In-loop flow [11C]CO2 fixation and radiosynthesis of N,N′-[11C]dibenzylurea

Article abstract of DOI:10.1002/jlcr.3568

Cyclotron-produced carbon-11 is a highly valuable radionuclide for the production of positron emission tomography (PET) radiotracers. It is typically produced as relatively unreactive carbon-11 carbon dioxide ([¹¹C]CO₂), which is most commonly converted into a more reactive precursor for synthesis of PET radiotracers. The development of [¹¹C]CO₂ fixation methods has more recently enabled the direct radiolabelling of a diverse array of structures directly from [¹¹C]CO₂, and the advantages afforded by the use of a loop-based system used in ¹¹C-methylation and ¹¹C-carboxylation reactions inspired us to apply the [¹¹C]CO₂ fixation “in-loop.” In this work, we developed and investigated a new ethylene tetrafluoroethylene (ETFE) loop-based [¹¹C]CO₂ fixation method, enabling the fast and efficient, direct-from-cyclotron, in-loop trapping of [¹¹C]CO₂ using mixed DBU/amine solutions. An optimised protocol was integrated into a proof-of-concept in-loop flow radiosynthesis of N,N′-[¹¹C]dibenzylurea. This reaction exhibited an average 78% trapping efficiency and a crude radiochemical purity of 83% (determined by radio-HPLC), giving an overall nonisolated radiochemical yield of 72% (decay-corrected) within just 3?minutes from end of bombardment. This proof-of-concept reaction has demonstrated that efficient [¹¹C]CO₂ fixation can be achieved in a low-volume (150?μL) ETFE loop and that this can be easily integrated into a rapid in-loop flow radiosynthesis of carbon-11–labelled products. This new in-loop methodology will allow fast radiolabelling reactions to be performed using cheap/disposable ETFE tubing setup (ideal for good manufacturing practice production) thereby contributing to the widespread usage of [¹¹C]CO₂ trapping/fixation reactions for the production of PET radiotracers.

Full text of DOI:10.1002/jlcr.3568

In-Loop Flow [11C]CO2 Fixation and Radiosynthesis
of N,N’-[11C]Dibenzylurea
Joseph Downey,‡ Salvatore Bongarzone,‡ Stefan Hader, Antony D. Gee*
Division of Imaging Sciences and Biomedical Engineering, King's College London, 4th Floor Lambeth Wing, London
SE1 7EH, UK
‡ These authors contributed equally to this work.
*Corresponding author
Abstract
Cyclotron-produced carbon-11 is a highly valuable radionuclide for the production of
positron emission tomography (PET) radiotracers. It is typically produced as relatively
unreactive carbon-11 carbon dioxide ([11C]CO2 ) which is most commonly converted into a
more reactive precursor for synthesis of PET radiotracers.
The development of [11C]CO2 fixation methods have more recently enabled the direct
radiolabeling of a diverse array of structures directly from [11C]CO2, and the advantages
afforded by the use of a loop-based system employed in 11C-methylation and 11C-
carboxylation reactions inspired us to apply the [11C]CO2 fixation “in-loop”. In this work we
developed and investigated a new ethylene tetrafluoroethylene (ETFE) loop-based [11C]CO2-
fixation method, enabling the fast and efficient, direct-from-cyclotron, in-loop trapping of
[11C]CO2 using mixed DBU/amine solutions. An optimised protocol was integrated into a
proof-of-concept in-loop flow radiosynthesis of N,N’-[11C]dibenzylurea. This reaction
exhibited an average 78% trapping efficiency and a crude radiochemical purity of 83%
(determined by radio-HPLC), giving an overall non-isolated radiochemical yield of 72%
(decay corrected) within just 3 min from end-of-bombardment.
This proof-of-concept reaction has demonstrated that efficient [11C]CO2 fixation can be
achieved in a low-volume (150 µL) ETFE loop, and that this can be easily integrated into a
rapid in-loop flow radiosynthesis of carbon-11 labeled products.
This new in-loop methodology will allow fast radiolabelling reactions to be performed using
cheap/disposable ETFE tubing setup (ideal for GMP production) thereby contributing to the
widespread usage of [11C]CO2 trapping/fixation reactions for the production of PET
radiotracers.
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/jlcr.3568
This article is protected by copyright. All rights reserved.

