Toward Matching Optically and NMR Active Volumes for Optimizing the Observation of Photo-Induced Reactions by NMR

Article abstract of DOI:10.1515/zpch-2016-0854

In the recent years photo-induced reactions are becoming increasingly popular in many fields of chemistry comprising biological conversions, material/environmental science and synthesis. NMR monitoring of such reactions has been shown being advantageous a

Full text of DOI:10.1515/zpch-2016-0854

Z. Phys. Chem. 2016; aop
Eduard Stadler, Anna Eibel, Dmytro Neshchadin
and Georg Gescheidt*
Toward Matching Optically and NMR Active
Volumes for Optimizing the Observation
of Photo-Induced Reactions by NMR
DOI 10.1515/zpch-2016-0854
Received July 13, 2016; accepted September 21, 2016
Abstract: In the recent years photo-induced reactions are becoming increasingly
popular in many fields of chemistry comprising biological conversions, material/
environmental science and synthesis. NMR monitoring of such reactions has been
shown being advantageous and several strategies of providing an efficient irradi-
ation of the NMR sample have been developed and reported. Here we show that
adjusting the optical properties of the investigated solution to the active volume
detected by the NMR experiment is valuable. This is shown with the help of three
examples comprising photo-isomerization, photo-induced polymerization and
CIDNP-detected bond cleavage. Adjusting the photo-active volume to the NMR-
detectable portion of the sample provides a substantially more realistic kinetic infor-
mation, background suppression and reduction of thermal and diffusional effects.
Keywords: CIDNP; NMR; photochemistry; photo-induced reactions.
Dedicated to: Kev Salikhov on the occasion of his 80th birthday.
1 Introduction
Photo-induced chemical reactions have undergone a substantial renaissance
within the very recent time. Photo-catalyzed and photo-triggered reactions
have been established as very valuable protocols for synthesis [1], redox cataly-
sis [2] and materials [3]. There are several approaches for following the course
of the corresponding conversions and establishing products in situ. However,
for monitoring these reactions with the advantage of obtaining a molecular
*Corresponding author: Georg Gescheidt, Institute of Physical and Theoretical Chemistry, TU
Graz, NAWI Graz, Stremayrgasse 9, A-8010 Graz, Austria, E-mail: g.gescheidt-demner@tugraz.at
Eduard Stadler, Anna Eibel and Dmytro Neshchadin: Institute of Physical and Theoretical
Chemistry, TU Graz, NAWI Graz, Stremayrgasse 9, A-8010 Graz, Austria
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢀꢀꢀꢁ
ꢀE. Stadler et al.
2ꢀ
fingerprint together with kinetic information, NMR spectroscopy is the most
valuable method. Accordingly, several experimental setups have been devel-
oped to achieve an efficient monitoring of photo-induced processes inside
NMR spectrometers. Sample illumination is, in most cases, accomplished by
coupling the light source to optical fibers, which are inserted into the NMR
sample tube [4, 5]. A rather extended overview was presented by Kuprov and
Hore [6], who developed a stepwise tapered optical fiber for a uniform illumi-
nation of samples. The photo-isomerizations of azobenzene derivatives have
been investigated using optical fibers [7] and recently different irradiation pro-
cedures were evaluated in terms of following environmental processes [8]. It is
crucial introducing a sufficient light intensity into the NMR sample. However,
one also has to bear in mind that the active volume of the photo-induced reac-
tion does not match the portion of the sample detected by NMR. Indeed, in
many cases, the light gets absorbed in a region very close to the area of irra-
diation and, accordingly, the NMR spectrum becomes substantially contami-
nated by a background stemming from signals of the unreacted/non-irradiated
portion of the sample (Figure 1).
A somehow reversed aspect was followed obtaining optical spectra in situ
inside NMR spectrometers. When the NMR sample possesses a rather high
optical density the optical path of the light has to be made smaller. In such
cases, polytetrafluoroethylene inserts were placed inside NMR tube (lower part)
to produce thin, optically-transparent layers for achieving a short path length
and obtaining reasonable optical spectra; the insert additionally served as a
mirror for a reflection probe [9].
