinal relaxation T;"('3C)
ABS, cross polarization and dipolar decoupling give, as a
of the mobile components and also in the case of selection
complement to the 90"-LPD measurement, the spectrum of
of the rigid components via a short cross polarization conthe styrene-acrylonitrile blocks. In analogous terms, Figure
tact (shorter TCHin the rigid component), use is made of
10 shows two spectra of a poly(styrene-butadiene-isothe molecular dynamics in order to make the distinction.
prene)
block copolymer.
The same applies to T,,(I3C), the relaxation time of the
I3C nuclei in the rotating coordinate system, which is
likewise a quantity that is local at the molecular level
and is determined by the dynamics in the median
kHz range (cf. Section 5.2). Nevertheless, in rigid systems,
Tl,('3C) i s shortened by spin-spin processes, so that
Ty:( "C)> c',y"'(13C)
is a p p l i ~ a b l e . ' ~ ~ " . ~ ~ ~
In cases where the association of rigid with crystalline
and mobile with amorphous does not apply, the relaxation
behavior of the protons must be used for selection. Spin
diffusion between the protons leads to a volume averaging
of the polarization state (cf. Section 4.4). Thus, for example, the question of the distribution of the ethyl branches
0
150
100
50
over the two phases of semi-crystalline polyethylene was
investigated by a TI,('H)-NMR e ~ p e r i m e n t [ ~ ~pre~.~~.~'~
ceding the cross-polarization as pre-contact delay[551(Fig.
9d). Since ethyl side chains within the crystalline regions
are certainly to be regarded as mobile, on account of their
defect property in comparison with the crystallized main
chain of the polyethylene, in this case a selection by means
of the local molecular mobility, i.e. the relaxation of the
150
100
50
0
13C nuclei, would lead to erroneous results.
-6
4.3. Elastomer Components
The very much more rapid longitudinal relaxation of
rubber components in a sample consisting of several components frequently permits the spectrum of the elastomer
components to be selectively recorded without special preparation of the samples, by means of 90 degree 13C
pulses.
I n order to achieve decoupling of the residual dipolar
interaction between 'H and I3C, it is sufficient to provide a
1476
Fig. 10. Solid-state I3C-NMR spectra of a poly(styrene-butadiene-isoprene)
rubber. a) CP 2 ms gives the polystyrene signal PS (17%) (ZOO00 recordings
FIDs broadened by 30 Hz with exponential multiplication). b) Polyisoprene
component, measured with 9O0-r-180"-r-AQ, r = 2.9 ms (6000 recordings,
waiting time 0.1 s, low power decoupling (LPD), inverse gating).
Figure 1 1 also follows the idea of the selective recording
of the soft phase by a 90"-LPD experiment, in this case with
the investigation of an unsaturated polyester resin (UP resin) which is toughened by rubber particles. The detection
of a rubber in such a matrix, which is also frequently filled
Angew. Chem. hi.Ed. Engl. 27 (1988) 1468-1483
with glass fibers, confronts other techniques with particular
problems: the matrix is insoluble and double bonds are invariably present in a UP resin on account of incomplete
conversion.
a1
200
150
-6
roo
molecule, in the Tipof the protons there is a volume property averaged over a distance of approximately 2 nm.Ih3]
The cause of this spatial averaging of TIP('H) is the small
mean distance between the ' H nuclei, since ' H is an abundant nucleus, in contrast to the "rare" nucleus I3C. As a
result of the spatial closeness of the protons to one another, a n exchange of spin energy of the individual protons
within the transverse relaxation time T,('H) takes place,
i.e. in rigid systems within less than 0.1 ms.
The residual 'H magnetization at the end of the spinlock time in the T,,('H)-NMR experiment is not measured
directly. Instead, a polarization transfer IH- I3C is undertaken and the ' H magnetization is detected indirectly via
the I3C-NMR signal (Fig. 12).
SO
Fig. I I. Detection of 5"/o rubber (poly(styrene-butadiene) with 20% styrene) in
a matrix consisting of unsaturated polyester (adipic acid/neopentyl glycol/
maleic acidlstyrene). a) Rubber signal (90000 recordings as in Fig. lob,
~ ~ 2ms,
. waiting
5
time 0.5 s). Attention is drawn to the high 1,2 linkage content (signals at 6 = 142.7 an 114.6). Only CHI groups of the matrix (&=21.8)
are included in the recording. b) Matrix signals (CP 1 ms, 4000 recordings,
dipolar (DD) decoupled with vIH= 50 kHz).
MAS-NMR not only detects the rubber (in the present
case 5% by weight), but also permits statements concerning
the chemical structure of the rubber employed (a polybutadiene with a considerable proportion of 1,2 linkages).
If the proportions of the soft phase amount to a few percent only, rapid 90" excitation and low power decoupling
are frequently not sufficient to suppress the matrix signal
entirely. Accordingly, in the examples given (Figs. 10 and
II), use was additionally made of the fact that under the
condition of weak decoupling the signals of the soft phase
are very much narrower (longer Tz)than those of the hard
phase: a 180" pulse of the 13Cnuclei within the longer free
induction decay of the rubber component refocuses only
the signals of the rubber, whereas the broad signals from
the matrix are already irreversibly relaxed. Thus, even in
the case of difficult samples, what is obtained is the rubber
signal without the broad matrix and blank signal background (Fig. 11).
Rubbers are usually measured by direct 90" excitation of
the I3C nuclei, since cross polarization is based on dipoledipole interaction and is therefore of low effectiveness, on
account of the high molecular mobility. Nevertheless,
cross polarization represents an alternative of interest even
for pure elastomers: by variation of the mismatch of the
Hartmann-Hahn condition (yC'cBjcf yHBIH)
o r alteration
of the speed of rotation of the rotor, statements can be
made on the degree of cross-linking of the rubber.[621
4.4. Molecular Neighborhood: Blends
While the I3C relaxation times TI(l3C),Tl,('3C) and TCH
are quantities which are individual to each position in the
Angew. Chem. In[.
