Real-time monitoring of tritium gas reactions using Raman spectroscopy

Article abstract of DOI:10.1002/jlcr.887

The utility of Raman spectroscopy for noninvasive, real-time monitoring of a range of tritium gas labeling reactions has been investigated, using deuterium gas as a model in most cases. Reaction types include organoiridium-catalyzed heteroatom-directed ex

Full text of DOI:10.1002/jlcr.887

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2004; 47: 983–995.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.887
Research Article
Real-time monitoring of tritium gas reactions using
Raman spectroscopy
J. Richard Heys1,*, Mark E. Powell1 and Don E. Pivonka2
1 Isotope Chemistry, CNS Chemistry, AstraZeneca Pharmaceuticals, 1800 Concord
Pike, Wilmington, DE 19850, USA
2 Vibrational Spectroscopy, CNS Chemistry, AstraZeneca Pharmaceuticals,
1800 Concord Pike, Wilmington, DE 19850, USA
Summary
The utility of Raman spectroscopy for noninvasive, real-time monitoring of a range of
tritium gas labeling reactions has been investigated, using deuterium gas as a model in
most cases. Reaction types include organoiridium-catalyzed heteroatom-directed
exchange (HDE), olefin hydrogenation and catalytic aryl dehalogenation. Five
examples of HDE reactions with several different substrate types were monitored by
observation of Raman vibrational bands sensitive to the isotopic substitutions.
Changes in peak intensities and/or frequencies associated with the course of labeling
are clearly observable at concentrations and reaction scales typical of tritium gas
reactions. Similarly, Raman bands sensitive to the chemical changes that occur during
catalytic deuterogenation of an olefin and to catalytic deuterium–bromine exchange of
an aryl bromide were successfully monitored. This methodology can provide
unprecedented real-time information, otherwise difficult to obtain, over the course
of such reactions. Copyright # 2004 John Wiley & Sons, Ltd.
Key Words: Raman spectroscopy; reaction monitoring; heteroatom-directed ex-
change; tritium–halogen replacement; deuteration; tritiation
Introduction
A major impediment to rapid development and optimization of tritium gas
reactions stems from the limited options available for monitoring their
progress. Such reactions generally use small quantities (20–200 mmol) of
tritium gas at subatmospheric pressure, in small vessels with the minimum
practical headspace. Investigators rarely monitor ongoing reactions by
physical sampling of the reaction mixture because of radiation safety issues
*Correspondence to: J. R. Heys, Isotope Chemistry, CNS Chemistry, AstraZeneca Pharmaceuticals, 1800
Concord Pike, Wilmington, DE 19850, USA. E-mail: richard.heys@astrazeneca.com
Copyright # 2004 John Wiley & Sons, Ltd.
Received 26 July 2004
Revised 11 August 2004
Accepted 16 August 2004

984
J. R. HEYS ET AL.
and the risk of perturbing the reaction’s atmosphere. In addition, incorpora-
tion of a septum sidearm into the design of a reaction vessel inevitably
increases the reaction headspace, leading to the need to use either larger
amounts of tritium or lower pressures. To date, there has been no published
method for noninvasive monitoring of such reactions to provide real-time
functional group selectivity. In lieu of real-time monitoring, it is common
practice either to conduct deuterium model reactions before embarking on a
tritiation, or to run tritiations for extended periods of time against the
possibility of slow reaction rates. Such practices require multiple experiments,
increase the risk of failed tritiations, and/or generate excessive byproducts or
radiation-induced decomposition from unnecessarily long reaction times.
It would be desirable to monitor pathways and progress for these kinds of
reactions without the need to physically contact the reaction mixtures. Raman
spectroscopy provides one method by which functionally specific data can be
obtained from a reaction mixture without the need for invasive sampling.1 The
ability to conveniently obtain spectra by aiming a small portable laser device
at a reaction vessel and detecting the scattered radiation renders in situ
reaction analysis practical in a tritium gas handling facility. A number of
functional groups often involved with tritium gas reactions are known to have
sufficiently sensitive Raman bands2 to enable observation of the spectral
features in relatively low concentration solutions commonly associated with
tritium based reactions. Olefin, acetylene, conjugated carbonyl, aromatic ring,
aliphatic C–H, C–D and C–T vibrations are examples of those suited to
investigation by Raman analysis.3
The present work demonstrates the successful application of Raman
spectroscopy for the noninvasive, real-time analysis of several common types
of tritium gas reactions, and the use of various modes of data presentation and
analysis. For reasons of cost and safety, most of the described reactions were
conducted using deuterium as a surrogate; however, several tritium reactions
were conducted to demonstrate that the methodology applies also to this
isotope.
Results and discussion
Raman monitoring of heteroatom directed exchange
The catalytic exchange of hydrogen in unactivated C–H bonds4 with isotopic
hydrogen gas has proven to be an important addition to the repertoire of
tritium (and deuterium) labeling methods. Typical of this class of reactions
is the conversion shown in Scheme 1. Under suitable organoiridium
catalysis, one or more aryl hydrogens ortho to an activating function
are replaced by tritium or deuterium, leaving the gross chemical structure
unchanged.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

