Photo-CIDNP studies of hydroxy ketones and higher-molecular-weight systems

Article abstract of DOI:10.1002/hlca.200690230

In situ NMR spectroscopy can be applied to investigate chemical reactions during which free radicals occur as intermediates. In chemical systems of low molecular weight, nuclear-spin polarization results from the spin selectivity of free-radical reactions, because a pair of radicals has to obey the exclusion principle according to Pauli; therefore, this system reacts in a spin-selective manner when forming a chemical single bond. As a consequence, strong transient absorption and emission lines occur in the NMR spectra acquired during free-radical reactions. This phenomenon is known as 'chemically induced dynamic nuclear polarization' (CIDNP). However, long correlation times associated with short proton spin-lattice relaxation times T₁ render it difficult to observe CIDNP in macromolecules. Therefore, in these and in many other cases, it can be attractive to utilize the simultaneously occurring heteronuclear polarization or to selectively transfer the ¹H polarization to heteronuclei, since their T₁ times can be substantially shorter than those of protons. In this paper, we present examples of how CIDNP can be observed both in low-molecular-weight systems as well as in systems exhibiting a rather macromolecular character. Also, CIDNP can assist in obtaining useful information about macromolecular systems that normally is very difficult to obtain otherwise. Since CIDNP is primarily a qualitative method by which free-radical intermediates may be identified, we have developed a procedure allowing the quantitative determination of the magnetic properties of the intermediate free radicals. This process is especially useful for very short-lived radicals, which are frequently elusive to ESR spectroscopy conducted in solution. In particular, g values of a variety of radical ions have been determined in this fashion by CIDNP-NMR spectroscopy. The data thus obtained provide an alternative to ESR and, hence, complement this traditional method (manuscript in preparation).

Full text of DOI:10.1002/hlca.200690230

2
522
Helvetica Chimica Acta – Vol. 89 (2006)
Photo-CIDNP Studies of Hydroxy Ketones and Higher-Molecular-Weight
Systems
a
b
by Joachim Bargon* ) and Lars T. Kuhn )
a) Institute of Physical & Theoretical Chemistry, University of Bonn, Wegelerstrasse 12, DE-53115 Bonn
phone: +49-228-732261; e-mail: bargon@uni-bonn.de)
b) Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1
QZ, UK
(
3
In memoriam Professor Hanns Fischer
In situ NMR spectroscopy can be applied to investigate chemical reactions during which free radicals
occur as intermediates. In chemical systems of low molecular weight, nuclear-spin polarization results
from the spin selectivity of free-radical reactions, because a pair of radicals has to obey the exclusion prin-
ciple according to Pauli; therefore, this system reacts in a spin-selective manner when forming a chemical
single bond. As a consequence, strong transient absorption and emission lines occur in the NMR spectra
acquired during free-radical reactions. This phenomenon is known as ꢀchemically induced dynamic
nuclear polarizationꢁ (CIDNP). However, long correlation times associated with short proton spin–lattice
relaxation times T render it difficult to observe CIDNP in macromolecules. Therefore, in these and in
1
many other cases, it can be attractive to utilize the simultaneously occurring heteronuclear polarization
1
or to selectively transfer the H polarization to heteronuclei, since their T times can be substantially
1
shorter than those of protons.
In this paper, we present examples of how CIDNP can be observed both in low-molecular-weight
systems as well as in systems exhibiting a rather macromolecular character. Also, CIDNP can assist in
obtaining useful information about macromolecular systems that normally is very difficult to obtain oth-
erwise. Since CIDNP is primarily a qualitative method by which free-radical intermediates may be iden-
tified, we have developed a procedure allowing the quantitative determination of the magnetic properties
of the intermediate free radicals. This process is especially useful for very short-lived radicals, which are
frequently elusive to ESR spectroscopy conducted in solution. In particular, g values of a variety of rad-
ical ions have been determined in this fashion by CIDNP-NMR spectroscopy. The data thus obtained
provide an alternative to ESR and, hence, complement this traditional method (manuscript in prepara-
tion).
Introduction. – Ever since its accidental discovery [1][2] in the form of intense emis-
sion and enhanced absorption lines in NMR spectra ca. 40 years ago, the nuclear-spin-
polarization phenomenon termed ꢀchemically induced dynamic nuclear polarizationꢁ
(CIDNP) has been used to investigate a great variety of free-radical reactions. Espe-
cially attractive is its application to determine structural details of biochemically impor-
tant molecules both in the steady state as well as during the progression of folding or
unfolding [3]. The observation of CIDNP requires in situ NMR spectroscopy, which
allows one to investigate chemical reactions during which free radicals occur as inter-
mediates. At least in chemical systems of low molecular weight, nuclear-spin polariza-
tion results from the spin selectivity of free-radical reactions, because a pair of radicals
ꢂ 2006 Verlag Helvetica Chimica Acta AG, Zürich

