3 October 1997
a',
CHEMICAL
PHYSICS
LEI"rER$
,!v:;m
ELSEVIER
Chemical PhysicsLetters277 (1997) 79-83
High-resolution oxygen-17 NMR spectroscopy of solids by
multiple-quantummagic-angle-spinning
Gang Wu, David Rovnyak, Philip C. Huang, Robert G. Griffin *
Francis Bitter Magnet Laboratory and Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
Received 12 May 1997; in finalform21 July 1997
Abstract
High-resolution solid state nTo NMR spectra of nTo-enficbed compounds were observed with multiple-quantum
magic-angle-spinning (MQMAS) experiments. The resolution of the 170 MQMAS spectra is approximately 30- to 150-fold
higher than that found in conventional 170 MAS spectra, making it possible to detect crystallographically distinct oxygen
sites. It is shown that into MQMAS studies are feasible for systems with 170 quadrupole coupling constants up to 7 MHz,
provided that sufficiently high radio-frequency field strengths (> 120 kHz) and spinning speeds (~ 20 kHz) can be
achieved. © 1997 Published by Elsevier Science B.V.
1. Introduction
The recent development of multiple-quantum
magic-angle-spinning (MQMAS) spectroscopy by
Frydman and co-workers [1,2] has made it possible
to obtain high-resolution solid state NMR spectra of
half-integer, quadrupolar nuclei ( S > 3 / 2 ) under
standard MAS conditions. Among quadrupolar nuclei, 170 (S = 5 / 2 ) has special importance because
oxygen is the key constituent of many chemically
and biologically important functional groups. For
example, oxygen is often directly involved in hydrogen bonding, which is fundamental to physical and
biological processes. In addition, 170 NMR parameters are sensitive to molecular structure and chemical
environment, suggesting that direct observation of
~70 spectra has the potential to yield valuable and
previously inaccessible information [3]. Oxygen-17
* Correspondingauthor.
chemical shifts span a range of ~ 1500 ppm and,
along with the quadrupole coupling parameters, provide detailed information on hydrogen bonding effects, crystallographic symmetry, and molecular
structure. However, among the rapidly growing number of MQMAS studies, there is a paucity of reports
concerning ~70 nuclei. In the two cases where 170
MQMAS spectra are reported [4,5], only compounds
with moderate 170 quadrupole coupling constants,
< 5 MHz, were investigated. This fact reflects one
of the major problems in 170 MQMAS studies,
namely, the low sensitivity of the technique in systems with large quadrupole coupling constants. While
27A1 (S = 5 / 2 ) sites with quadrupole couplings of
8-9 MHz have been observed in MQMAS spectra
[6,7], the large gyromagnetic ratio of 27A1 (e.g.,
7(27A1)/7(170)----1.92) provides a factor of 5 in
signal-to-noise improvement in the directly detected
dimension of 27A1 MQMAS as compared to 170. In
this Letter, we demonstrate that the problems en-
0009-2614/97/$17.00 © 1997 Published by ElsevierScience B.V. All rightsreserved.
PI1 S 0 0 0 9 - 2 6 1 4 ( 9 7 ) 0 0 8 7 3 - 7
G. Wu et al. / Chemical PhysicsLetters 277 (1997) 79-83
80
countered in 170 MQMAS experiments can be circumvented for an important class of 170 sites, i.e.,
H2170 and -17OH groups, with a combination of
high-speed spinning ( ~ 20 kI-Iz) and large B I fields
( > 120 kHz).
The issue of sensitivity in MQMAS experiments
has been addressed in several recent studies.
Amoureux et al. [8] performed a theoretical analysis
of the sensitivity of MQMAS experiments using
nutation pulses. In agreement with early studies [911], it was shown that strong B 1 field strengths are
necessary for studying systems with large quadrupole
coupling constants. Wu et al. [12] demonstrated the
utility of a technique termed as rotation-induced
adiabatic coherence transfer (RIACT) in S = 3 / 2
systems. It was found that RIACT leads to quantitative MQMAS spectra and also enhances the overall
sensitivity of the MQMAS experiment, particularly
in systems with large quadrupole coupling constants
(e.g., e 2 q Q / h > 4 MHz for S = 3/2). However,
both the nutation and RIACT methods require strong
B 1 fields to obtain reasonable multiple quantum coherence excitation and transfer efficiencies. Finally,
the well-known technique of rotor-synchronized acquisition [13] can also be used to improve the sensitivity of MQMAS eXl~lfiments [7,14]. Here we combine these features in O MQMAS experiments and
report results for PO 4, H 2 0 and - O H containing
compounds.