REAL-TIME MONITORING
985
D,T
Y
X
H
H
Y
X
General conditions:
[(cod)IrL1L2]+,
<1 atm D2 or T2,
Y
Y
D,T
X
z
X
z
CH2Cl2, rt
Scheme 1.
Published literature detailing the effects of deuterium substitution on the
Raman spectra of arenes is limited yet consistant with theory by illustrating
that the additional mass of the deuterium vs hydrogen atom results in an
attenuation of the observed vibrational frequency. For example, the 1002 cmꢀ1
aryl ring band of benzophenone is shifted to 960 cmꢀ1 in perdeuterobenzo-
phenone, and the 1599 cmꢀ1 peak to 1568 cmꢀ1;5 the imidazole ring vibration
of 9-b-d-arabinofuranosyladenine at 1518 cmꢀ1 is shifted to 1488 cmꢀ1 upon
deuteration of the purine at C8.6
As a preliminary feasibility study, Raman spectra of 3,5-difluorobenzoyl
pyrrolidine (1a) and 2,6-dideutero-3,5-difluorobenzoyl pyrrolidine (1b) were
collected from methylene chloride solutions. An aromatic ring band of
compound 1a was observed at 1000 cmꢀ1. For compound 1b this band was
shifted to 973 cmꢀ1. At an instrumental resolution of 4 cmꢀ1, these peaks were
baseline resolved and exhibited no interfering bands from either solvent or
other vibrational modes of the compound itself. Results of this experiment
further reinforced the potential utility of Raman spectroscopy for monitoring
the course of an ortho-deuteration or tritiation reaction of 1a to 1b.
R
O
F
N
1a: R = H
1b: R = D
R
F
An HDE reaction of 1a (12.5 mg in 0.5 ml CH2Cl2, 3.9 mol% of
[(cod)Ir(PPh3)2]BF4, 3.5 equiv of D2) was run in a cylindrical reaction vessel
(5 mm ID ꢁ 7 cm). The vessel was attached to the gas manifold by a Swagelok
fitting and equipped with a small magnetic spinvane for stirring. The charged
reaction vessel contained a residual headspace volume of approximately 5 ml.
The laser was focused through the vessel wall into the solution about 5 mm
above the spinvane, to avoid disturbance of the beam as the solution was
stirred.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

986
J. R. HEYS ET AL.
Figure 1. Waterfall view of the deuteration of 3,5-difluorobenzoylpyrrolidine
Spectra were collected at 3 min intervals throughout the reaction to provide
the real-time feedback illustrated in the 1070-940 cmꢀ1 waterfall plot of
Figure 1. The progress of the reaction is clearly apparent, as the 1000 cmꢀ1
peak of 1a diminished and was replaced by that of 1b at 973 cmꢀ1 over the
course of about 2.5 h. A peak at 990 cmꢀ1 was observed to appear and then
diminish through the course of the reaction. This peak has been assigned to the
monodeutero intermediate. The reaction was worked up after about 2.5 h.
LCMS analysis of the product showed the isotopic distribution to be 4%d0,
32%d1, 64%d2, consistent with the extent of reaction indicated by the spectral
record.
A tritiation of 1a at much lower substrate concentration (2 mg 1a in 0.5 ml
CH2Cl2, 15 mol% [(cod)Ir(PPh3)2]BF4, 2.3 equiv. T2 at partial pressure of
about 100 mbar) was conducted using a vessely designed for more efficient gas/
liquid exchange; the laser was focused on the solution inside the lower
projection. Analogous to the deuterium experiment, diminution of
the1000 cmꢀ1 peak was observed throughout the reaction (Figure 2, shown
in overlay mode). Growth of peaks at 968 cmꢀ1 and 946 cmꢀ1 was attributed
to 2-tritio-1a and 2,6-ditritio-1a, respectively. The faster kinetics of the
tritiation compared to the previous reaction are attributed to the higher
catalyst loading and the more efficient gas/liquid exchange in this reaction
y
0.5 mL fill level
3 mm ID
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