Helvetica Chimica Acta – Vol. 89 (2006)
2523
has to obey the exclusion principle when forming a chemical single bond. Also, the
CIDNP phenomenon is suited to identify both the nature of intermediate radicals as
well as their reaction products in macromolecular systems. However, observing
CIDNP in systems of high molecular weight is intrinsically more difficult, mainly
because of the very short relaxation times of the nuclei within or even in the mere pres-
ence of macromolecules. This is especially true for situations where the reaction prod-
ucts themselves have a high molecular weight, and it is worst in cases where the product
molecules contain long molecular chains, e.g., alkyl chains. The situation is, however,
not totally hopeless provided the NMR equipment is adapted to suit the required boun-
dary conditions for investigating such reactions.
CIDNP is an indirect method to observe and characterize free radicals by means of
the polarization observed in the corresponding reaction products. Since the extent of
CIDNP is inversely proportional to the lifetime of the interacting paramagnetic inter-
mediates, this spectroscopic method complements direct electron-spin-resonance
(ESR) methods to investigate free radicals both in its applicability as well as in the
type of results obtained.
The magnitude of CIDNP enhancements in reaction products of low molecular
weight is governed by the ratio of the magnetic moment of the electron vs. the magnetic
moments of the nuclei being observed. Thus, for H-atoms, the experimentally observed
enhancements typically range in regions of one to almost three orders of magnitude,
and for heteronuclei even higher. The related CIDEP effect in ESR is much smaller
than typical CIDNP-NMR enhancements. The CIDEP enhancements are of the
order of unity in steady-state spectra, but can reach values of up to two orders of mag-
nitude. To detect these higher CIDEP enhancements, ESR spectrometers have to be
modified to allow observation of the radicals within microseconds after their formation.
Experimental. – To observe and record CIDNP spectra, no modification of conventional NMR spec-
trometers operating in either FT or CW mode is required. The latter mode used to be typical for NMR
spectrometers prior to the advent of the Fourier transform (FT) technology, which requires computers to
generate the NMR spectrum from a free induction decay (FID). Accordingly, we have used the CW
mode when recording spectra at a lower magnetic-field strength of either 60 or 100 MHz due to the avail-
1
ability of older CW spectrometers in our laboratory. All H-NMR spectra recorded at 80 or 200 MHz (or
higher), however, were obtained using contemporary FT-NMR spectrometers.
Photochemical generation of free radicals requires appropriate attachments to conventional NMR
spectrometers. Appropriately modified probes were either home-built or available commercially,
whereby the required light is admitted either along the axis of the static magnetic field or from the
side of the spinning NMR tube. In the latter case, mirrors were used to deflect and admit the radiation
through an appropriately shaped and spaced radio-frequency coil.
However, the required modifications can be minor, especially if an appropriately shaped UV-light
rod built from fused quartz is used, as in our case. Such a rod of an outer diameter between 4 and 10
mm (depending on the inner diameter of the NMR tube) extends into the open spinning sample from
the top with its tapered end, cut at an angle of ca. 458, which, in turn, extends ca. 1 cm into the liquid
sample. There it introduces efficient stirring agitated by the spinning, in addition to allowing sufficient
13
UV irradiation of the continuously stirred solution. This stirring effect is especially beneficial in C-
FT-NMR spectrometers and has its merit even without accompanying UV illumination. Recording of
CIDNP spectra with FT-NMR spectrometers in combination with a pulsed-light source (laser) has certain
advantages over a CW spectrometer in combination with a Hg/Xe arc lamp, because, in this way, the
effects of NMR relaxation can be greatly reduced or even eliminated.

