Heteronuclear polarization transfer using selective pulses during hydrogenation with parahydrogen

Full text of DOI:10.1006/jmra.1996.0109

JOURNAL OF MAGNETIC RESONANCE, Series A 120, 129–132 (1996)
ARTICLE NO. 0109
Heteronuclear Polarization Transfer Using Selective Pulses
during Hydrogenation with Parahydrogen
J
ENS B BARGON,* HELMUT S RAY FREEMAN‡
ARKEMEYER,*,† JOACHIM ENGSTSCHMID,‡,§ AND
*Institute of Physical Chemistry, University of Bonn, Wegelerstrasse 12, D-53115 Bonn, Germany; and ‡Department of Chemistry,
University of Cambridge, Lensfield Road, Cambridge CB2 1EW, United Kingdom
Received February 21, 1996
In recent years, a new NMR technique called parahydro-
s
(t0 )
Å
0.25
0
I1 I2 .
[1]
gen-induced polarization (PHIP) (1) has been developed to
study homogeneous catalytic hydrogenations in situ. This
method uses parahydrogen to generate a large signal en-
hancement in the resulting 1H NMR spectra. In this Commu-
nication, we introduce new methods of transferring this po-
larization of the PHIP effect to heteronuclei, particularly to
those with a low gyromagnetic ratio and low natural abun-
dance, like 13C. These techniques, which are based on the
recently introduced method called SEPP (selective excita-
tion of polarization using PASADENA), generate in-phase
signals of the heteronuclei. Normally, the anti-phase nature
of the polarization signals precludes decoupling during ac-
quisition, but after SEPP, the in-phase character of the in-
duced signals permits decoupling, thereby increasing the sig-
nal-to-noise ratio by a factor of 3 to 4.
The PHIP technique was introduced in 1986 by Bowers
and Weitekamp (2, 3) and has since been used as a method
for the examination of homogeneous hydrogenation reac-
tions in situ. Widespread applications have been found, for
example, the detection of reaction intermediates, examina-
tion of reaction mechanisms, and kinetic studies (4–9). The
PHIP method takes advantage of the selective overpopula-
tion of spin states of the hydrogenated products, due to the
addition of parahydrogen to double or triple bonds of organic
compounds, or directly to transition metal compounds.
If both hydrogen atoms from the parahydrogen retain their
phase correlation during the transfer reaction, then the corre-
sponding spin levels of the product (i.e., the ab and ba
states) become highly overpopulated. Accordingly, the dif-
ference in population of individual energy levels can be as
much as five orders of magnitude higher than those resulting
from the thermal Boltzmann distribution.
After the development under the Hamiltonian of the prod-
uct spin system in which these protons form an AX system,
and after averaging over the whole evolution time, the spin-
density matrix becomes
s
(tave
)
Å
0.25
0
IAz IXz .
[2]
In short, addition of parahydrogen in the presence of a
static magnetic field leads to a product-operator term corre-
sponding to longitudinal two-spin order for the hydrogena-
tion product. This abnormally high population, in the follow-
ing termed ‘‘polarization,’’1 persists for a time on the order
of the spin–lattice relaxation time. Up to now, most applica-
tions of this method exploit proton polarizations; only a
few publications have dealt with the subsequent transfer of
polarization to heteronuclei (4, 9, 10).
Two simple pathways for heteronuclear polarization trans-
fer have been found, which can be detected after application
of a simple hard 90
Њ pulse to the heteronucleus. The first
mechanism is that of the well-known nuclear Overhauser
effect, which transfers polarization by cross relaxation from
the added hydrogen atoms to the heteronucleus. The first
example of using PHIP for this purpose was signal enhance-
ment in a 31P spectrum (4). In the case of generating a
reaction product via hydrogenation whereby the former para-
hydrogen protons end up in an AA؅ or AB part of a spin
system, even stronger enhancements can be achieved owing
to higher-order effects (11). More recently, we were able
to detect a signal enhancement of up to 2500 in a 13C spec-
trum as a result of hydrogenation of acetylenedicarboxylic
dimethyl ester (10). A second method of transferring polar-
ization to nuclei with low gyromagnetic ratio and low natural
In terms of product-operator formalism, the density matrix
of parahydrogen can be described as
abundance is the application of the INEPT
/
sequence to the
† While at Cambridge on leave from the Institute of Physical Chemistry,
University of Bonn, Germany.