REAL-TIME MONITORING
987
Figure 2. Overlay view of the tritiation of 3,5-difluorobenzoylpyrrolidine
1.7 mol% cat. 4.1 mol% cat.
3000
starting material
2500
2000
monodeuterated
intermediate
dideuterated
product
1500
1000
500
0
0
5
10
15
20
Time (hours)
Figure 3. Graphic view of the deuteration of 2-phenylimidazole
vessel. Spectral changes had come to a stop within 30 min, indicating cessation
of the reaction. Consistent with the Raman data, the isotope distribution of
[3H]1 as measured by LCMS was 6% t0, 37% t1 and 57% t2.
The deuteration reactions of several other model compounds were
monitored by Raman, with similar results. Deuteration of 2-phenylimidazole
(2) with [(cod)Ir(PPh3)2]BF4 resulted in consumption of a 1001 cmꢀ1 aromatic
band and generation of new peaks at 986 and 974 cmꢀ1 which were attributed
to 2-(2-deuterophenyl)imidazole and 2-(2,6-dideuterophenyl)imidazole, re-
spectively. The reaction, initiated with 1.7 mol% catalyst, progressed slowly,
with no detectable dideuterated product observed after 3 h as illustrated in
Figure 3 (shown in graphic mode). Addition of 2.4 mol% of catalyst at 3.1 h
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

988
J. R. HEYS ET AL.
significantly accelerated the reaction, which asymptotically approached
equilibrium over the next 17 h.
Deuteration of 2-phenylpyridine was monitored by observation of the peaks
at 1001 cmꢀ1 (starting material), 990 cmꢀ1 (20-deutero) and 972 cmꢀ1 (20,60-
dideutero). Likewise, labeling of 3 in the benzoyl ring using [(cod)
Ir(PPh3)2]BF4 or [(cod)Ir(PCy3)(py)]PF6 could be monitored (unlabeled,
1001 cmꢀ1; 2-deutero, 988 cmꢀ1; 2,6-dideutero, 975 cmꢀ1).
OMe
O
N
O
N
N
H
Me
Cl
3
4
2
The Raman spectrum of 4-chlorophenyl cyclopropyl ketone (4) includes a
peak at 1592 cmꢀ1 which is sensitive to deuteration in the ring. As illustrated in
Figure 4, the ortho-deuteration of this substrate is indicated by diminution of
the 1592 peak and development of signals at 1585 cmꢀ1 (monodeutero) and
1577 cmꢀ1 (dideutero).
Raman monitoring of heterogeneously catalyzed tritiations
Many tritiation reactions utilize heterogeneous catalysts, such as palladium on
carbon, in stirred solutions of substrates in organic solvents. Typical of such
reactions are tritium–halogen exchange reactions of aryl halides and
tritiogenation of olefins and acetylenes. The experiments described below
show that the presence of suspended Pd/C does not seriously impede the
Figure 4. Overlay view of the deuteration of 4-chlorophenyl cyclopropyl ketone
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