2
524
Helvetica Chimica Acta – Vol. 89 (2006)
Results and Discussion. – 1. CIDNP in Macromolecular Systems. The dilemma with
CIDNP in macromolecular systems is relaxation: macromolecules, in particular those
containing long alkyl chains, have notoriously short relaxation times. When investigat-
ing macromolecular systems using conventional NMR spectroscopy it is, thus, common
practice to increase the temperature, typically to values exceeding 1008C, to obtain rel-
atively narrow resonance lines thanks to the lowered viscosity of the solution. For bio-
logical systems, however, this option is not applicable because under such conditions
they would collapse and loose there activity.
An early attempt to explore the seriousness of this situation and the consequences
of such relaxation processes for the nuclear-spin polarization derived from free radicals
via CIDNP was the investigation of a number of macromolecular model systems [4],
whereby the required free radicals or pairs thereof were generated photochemically
according to Scheme 1.
Scheme 1
The systems (mainly ketones) studied contained appropriate chromophores to
absorb UV light at a wavelength of ca. 300 nm. Their C=O groups can be photo-
excited, upon which they undergo a variety of photoreactions, including a-cleavage
and photoreduction.
In one set of experiments, we investigated the photo-induced a-cleavage of poly-
(
3
methylisopropenyl ketone) (PMIK) and its low-molecular-weight model compounds
,3-dimethylbutan-2-one (DMB) and 3-ethyl-3-methylpentan-2-one (EMP). Scheme 1
illustrates the corresponding photoreactions of these systems.
Solutions of the ketones in CD CN, C D , CDCl , or CCl , each containing 0.1M of
A
H
R
U
G
ACHTREUNG
3
6
6
3
4
the model compound or 10 g/l of the polymer, respectively, were irradiated in 5-mm
quartz tubes inside a modified probe of a CW-NMR spectrometer operating at 60
MHz. The probe admitted UV irradiation through a quartz light pipe from the rear.
A 5-kW high-pressure Hg/Xe lamp served as the light source. The optical system con-
sisted of a spherical mirror, two quartz lenses, and a filter of 10-cm path length contain-
ing an aqueous solution of NiSO ·7 H O (250 g/l), CoSO ·6 H O (50 g/l), and H SO
4
2
4
2
2
4
(1 g/l) to remove the IR and most of the VIS radiation. Tetramethylsilane (TMS) sealed
in a capillary provided a magnetic-field lock. Chemical shifts d were measured against
internal TMS in separate runs. Both aerobic and anaerobic (deoxygenated) solutions
gave identical results.
In Fig. 1, the low-field portions of the CIDNP spectra observed during the photo-
A
C
H
T
R
E
U
N
G
lysis of DMB (a), EMP (b), and PMIK (c) in CD CN are shown. The high-field portions
A
T
E
N
3
of the spectra were omitted because of overlapping absorption lines of the starting
materials at d(H) 2.5ꢀ0 ppm. The sections of the spectra showed no lines in the
absence of irradiation, but on admission of UV light, the following resonances

Helvetica Chimica Acta – Vol. 89 (2006)
2525
Fig. 1. CIDNP Spectra during the photolysis of a) DMB, b) EMP, and c) PMIK in CD CN solution
3
appeared and disappeared after termination of the irradiation. An intense quadruplet
(q) centered at d(H) 9.7 occurred in all three traces of Fig. 1, which can be attributed to
the aldehyde H-atom of acetaldehyde. Furthermore, a strong multiplet (m) was
observed at d(H) 4.7 (Fig. 1,a) caused by the methylene protons of isobutylene (=2-
methylprop-1-ene). In Fig. 1,b, a similar, but much weaker, multiplet was found, with
a broad and even weaker absorption band at d(H) 5.2–6.2. In Fig. 1,c, however, no cor-
responding resonance was observed in the olefinic region.
The starting materials DMB and EMP showed CIDNP themselves during irradia-
tion, in contrast to PMIK. In C6
tical results were obtained. Table 1 lists the products identified during the photolysis of
PMIK in C D solution.
A
H
R
U
G
AHCTREUNG
6
6
Using known values for the g parameters and the proton hyperfine coupling con-
stants of the radicals occurring in Scheme 1, it can be concluded, according to the
well-established CIDNP-sign rules [5], that photodecomposition of DMB yields ace-
tyl/tert-butyl radical pairs with predominant triplet character. Accordingly, DMB
undergoes a Norrish type-I cleavage from its triplet state.