1 ‘‘Polarization,’’ strictly speaking, is restricted to a spin system for which
§ On leave from the Institute of Organic Chemistry, University of Graz, a spin temperature can be defined. An alternate, more general expression,
Austria. would be ‘‘alignment.’’
129
1064-1858/96 $18.00
Copyright
᭧
1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

130
COMMUNICATIONS
polarized hydrogenation products, which in the case of an be used for efficient polarization transfer to other nuclei by
first transforming the IAz IzX magnetization of the PHIP-polar-
ized protons into in-phase magnetization, followed by an
iridium catalyst generates a 160-fold enhancement of the
signal strength of a 13CO group (9).
All these techniques of polarization transfer have some
disadvantages with respect to a general in situ NMR experi-
ment. The transfer by way of cross relaxation is dependent on
the magnitude of the dipolar couplings and may be severely
decreased by competing relaxation processes. Cross relax-
INEPT or DEPT pulse sequence. (Note that the hetero-
nuclear polarization-transfer step prepares the proton signals
in an anti-phase configuration with respect to the 13C–H
coupling, not with respect to the H–H coupling.)
The experiments were carried out in a spinning 10 mm
NMR tube in a 400 MHz Varian spectrometer (VXR-400).
Parahydrogen was enriched to about 50% by passing ordi-
nary hydrogen over charcoal in a U-tube cooled to 77 K in
liquid nitrogen. This gas was bubbled into a solution con-
taining 3.1 ml CDCl3 , 20 mg of the catalyst [Rh(norborna-
ation and the application of the INEPT
/
sequence both lead
to anti-phase signals. Consequently, broadband decoupling
of the protons cannot be used to increase the signal-to-noise
ratio of the signals from the hydrogenation product. Both
techniques lead to signals that are anti-phase with respect
to long-range couplings, particularly when applied to 13C
spectroscopy. As these splittings are often on the same order
as the experimental linewidths, the intensities of the signals
may suffer significantly. Finally, only very few hydrogena-
diene)(PPh3 )2 ], and 100 ml of 1-hexyne or phenylacetylene
(13). In order to carry out the experiments while spinning
the sample tube, a long glass tube with an outer diameter of
5 mm was fixed centrally inside the magnet bore. One end
of this glass tube (the gas inlet) extended about 3 cm outside
the magnet; the other end reached about 2 cm inside the 10
tion reactions lead to an AA
؅ or an AB spin system.
The recently introduced SEPP technique (12) transforms
the longitudinal two-spin order term IAz IzX into in-phase mag- mm NMR sample tube. Its purpose was to hold in place a
netization on one of the protons originating from the parahy- long polyethylene tube which carried the parahydrogen gas
drogen. In this publication, we show that this technique can to a 1 mm glass capillary that dipped into the reaction solu-
FIG. 1. The three proposed pulse sequences; in each case, the delay t1 is adjusted to 1/(2JCH ). (a) SEPP-INEPT with refocusing, (b) SEPP-INEPT ,
/
and (c) SEPP-DEPT.

COMMUNICATIONS
131
tor is available. During an evolution time of 1/(2JAM ), the
2IAy IMz magnetization converts into IAx , if a simultaneous
180x proton pulse is applied after an interval 1/(4JAM ) to
refocus the 13C–H coupling. Once the in-phase magnetiza-
tion
IAx has been created, signal enhancement of the ole-
0
Њ
0
finic carbon sites can be achieved by the well-known tech-
niques of heteronuclear polarization transfer (Fig. 1) (16).