REAL-TIME MONITORING
989
ability to acquire a Raman signal, so that Raman spectroscopy can be used to
follow the course of such reactions.
The first reaction investigated under heterogeneous catalyst conditions was
the catalytic deuterogenation of methyl cinnamate. The Raman spectrum of
methyl cinnamate includes two pairs of peaks, one pair at 1640 cmꢀ1 (s) and
1602 cmꢀ1 (m) and the other pair at 1205 cmꢀ1 (s) and 1183 cmꢀ1 (m). These
peaks are absent in the spectrum methyl 3-phenylpropionate. As illustrated in
Figure 5, the deuterogenation of methyl cinnamate could be monitored by
observing the disappearance of the two pairs of peaks. All other peaks in the
spectrum remained unchanged, indicating that the disappearance of these
peaks correlated with a chemical change in the substrate. Figure 5 data
indicate that the reaction was complete within 30 min. Analysis of the product
showed clean conversion to methyl phenyl[2,3-2H2]propionate.
The catalytic reduction of N-acetyl-4-bromophenylalanine with deuterium
gas was investigated as a representative of the tritium–halogen replacement
class of reactions. The Raman spectrum of acetylphenylalanine includes an
important peak at 1003 cmꢀ1 which is absent in N-acetyl-40-bromophenyla-
lanine. Raman monitoring of the deuterium-bromine exchange reaction
(Figure 6) shows the progress from starting material to product, as a peak
emerges in the region of interest. Significantly, the product peak is at 987 cmꢀ1
rather than at 1003 cmꢀ1, the frequency of unlabeled N-acetylphenylalanine.
This shift is attributed to the presence of deuterium at C4 of the ring. NMR
and mass spectrometric analyses were consistent with the expected product, N-
acetyl-40-deuterophenylalanine.
Mechanistic investigation
HDE of unsymmetrically substituted substrates offers the possibility that two
different monodeutero isomers may be produced. Although this may be of
little interest in a tritiation where the goal is to maximize tritium
incorporation, the ability to measure the relative rates of exchange at different
Figure 5. Staggered overlay views of methyl cinnamate deuterogenation
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

990
J. R. HEYS ET AL.
sites could provide mechanistic information about the labeling process.
In order to investigate the consequences of this for Raman monitoring of
HDE, we selected 3-methoxybenzoyl pyrrolidine (5). Authentic samples
of 2-deutero-3-methoxybenzoyl pyrrolidine (5a) (85% deuterated), 6-deutero-
3-methoxybenzoyl pyrrolidine (5b) (82% C6-d, 12% d2) and 2,6-dideutero-3-
methoxybenzoyl pyrrolidine (5c) (75% d2, 19% d1), were separately prepared
as aids to spectral assignment. Relevant regions of their Raman spectra are
shown in Figure 7. Monodeutero and dideutero isotopomers of 5 are easily
differentiated from unlabeled 5 in the ꢂ1000 cmꢀ1 region, but the two
monodeutero isomers 5a and 5b are indistinguishable in this region, as their
peak maxima differ only by about 1 cmꢀ1. As shown in Figure 7, the spectra of
5a and 5b do differ enough at lower wave numbers that they could be
Figure 6. Staggered overlay view of N-acetyl-40-bromophenylalanine deuter-
odebromination
Figure 7. Isotope-sensitive bands of 3-methoxybenzoylpyrrolidine isotopomers
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

REAL-TIME MONITORING
991
distinguished under favorable conditions. Unfortunately, these peaks were not
very distinct from solvent background, so we were not able to distinguish
during an actual labeling experiment between the rates of deuteration at C2
and at C6 in this substrate. Nevertheless, the principle has been demonstrated.
R1
O
5: R1 = R2 = H
5a: R1 = D; R2 = H
5b: R1 = H; R2 = D
5c: R1 = R2 = D
MeO
N
R2
Catalysts
In control experiments, Raman observations were made during the treatment
of HDE catalysts with hydrogen or deuterium in the absence of substrate, in
order to investigate the potential for interfering Raman spectral activity. This
knowledge is important because high HDE catalyst loadings are sometimes
used. Stirred solutions of [(cod)Ir(PPh3)2]BF4 and [(cod)Ir(PCy3)(py)]PF6 were
exposed with stirring to hydrogen or deuterium gas, and their Raman spectra
were monitored for at least 12 h, over which period the organometallic
complexes are first converted to active catalytic species, then to catalyst
decomposition products analogous to those likely formed during actual
labeling experiments. The spectrum of [(cod)Ir(PCy3)(py)]PF6 includes a weak
peak at 1015 cmꢀ1. During the experiments, new peaks grew in at 990 and
1020 cmꢀ1 with hydrogen treatment, and at 996 and 1019 cm–1 with deuterium
treatment. Different changes may be produced upon treatment with tritium.
Such changes could make it difficult to distinguish spectral changes associated
with substrate, especially in those cases when high catalyst loadings are used.
The spectrum of [(cod)Ir(PPh3)2]BF4 includes a moderate-intensity peak at
1002 cmꢀ1 but it remained unchanged throughout the experiment, so that with
this catalyst any spectral changes observed in this region can more safely be
attributed to substrate.
Raman monitoring of experiments using (cod)Ir(1,1,1,5,5,5-hexafluoropen-
tane-2,4-dionate)7 to catalyze the deuterium exchange labeling of benzylamine
and of 3-methoxy-5-trifluoromethylaniline (DMF, D2) were unsuccessful
because the Raman signal weakened drastically as the reactions turned black
with precipitated or colloidal iridium. These results contrasted with the ability
to obtain good signals throughout the course of reactions containing Pd/C.
Solvents
Several solvents were profiled for their utility in Raman-monitoring of isotopic
labeling reactions. Methylene chloride is a highly suitable solvent since it lacks
peaks in the regions 900–1100 and 1500–1700 cmꢀ1, where important substrate
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