2
526
Helvetica Chimica Acta – Vol. 89 (2006)
Table 1. The Photoproducts and Their CIDNP Observed During Irradiation of Poly(methylisopropenyl
Ketone) (PMIK)
Product
d(H) [ppm]
T [s]
1
Multiplicitya)
CIDNPb)
(
(
(
CH C(Me)COMe)n
1.5
1.1
2.1
1.49
9.7
5.2–5.6
1.5
4.8
1.79
7.2–
<0.5
<0.5
<0.5
<20
<35
<0.5
<0.5
<0.5
<4
br.
br.
br.
d
N
N
N
2
CH C(Me)COMe)n
2
CH C(Me)COMe)n
2
MeCHO
MeCHO
E (+A/E)
q
A
N
N
N
A
ꢀ
CH=C(Me)
CH=C(Me)
CH (C=CH )ꢀ
br.
br.
br.
s
ꢀ
ꢀ
2
2
MeCOꢀCOMe
HD (solvent)
C
A
C
H
T
R
E
U
N
G
–
N
6
5
a) Abbreviations: br, broad; d, doublet; q, quadruplet; s, singlet. ) A, enhanced absorption; E, emission;
N, no CIDNP; A/E, superimposed multiplet effect with low-field components showing enhanced absorp-
tion (A) and high-field components showing emission (E).
b
If we assume that the other two radicals in Scheme 1 have the same g value as Me3CC
A
H
E
N
and a positive hyperfine coupling constant for all their b-H-atoms, it follows from the
CIDNP absorption of the universal photoproduct acetaldehyde (observed throughout
in Fig. 1) that PMIK as well as the two model compounds are cleaved in inert solvents
predominantly in their photo-excited triplet state. The CIDNP spectra obtained from
the polymer exhibit fewer lines than expected, which is outlined in the following: acet-
aldehyde is one of two simultaneously formed disproportionation products of the Nor-
rish type-I decomposition of the ketones described. Hence, one would expect the H-
atoms of the complementary disproportionation product (last radical in Scheme 1) to
display CIDNP as well. However, there is no CIDNP signal from an olefin in Fig.
1
,c, whereas in Fig. 1,a and Fig. 1,b the products show CIDNP for their olefinic H-
atoms. Furthermore, because the acetaldehyde H-atom is in b-position in the inter-
mediate radical and since the remaining equivalent b-H-atoms become the olefinic
ones in the last olefin, both the aldehyde H-atom and the olefinic H-atoms should
become equally strong when polarized. This, however, was not the case. Therefore,
the missing CIDNP signal of the olefinic H-atoms in Fig. 1,c can be rationalized only
by a relaxation phenomenon: once the initially equivalent radical nuclei become part
of different reaction products, their spin polarization will decay with different rates.
These rates can be measured experimentally by NMR with the aid of the proton nuclear
spin–lattice relaxation times T . It is well-known that H-atoms of different chemical
1
groups of the same molecule may have different relaxation times. Hence, CIDNP-
intensity correlations will even be lost for nuclei that are equivalent in the intermediate
radicals, but become nonequivalent in the reaction products.
For reactions yielding macromolecular products, it has to be considered that long-
chain molecules show decreasing T values with increasing chain lengths until a critical
1
polymerization degree is reached, above which T is constant. This phenomenon has
1
been attributed to segmental motions of the macromolecules. Measurements for aque-
ous solutions of polyethylene oxide as a function of the polymerization degree n yielded