Ed. Engf. 27(1988) 1468-1483
0
1
2
3
4
5
fimsl
6
7
8
__t
Fig. 12. Indirect measurement of /',,,('HI of the protons 01 aolid polybutyl
acrylate via the "C-NMR spectrum (298 K). The proton magnetization is
"locked" after 90" excitation for a variable time f in the radio-frequency
field, and the remaining magnetization is detected by cross polarization with
a fixed contact time. At room temperature, the T, of the protons is too long
to produce a common T,,('H) by spin diffusion; individual values of T,,('H)
[msl in the formula (10240 recordings in each instance, C P I ms, ordinates
arbitrarily shifted, vr=3000 Hz, v l H=SO kHz, D,(contact time, already given
as C P 1 ms) = 1 ms). n measured values for skeleton C atoms.
This process appears at first sight to be awkward and of
little advantage, since the experiment becomes more complicated on account of the indirect detection of the ' H
magnetization. Moreover, there is a loss by a factor exceeding 1000 in the signal-to-noise ratio. However, the
benefit of the indirect detection of the T I , of the protons
via I3C resides in that this process combines the high resolution of the MAS-I3C-NMR spectrum with a volume
property, namely TID('H).In view of the fact that the remaining ' H magnetization is detected at the end of a specified ' H spin-lock time via the signal amplitude of a I3CN M R signal, which is to be correlated with a quite specific
1477
chemical structure, the T,,('H) can be interrogated in the
environment of this
If, for example, a blend of
two polymers A and B is present, which differ in their
solid-state I3C-NMR spectra, the T,,('H) can be separately
determined in the environment of molecules of type A
and B respectively by evaluation of the signals from A and
B respectively. If there is molecular miscibility of A with B
in the sense of a spacing of less than 2 nm, what is expected is a common TtP('H), or otherwise, depending
upon the T,,('H) of the homopolymers, two separate
T?,('H) and T&('H) values. This process was employed in
the first instance with blends of polystyrene and polyphenylene oxide[651and later with polystyrene/polyvinyl methyl
ether[h61and polyvinylidene fluoride/polymethyl metha~ry1ate.l~''
Two different measurement processes are possible for
the measurement of Tl,('H) via 13C.[55.65b1
The more frequently employed pulse program varies the length of the
C P contact and determines the T,,('H) from the I3C-NMR
signal amplitudes for long contact times165b1(cf. Fig. 5).
The advantage is that it is possible to determine the magnetization transfer rate l/TcH with short contact times by
the same measurement program. However, the evaluation
presupposes TCH< TlP(l3C)<T,,('H) and moreover the
beginning of the T,,('H) curve is absent. During the spin
lock time the ' H and the I3C nuclei are irradiated simultaneously, so that the sample head has to cope with a relatively high power. As a n alternative to the T,,('H) measurement via variable length of the contact time, it is possible in the first instance to carry out the 'H spin-lock experiment, and only at the end of the ' H spin-lock time to
cross-polarize with a fixed contact time as usual.L551The
latter process (Fig. 12) was used, which avoids the disadvantages mentioned above and is more reliable in terms of
evaluation. Last but not least this procedure permits the
selective recording of a specified phase by means of a suitable contact time in the case of heterogeneous systems.
A system which is of interest for such investigations
with regard to molecular miscibility comprises dispersions,
made into films, obtained by emulsion copolymerization.
This may be, for example, a butyl acrylate/acrylic acid copolymer, which was produced by emulsion polymerization. Such films from emulsions are spatially heterogeneous with one phase predominantly containing the hydrophobic building block (in this case the butyl acrylate), and
other regions which principally consist of a polymer of the
hydrophilic monomer (in this case, acrylic acid). The question of how sharp the separation is, i.e. to what extent, for
example, the acrylic acid is also present in the phase rich
in butyl acrylate, can be investigated by such T,,('H)
measurements. What is important is to select the correct
temperature for the measurement: in the case of elastomers such as, for example, polybutyl acrylate (TG=
-46"C), at room temperature the proton line width under
MAS conditions amounts to only approximately 150 Hz.
The proton-proton dipolar interaction is averaged out by
the molecular mobility to such an extent that an individual
Tl,,('H) is measured for the protons at each position in the
butyl radical of the polybutyl acrylate homopolymer (see
values [ms] quoted in the structural formula in Fig. 12). As
is expected, the ' H line width increases rapidly with de1478
creasing measurement temperature. At - 25 "C it reaches
approximately 13 kHz, and thus leads to a common
T,,('H) (Fig. 13). When acrylic acid is incorporated as co-
I
InA
-CH,
-C_H2-CH,
-O-CH,-LH,-0-C_H,-CH,-
-0
-'-O-Bu
0
I
,
1
1
2
3
I
4
tlrnsl
I
5
6
7
Fig. 13. T,,('H) via "C in a diapeiaivn mdde iniv rl lilin (emulsion copolymer
consisting of 85% butyl acrylate (BA) and 15% acrylic acid (AA)). At 248 K,
spin diffusion leads to a unitary T,,('H) for all protons (768 recordings in
each instance, otherwise conditions as in Fig. 12). 0 measured values for
skeleton C atoms.
monomer into the phase rich in polybutyl acrylate, the
glass transition temperature should rise in comparison
with the homopolymer consisting of polybutyl acrylate.
This is why a temperature of - 25°C satisfies the requirements with regard to adequate spin diffusion for the copolymer as well. Therefore T,,('H) is a volume property.
The evaluation of the signal amplitudes as a function of
the proton spin lock time is, in fact, purely exponential
and gives a unitary T,,('H) for all signals. This T,,('H) is
between the values for the homopolymers polybutyl acrylate and polyacrylic acid respectively, and shows that
acrylic acid and butyl acrylate in the phase rich in butyl
acrylate are present alongside one another at a spacing of
less than 10 (Fig. 14).