992
J. R. HEYS ET AL.
signals often appear. Ethanol has interference-free windows 905–1025 and
above 1545 cmꢀ1. DMF displays peaks at 1011 cmꢀ1 (weak) and 1660 cmꢀ1
(moderate); DMA 960, 1020 and 1636 cmꢀ1 (all weak-moderate). These bands
did not interfere with observation of the aryl breathing mode vibrations of
undeuterated (1002 cmꢀ1), 2-deutero (989 cmꢀ1) and 2,6-dideutero (975 cmꢀ1
)
isotopomers of benzylamine at concentrations of 24 mg/ml. The spectrum of
ethyl acetate contains several peaks in the 900–1100 cmꢀ1 region which might
prevent such observations, but its carbonyl stretch at 1737 cmꢀ1 (medium)
provides an alternative to the preceding solvents for use with substrates which
have peaks of interest in the 1600–1700 cmꢀ1 range. Most solvents lack peaks
in the alkyne region (2100–2250 cmꢀ1) where strong substrate Raman signals
may be used to monitor reactions such as Lindlar reductions. C–D and C–T
vibrations stretching vibrations also occur near this region.
Conclusion
The results of initial studies demonstrate the successful application of Raman
spectroscopy to noninvasive, real-time analysis of several different classes of
labeling reactions in which tritium gas is normally used as the isotope source.
These results indicate that Raman can be a useful tool for observing the
progress of reactions, including disappearance of starting materials, formation
of products, and in suitable cases the formation and consumption of
isotopomeric intermediates. Further investigations continue into the applica-
tions of Raman spectroscopy to isotope labeling chemistry.
Experimental
Raman spectra were obtained on a Kaiser Optical Systems RXN1 spectro-
meter operating at 785 nm. Spectra were acquired using the integral software
package HoloGRAMS 4.0 and processed with HoloReact in Matlab 6.5.
Further software analysis of spectra was carried out using Nicolet OMNIC 6.0.
Ambient room light was reduced and reactions shielded with a drape in order to
minimize the appearance of spurious peaks in the spectra. Data collection
parameters typically included 3-s acquisitions with 20 acquisitions averaged/
spectrum. Deuterium exchange reactions were conducted using standard
glassware attached to a glass vacuum manifold for manipulation of the reaction
atmosphere. Tritiations and reactions modeling tritiations were conducted in
specially fabricated glass vessels attached, via a Swagelok fitting, to a stainless
steel manifold system (RC TRITEC AG, Teufen, Switzerland). Unless
otherwise specified, all solvents were commercially available anhydrous reagent
grade and all chemicals were ACS reagent grade or better from Aldrich.
3,5-Difluorobenzoyl pyrrolidine (1a). Pyrrolidine (2.83 ml, 34 mmol) was
dissolved in diethyl ether (50 ml) with stirring in a round bottom flask under
argon. 2,5-Difluorobenzoyl chloride (2.0 ml, 16 mmol) was added dropwise via
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