Helvetica Chimica Acta – Vol. 89 (2006)
2527
T values of 5.0, 1.0, and 0.6 s for model compounds with n=1, 6, and 20–20,000,
1
respectively. Similar data have been reported for the H-atoms of unbranched alkanes.
From these data, it can be extrapolated that T <0.5 s for the products of Scheme 1.
1
The CIDNP signals for molecules with such short T values could at best be weak,
1
but are most likely undetectable. This conclusion follows from the early studies of Bar-
gon and Fischer [6], who showed that, on shortening T by addition of increasing con-
1
centrations of paramagnetic ions to systems showing CIDNP, the signal enhancements
of the products could be decreased and eventually suppressed.
In the systems studied here, the proton T values of the simultaneously formed dis-
1
proportionation products of Scheme 1 differ by a factor of ca. 2for those derived from
DMB, but by two orders of magnitude for those from PMIK. The CIDNP intensities of
these products cannot be correlated in a simple way with their relative chemical yields,
which is evident from Fig. 1,a and Fig. 1,c.
Nevertheless, the fact that the reaction product acetaldehyde shows CIDNP, inde-
pendent of the type and the size of its partner radical in Scheme 1, suggests that, in prin-
ciple, CIDNP can be generated in macromolecular systems as well as in reactions of
small radicals. Thus, it appears that in the intermediate free radicals the proton and
electron spin–lattice relaxation times are quite insensitive to the chain length of the
substituent R. Once the CIDNP is transferred to the reaction products, however, the
nuclei in the macromolecules will experience a fast decay of their polarization, whereas
those in small molecules retain it much longer. Because of this, it is not surprising that
CIDNP spectra recorded during photolysis of polymers do not reflect the relative sig-
nificance of alternate decomposition pathways correctly. In PMIK, it is known that
ꢀ
side-group scissionsꢁ (for which we found exclusive evidence) are indeed of only
minor significance. The main photoreaction of PMIK is instead ꢀmain-chain cleavageꢁ,
which typically yields only long-chain molecules and, hence, no CIDNP. Furthermore,
the reaction mechanism of this main-chain cleavage is not known in detail, but is
thought to proceed via biradical intermediates and, therefore, may not cause CIDNP
in the first place. CIDNP from 1,4-biradical intermediates is unlikely to be detected
at the field strength of 1.4 T of our NMR spectrometer, because the exchange interac-
tion does not match with the electronic Zeeman energies.
2
. Photolysis in Reactive Solvents. – Small product molecules alone can, neverthe-
less, reflect useful information in CIDNP spectra. To demonstrate this, we have inves-
tigated the photolysis of PMIK and its model compounds in CDCl solution. In Fig. 2,
3
the corresponding CIDNP spectra observed during irradiation are shown for the region
d(H)>2.5 ppm. Again, the sections 0<d<2.5 of the spectra were masked by the
intense CIDNP lines of the ketones. Thus, PMIK shows three broad resonances in
CDCl : at d(H) 1.00 the Me groups, at d(H) 1.86 the CH groups, and at d(H) 2.12
3
2
the Ac H-atoms. The onset of the polymer absorption is evident at the right end of
Fig. 2,c. The photoproducts identified are acetaldehyde (d(H) 9.8), CHCl (7.3),
3
CDHCl (5.3), AcCl (2.7), AcCCl (2.5), and individual olefins (e.g., H C=CMe at
2
3
2
2
d(H) 5.3; Fig. 2,a).
From these products, it can be concluded that PMIK, as well as the two model
ketones, undergoes a-cleavage yielding acetyl and tertiary-alkyl radicals. These pri-
mary radicals attack the solvent (CDCl ) and abstract D- or Cl-atoms. Thus, the CCl
3
3
and CDCl moieties being generated encounter other free radicals in the system yield-
2

2
528
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 2. CIDNP Spectra during the photolysis of a) DMB, b) EMP, and c) PMIK in CDCl solution
3
ing a variety of secondary radical pairs. Many of these pairs contain a tertiary-alkyl rad-
ical as one component. All can produce one identical alkene by disproportionation, and
for the precursor DMB, all radical pairs containing tert-butyl radicals can yield isobu-
tylene. Therefore, the CIDNP found for the isobutylene H-atoms stem from a variety
of radical pairs and, hence, correspond to the sum of individual contributions from the
primary t-Bu/Ac radical pair and all appropriate secondary radical pairs. For all systems
studied here, it has been deduced from the CIDNP studies in inert solvents (Fig. 1) that
both primary as well as secondary radical pairs have predominantly electronic triplet
multiplicities. If we assume that the same holds true for the reactive solvent (CDCl ),
3
the CIDNP phase of the isobutylene H-atoms depends only on the g values of the part-
ner radicals (S) in the pairs of the type t-Bu/S. A series of combinations are possible
giving rise to the following phase patterns, where B and S refer to tertiary-alkyl and
partner radical, respectively:
a) g(S)<g(B) ! absorption
b) g(S)>g(B) ! emission
c) g(S)=g(B) ! multiplet effect