On the other hand, a non-exponential decline of the signal amplitude in the T,,('H)-NMR experiment cannot be
taken as unambiguous evidence leading to the opposite
conclusion as to a multiphase system.[681
In addition to the T,,('H) measurement via I3C, there
are certain other CP/MAS NMR techniques for detecting
molecular adjacency. These include the intermolecular
magnetization transfer from a molecule containing ' H to a
deuterium-containing polymer. However, this presupposes
the availability of a corresponding deuterium-containing
polymer, and is accordingly a process which is not suitable
for the investigation of commercial products. But it can
certainly answer questions concerning the fundamental
understanding of, for example, the functioning of block
copolymers as compatibility promotors in blends. Thus,
the formation of a mixture of the styrene block from a
A
Angew. Chem. Inr. Ed. Engl. 27 (1988) 1468-1483
poly(styrene-butadiene) block copolymer with the surrounding styrene matrix was demonstrated,’6y1but also
block copolymers of polyether/polyester segments[701and
the miscibility of polystyrene with polymethyl methacrylate or polyvinyl methyl etherl7’1 were investigated. Furthermore, the compatibility of low-molecular-weight additives with the polymer matrix and selective interactions of
these additives with functional groups of the polymer are
detectable by intermolecular cross polarization, as has
been demonstrated for a photoconductive compound in a
polycarbonate matrix[721and a plasticizer in PMMA.[701
their intensity
Whilst here, as a result of the
deuteration, ’H becomes the “rare” nucleus, I3C can conversely also be abundant, as a result of ”C enrichment, so
that intermolecular spin diffusion from the I3C-labeled
molecule (e.g. a plasticizer) to the unlabeled neighbor (the
polymer) becomes possible.[751If both components are I3Clabeled, as in the case of a polyethylene terephthalatelpolycarbonate
it is finally possible for spin diffusion between the I3C nuclei to be used for the investigation
of blends in the same way as in the case of ‘H multipulse
MAS-NMR experiments.[771
5. Molecular Dynamics
Relaxation time measurements under CP/MAS conditions give individual 13C-relaxation times for the positions
in the molecule which are resolved in the spectrum.[641This
gives rise to the hope that it will be possible to investigate
the molecular dynamics of the pertinent group in this
way.
5.1. Longitudinal Relaxation Time T1(”C)
\
’\ Poly-BA
2.0
0
0
1
2
3
4
5
6
7
t Imsl
Fig. 14. TIc,(’H)via “ C OF polyacrylates for the detection of molecular adjacency. The emulsion copolymer, made into a film, consisting of BA/AA (85/
15% by weight) shows a T,,(’H) between the corresponding values for the
homopolymers (P-BA/AA: 3.2 ms, P-BA: 2.0 ms, P-AA: 5.8 ms). The phase
rich in butyl acrylate contains polymerized acrylic acid at the molecular level
(presumably as comonomer) (conditions of measurement as in Fig. 13).
x -COO-,
skeleton C atoms.
Instead of using the l/r3 dependence of the C-H dipolar
interaction upon the C-H spacing r for the detection of
adjacency by intermolecular cross polarization, it is possible to use the dipolar interaction with a nucleus, e.g. 19F,
which, in contrast to ’H, is not decoupled, in order to demonstrate adjacency: this is true of suitable systems, as has
been proved for a number of polyvinylidene(PVF,) blends.
In the event of miscibility, the adjacency of I9F nuclei
broadens the 13C-NMR signal of the admixed components
at 13C-’yFspacings of ca. 0.1 nm, in such a manner that
the I3C nuclei adjacent to the 19Fnuclei are no longer recorded in the MAS-NMR experiment and a reduction of
the signal intensities results, as has been shown for PVF2/
PMMA and other PVF2
Finally, in order to demonstrate adjacency, it is possible
to use the homonuclear dipole-dipole interaction and to
measure directly the residual protons in an (almost) deuterated matrix. If a ’H-containing polymer is admixed,
only isolated ’H nuclei are measured as sharp lines, and
Angew Chem. Inr Ed. Engl. 27 (1988) 1468-1483
measurement^"^] of the longitudinal relaxation time
T,(I3C) are used to detect motions with a spectral density
in the region of the Larmor frequency, i.e. in the MHz
range. In the case of polymers in the glassy state, by reason
of this high frequency, it is frequently a methyl group
which determines the relaxation behavior, in the event that
such a group is present: the dipole field which is generated
by the protons of the rapidly rotating methyl group contributes most strongly to the fluctuations in the range of
high frequencies at the location of the neighboring 13C nuclei, although the effectiveness of the dipole field for the
relaxation decreases strongly with the distance r (- l / r 6 ) .
In polymethyl methacrylate, the a-CH, group, for which
room temperature is above the TI minimum,[79a1
determines
the Ti behavior of the polymer chain as
The mobility of the polymer chain does not contribute to the spectral density in the MHz range, so that, even when the glass
point is exceeded, the effects on the Tl value of the carbon
atoms in the chain are small.[79c1
For the same reason, plasticizer effects in the case of PMMA[7ybJ
are not detectable
via TI(l3C), in contrast to polycarbonate.[801However, in
the case of PMMA,[8’a1just as in the case of polyethylene
oxide,”lbJ TI can be employed for the selective recording
of the amorphous component (see Section 4.2).
When elastomers are analyzed, individual TI(I3C) relaxation times can also be measured using a stationary sample. The position of the TI(l3C)minimum in the case of a
stationary measurement but with variable temperature is
dependent upon the content of 1,2 linkages, in the case of
polybutadiene. If it is borne in mind that the TI minimum
occurs at correlation times which correspond to the Larmor frequency, then good agreement is found between the
temperature of the Tl minimum and the glass point of the
samples.[821Stationary investigations were likewise carried
out above the glass point on polyterephthalic acid esters
and polysuccinic acid esters respectively.[831
1479
5.2. Relaxation in the Rotating Coordinate System,
T,p(’3c)
A matter of particular interest for the investigation of
polymers is the relaxation time in the rotating coordinate
system, T1,(13C)[641
(Fig. 15). The amplitude of the radio-
It is indeed possible for fluctuating dipole fields to be generated not only by molecular dynamics but also by a
change of the spin state of the proton adjacent to the I3C
nucleus. In this case, the dipole field at the location of the
13
C nucleus changes, without a change of the space coordinates of the molecule being associated with this. Thus, relaxation due to such spin-spin processes does not give any
information on the molecular dynamics. Before measured
T&(”C) values can be interpreted in terms of molecular
dynamics, it must be certain that the spin-spin relaxation
time TEH Tl,(’3C), i.e. the spin-lattice relaxation
T1,(’3C) determines the measured T&.[84”1TgH can be
measured ~ e p a r a t e l y . [ ’ *It~ ~becomes
~~~
evident that the
condition TzH Tlp(l3C) in the case of amorphous polymers is also satisfied in the glassy state, as is the case for
plastic crystals at room temperature. In amorphous polymers, the measured TTP(l3C)is predominantly (^I 80%) determined by spin-lattice processes, and may thus be interpreted in terms of molecular
Polystyrene,
polymethyl methacrylate and polyphenylene oxide, as well
as polyethylene terephthalate and polycarbonate, may
serve as examples.[z8~54b~64~84b1
However, when rigid polymers with strongly coupled protons are analyzed, and particularly when crystalline samples are used, spin-spin relaxation is predominant, as for example in the case of polyethylene[s4a1
and polyoxymethylene.[’sl Besides the homopolymers, investigations were also carried out on poly(etherester) block copolymers[8s1and anti-plasticizer effects on
polycarbonate.‘861 Measurements on polypropylene with
variable measurement temperature^^^'^ d o indeed indicate
spin-lattice processes determining the relaxation at a low
temperature, in which connection, however, the movement
of the CH, groups contributes to relaxation. For this reason, statements concerning the dynamics of the main chain
were not possible. Relaxation due to the rotation of the
CH3 groups is likewise found in PMMA.[881Furthermore, a
cured epoxy resin was investigated over a broad temperature range, and the relaxation was at least predominantly
ascribed to spin-lattice processes.[891 Apparently a very
broad correlation time distribution is present, since
TIC,(
I3C) hardly varied at all with the temperature.