2. E x p e r i m e n t a l
All solid state 170 NMR spectra were obtained on
a custom-designed NMR spectrometer opelfating at
53.93 MHz for 170 nuclei (9.4 T). Static O NMR
spectra were obtained using the echo sequence described by Kunwar et al. [15]. The custom designed
MAS probe was equipped with a 3.2 mm spinning
assembly (Chemagnetics, Inc., Fort Collins, Colorado) which allowed sample spinning at 20-25
kHz. The B l field strength at the 170 frequency was
100-135 kHz. All 170 triple-quantum (3Q) M_AS
experiments were performed with the two-pulse sequence with a 24-step phase cycling scheme [1,2].
Typical pulse widths for 3Q excitation and 3Q-to-lQ
conversion were 5.5 and 1.8 Is,s, respectively. For I H
containing compounds, proton decoupling was employed during both evolution and acquisition periods.
All NMR interaction parameters presented in Table 1
were obtained from comparison of MAS lineshapes
and MQMAS line positions [4,16]. All 170 chemical
shifts were referenced to H2170 with an external
sample. The samples of 170-labeled phosphates and
Table 1
Oxygen-17 NMR parametersobtained from solid state '70 NMR spectra
compound
peak
8i~o a(ppm)
e2qQ/h b(MHz)
7/c
into line width (kHz)
static
MAS
MQMAS
Cas(plTo4)3(OH)
1
2
108
115
4.0
4.1
0.00
0.I0
17.3
5.2
0.20
CaHpI704 • 2H20
1
2
98
4.2
0.00
21.0
5.8
0.15
96
4.3
0.00
KH2PI704
92
5.2
0.55
25.0
7.5
0.19
NH4H2pI704
93
5.1
0.55
23.5
7.2
0.21
Ba(CIO3) 2 • H2170
22
6.8
1.00
51.0
15.0
0.10
CaPTOH)2
62
6.5
0.00
26.0
10.5
0.10
Chemical shifts are referenced to external H2170 liquid.
a Errors in the isotmpic chemical shifts are estimated to he + 2 ppm. Note that reported isotropicchemical shifts were derived from line
~sitions in MQMAS spectra, which are weighted sums of isotropic chemical shift and quadrupole coupling paran~ters [4,16].
Errors in the quadrupole coupling constants are estimated to he +0.2 MHz.
c Errors in the asymmetryparametersare estimated to he :t:0.10.
G. Wu et al. / Chemical Physics Letters 277 (1997) 79-83
Ca(17OH)2 were prepared by the literature methods
[17-19] with 170-enriched H20 (containing 34%
170). B a ( C l O 3 ) 2 • H 2 1 7 0 w a s prepared by recrystalizing the compound from 170-euricbed water (containing 50% 170). Enriched water was subsequently
recovered on a vacuum line. Oxygen-17 enriched
water was obtained from ISOTEC Inc. (Miamisburg,
Ohio).
81
(b)
3. Results and discussion
Oxygen-17 MAS NMR spectra of four 170-enricbed phosphate samples are shown in Fig. 1. Unique
NMR line shapes arising from the second-order
quadrupolar interaction are observed, with line widths
ranging from approximately 5 to 7 kHz. Although it
is possible to estimate 170 quadrupole parameters
from the MAS spectra, no information concerning
the crystallographic equivalence of the oxygen atoms
can be obtained. In contrast, as shown in Fig. 2, the
170 MQMAS spectra of the samples consist of
isotropic peaks significantly narrower than the sec-
(a)
(b)
(c)
kHz
Fig. 1. Oxygen-17 MAS spectra of (a)
Cas(PITO4)3(OH), (b)
CaHpIToa'2H20, (c) KH2PI704, and (d) NI-I4H2PI704 . The
sample spinning frequency was 19.8 kHz.
(d)
kHz
Fig. 2. Oxygen-17 3QMAS spectraof (a) Cas(PITO4)3(OH), (b)
CaHpITo4.2H20, (C) KH2PI704, and (d) NH4H2PITO4. The
sample spinning frequency was 19.8 kHz in all the experiments.
The t 1 increment was synchronized with the sample spinning
period. 50.5 p.s. For each t I increment, 552 transients were
accumulated. A total of 40 tl increments were collected. Each of
the MQMAS experiments took ~ 6 h.