We have successfully applied SEPP-INEPT (with refo-
cusing), SEPP-INEPT , and SEPP-DEPT (12, 15) to ole-
/
finic products of the hydrogenation of phenylacetylene and
1-hexyne. These new techniques have been compared with
INEPT
of a hard proton 90
/
and a DEPT sequence, modified by the addition
pulse after the last refocusing delay.
Њ
All the results have also been compared with a simple 13C-
pulse–acquire experiment. The initial selective pulse (E-
BURP-2) was applied to one or the other of the newly polar-
ized protons (from parahydrogen) added across the triple
bond. The final refocusing interval of the refocused INEPT
FIG. 2. 13C-NMR spectra recorded during hydrogenation of phenylacet-
ylene to styrene with parahydrogen, starting with an E-BURP-2 pulse on
M (at 5.13 ppm) and generating a signal from the 13C nucleus directly
bound to A. (a) SEPP-DEPT (
(b) Proton-coupled SEPP-DEPT (
through 54 ) with refocusing and decoupling. (d) Proton-coupled SEPP-
INEPT (evolution through 45 ), starting with a soft pulse on A (6.61
u
Å
54
Њ
) with decoupling during acquisition.
90Њ). (c) SEPP-INEPT (evolution
u
Å
Њ
and the INEPT
tral editing (16) according to the number of directly bound
protons at the 13C site, i.e., 45
of relative precession in the
case of two protons, 90 precession for one proton, and
54 to achieve the optimum for both sites. For the DEPT
sequences, the proton editing pulse was modified in a corre-
sponding manner.
Almost all the selective pulse experiments gave a higher
signal-to-noise ratio than the nonselective sequences. The
resulting spectra (Figs. 2 and 3) show stronger signals than
/
sequences was optimized to achieve spec-
/
Њ
ppm). (e) The same as (d) but decoupled.
Њ
Њ
Њ
tion. This could be retracted just before the radiofrequency
pulse sequence so that the field homogeneity was not dis-
turbed by the hydrogen bubbles.
To permit a proper comparison of the performance of the
different pulse sequences, the hydrogenation conditions must
be the same in order to achieve the same degree of initial
polarization. Consequently, we carried out a series of experi-
ments using six different pulse sequences. Any individual
experiment in this series involved 35 s of hydrogenation
(with the capillary inside the solution), followed by an inter-
val of 3 s after the capillary had been removed, before start-
ing the pulse sequence. Two dummy experiments were car-
ried out before each run. A two-step phase cycle was used
to cancel spurious signals.
the INEPT
cause of the strong cross relaxation between the directly
bonded protons, application of a single hard 90 pulse gener-
ates two- to threefold stronger signals than the nonselective
INEPT or modified DEPT sequence. In these cases, the
/
sequence or the modified DEPT sequence. Be-
Њ
/
NOE is more effective than the transfer by the nonselective
pulse sequences, which do not convert the NOE enhance-
ment into detectable magnetization. However, the applica-
tion of pulse sequences starting with SEPP leads to signals
Consider an AMX spin system in which the parahydrogen
is transferred to spins A and M, while X is a 13C spin that
is directly bonded to A but two bonds away from M:
In order to transform the 2IAz IzM magnetization of the highly
polarized proton spins into IAx magnetization, we apply a
90
Immediately afterward, a hard 90x
Њ
E-BURP-2 pulse (14) to spin M, generating 2IAz IyM .
Њ
proton pulse converts
this into 2IAy IMz . This avoids having to apply separate selec-
tive pulses to excite the outer 13C satellites of the A proton.
Simulations (15) indicate that the side effects due to evolu-
tion during a simple selective pulse (for example, a rectangu-
lar pulse) are less serious than those that occur in more
sophisticated (and, therefore, longer) shaped pulses. This
FIG. 3. 13C-NMR spectra obtained during hydrogenation of 1-hexyne
with parahydrogen, starting with a selective pulse on proton A (5.78 ppm)
followed by selective DEPT (
u
Å
54Њ) with decoupling (a) and coupled
makes the experiment feasible even if no pulse shape genera- (b). Note the eightfold expansion of the frequency scale in (b).