Figure 16 shows an example with a marked T,D(13C)
minimum, as is expected for spin-lattice-determined
TlP(l3C).On account of the measurement frequency of 25
kHz, the minimum is situated approximately 30 K above
the glass point ( T, = - 46°C). In the case of copolymerization of the butyl acrylate with acrylic acid, the position of
the minimum is shifted by approximately 10 K (Fig. 17); a
value of 9 K was measured by differential scanning calorimetry.
With an adequate radio-frequency field strength a clear
dominance on the part of spin-lattice processes can be expected, particularly in the vicinity of the TTp minimum.
This dominance also extends into the temperature range
above the minimum provided that the rotational diffusion
is not excessively
This likewise applies below the minimum, at least for the quaternary carbons
which are less strongly coupled to ‘ H and for methyl
*
*
\.
P
P
-CH,-CH,-CH,
\
T,,(”C)Lmsl
0.65 2.0 4.1 7.6
0
0
1
2
3
4
f lmsl
5
6
-
7
0
8
Fig. 15. Tr,,(“C) of polybutyl acrylate at 2 x 3 K. The usual generation of the
signal via a CP contact is followed by a “spin locking’’ of the “C nuclei for a
variable time f. The plot of the signal amplitudes as a function o f f gives, as a
reciprocal of the gradients of the straight lines, individual relaxation times
T,0(’3C)for the various positions in the molecule (768 recordings in each
instance, CP 0.75 ms, v,=3000 Hz, vlr=25 kHz.
frequency field which is applied during the relaxation and
which is parallel to the magnetization determines the frequency vlc=wlc/(2n) of the spins within the rotating
coordinate system. Molecular motions having a spectral
component at this frequency are of crucial importance for
the relaxation. I n the case of the conventional radio-frequency field strengths, this is 25 kHz-75 kHz. Now, the
median kHz range is of particular interest in the case of
polymers, since the mechanical properties are also governed by the molecular dynamics in this frequency range.
The interpretation of TIP(I3C) data, however, is complicated by the fact that the relaxation in the rotating coordinate system is caused by the fluctuating dipole fields of the
protons.
This fluctuation can arise on account of changes of the
C-H bond vector r, e.g. a change of direction in the case of
the rotational diffusion of a methine group. This phenomenon (i.e. the change of r) is referred to as spin-lattice relaxation, and it is possible to draw conclusions as to the
molecular dynamics from T1,(13C). However, the measured TTP(l3C)[Eq. (e)] consists of two possible components, the spin-lattice TlP(l3C)just mentioned and a possible spin-spin contribution T&.114h.28.54.641
1480
Angew. Chem. I n l . Ed. Engl. 27 (1988) 1468-1483
groups, both of which show an appropriately marked
T$("C) minimum (Fig. 16). At the minimum, the correlation time z of the motion causing the minimum can be inferred from oIc7 = 1. The depth of the minimum is determined by the magnitude of the second moment, modulated
by the variation of the C-H dipolar i n t e r a c t i ~ n . " Ac~~.~~~~
cordingly, the expectation for all CH, groups is the same
[TTp('3C)],,n and that for quaternary C is a longer
[TTo(i3C)]m,n,as, in fact, is verified (Fig. 16). The large
1 '"1
'
0
\\\
\
\
:. ...._..?.
T,; lmsl
\
..
"=
!
\
5-
0
-80 -60 - 4 0
20-
-20
TI"C1
0
20
40
c
Fig. 17. T&("C) of the BA-carboxy group of the BA-homopolymer ( x ) and
of a copolymer (0)consisting of BA/AA (S5/15% by weight). In the copolymer, the position of the minimum is shifted by approximately 10 K, corresponding to the rise in the glass transition temperature. The greater
T7p('3C)at the minimum in the case of the copolymer is presumably due to a
higher degree of anisotropy of the carboxy group motion due to hydrogen
bonds.
15-
T,; l m s l
10-
5.3. Transverse Relaxation T2(13C)
5-
,
, ,
- 80 - 6 0
-LO
o ! ,
,
~_"_
, , ,
-20
0
T l"C1
,
I
,
20
,
,
-
40
Fig. 16. TT,,("C) of polybutyl acrylate as a function of temperature. At the
minimum, the correlation time T = 1/(2xv,,) of the motion ( ~ = 6 . 4ps) is obtained from the "spin lock" frequency vIc. As a result of the frequency of
v,,=25 kHz, the position of the minimum is shifted by approximately 30 K
above the glass-transition 'temperature. vr=3000 Hz.
long T7,('3C)of the CH3 groups at the minimum is due to
the rotation which is still present about the ternary axis,
and corresponds to the finding that the cross-polarization
behavior of CH3 groups is also rather similar to that of CH
groups (cf. Fig. 5 and the l/TcH value in Table 1).