REAL-TIME MONITORING
993
syringe, resulting in the formation of a white precipitate. The reaction mixture
was stirred for 5 h then filtered to remove the white precipitate. The filtrate was
washed with water (1 ꢁ 25 ml), 1N HCl (2 ꢁ 25 ml), H2O (1 ꢁ 25 ml), and
saturated NaHCO3 (2 ꢁ 25 ml) and then was dried with MgSO4. The solvent
was evaporated by rotary evaporation and the residue dried in vacuo (1 Torr,
1
rt, 30 min) to give 2.85 g (80% yield) of 1a as a colorless oil. H NMR
(300.132 MHz, DMSO-d6) d 7.34 (dt, J ¼ 10:4, 2.3 Hz, 1 H), 7.25 (dd, J ¼ 4:8,
2.1 Hz, 2 H), 3.43 (m, 4 H), 1.84 (m, 4 H); MS (ES+) m/Z 212.
3,5-Difluoro[2,6-2H2]benzoyl pyrrolidine (1b). A 100 mg portion of 1a and
10 mg of [(cod) Ir(PPh3)2]BF4 were dissolved in 10 ml CH2Cl2 in a 50 ml round
bottom flask containing a magnetic stir bar. The solution was frozen using a
liquid nitrogen cooling bath, the flask was evacuated, and 1 atm of D2 was
introduced. The flask was allowed to warm to room temperature as expansion
gases were allowed to escape through a bubbler, and the mixture was stirred
for 16 h. The solvent was evaporated in vacuo and the residue was triturated
with diethyl ether (2 ꢁ 5 ml), the supernatant was filtered through a syringetip
PFTE membrane filter, and the filtrate evaporated in vacuo to provide 85 mg of
1
colorless oil. H NMR (300.132 MHz, DMSO-d6) d 7.34 (t, J ¼ 9:3 Hz, 1 H),
7.25 (m, trace), 3.43 (m, 4 H), 1.84 (m, 4 H); MS (ES+, uncorr) m/Z 213 (9%),
214 (100%), 215 (14%).
N-(4-Methoxyphenyl)-N-methyl-benzamide (3). (4-Methoxyphenyl)methyla-
mine (1 g, 7.3 mmol) was dissolved in anhydrous ether under argon in a 50 ml
round bottom flask equipped with a stir bar. Benzoyl chloride (0.43 ml,
3.6 mmol) was added dropwise via a syringe, resulting in the formation of a
white precipitate. The reaction mixture was stirred for 1 h and then filtered to
remove the white precipitate. The filtrate was washed with 1N HCl (3 ꢁ 20 ml),
H2O (2 ꢁ 20 ml), and saturated NaHCO3 (3 ꢁ 20 ml) and was dried with
MgSO4. The solvent was evaporated to give 0.83 g of solid. 1H NMR
(300.132 MHz, CDCl3) d 7.22 (m, 5 H), 6.95 (d, J ¼ 8:7 Hz, 2 H), 6.73 (d,
J ¼ 8:7 Hz, 2 H), 3.74 (s, 3 H), 3.45 (s, 3 H); MS (ES+) m/Z 242.
3-Methoxybenzoyl pyrrolidine (5). Pyrrolidine (1.07 ml, 12.9 mmol) was
dissolved in diethyl ether (50 ml) with stirring in a round bottom flask under
argon. 3-Methoxylbenzoyl chloride (1.0 g, 5.8 mmol) was added dropwise via
syringe, resulting in the formation of a white precipitate. The reaction mixture
was stirred for 1 h and then filtered to remove the white precipitate. The filtrate
was washed with water (1 ꢁ 20 ml), 1 N HCl (2 ꢁ 20 ml), H2O (1 ꢁ 20 ml), and
saturated NaHCO3 (2 ꢁ 20 ml) and was dried with MgSO4. The solvent was
evaporated. The product was purified by flash chromatography (0–5%
1
methanol in dichloromethane over 15 min) to give 0.38 g of colorless oil. H
NMR (300.132 MHz, C6D6) d 7.36 (d, J ¼ 0:9 Hz, 1 H), 7.16 (t, J ¼ 7:8 Hz,
1 H), 7.25 (t, J ¼ 7:8 Hz, 1 H), 6.91 (dt, J ¼ 7:8, 0.9 Hz, 1 H), 3.66 (m, 2 H),
3.39 (s, 3 H), 3.05 (m, 2 H), 1.33 (m, 4 H); MS (ES+) m/Z 206.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