Helvetica Chimica Acta – Vol. 89 (2006)
2529
In Fig. 2,a, the isobutylene H-atom resonances reflect a pure multiplet effect, with
no superimposed absorption or emission. We rationalize this as resulting from a cancel-
lation of an equally strong emission and enhanced absorption from both pairs of pat-
terns a and b. Hence, the remaining multiplet effect must stem from pairs of type c.
The only choice for a partner radical S to give radical pairs of type c is another
tertiary-alkyl radical (S=B). All other radicals have inappropriate g values, as can
be seen from Table 2. Similar conclusions have been reached when studying the photo-
reactions of di(tert-butyl) ketone (=2,2,4,4-tetramethylpentan-3-one) in CDCl solu-
3
tion. From the cancellation of the CIDNP stemming from pairs a and b, one is tempted
to determine their relative concentrations. However, because the magnitude of a
CIDNP signal is a complex function of numerous parameters, including a series of
almost unknown rate constants, this effort has to be postponed. According to our
rough estimate, the concentration of the a-type radical pairs appears to dominate
over that of type b.
Table 2. ESR Data of Selected Free Radicals
Radical
g Value
Proton hyperfine
coupling (aH)
AcC
2.005
5.3
22.72
ꢀ16.79
–
Me
A
H
R
U
G
2.0025
2.0083
2.0091
3
Cl (H)CC
2
Cl CC
3
3
. CIDNP from the Photolysis of Steroids. During the photolysis of steroids contain-
ing appropriate chromophores such as an a-hydroxy ketone, enhanced absorption was
observed for the aldehyde H-atom of the low-molecular-weight fragment glycolalde-
hyde. In contrast, the complementary fragment, namely the steroid framework contain-
1
ing a C=C bond, failed to exhibit any detectable CIDNP, at least not in its H-NMR
spectrum. This failure to detect any evidence of nuclear-spin polarization from the ste-
roid has been attributed to efficient relaxation processes quenching any polarization.
4
. CIDNP Studies of Low-Molecular-Weight a-Hydroxy Ketones. Upon investigat-
ing the photolysis of glycolaldehyde (=2-hydroxyacetaldehyde) in an inert, aprotic sol-
ution (CDCl ), we found that, in this case, the aldehyde H-atom displays emission upon
3
1
UV irradiation at 300 nm (Scheme 2). In Fig. 3, the effects observed in the H-NMR
CDNP spectra during the photolysis are shown, recorded at a resonance frequency
of 100 MHz. Photolyzing glycolaldehyde in, e.g., D O, gave rise to much more compli-
A
H
R
U
G
2
cated results because the glycolaldehyde may either occur in the form of different iso-
meric dimers or in its hydrated form.
Scheme 2. Photolysis of Glycolaldehyde

2
530
Helvetica Chimica Acta – Vol. 89 (2006)
1
Fig. 3. 100-MHz H-CIDNP Spectrum recorded during the photolysis of glycolaldehyde
Similar results were obtained in the photolysis of higher a-hydroxy ketones, as is
1
evident from the H-NMR CIDNP spectrum recorded during the photolysis of 1,3-
dihydroxyacetone (=1,3-dihydroxypropan-2-one; Fig. 4).
The higher homologues of a-hydroxy ketones frequently yield the enols of a variety
1
1
of ketones via disproportionation of the intermediate radicals ). Fig. 5 displays the H-
NMR CIDNP spectrum of the enol form of acetone obtained in this fashion from an
appropriate precursor during the photolysis in CD CN solution. Normally, the NMR
A
H
R
U
G
3
spectra of enols, especially of the simplest ketones, are elusive since the concentration
of the tautomeric enols typically remains below the NMR detection limit. Whether the
resonances of these enols appear in either enhanced absorption (A) or emission (E), or
whether they display the multiplet effect, depends on the types of radical pairs from
which they result. When the g values of the two radicals have the same value, the mul-
tiplet effect results (Fig. 5,a), but when the g values differ, the resonances display pure
emission or absorption (Fig. 5,b).
1
Conclusions. – The examples outlined above demonstrate convincingly that H-
CIDNP can be observed very readily in low-molecular-weight products, even when
these products result from macromolecular systems due to the cleavage of a side
group. The complementary macromolecular products, even when formed from the
very same radicals as the low-molecular-weight fragments, typically fail to display
1)
CIDNP has also been used extensively by the late Prof. Hanns Fischer and his group to detect short-
lived enols arising during the radiation of various ketones, see for example [7].