ond-order quadrupolar line shapes observed in the
M A S spectra. For Cas(pl704)3(OH), the 170 M Q M A S spectrum exhibits two isotropic peaks. This
doublet structure was not observed in our previous
study [4] due to constrained resolution in the t1
dimension. The significant improvement of resolution and sensitivity in this work results from the
combination of fast sample spinning (~ 20 kHz)
with rotor-synchronized tI acquisition [7,14]. The
advantage of the rotor-synchronized t~ acquisition is
that signal intensitiesin the sidebands are folded into
the center band. The total spectral width in the
isotropic ~VO M Q M A S
spectra is determined by
uR/(l + k), where u R is the spinning frequency and
k equals 19/12 for S -- 5/2 nuclei. Therefore, rapid
sample spinning (~ 20 kHz) is required to obtain a
sufficient spectral window in the isotropic dimension. The crystal structure of Cas(PO4)3(OH) indicates that the PO4 tetrabedron consists of three crystallographically distinct oxygen atoms, 01, 02, and
O 3 with a population ratio of 1:1:2 in the unit cell
[20]. Examination of the P - O bond distances reveals
82
G. Wu et a L / Chemical Physics Letters 277 (1997) 79-83
that 01 and 0 2 are similar and distinctively different
from 03 (P-O~ = 1.533 A, P - O 2 --1.544 A, and
P - O 3 = 1.514 A). The quadrupole coupling parameters for the three environments are nearly identical
and so the relative intensities in the MQMAS spectrum can be taken to be quantitative. Thus our
observation of a 1:1 doublet in the 170 MQMAS
spectrum is consistent with the crystal structure.
The crystal structure of CaHPO 4 • 2H20 belongs
to the space group, Ia, which yields four non-equivalent oxygen atoms in the PO 4 tetrabedron with the
P - O distances of 1.69, 1.58, 1.69, and 1.34 A [21].
In the 170 MQMAS spectrum o f C a H p I 7 0 4 • 2H20,
two isotropic peaks are observed with an approximate 1:3 intensity ratio. Our observation supports the
conclusion of the non-centrosymmetric space group
Ia. Both KH2PO 4 and N H j H 2 P O 4 crystallize in the
tetragonal space group I42d at room temperature
[22]. In each of the two compounds, all four oxygen
atoms of the PO 4 tetrahedron are crystallographically
equivalent. Indeed, single resonances are observed in
the 170 MQMAS spectra of these two compounds.
Since a potentially important application of highresolution solid state ~70 NMR is the study of hydration of biological macromolecules, we also investigated the efficiency of MQMAS in obtaining spectra
of crystalline hydrates and hydroxides, where 170
quadmpole couplings are larger ( ~ 7 MHz) than in
the phosphates. Fig. 3 shows 170 NMR spectra of
Ba(CIO3) 2 • H2170 obtained under static, MAS and
MQMAS conditions. As expected, the static and
MAS 170 NMR spectra of Ba(CIO3) 2 • H2170 arc
significantly broader than those of the phosphates. It
is seen in Fig. 3b that rapid sample spinning at 20
kHz significantly reduces the rotational sideband intensities, thus enhancing the sensitivity. Analysis of
the static and MAS 170 NMR spectra of Ba(C103) 2 •
H2170 yields e2qQ/h= 6.8 MHz and ~ = 1.0. This
value of e2qQ/h is somewhat smaller than the 7.61
MHz determined at 77 K by NQR [23]. As seen in
Figl.7 3a, the 170 MQMAS spectrum of Ba(C103) 2 •
H 2 0 exhibits an isotropic peak indicating only one
water molecule in the asymmetric unit cell, consistent with the neutron diffraction structure [24]. The
100 Hz line width of the 170 MQMAS spectrum is
about 150-fold narrower than that of the 170 MAS
spectrum. Similarly, the l~O MQMAS spectrum of
Ca(17OH)2 (not shown) exhibits an isotropic line
o
o
(a) , ~ ,
4
2
40
20
0
kHz
-2
-4
(b)
(c)
(d)
0
-20
-40
kHz
Fig. 3. Oxygen-17 (a) MQMAS, (b, c) MAS, and (d) static NMR
spectra of Ba(CIO3)2 .H2170. The sample spinning frequency was
20 and 10 kHz in (b) and (c), respectively. Note the different
scales used in (a) and (b)-(d).
narrower than that of the corresponding MAS spectrum by a factor of 100, and is consistent with one
crystallographically distinct hydroxyl group [25].
All '70 NMR parameters obtained for compounds
studied in this work are listed in Table 1. It should
be noted that a significant portion of the line width
observed in the 170 MQMAS spectra of the phosphates arises from 1j(170,31p), which is ~ 90 Hz for
the [PO4] 3- group [26]. Indeed, J-coupling has been
observed in the 11B (S = 3 / 2 ) MQMAS spectra of a
borane phosphite adduct [27]. In the 170 (S = 5 / 2 )
triple-quanmm-MAS spectra, however, the spectral
splitting from 1j(170,31p) would be reduced to 0.55
1j(170,31P) [27]. Therefore, the true line width in the
170 MQMAS spectra of the phosphates is estimated
to be on the order of 100 Hz ( ~ 1.8 ppm at 9.4 T), a
value comparable to the line widths observed for
Ba(CIO3) 2 • H2170 and Ca(17OH)2.