In our projects, we have been performing experiments which require an
efficient irradiation of the sample inside the NMR spectrometer [10–14]. In
particular, photo-CIDNP experiments necessitate sufficient pre-saturation, to
reduce background signals. However, artifacts stemming from resonances from
Fig. 1:ꢀNMR tube and irradiated region (cross section). Left: irradiation source from the side
outside the NMR tube, as used in this work, right: irradiation source inside the NMR tube
e.g. by optical fibers.
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢁꢀꢀꢀ
ꢂ3
Optimizing the Observation of Photo-Induced Reactions by NMRꢂ
non-irradiated regions of the NMR tube may occur. We have, therefore, tested a
rather simple setup, which can be combined with various irradiation sources
(see Figure 1 and reference [6]). It consists of a standard NMR tube with a nar-
rower tube inserted into it (double-tube setup, DTS, Figure 2). The solution that
is investigated occupies the space between the two tubes. The advantage of this
setup is that the diameter of the inserted tube can be adjusted to the optical
density of the reaction solution thus providing that the irradiated volume
matches that detected by NMR. In the present study a quartz rod serves as a light
guide, and the sample is irradiated from the side. We will show the advantages of
this setup on three examples. The usability is not just limited to CIDNP but also to
photo-induced processes followed by NMR. Accordingly, the first example is the
well-described photo-isomerization of azobenzene from its E isomer towards its
Z isomer [7], the second indicates the product formation within a photo-induced
radical polymerization and finally we describe CIDNP spectra taken upon photo-
cleavage of 2-hydroxy-2-methyl-1-phenylpropan-1-one, a commercially-available
photoinitiator (Darocur 1173) for radical polymerization.
2 Results and discussion
2.1 E/Z-Azobenzene
1
The H-NMR spectrum of E-azobenzene (Scheme 1) presents a multiplet around
7.9 ppm (ortho protons), a multiplet at 7.55 ppm (para, meta protons). For the Z
isomer a multiplet around 7.2 ppm (para and meta protons) and a multiplet at
6.8 ppm emerge.
Fig. 2:ꢀ(a) DTS setup (b) Inner tube, lifted sample volume around 150 μL.
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢀꢀꢀꢁ
ꢀE. Stadler et al.
4ꢀ
Scheme 1:ꢀE- and Z-isomers of azobenzene.
Fig. 3:ꢀ1H-NMR monitoring of the photo isomerization of E/Z-azobenzene. Left: Spectra taken
after (a) 30 s, (b) 60 s and (c) 7 min in DTS samples; right: corresponding spectra taken in
conventional NMR tubes.
In Figure 3 we compare the NMR spectra obtained upon irradiation for 30 s,
60 s, and 7 min. Clearly, the “double-tube setup” (DTS) reveals well distinguish-
able conversion after 30 s whereas, the “usual” setup (i.e. conventional NMR tube
instead of the DTS) indicates a 3-times lower conversion (Figure 3, right trace).
These differences can be straightforwardly traced back to the substantially
more pronounced background signal of non-reacted trans-azobenzene, i.e. the
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢁꢀꢀꢀ
ꢂ5
Optimizing the Observation of Photo-Induced Reactions by NMRꢂ
dominating segment of the sample, which is not irradiated. At longer irradiation
periods, the differences between the DTS and the “usual” setup become less pro-
nounced. This can be traced back to diffusional processes and heat transfer.
2.2 Photo-initiated radical polymerization
Photo-initiated radical polymerization has been an exceedingly successful
technique for obtaining rapidly curing surfaces and for 3D structures [15–21].
Such reactions can be monitored by thermal analysis (e.g. photo DSC) and infra-
Fig. 4:ꢀ1H-NMR spectra obtained upon the photo-reaction of butyl acrylate initiated by BAPO
(educts, bottom) leading to a polymer (top). The assignment of the signals follows the labels in
the reaction scheme.