In the copolymer of butyl acrylate with 10%by weight of
acrylic acid, all butyl positions have a minimum of the
same depth as in the homopolymer. O n the other hand, the
position of the minimum is shifted by 10 K towards a
higher temperature. An exception with regard to the magnitude of [TTp('3C)],,, is the C nucleus of the carboxy
group, whose [TTp(13C)]m,nrises from approximately 2 ms
(homopolymer) to 3.2 ms in the copolymer (Fig. 17). A
broader distribution of the correlation times can be ruled
out as a possible reason, since the shape of the TTP(l3C)/
temperature curve is, in the case of the copolymer, the
same as in the case of the homopolymer. There remains a
greater anisotropy of the motion as a possible cause, presumably on account of hydrogen bonds between the ester
and acid radicals.
Angew. Chem. l n t . Ed. Engi. 27 (1988) 1468-1483
The transverse relaxation rate [Tz('3C)]-1 is affected by
the spectral density at the frequency w I H of the protons in
the rotating coordinate system under the influence of the
decoupling field. If n o other effects determine the line
width Avc of the I3C-NMR signal, then Avc- l/TZ(l3C)is
applicable, and Avc passes through a maximum when the
stochastic dynamics possesses the same frequency as the
coherent decoupling. In the event of equality of frequency,
the latter becomes inefficient.[231In CP/MAS-NMR spectroscopy, the Hartmann-Hahn
w I C= w l H is
usually satisfied for the purpose of cross polarization, and
the angular frequency in the radio-frequency field is equal
for 'H and I3C (Section 2.2). In the Tl,('3C)-NMR experiment, the spectral density is detected at wIc (cf. Section
5.2). T2(I3C)(or the line width Avc of the I3C-NMR signal)
is determined by the spectral density at w I H . Thus, for
w I c = w I Hin CPIMAS-NMR experiments, both TI,(l3C)
and also T2(I3C) are affected by the same spectral density.r201Accordingly, the Iine width of the I3C-NMR signal
should pass through a maximum at the TiP(l3C)minimum,
if the molecular dynamics determine the two quantities.
This is in fact the case for polybutyl acrylate (compare Fig.
16 with Fig. 4).
As the measurement of relaxation times only permits
statements on the spectral density at the measurement frequency no information on the geometry of the movement
can be obtained for isotropic samples. If, for example, we
add a plasticizer to polystyrene,[901no distinction can be
made as to whether the reduction of the Tlp(l3C)is based
on a change in the frequency of the motion o r its amplitude. Only processes which utilize the anisotropy of a magnetic interaction permit statements on the geometry of the
motion. They include the two-dimensional methods[lo1of
the "dipolar rotational s p i n - e ~ h o " [ and
~ ' ~ o f the "chemical
exchange"[921 under MAS conditions. The "dipolar rotational spin-echo'' experiment is based on the dipolar inter-
I481
action of a C-H pair, in which connection the homonuclear interaction of the protons with one another must be
suppressed by multiple-pulse sequences. Investigations
were carried out principally on p o l y c a r b ~ n a t e [ ~and
'~]
polystyrene.'"'' The selectivity with respect to a specified
position in the molecule is achieved, in the case of the "dipolar rotational spin-echo", by MAS-NMR spectroscopy
via the chemical shift. In contrast to this, for the evaluation
of powder spectra of stationary samples"31 it is necessary
to have 13C-labeled substances, i.e. the selectivity is obtained by a preparative route, as in the case of deuteronNMR spectroscopy.
In two-dimensional exchange s p e c t r o s ~ o p y linforma~~~
tion on jump angles and correlation times is obtained from
the anisotropy of the chemical shift. The possibilities offered by this process regarding the investigation of slow
motiondrol with correlation times z, in the range
T, > rC> 1 ms have been demonstrated with reference to
the example of poIyoxymethylene.lg2"J
6. Summary and Outlook
The recording of solid-state "C-CP/MAS-NMR spectra
is now no longer a problem; however, when measurement
at variable temperature is concerned, this statement is only
partly true. If we have sufficient prior knowledge of the
sample, the chemical structures in the compounds which
exhibit low solubility or are insoluble are accessible by
means of this technique. When dealing with partially crystalline samples, familiarity with the particular solid-state
effects on the spectrum is required. Where soluble compounds are concerned and where chemical analysis only is
involved, recourse will also be had in future to NMR spectroscopy in solution, since the line widths achievable in solution are smaller by more than an order of magnitude.
Quite apart from the analytical application in the case of
insoluble samples, the strength of solid-state CP/MAS
NMR spectroscopy becomes clear in two areas: it becomes
evident in the techniques for correlating chemical shifts
(= chemical structure) with the relaxation behavior (= dynamics, adjacency). By this means, it is possible to make
very detailed statements on the dynamics of individual positions in the molecule. A second strength resides in the
possible selectivity which in the case of heterogeneous
samples can be achieved by utilizing the relaxation behavior. Simply by selecting the measurement conditions, i.e.
without sample preparation, it is possible to investigate selectively the chemical structure, molecular dynamics and
possibly the orientation of a specified component of the
sample. It is precisely this selectivity which is of great advantage in connection with the frequently very complex
composition of industrial products.
By way of analogy with N M R spectroscopy in solution,
future development gives rise to the expectation of a diversification from I3C nuclei to other nuclei; even at the present time, 29Siis playing an important part in certain areas
of research, and there is increasing interest in "N. In specific cases outside the field of pure analysis, solid-state 'HNMR spectra could also be used to an increasing extent.
Both the recording of "N-NMR spectra and also increas1482
ing utilization of variable temperature measurements, relaxation time measurements and a greater range of 2-D
N M R experiments in the solid state will result in an increasing measurement time requirement for solid-state
N M R spectroscopy.
Two-dimensional solid-state NMR spectroscopy1"] is
still at the start of its development. In this instance; additional information is to be expected in the future concerning molecular structure, e.g. bond lengths, concerning molecular dynamics, such as the nature and time scale of motions, and concerning molecular order.
Received: February 13, 1988 [A 695 IE]
German version: Angew. Chem. 100 (1988) 1525
[ I ] J. Schaefer, E. 0. Stejskal, R. Buchdahl, Macromolecules 8 (1975) 291.
[21 a) F. Bloch, Phys. Rev. 1 1 1 (1958) 841; b) L. R. Sarles, R. h4. Cotts, 2nd.