4. Conclusions
Our results clearly demonstrate the practicality of
obtaining high-resolution solid state 170 NMR spec-
G. Wu et al. / Chemical Physics Letters 277 (1997) 79-83
tra with MQMAS. We have shown that the resolution in the 170 MQMAS spectra is approximately 30to 150-fold higher than that found in the 170 MAS
spectra, permitting the detection of crystallographically distinct oxygen sites. Such spectral details are
not observable in the 170 MAS spectra because of
the second-order quadrupolar broadening. We have
also shown that the sensitivity of I~O MQMAS
experiments can be greatly enhanced by rotor-synchronized t I acquisition with very rapid MAS. Since
MQMAS can be readily implemented, extensive
high-resolution 170 NMR studies of solid materials
should be feasible. We anticipate that, with a combination of high magnetic fields (e.g., 750 MHz), fast
sample spinning (e.g., > 25 kI-Iz) and NMR probes
that can deliver high RF power, 170 MQMAS NMR
will be useful for studies of a variety of solid materials including biologically important macromolecules.
Further investigations along this line are in progress.
Acknowledgements
GW is grateful to the Natural Sciences and Engineering Research Council (NSERC) of Canada for
awarding a postdoctoral fellowship. PCH thanks the
Undergraduate Research Opportunity Program
(UROP) of MIT. This research was supported by
grants from the National Institutes of Health (GM23403 and RR-00995).
References
[1] L. Frydman, J.S. Harwood, J. Am. Chem. Soc. 117 (1995)
5367.
[2] A. Medek, J.S. Harwood, L. Frydman, J. Am. Chem. Soc.
117 (1995) 12779.
83
[3] D.W. Boykin, Oxygen-17 NMR Spectroscopy in Organic
Chemistry, CRC Press, Boca Raton, FL, 1991.
[4] G. Wu, D. Rovnyak, B. Sun, R.G. Griffin, Chem. Phys. Lctt.
249 (1996) 210; 257 (1996) 414.
[5] P.J. Dirken, S.C. Kohn, M.E. Smith, E.R.H. van Eck, Chem.
Phys. LeU. 266 (1997) 568.
[6] J.H. Baltisberger, Z. Xu, J.F. Stebbins, S.H. Wang, A. Pines,
J. Am. Chem. Soc. 118 (1996) 7209.
[7] D. Massiot, J. Magn. Resort. Set. A 122 (1996) 240.
[8] J.-P. Amoureux, C. Fernandez, L. Frydman, Chem. Phys.
Leu. 259 (1996) 347.
[9] S. Vega, T.W. Shattuck, A. Pines, Phys. Rev. Lett. 37 (1976)
43.
[10] S. Vega, A. Pines, J. Chem. Phys. 66 (1977) 5624.
[11] S. Vega, Y. Naor, J. Chem. Phys. 75 (1981) 75.
[12] G. Wu, D. Rovnyak, R.G. Griffm, J. Am. Chem. Soc. 118
(1996) 9326.
[13] M.M. Maricq, J.S. Waugh, Chem. Phys. Leu. 47 (1977) 327.
[14] J.-P. Amoureux, C. Fernandez, S. Steuernagei, J. Magn.
Resort. Ser. A 123 (1996) 116.
[15] A.C. Kunwar, G.L. Turner, E. Oldfield, J. Magn. Reson. 69
0986) 124.
[16] D. Massiot, B. Touzo, D. Trumean, J.P. Coutures, J. Virlet,
P. Florian, PJ. Grandinetti, Solid State Nucl. Magn. Reson. 6
(1973) 73.
[17] A. Perloff, A.S. Posner, Science 124 (1956) 583.
[18] D.G. Ott, Syntheses with Stable Isotopes of Carbon, Nitrogen, and Oxygen, Wiley, New York, 1981, p.202.
[19] T.H. Walter, G.L. Turner, E. Oldfield, J. Magn. Reson. 76
(1988) 106.
[20] A.S. Posner, A. Pedoff, A.F. Diorio, Acta Crystallogr. 11
(1958) 308.
[21] D.W. Jones, J.A.S. Smith, J. Chem. Soc. (1962) 1414.
[22] R.W.G. Wyckoff, Crystal Structures, vol. 3, Wiley, New
York, 1963, p.160.
[23] M. Shporer, A.M. Achlama, J. Chem. Phys. 65 (1976) 3657.
[24] S.K. Sikka, S.N. Momin, H. Rajagopal, R. Chidambaram, J.
Chem. Phys. 48 (1968) 1883.
[25] W.R. Busing, H.A. Levy, J. Chem. Phys. 26 (1957) 563.
[26] J.A. Gerlt, P.C. Demon, S. Mehdi, J. Am. Chem. Soc. 104
(1982) 2848.
[27] G. Wu, S. Kroeker, R.E. Wasylishen, R.G. Griffin, J. Magn.
Resort. 124 (1997) 237.