994
J. R. HEYS ET AL.
2-Deutero-3-methoxylbenzoyl pyrrolidine (5a). A round bottom flask
equipped with a stir bar was charged with 5 (0.1 g, 0.488 mmol), THF
(2.5 ml), and TMEDA (81 ml, 0.536 mmol) under argon. This solution was
cooled to –788C. Sec-butyllithium (0.38 ml, 0.536 mmol) was added dropwise
via syringe and the solution turned bright yellow. The reaction mixture was
stirred 30 min, EtOD (287 ml, 0.488 mmol) was added dropwise, and the yellow
color disappeared. The reaction was allowed to warm to room temperature.
The solution was evaporated and the resulting residue was dissolved in ether,
washed with H2O (3 ꢁ 5 ml), and dried with MgSO4. The ether was evaporated
and the product purified by flash chromatography (0–5% methanol in
dichloromethane over 15 min) to give 0.07 g of colorless oil. 1H NMR
(300.132 MHz, C6D6) d 7.24 (dd, J ¼ 7:5, 0.8 Hz, 1 H), 7.16 (t, J ¼ 7:8 Hz,
1 H), 6.92 (dd, J ¼ 7:5, 0.8 Hz, 1 H), 3.66 (m, 2 H), 3.39 (s, 3 H), 3.05 (m, 2 H),
1.33 (m, 4 H); MS (ES+, uncorr) m/Z 206 (15%), 207 (100%), 208 (12%).
6-Deutero-3-methoxylbenzoyl pyrrolidine (5b). A round bottom flask
equipped with a stir bar was charged with 5c (0.1 g, 0.488 mmol), THF
(2.5 ml), and TMEDA (81 ml, 0.536 mmol) under argon. This solution was
cooled to –788C. Sec-butyllithium (0.38 ml, 0.536 mmol) was added dropwise
via syringe and the solution turned bright yellow. The reaction mixture was
stirred 30 min, EtOH (287 ml, 0.488 mmol) was added dropwise, and the yellow
color disappeared. The reaction was allowed to warm to room temperature.
The solution was evaporated and the resulting residue was dissolved in ether,
washed with H2O (3 ꢁ 5 ml), and dried with MgSO4. After removing the ether,
the product was purified by flash chromatography (0–5% methanol in
dichloromethane over 15 min) to give 0.06 g of colorless oil. 1H NMR
(300.132 MHz, C6D6) d 7.37 (d, J ¼ 2:6 Hz, 1 H), 7.16 (d, J ¼ 8:2 Hz, 1 H),
6.92 (dd, J ¼ 8:2, 2.6 Hz, 1 H), 3.66 (m, 2 H), 3.39 (s, 3 H), 3.05 (m, 2 H), 1.33
(m, 4 H); MS (ES+, uncorr) m/Z 206 (7%), 207 (100%), 208 (28%).
2,6-Dideutero-3-methoxybenzoyl pyrrolidine (5c). A 250 ml round bottom
flask equipped with a stir bar was charged with 5 (0.5 g, 2.44 mmol), Crabtree’s
catalyst (39 mg, 0.0488 mmol), and dichloromethane (49 ml). The reaction
mixture was stirred under an atmosphere of D2 for 4 h at which time
additional catalyst (20 mg) in dichloromethane (3 ml) was added. This solution
was stirred overnight under D2. The solvent was evaporated and the resulting
yellow oil was triturated with ether to precipitate the catalyst. The solid was
filtered and the solvent was removed from the filtrate. The resulting oil was
triturated and filtered again to leave a yellow oil. The product was purified by
flash chromatography (0–5% methanol in dichloromethane over 15 min) to
1
give 0.38 g of colorless oil. H NMR (300.132 MHz, C6D6) d 7.37 (0.2 H
residual), 7.23 (0.2 H residual), 7.16 (d, J ¼ 8:2 Hz, 1 H), 6.92 (d, J ¼ 8:2 Hz,
1 H), 3.66 (m, 2 H), 3.39 (s, 3 H), 3.05 (m, 2 H), 1.33 (m, 4 H); MS (ES+,
uncorr) m/Z 206 (0%), 207 (32%), 208 (100%), 209 (12%).
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995