111 (1958) 853.
131 E. R. Andrew, Prog. Nucl. Mugn. Reson. Spectrosc. 8 (1971) I .
141 a) S. R. Hartmann, E. L. Hahn, Phys. Reu. 128 (1962) 2042; b) A. Pines,
M. G . Gibby, J . S. Waugh, J . Chem. Phys. 59 (1973) 569.
151 a) J. Schaefer, E. 0 . Stejskal, Top. Curbon-I3 N M R Spectrosc. 3 (1979)
283; b) B. C. Gerstein, rbid. 4 (1984) 123; c ) C. S . Yannoni, Acc. Chem.
Res. 15 (1982) 201; d) D J. ODonnell, ACS Symp. Ser. 247 (1984) 21.
161 a) M. Mehring: Principles of High Resolution N M R in Solids, Springer,
Berlin 1983; b) C. A. Fyfe: Solid Srute N M R for Chemists, CFC Press,
Guelph, Canada 1983; c) R. A. Komoroski (Ed): High Resolufion N M R
Spectroscopy of Synthetic Polymers in Bulk, VCH Publishers, Deerfield
Beach, FL, U S A 1986.
[?I a) J. R. Lyerla, Contemp Top. Polym. Sd. 3 (1979) 143; b) A. N. Garroway, W. B. Moniz, H. A. Resing, ACS Symp. Ser. 103 (1979) 67; c) L. W.
Jelinski, J. J. Dumais, F. C. Schilling, F. A. Bovey, Pure Appl Chem. 54
(1982) 345; d) E. 0. Stejskal, J. Schaefer, M. D. Sefcik, G . S. Jocob, R.
A. McKay, ibid. 54 (1982) 461; e) H. A. Resing, A. N. Garroway, D. C.
Weber, J. Ferraris, 0. Slotfeldt-Ellingsen, ibid. 54 (1982) 595; f ) D. E.
Axelson, K. E. Russel, Prog. Polym. Sci. 11 (1985) 221; g) J. Schaefer, E.
0. Stejskal, M. D. Sefcik, R. A. McKay, Phrlos. Trans. R . Sac. London
A 2 9 9 (1981) 593; h) D. R. Bduer, Prog. Org. Coaf. 14 (1986) 45.
[8] J. J. Lindberg, B. Hortling, Ado. Polym. Sci. 66 (1985) I .
[9] S. J. Opella, J. G. Hexem, M. H. Frey, T. A. Cross, Philos. Trans. R. Soc.
London A299 (1981) 665.
[I01 B. Bliimich, H. W. Spiess, Angew. Chem. 100 (1988) Nr. 12; Angew.
Chem. I n ( . Ed. Enyl. 27 (1988) Nr. 12.
[ I I] a ) M. Mehring, H. Weber, W. Muller, G. Wegner, Solid Stme Commun.
45(1983) 1079; b) M. M. Maricq, J . S. Waugh, J. Chem. Phys. 70 (1979)
330; c) J. Herzfeld, A. E. Berger, ibid. 73 (1980) 6021.
[I21 a) G. S. Harbison, H. W. Spiess, Chem. Phys. Lett. 124 (1986) 128; b) G.
S. Harbison, V.-D. Vogt, H. W. Spiess, J . Chem. Phys. 86 (1986) 1206.
[I31 a) W. T. Dixon, J. Chem. Phys. 77 (1982) 1800: b) M. A. Hemminga, P.
A. De Jager, K. P. Datema. J. Breg, J. Mugn. Reson. 50 (1982) 508: c) D.
P. Raleigh, E. T. Olejniczak, S. Vega, R. G. Griffin, ibid. 72 (1987) 238.
[I41 a) D. L. Van Der Hart, W. L. Earl, A. N. Garroway, J. Magn. Reson. 44
(1981) 361; b) A. N. Garroway, D. L. Van Der Hart, W. L. Earl, Phrlos.
Truns. R. Soc. London A 299 (198 I ) 609.
1151 J. S. Frye, G. E. Maciel, J. Mugn. Reson. 48 (1982) 125.
1161 G. E. Balimann, C. 1. Groombridge, R. K. Harris, K. J. Packer, B. J. Say,
S . F. Tanner, Philos. Truns. R . Soc. London A299 (1981) 643.
1171 a) H. R. Kricheldorf, D. Miiller, Colloid Polym. Sci. 262 (1984) 856; b) A.
Shoji, T. Ozdki, H. Saito, R. Tabeta, 1. Ando, Macromolecules 17 (1984)
1472.
[IS] a) H.-J. Cantow, D. Emeis, W. Gronski, A. Hasenhindl, D. Lausberg, M.
Moller, Y. Shahab, Makromol. Chem. Suppl. 7 (1984) 63; b) 1. Ando, T.
Sorita, T. Yamanobe, T. Komoto, H. Sato, K. Deguchi, M. Imanari, Polymer 26 (1985) 1864.
[I91 A. N. Garroway, W. M. Ritchey, W. B. Moniz, Macromolecules I5 (1982)
1051.
[20] F. LauprBtre, L. Monnerie, J. Virlet, Mucromolecule.7 17 (1984) 1397.
1211 A. B u m , M. E. A. Cudby, R. K. Harris, K. J. Packer, B. J. Say, J. Chem.
Sac. Chem. Commun. 1981, 15.
(221 T. Terao, S. Maeda, A. Sdika, Macromolecules I6 (1983) 1535.
[23] W. P. Rothwell, J. S. Waugh, 1.Chem. Phys. 74 (1981) 2721.
[24] R. A. Komoroski, J. Polym. Sci. Polym. Phys. Ed. 21 (1983) 2551.
I251 J. J . Dechter, R. A. Komoroski, D. E. Axelson, L. Mandelkern, J . Polym.
Sci. Polym. Phys. Ed. 19 (1981) 631.
I261 A. P. M. Kentgens, W. S. Veeman, J. van Bree, Macromolecules 20 (1987)
1234.
[27] a) A. Bax, N. M. Szeverenyi, G. E. Maciel, J. Magn. Reson. 50 (1982)
227: b) P. Caravatti, G. Bodenhausen, R. R. Ernst, Chem. Phys. Lett. 89
(1982) 363.
Angew. Chem.
Int.