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢀꢀꢀꢁ
ꢀE. Stadler et al.
6ꢀ
red spectroscopy, however only NMR is the method of choice when rather precise
information at a molecular resolution is required. Such investigations have to
be performed at early stages of the polymerization since the products should
be soluble to provide a straightforward NMR fingerprint, which is not deterio-
rated by intense line broadening. It is also important to avoid too exaggerating
irradiation since follow-up reactions of the polymerized material may occur. We
have, therefore chosen following a standard photopolymerization at a seconds-
to-minutes time scale. To this end we have photolyzed a mixture of butyl acrylate
(see Figure 4, bottom) and the bisacylphosphine oxide (phenylphosphoryl)
bis(mesitylmethanone) (BAPO) in acetonitrile. With the extinction coefficient of
the BAPO being 795 L*mol−1*cm−1 (55 mM) ca. 90% of the light is absorbed by
the thin (1 mm) liquid layer. The conversion was monitored using both sample-
tube setups. Figure 4 presents the conversion of the educts (bottom, see the color-
coded assignment of the signals) to the polymer. Here the proton resonance at
3.9 ppm assigned to the two protons adjacent to the oxygen of the ester group
attached to the growing polymer chain and the decaying signals of the alkene
protons between ca. 6.3 and 5.6 ppm are the most prominent and clearly distin-
guishable signals. The integrals of these latter resonances vs. time are presented
in Figure 5. Whereas only an initial conversion can be detected after 5 min in the
usual setup, the DTS indicates basically full conversion. The plateau observed in
the usual setup is probably caused by an initial drop of the signal intensity due
to relaxation effects of photo-initiator radicals. This translates into a half-life (t1/2)
of butyl acrylate being ca. 8 s for the DTS whereas irradiation in the conventional
Fig. 5:ꢀTime profile for the depletion of the NMR signals attributed to Ha,b,c of butyl acrylate
using a conventional NMR tube (blue) and the DTS (red). The insert shows the decay for the DTS
sample at an extended scale.
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢁꢀꢀꢀ
ꢂ7
Optimizing the Observation of Photo-Induced Reactions by NMRꢂ
NMR tube suggests t1/2ꢀ=ꢀ414 s; a factor of 51! (Figure 5) Accordingly the DTS reflects
the kinetics for the formation of the photopolymer to a much more realistic extent.
As for the first example, the particular advantage of DTS is that the photoreaction
is confined to a volume basically compatible with the NMR-active area.
2.3 Photo CIDNP
Photo CIDNP spectroscopy has been a valuable tool for observing rather rapid
photo-triggered chemical reactions [22]. Many investigations have been dedicated
to early events in photopolymerizations particularly, the mechanism and the
follow-up reactions of the initiating radicals (Scheme 2) [23–25]. We have tested the
DTS for the well-established cleavage of 2-hydroxy-2-methyl-1-phenylpropan-1-one.
CIDNP spectroscopy requires an efficient saturation of the NMR spectrum
before irradiation with light produces a radical (ion) pair leading to polarized
NMR resonances (Scheme 3).
As a result, the observed CIDNP intensities may be influenced (beside the well-
known spin dynamics) [22, 26–28] by incomplete saturation and by a substantial
portion of unreacted substrate(s). In Figure 6, we compare the DTS with CIDNP
performed with the use of a conventional NMR tube. The extinction coefficient
of Darocur 1173 in acetonitrile-d3 is 22 L*mol−1*cm−1. This results in absorbance of
90% light intensity in a layer of 1.4 mm at the used concentration of 330 mM. In
our DTS setup, the cross section of the liquid layer was 0.15 mm; accordingly the
light beam passed through the entire solution.
O
O
OH
photoproducts with
polarized spins
hv
+
OH
Scheme 2:ꢀCleavage of 2-hydroxy-2-methyl-1-phenylpropan-1-one.
„waltz16“
FID
1
2
Scheme 3:ꢀCIDNP sequence (for τ1 and τ2, see Experimental part).