Helvetica Chimica Acta – Vol. 89 (2006)
2531
1
Fig. 4. 100-MHz H-CIDNP Spectrum recorded during the photolysis of 1,3-dihydroxyacetone
1
CIDNP, at least in the H-NMR spectra. We attribute this phenomenon to the different
T nuclear-relaxation times, which are known to be short for macromolecular systems.
1
1
A quite promising, potential alternative is to transfer the CIDNP-derived initial H spin
polarization to heteronuclei, in particular to C. For certain heteronuclei, notably for
C, the T nuclear-relaxation times are considerably longer than those of the corre-
13
13
1
sponding protons. Hence, we have investigated the feasibility and rules for transferring
1
13
19
nuclear polarization from H to C or to F, respectively. For this purpose, we have
1
used another and much more efficient source of initial H spin polarization, namely
the so-called ꢀparahydrogen-induced polarizationꢁ. In this latter case, the mechanisms
leading to nuclear polarization are quite different. However, the subsequent behavior
of the polarized spin system is identical to those where the initial polarization was
attained by exploiting the CIDNP phenomenon. Various systems have been investi-
1
gated in great detail, where we have successfully transferred the initial H polarization
13
19
to C and, alternatively, to F. For this purpose, we have also developed the appropri-
ate pulse sequences, which differ significantly from those used typically for nonpolar-
ized spin systems. The results of all of these studies have been published [8–13].
L. T. K. thanks the Studienstiftung des Deutschen Volkes for constant support and for generous fund-
ing throughout his studies. Further financial support from the Deutsche Forschungsgemeinschaft (DFG)
and the Theodor-Laymann-Stiftung (to L. T. K.) is also gratefully acknowledged.

2
532
Helvetica Chimica Acta – Vol. 89 (2006)
1
Fig. 5. 100-MHz H-CIDNP Spectra of the enol form of acetone displaying the multiplet effect (a) or
appearing in enhanced absorption (b)
REFERENCES
[
[
[
[
[
[
[
1] J. Bargon, H. Fischer, U. Johnsen, Z. Naturforsch. A 1967, 20, 1551.
2] H. R. Ward, R. G. Lawler, J. Am. Chem. Soc. 1967, 89, 5518.
3] R. Kaptein, K. Dijkstra, K. Nicolay, Nature 1978, 274, 293.
4] J. Bargon, J. Polym. Sci. Polym. Chem. Ed. 1978, 16, 2747.
5] R. Kaptein, J. Chem. Soc., Chem. Commun. 1971, 732.
6] J. Bargon, H. Fischer, Z. Naturforsch. A 1967, 22, 1556.
7] B. Blank, H. Fischer, Helv. Chim. Acta 1973, 56, 506; G. P. Laroff, H. Fischer, Helv. Chim. Acta 1973,
56, 2011; B. Blank, A. Henne, G. P. Laroff, H. Fischer, Pure Appl. 1975, 41, 475.
[
[
8] J. Natterer, J. Bargon, Prog. Nucl. Magn. Reson. Spectrosc. 1997, 31, 293.
9] M. Haake, J. Natterer, J. Bargon, J. Am. Chem. Soc. 1996, 118, 8688.
[
[
[
10] J. Barkemeyer, J. Bargon, H. Sengstschmid, R. Freeman, J. Magn. Reson. A 1996, 120, 129.
11] J. Barkemeyer, M. Haake, J. Bargon, J. Am. Chem. Soc. 1995, 117, 2927.
12] M. Stephan, O. Kohlmann, H. G. Niessen, A. Eichhorn, J. Bargon, Magn. Reson. Chem. 2002, 40,
157.
[
13] L. T. Kuhn, U. Bommerich, J. Bargon, J. Am. Chem. Soc. 2006, 110, 3521.
Received August 14, 2006