Ed Engl. 27 (1988) 1468-1483
1281 E. 0. Stejskal. J. Schaefer, T. R. Steger, Faraday Symp. Chem. SOC.13
(1979) 56.
1291 a) B. Schroter, A. Posern, Macromol. Cltem. Rapid Commun. 3 (1982)
623; b) I). L. Van der Hart, F. Khoury, Polymer 25 (1984) 1589.
[30] 1. S. Chuang, G. E. Maciel, G. E. Myers, Macromolecules 17 (1984) 1087.
1311 J. R. Havens, M Thakur, J. B. Lando, J. L. Koenig, Macromolecules 17
(1984) 1071.
[32] a) M. D. Sefcik, E. 0. Stejskal, R. A. McKay, J. Schaefer, Macromolecules 12 (1979) 423; b) A. C. Wong, A. N. Garroway, W. M. Ritchey,
ibid. 14 (1981) 832.
[33] a) A. D. English, D. B. Chase, H. J. Spinelli, Macromolecules 16 (1983)
1422: b) I). R. Bauer, R. A. Dickie, J. L. Koenig, J . Polym. Sci. Polym.
Ph-ys. Ed. 22 (1984) 2009: c) P. E. Allen, G. P. Simon, D. R. G. Williams,
E. H. Williams, Eur. Polym. J . 22 (1986) 549; d) R. G . Earnshaw, C. A.
Price, J. H. O’Donnell, A. K. Wittaker, J. Appl. Pulym. Sci. 32 (1986)
5337.
1341 a) G. E. Maciel, N. M. Szeverenyi, T. A. Early, G. E . Myers, Macromolecules 16 (1983) 598; b) F. S. Chuang, B. L. Hawkins, G. E. Maciel, G.
E. Myers, rbid. 18 (1985) 1482.
1351 a) C. Fyfe, M. McKinnon, A. Rudin, W. Tchir, Macromolecules 16 (1983)
1216; b) S. So, A. Rudin, J . Polym. Scr. Polym. Lett. Ed. 23 (1985) 403: c)
G. E. Hatfield, G. E. Maciel, Macromolecules 20 (1987) 608.
[36] a) H. A. Resing, W. B. Moniz, Macromolecules 8 (1975) 560: b) M.-F.
Grenier-Loustalot, P. Grenier, Eur. Polym. J. 22 (1986) 457; c) E. Metzel,
J. L. Koenig, Adu. Polym. Sci. 75 (1986) 73.
[37] D. M. Grant, E. G . Paul, J. A m . Chem. Soc. 86 (1964) 2984.
1381 A. E. Tonelli, J. Am. Chem. Soc. 102 (1980) 7635.
[39] A. Bunn, M. E. A. Cudby, R. K. Harris, K. J. Packer, B. J. Say, Polymer
23 ( 1982) 694.
1401 R. P. Veregin, C. A. Fyfe, R. H. Maressault, Macromolecules 19 (1986)
2379.
1411 M.-F. Grenier-Loustalot, G. Bocelli, Eur. Polym. J. 20 (1984) 957.
[42] H. R. Kricheldorf, D. Miiller, Macromolecules 16 (1983) 615.
1431 L. A. Belfiore, F. C. Schilling, A. E. Tonelli, A. J. Lovinger, F. A. Bovey,
Macromolecules 1 7 (1984) 2561.
1441 M. C. Sacchi, 1. Tritto, P. Locatelli, L. Zetta, D. R. Ferro, H. Forster,
Macromolecules 19 (1986) 1634.
[45] W. L. Earl, D. L. Van Der Hart, Macromolecules 12 (1979) 762.
[46] M. D. Sefcik, J . Schaefer, E. 0. Stejskal, R. A. McKay, Macromolecules
13 (1980) 1132.
1471 F. C. Schilling, F. A. Bovey, A. E. Tonelli, S. Tseng, A. E. Woodward,
Macromolecules I 7 (1984) 728.
1481 A. L. C‘holli, W. M. Ritchey, J. L. Koenig, Spectrosc. Letr. 16 (1983) 21.
[49] a) W. W. Fleming, C. A. Fyfe, R. D. Kendrick, J. R. Lyerla, H. Vanni, C.
S. Yannoni, A C S Symp. Ser. 142 (1980) 193; b) J. J. Dechter, J. Polym.
Sci. Polvm. Lett. Ed. 23 (1985) 261.
[SO] R. Kitamaru, F. Horii, K. Marayama, Polym. BUN. 7 (1982) 583.
1511 M. Moller, W. Gronski, H.-J. Cantow, H. Hocker, J. A m . Chem. Sac. 106
(1984) 5093.
1521 a) W. S. Veeman, E. M. Menger, H. H. C. de Moor, Proc. I.U.P.A.C.
Macromol. Symp. 28th. 1982, 2: b ) F. C. Schilling, F. A. Bovery, A. E.
Tonelli, Pulyni. Muter. Sci. Eng. 50 (1984) 256.
1531 D. C. McFaddin, K. E. Russell, E. C. Kelusky, Polym. Commun. 27
(1986) 204.
1541 a) D. L. Van Der Hart, A. N. Garroway, J. Chem. Phys. 71 (1979) 2773:
b ) J. Schaefer, E. 0. Stejskal, T. R. Steger, M. D. Sefcik, R. A. McKay,
Macromolecules 13 (1980) 1121.
[55] R. S. Aujla, R. K. Harris, K. J. Packer, M. Parameswaran, B. J. Say, A.
Bunn. M. E. A. Cudby, Polym. Bull. 8 (1982) 253.
I561 a) D. L. Van Der Hart, E. Perez, Macromolecules 19 (1986) 1902; b) E.
Perez, D. L. Van Der Hart, B. Crist, P. R. Howard, ibid. 20 (1987) 78.
[57] F. Laupretre, L. Monnerie, L. Barthelemy, J. P. Vairon, A. Sauzeau, D.
Roussel, Polym. Bull. I5 (1986) 159.
1581 a) R. A. Komoroski, Rubber Chem. Techno/. 56 (1983) 959: b) D. J Patterson, J. L. Koenig, J. R. Shelton, ibid. 56 (1983) 971.
(591 P. F. Barron, F. Horii, K . Murayama, Polym. 24 (1983) 1252.