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢀꢀꢀꢁ
ꢀE. Stadler et al.
8ꢀ
Fig. 6:ꢀNMR spectrum and CIDNP spectra of 2-hydroxy-2-methyl-1-phenylpropan-1-one upon
photolysis at 355 nm (for details, see Experimental part). The CIDNP and dummy spectra are
depicted on the same scale (compensated for different receiver gains). Squares indicate sub-
stantial differences in the spectra using conventional tubes and the DTS.
The particular advantage of the DTS can be easily distinguished by comparing
the “dummy” experiments with the irradiation source blocked representing the
quality of the background signal (in terms of pre-saturation). Figure 6 shows that
in particular, the signals of the aliphatic protons can hardly be saturated in the
conventional NMR tube at concentrations around 100 mM. However, application
of the DTS markedly decreases the background (Figure 6, bottom) producing better
distinguishable lines in the CIDNP spectrum (Figure 6, bottom left). Our example
represents an ideal case for the CIDNP technique presenting strong polarizations,
but when the polarization pattern gets more complex and polarization patterns
become weaker the quality of the background signal becomes more critical.
3 Summary and conclusions
Our investigations reveal that the “tube in the tube” (DTS) technique presents
a simple but valuable approach of getting NMR information of photo-induced
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢁꢀꢀꢀ
ꢂ9
Optimizing the Observation of Photo-Induced Reactions by NMRꢂ
reactions and their products. When the DTS is used, the active photochemical
volume can be adjusted to the NMR-detection volume. This procedure clearly
does not increase the signal-to-noise ratio of the NMR signal but leads to a
substantial decrease of background effects. Our kinetic investigation follow-
ing the photo-induced radical polymerization of butyl acrylate well illustrates
this effect. This advantage is accompanied by the fact that a substantially
shorter time of irradiation is necessary for clearly distinguishing NMR reso-
nances of photoproducts since they are less masked by those of the educts.
Moreover since less energy has to be transferred to the sample, thermal effects
are basically undetectable. The chemical advantage of this approach is that
it becomes less likely the primary products undergo undesired follow-up
reactions.
We have tested our setup using a wide-bore probe head and irradiation from
one side (Figure 1), which is of advantage, since the sample can be rotated during
the experiment. However, the DTS can easily be applied also in spectrometers
with a standard probe head with inserted fibers of any shape. The DTS could also
be used for implementing UV-VIS detection in NMR spectroscopy; if the inner
tube is coated or replaced by a material that reflects light into an optical fiber that
is connected to a detector as developed by Tolstoy et al. [9]. This approach would
also ensure the matching of the UV-VIS and NMR detected volumes. Generally
this setup allows adjusting the size of the inner tube to the optical density of the
investigated solution providing a close to ideal optical path for the irradiation
experiment.
4 Experimental
4.1 Materials
Azobenzene was obtained from Sigma Aldrich (purity 98%) and was used without
further purification. Twenty millimolar samples of azobenzene in acetonitrile-d3
(Euriso-top, used as received) were prepared. Darocur® 1173 was purchased from
MERCK and used as delivered. BAPO (commericialized as Irgacure 819, Ciba) and
butyl acrylate (Fluka, filtered with aluminum oxide powder shortly before use)
were used as obtained. NMR measurements were performed either in conven-
tional NMR tubes or in the double tube setup. In the double tube setup, an EPR
tube filled with the solvent (acetonitrile-d3) is placed inside the NMR tube, which
contains the sample.
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢀꢀꢀꢁ
ꢀE. Stadler et al.
10ꢀ
4.2 Azobenzene
1H NMR spectra were recorded on a 200 MHz Bruker AVANCE DPX spectrometer.