132
COMMUNICATIONS
2. C. R. Bowers and D. P. Weitekemp, Phys. Rev. Lett. 57, 2645
(1986).
with a comparable signal-to-noise ratio, even if no decou-
pling is carried out (Figs. 2b, 2d, and 3b). Moreover, these
signals are in phase and therefore are easier to interpret in
cases where multiple couplings lead to more complicated
signals. In addition, phase cycling in the SEPP experiments
permits the suppression of artifacts and undesirable conven-
tional signals not generated by the PHIP effect. But the
most striking feature of this technique is the possibility of
decoupling during acquisition (Figs. 2a, 2c, 2e, and 3a),
which leads to spectra with a signal-to-noise ratio as high
as 400 after only two scans. This gain is between three
3. C. R. Bowers and D. P. Weitekamp, J. Am. Chem. Soc. 109, 5541
(1987).
4. T. C. Eisenschmid, J. McDonald, R. Eisenberg, and R. G. Lawler,
J. Am. Chem. Soc. 111, 7267 (1989).
5. C. R. Bowers, D. H. Jones, N. D. Kurur, J. A. Labinger, M. G.
Pravica, and D. P. Weitekamp, in ‘‘Advances in Magnetic Reso-
nance’’ (W. S. Warren, Ed.), Vol. 14, p. 269, Academic Press, San
Diego, 1990.
6. R. Eisenberg, Acc. Chem. Res. 24, 110 (1991).
7. M. S. Chinn and R. Eisenberg, J. Am. Chem. Soc. 114, 1908
(1992).
and four times higher than that after a hard 90Њ pulse. The
attractiveness of parahydrogen as a tool for the examination
of catalytic hydrogenation in situ will benefit from the in-
creased sensitivity, especially in cases where the resulting
hydrogenation products exhibit polarization signals that are
split by many couplings and where the initial polarization
is weak.
8. R. Eisenberg, T. C. Eisenschmid, M. S. Chinn, and R. U. Kirss, in
‘‘Advances in Chemistry’’ (W. R. Moser and D. W. Slocum, Eds.),
Vol. 230, Washington, DC, 1992.
9. S. B. Duckett, C. L. Newell, and R. Eisenberg, J. Am. Chem. Soc.
115, 1156 (1993).
10. J. Barkemeyer, M. Haake, and J. Bargon, J. Am. Chem. Soc. 117,
2927 (1995).
ACKNOWLEDGMENTS
11. H. J. Bernstein, J. A. Pople, and W. G. Schneider, Can. J. Chem.
35, 65 (1957).
H. Sengstschmid thanks the Austrian government as well as the govern-
ment of the Federal State of Upper Austria for financial support. J. Barke-
meyer thanks Professor C. Griesinger and Dr. S. J. Glaser, University of
Frankfurt am Main, Germany, for providing their simulation program
SIMONE, and the Graduiertenkolleg of the Deutsche Forschungsgem-
einschaft Bonn, Germany, for financial support.
12. H. Sengstschmid, J. Barkemeyer, R. Freeman, and J. Bargon, J.
Magn. Reson., in press.
13. J. Bargon, J. Kandels, and K. Woelk, Z. Phys. Chem. 180, 65
(1993).
14. H. Geen and R. Freeman, J. Magn. Reson. 93, 93 (1991).
REFERENCES
15. S. J. Glaser and G. P. Drobny, in ‘‘Advances in Magnetic Reso-
nance’’ (W. S. Warren, Ed.), Vol. 14, p. 35, Academic Press, San
Diego, 1990.
1. J. C. Eisenschmid, R. U. Kirss, P. P. Deutsch, S. I. Hommeltoft, R.
Eisenberg, J. Bargon, R. G. Lawler, and A. L. Balch, J. Am. Chem.
Soc. 109, 8089 (1987).
16. O. W. Sørensen and R. R. Ernst, J. Magn. Reson. 51, 477 (1983).