[60] a) L. W. Jelinski, F. C. Schilling, F. A. Bovey, Macromolecules 14 (1981)
58 I : h ) F. A. Bovey, Pure Appl. Chem. 54 (1982) 559.
Angew. Chem. In!. Ed. Engl. 27 11988) 1468-1483
1611 L. W. Jelinski, J. J. Dumais, P. T. Watnick, S. V. Bass, L. Shephard, J.
Polym. Sci. Polym. Chem. Ed. 20 (1982) 3285.
1621 S. A. Curran. A. R. Padwa, Macromolecules 20 (1987) 625.
V. J. McBrieAy, D. C. Douglas, Macromol. Rev. 16 (1981) 295.
J. Schaefer, E. 0. Stejskal, R. Buchdahl, Macromolecules I0 (1977) 384.
a) J. Schaefer, M. D. Sefcik, E. 0. Stejskal, R. A. McKay, Pobm. Prepr.
Am. Chem. SOC.Diu. Polym. Chem. 20 (1979) 247; b) E. 0. Stejskal, J.
Schaefer, M. D. Sefcik, R. A. McKay, Macromolecules 14 (1981) 275.
S. Kaplan, Polym. Prepr. A m . Chem. Soc. Din Polym. Chem. 25 (1984)
356.
a) T. S. Lin, T. C. Ward, Polym. Prepr. A m . Chem. Sac. Div. Polym.
Chem. 24 (1983) 136; b) P. Tekely, F. Laupritre, L. Monnerie, Pulymer
26 (1985) 1081.
a) J. I. Kaplan, A. N. Garroway, J . Magn. Reson. 49 (1982) 464: b) N.
Zumbulyadis, ibid. 53 (1983) 486.
J. Schaefer, M. D. Sefcik, E. 0. Stejskal, R. A. McKay, Macromolecules
14 (1981) 188.
a) L. A. Belfiore, Polym. Mater. Eng. 51 (1984) 223; b) Polymer 27 (1986)
80
G . C. Gobbi, R. Silvestri, T. P. Russell, J. R. Lyerla, W. W. Fleming, T.
Nishi, J. Polym. Sci. Polym. Letr. Ed. C25 (1987) 61.
J. M. Hewitt, P. M. Henrichs, M. Scozzafava, R. Scaringe, M. Linder, L.
J. Sorriero, Macromolecules I7 (1984) 2566.
T. C. Ward, T. S. Lin, Adu. Chem. Ser. 206 (1984) 59.
D. L. Van Der Hart, W. F. Manders, R. S. Stein, W. Herman, Macromolecules 20 (1987) 1724.
A. K. Roy, P. T. Inglefield, J. H. Shibata, A. A. Jones, Macromolecules 20
(1987) 1434.
M. Linder, P. M. Henrichs, J. M. Hewitt, D. J. Massa, J . Chem. Phys. 82
(1985) 1585.
a) P. Caravatti, P. Neuenschwander, R. R. Ernst, Macromolecules 18
(1985) 119; b) rbid. 19 (1986) 1889.
D. A. Torchia, J. M a p . Reson. 30 (1978) 613.
a) B. Gabrys, F. Horii, R. Kitamaru, Macromolecules 20 (1987) 175; b)
H. T. Edzes, W. S. Veeman, Polym. Bull. 5 (1981) 255; c) H. T. Edzes, A.
H. Montree, W. S. Veeman, E. de Boer, Proc. I.U.P.A.C. Macromul.
Symp. 28th 1982. 13.
M. K. Gupta, J. A. Ripmeester, D. J. Carlsson, D. M. Wiles, J . Polym.
Sci. Polym. Lett. Ed. 21 (1983) 211.
a) W. S. Veernan, E. M. Menger, Bull. Magn. Reson. 2 (1981) 77; b) W. S.
Veeman, E. M. Menger, W. Ritchey, E. de Boer, Macromolecules 12
(1979) 924.
S. Ni, L. Shen, F. Yu, B. Quian, Polym. Commun. 27 (1986) 318.
F. Horii, A. Hirai, K. Murayama, R. Kitamaru, T. Suzuki, Macromolecules 16 (1983) 273.
a) J. Schaefer, M. D. Sefcik, E. 0. Stejskal, R. A. McKay, Macrumolecules 17(1984) 1118; b)ibid. 14(1981)280.
L. W. Jelinski, J. J. Dumais, P. I. Watnick, A. K. Engel, M. D. Sefcik,
Macromolecules 16 (1983) 409.
T. R. Steger, J. Schaefer, E. 0. Stejskal, R. A. McKay, Macromolecules 13
(1980) 1127.
a) J. R. Lyerla, C. S. Yannoni, C. A. Fyfe, Acc. Chem. Res. 15 (1982) 208;
b) W. W. Fleming, J. R. Lyerla, C. S. Yannoni, ACS Symp. Ser. 247
(1984) 83.
M. Wobst, Acta Polym. 36 (1985) 492.
A. N. Garroway, W. B. Moniz, H. A. Resing, Faraday Symp. Chem. Soc.
13 (1979) 63.
L. A. Belfiore, S. L. Cooper, Polym. 25 (1984) 645.
a) J. Schaefer, R. A. McKay, E. 0. Stejskal, W. T. Dixon, J. Magn. Reson. 52 (1983) 123; b) J. Schaefer, E. 0 . Stejskal, R. A. McKay, W. T.
Dixon, Macromolecules 1 7 (1984) 1479; c) J. Schaefer, M. D. Sefcik, E.
0. Stejskal, R. A. McKay, W. T. Dixon, R. E. Cais, ibid. 17 (1984)
1107.
a) A. F. De Jong, A. P. M. Kentgens, W. S. Veeman, Chem. Phvs. Lett.
I09 (1984) 337; b) A. P. M. Kentgens, A. F. De Jong, E. de Boer, W. S.
Veeman, Macromolecules 18 (1985) 1045.
A. K. Roy, A. A. Jones, P. T. Inglefield, Macromolecules 19 (1986)
1356.
1483
Документ
Категория
Без категории
Просмотров
1
Размер файла
1 676 Кб
Теги
analytical, polymer, spectroscopy, resolution, solis, nmr, 13c, high, method, state, new
1/--страниц
Пожаловаться на содержимое документа