A Hg−Xe UV lamp (Hamamatsu Lightningcure LC4, 3500 mW/cm2, λmaxꢀ=ꢀ365 nm)
1
was guided into the NMR spectrometer using a quartz rod. 1D H NMR experi-
ments (32 scans) were performed after different irradiation times (between 30 s
and 7 min). Chemical shifts (δ) are reported in ppm relative to tetramethylsilane
(TMS) using the residual deuterated solvent signals as an internal reference (ace-
tonitrile-d3, δHꢀ=ꢀ1.94 ppm).
4.3 Polymerization experiments
Butylacrylate was pushed through a syringe filled with AlO powder to remove any
stabilizers prior to the photoexposure. A solution of 3ꢀ+ꢀ7 (vꢀ+ꢀv) of butylacraylate
and acetonitrile-d3 containing BAPO (55 mM) was prepared. All samples were
measured under atmospheric conditions, i.e. were not degassed.
The inner tube was filled with acetone-d6 for better shimming and for the
spectrometer lock. The NMR tube was irradiated during the NMR experiments
continuously with a low power LED (2 mW) with an emitting wavelength of 365 nm
operated at a current of 30 mA. The LED was put in direct contact with the quartz
rod. (Acquisition and processing parameters: NMR tube: One spectrum was
recorded every 20 s. Parameters: aqꢀ=ꢀ2 s, p1ꢀ=ꢀ0.33*4.2, rgꢀ=ꢀ32, data pointsꢀ=ꢀ4 k,
filled to 8 k. DTS: One spectrum was recorded every 2.5 s. Parameters: aqꢀ=ꢀ2 s,
p1ꢀ=ꢀ0.33*4.2, rgꢀ=ꢀ400, data pointsꢀ=ꢀ4 k, filled to 8 k.)
For obtaining the kinetic traces the pseudo 2D spectra were displayed
in absolute value mode and integrated from 5.5 to 6.5 ppm (protons of acryl
moiety, 3 H).
4.4 CIDNP of 2-hydroxy-2-methyl-1-phenylpropan-1-one
The concentration was 1ꢀ+ꢀ19 (vꢀ+ꢀv, photoinitiatorꢀ+ꢀacetonitrile-d3,ꢀ≈ꢀ0.33 M). For
the conventional setup the sample volume is 700 μL, for the inner tubing system
(DTS) the needed volume is 150 μL. The extinction coefficient of Darocur 1173 at
the wave lenght of the used laser (355 nm) was determined to be 22 L*mol−1*cm−1.
(Acquisition and processing parameters: presaturation periodꢀ=ꢀ15 s, presatura-
tion pulse power 42 dB, pulse lengthꢀ=ꢀ440 μs, nsꢀ=ꢀ1, rgꢀ=ꢀ20 (normal NMR tube)
and rgꢀ=ꢀ400 (DTS), p1ꢀ=ꢀ4.1 μs, atꢀ=ꢀ3.42 s, data points: 8 k, zero filled to 16 k, added
linebroadening LBꢀ=ꢀ1 Hz. All spectra were baseline corrected. Delay between
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢁꢀꢀꢀ
Optimizing the Observation of Photo-Induced Reactions by NMRꢂ
ꢂ11
presaturation period and laser flash t1ꢀ=ꢀ252 μs, delay between laser flash and rf-
pulse t2ꢀ=ꢀ70 μs).
4.5 UV-VIS spectroscopy
UV-Vis spectroscopy was performed using a Shimadzu UV-3101PC UV-Vis-NIR
scanning spectrophotometer. The concentration of azobenzene was 1 mM in ace-
tonitrile. The measurement was performed at ambient temperature.
References
1. E. Meggers, Chem. Commun. 51 (2015) 3290.
2. X. Lang, J. Zhao, X. Chen, Chem. Soc. Rev. 45 (2016) 3026.
3. N. Corrigan, S. Shanmugam, C. Boyer, Chem. Soc. Rev. (2016). DOI 10.1039/C6CS00185H.
4. J. E. Scheffler, C. E. Cottrell, L. J. Berliner, J. Magn. Reson. 63 (1985) 199.
5. C. Feldmeier, H. Bartling, E. Riedle, R. M. Gschwind, J. Magn. Reson. 232 (2013) 39.
6. I. Kuprov, P. J. Hore, J. Magn. Reson. 171 (2004) 171.
7. K. M. Tait, J. A. Parkinson, S. P. Bates, W. J. Ebenezer, A. C. Jones, J. Photochem. Photobiol. A
Chem. 154 (2003) 179.