REAL-TIME MONITORING
995
N-acetyl-4-bromophenylalanine deuterium-halogen replacement. N-acetyl-4-
bromophenylalanine (0.05 g, 0.17 mmol), triethylamine (0.15 ml, 1 mmol), 5%
Pd/C (10 wt%, 5 mg), and absolute ethanol (2.5 ml) were combined in a 25 ml
pear-shaped flask equipped with a stir bar. Three cycles of evacuation and
deuterium filling were performed. A Raman spectrum was acquired before
stirring was started. After 50 min, the reaction mixture was filtered through a
syringetip PTFE membrane to remove the catalyst. The solvent was removed
1
to leave a white solid. H NMR (300.132 MHz, C6D6) d 8.15 (d, J ¼ 8:0 Hz,
1 H), 7.32 (m, 4 H), 4.40 (td, J ¼ 8:7, 4.8 Hz, 1 H), 2.91 (m, 2 H), 1.78 (s, 3 H);
MS (ES+) m/Z 209; 52% content of m/Z 208.
Deuterogenation of methyl cinnamate. A 64 mg portion of methyl cinnamate
was dissolved in 5 ml of absolute ethanol in a 25 ml pear-shaped flask equipped
with a magnetic stir bar. A 7 mg portion of 5% Pd/C was added and the flask
was attached to the glass manifold. Stirring was initiated, and Raman
acquisition was begun in order to record a starting spectrum. Then in quick
succession three cycles of partial evacuation and deuterium gas introduction
were conducted; reaction stirring and spectra acquisition then continued for
the next 30 min. At this time the catalyst was removed by passage of the
reaction mixture through a syringetip PTFE membrane, and the filtrate was
1
evaporated to give 58 mg of colorless oil. H NMR (CD2Cl2) was consistent
with clean deuterogenation: d 7.58–6.96 (m, 5 H, aryl H), 3.68 (s, 3 H,
COOCH3), 2.94 (br s, 1.0 H, PhCDH-), 2.64 (br s, 1.0 H, -CDHCOOR); MS
(ES+) m/Z 107 (base peak, PhCDHCDH), 135 (40%, PhCDHCDHCO).
References
1. Shackman JG, Giles JH, Ennton MB. Special Publication – Royal Society of
Chemistry (Further Developments in Scientific Optical Imaging) 2000; 254: 186.
2. Colthup NB, Daly LH, Wiberley SE. Introduction to Infrared and Raman
Spectroscopy (3rd edn). Academic Press: San Diego, 1990.
3. Lin-Vien D, Colthup NB, Fateley WG, Grasselli JG. Infrared and Raman
Characteristic Frequencies of Organic Molecules. Academic Press: San Diego, 1991.
4. Heys JR. Chem Commun 1992; 681; Shu AYL, Saunders D, Levinson SH,
Landvatter SW, Mahoney A, Senderoff SG, Mokhallalati MK, Heys JR. J Label
Compd Radiopharm 1999; 42: 797; Shu AYL, Heys JR. Tet Lett 2000; 41: 9015;
Hesk D, Gignan G, Lee F, Yang J, Voronin K, Magatti D, McNamara P,
Koharski D, Hendershot S, Saluja S, Wang S. J Label Compd Radiopharm 2002;
45: 145; Ellames GJ, Gibson JS, Herbert JM, Kerr WJ, McNeill AH. J Label
Compd Radiopharm 2004; 47: 1; Hickey MJ, Jones JR, Kingston LP, Lockley WJS,
Mather AN, McAuley BM, Wilkinson DJ. Tet Lett 2003; 44: 3959.
5. Kolev T, Nikolova B, Jordanov B, Juchnovski I. J Mol Struc 1985; 129: 1.
6. Hernandez B, Navarro R, Hernanz A, Vergoten G. Biopolymers 2002; 67: 440.
7. Hickey MJ, Jones JR, Kingston LP, Lockley WJS, Mather AN, McAuley BM,
Wilkinson DJ. Tet Lett 2003; 44: 3959.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 983–995