8. L. Bliumkin, R. D. Majumdar, R. Soong, A. Adamo, J. P. D. Abbatt, R. Zhao, E. Reiner, A. J.
Simpson, Environm. Sci. Technol. 50 (2016) 5506.
9. P. M. Tolstoy, B. Koeppe, G. S. Denisov, H. H. Limbach, Angew. Chem. Int. Ed. 48 (2009)
5745.
10. D. Neshchadin, G. Gescheidt, J. Phys. Org. Chem. 26 (2013) 737.
11. D. Neshchadin, F. Palumbo, M. S. Sinicropi, I. Andreu, G. Gescheidt, M. A. Miranda, Chem.
Sci. 4 (2013) 1608.
12. S. Rahimipour, G. Gescheidt, I. Bilkis, M. Fridkin, L. Weiner, Appl. Magn. Reson. 37 (2010)
629.
13. D. Neshchadin, S. N. Batchelor, I. Bilkis, G. Gescheidt, Angew. Chem. Int. Ed. Engl. 53
(2014) 13288.
14. D. Neshchadin, A. Rosspeintner, M. Griesser, B. Lang, S. Mosquera-Vazquez, E.
Vauthey, V. Gorelik, R. Liska, C. Hametner, B. Ganster, R. Saf, N. Moszner, G. Gescheidt,
J. Am. Chem. Soc. 135 (2013) 17314.
15. N. Moszner, F. Zeuner, I. Lamparth, U. K. Fischer, Macromol. Mater. Eng. 294 (2009) 877.
16. Y. Yagci, S. Jockusch, N. J. Turro, Macromolecules 43 (2010) 6245.
17. K. Dietliker, T. Jung, J. Benkhoff, H. Kura, A. Matsumoto, H. Oka, D. Hristova, G. Gescheidt,
G. Rist, Macromol. Symp. 217 (2004) 77.
18. G. Mueller, M. Zalibera, G. Gescheidt, A. Rosenthal, G. Santiso-Quinones, K. Dietliker, H.
Gruetzmacher, Macromol. Rapid Commun. 36 (2015) 553.
19. K. Dietliker, A. Braig, A. Ricci, Photochemistry 38 (2010) 344.
20. P. Derboven, D. R. D’hooge, M.-F. Reyniers, G. B. Marin, C. Barner-Kowollik, Macromolecules
48 (2015) 492.
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM

ꢀꢀꢀꢁ
ꢀE. Stadler et al.
12ꢀ
21. J. Kumpfmueller, K. Stadlmann, Z. Li, V. Satzinger, J. Stampfl, R. Liska, Des. Monomers
Polym. 17 (2014) 390.
22. M. Goez, Annu. Rep. NMR Spectrosc. 66 (2009) 77.
23. D. Hristova, I. Gatlik, G. Rist, K. Dietliker, J. P. Wolf, J. L. Birbaum, A. Savitsky, K. Mobius, G.
Gescheidt, Macromolecules 38 (2005) 7714.
24. A. Borer, R. Kirchmayr, G. Rist, Helv. Chim. Acta 61 (1978) 305.
25. U. Kolczak, G. Rist, K. Dietliker, J. Wirz, J. Am. Chem. Soc. 118 (1996) 6477.
26. G. L. Closs, Adv. Magn. Reson. 7 (1974) 157.
27. J. Bargon, H. Fischer, U. Johnsen, Z. Naturforsch. A 22 (1967) 1551.
28. J. B. Pedersen, Theories of Chemically Induced Magnetic Polarization, Odense University
Press, Odense (1979).
Brought to you by | Cornell University Library
Authenticated
Download Date | 10/18/16 11:10 AM