Introduction
One of the most prevalent and important radionuclides used in positron emission tomography
(PET) is carbon-11 (11C, t1/2 = 20.4 min), it has been incorporated into a wide variety of
exogenous and endogenous ligands, used for both diagnostic and research purposes.1–5 The
short half-life potentially allows the administration and scanning of multiple PET radiotracers
in the same patient on the same day, thereby combining the advantages of imaging multiple
biochemical pathways; in addition, it enables the labelling of biologically relevant molecules,
without changing their pharmacodynamic and pharmacokinetic properties, by virtue of its
isotopology with carbon-12.
Carbon-11 is produced as [11C]CO2 (primary precursor), by the cyclotron proton-
bombardment of nitrogen-14 via the 14N(p,α)11C nuclear reaction. The relatively low
reactivity and solubility of [11C]CO2 leads to its rapid conversion to more reactive secondary
precursors.1 The most prevalent of these secondary precursors is [11C]CH3I for 11C-
methylation reactions.1 While these reactions are used to produce the vast majority of carbon-
11 radiotracers, in the time taken to convert [11C]CO2 to [11C]CH3I, significant amounts of
the starting radioactivity can be lost through multistep synthesis and radioactive decay.6 In
11
addition, C-methylation can somewhat limit the chemical space available for radiolabeling.
As such, there have been a variety of alternative secondary and tertiary 11C-precursors
11
developed to convert the poorly reactive [11C]CO2 into a more versatile toolbox for the C-
radiochemist: [11C]CO, [11C]HCN, [11C]CS2, [11C]CH3OTf and [11C]COCl2 are just a few
examples.1
In an effort to avoid this time consuming conversion to more reactive precursors,
[11C]CO2 has been reacted directly with Grignard reagents (alkyl magnesium bromide) to
synthesise 11C-carboxylic acid derivatives such as [11C]acetate, [11C]palmitic acid and 2-
[11C]octynoic acid.7–9 In an effort to easily automate these syntheses, these reactions have
been performed “in-loop” via a captive solvent system; whereby the walls of a small tubing
loop are coated with the highly-reactive Grignard reagents, then [11C]CO2 is flowed through
11
the loop, where it reacts “in-loop” to form the C-labeled carboxylic acid product.10,11 This
“in-loop” synthetic approach has also been employed in 11C-methylation reactions using
gaseous [11C]CH3I, with the synthesis of [11C]raclopride as a notable example.12–16 This
synthetic methodology lends itself particularly well to the specific demands of 11C-
radiochemistry, and it has a number of advantages over the standard vial-based methods. For
these mixed gas-liquid phase reactions, the high surface area provided by the walls of the
loop maximises gas-liquid contact, increasing the efficiency and thus decreasing the reaction
times required versus the standard vial-based setup. In addition, since loop synthesis
eliminates reaction vials, transfer losses are kept to a minimum; and since this, in effect,
miniaturises these reactions, precursor quantities can be greatly reduced. Finally, these
reactions are easily automated which is advantageous when considering the translational
potential of a tracer.12,13
More recently, there has been a resurgence of interest into the direct use of [11C]CO2
in radiolabeling using a technique called [11C]CO2-fixation.6,17 This new chemistry is a by-
product of the “green-chemistry” movement and the corresponding interest in new chemical
agents for carbon-capture;6,17–19 the most commonly used compounds being the amidine base
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU),20 and the phosphazene base, 2-tert-butylimino-2-
This article is protected by copyright. All rights reserved.

diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorine (BEMP).21 While there is still
some uncertainty regarding their trapping/fixation mechanisms,22,23 solutions of these bases
with an amine substrate exhibit efficient [11C]CO2 trapping and transfer to form amine
carbamate intermediates.20 This [11C]CO2 fixation has therefore enabled many new 11C-
carbonylation reactions starting from [11C]CO2, and many new syntheses have been
developed based on this fixation strategy (11C-carbamates, 11C-ureas, 11C-amides and 11C-
oxazolidinones, among others).20,24–29 Within our group, we have developed a Mitsunobu-
based method for the synthesis of 11C-ureas.26,27 This reaction involves the fixation of
[11C]CO2 by DBU and amines to form the 11C-amine carbamate intermediates; Mitsunobu
11
reagents (PBu3/DBAD) are added to form an C-isocyanate intermediate, which reacts with
another amine molecule to form an 11C-labeled urea product, in just 5 min total reaction
time.26,27
The recent interest in [11C]CO2 fixation chemistry has the potential to revolutionise
the field of carbon-11 radiochemistry, and the advantages afforded by the use of a loop-based
system employed in 11C-methylation and 11C-carboxylation reactions inspired us to apply the
[11C]CO2 fixation “in-loop”.
In this work, our aim was therefore to: i) establish a novel in-loop trapping/fixation of
[11C]CO2 for the investigation and optimisation of different trapping solutions; and ii)
11
implement the setup in an in-loop flow-radiosynthesis of C-labeled urea functional groups
from [11C]CO2.
Results and Discussion
Three-component [11C]CO2 loop trapping
To assess the potential for in-loop trapping/fixation of [11C]CO2 we designed a novel
prototype three-component loop trapping apparatus (Figure 1). In designing the system, good
manufacturing practice (GMP) requirements led us to opt for a single-use disposable ethylene
tetrafluoroethylene (ETFE) based loop system since it obviates the need for rigorous cleaning
etc. The system consists of three easily-separable components: a trapping loop (1/16” O.D.
ETFE tubing, 150 µL volume), the walls of which will be coated with a trapping solution; a
crimped glass waste vial; and a trapping cartridge able to fix all unreacted [11C]CO2 (ascarite
trap, Figure 1). On passing [11C]CO2 through this system, the [11C]CO2 is totally trapped
within these three components. The components are then simply separated and the
radioactivity levels of the loop, waste vial and ascarite trap (Rloop, Rwaste and Rtrap
respectively) are measured. Comparison of these values allows the calculation of trapping
efficiencies for a given solution.
To pre-coat the walls of the loop, it is half filled with 75 µL trapping solution, then it
is connected to the cyclotron delivery line and flushed with helium gas through to the waste
vial. The waste vial contains the trapping solution that was not retained by the loop.
Therefore when [11C]CO2 is passed through the three-component system, it is trapped within
the loop and any untrapped [11C]CO2 will be fixed in the solution within the waste vial and
the trapping cartridge.
This article is protected by copyright. All rights reserved.

To assess a solution’s overall trapping efficiency as well as its suitability for our
applications, we extracted two values from the acquired data: these are termed total-solution
trapping (Tsol) and loop-trapping (Tloop).
Tsol (Tsol = (Rloop + Rwaste)/(Rloop + Rwaste + Rtrap)) gives an insight into a solution’s
ability to trap [11C]CO2 in bulk and is a proxy measure for the chemical trapping efficiency of
a given solution. While a low Tsol indicates that a solvent mixture is unsuitable for application
within our in-loop [11C]CO2 trapping/fixation method, a high Tsol does not conversely
guarantee success for our work. Since a highly efficient trapping solution may still have poor
loop-retention – the degree to which the trapping solution is retained on the walls of the loop
– and thus [11C]CO2 trapping/fixation will not occur in-loop but in the waste vial.
Instead Tloop (Tloop = Rloop/(Rloop + Rwaste + Rtrap)) accounts both for the chemical
trapping ability of a solution, but also the degree to which it is retained within our 150 µL
ETFE loop. High values of Tsol and Tloop should guarantee success in developing an in-loop
[11C]CO2 fixation methodology.
Our model trapping solutions contained varying concentrations of benzylamine and
DBU dissolved in acetonitrile (MeCN). DBU was chosen instead of BEMP, since in previous
work using these Mitsunobu reagents, both compounds gave good [11C]CO2 trapping but only
DBU led to 11C-urea formation.27
To ensure MeCN does not exhibit any [11C]CO2 trapping itself, pure MeCN trapping
experiments were attempted and – as expected considering the low solubility of [11C]CO2 in
MeCN – negligible trapping was seen for this experiment (Tsol = 6.5 ± 0.1%, Tloop = 0.2 %, n
= 2).
Firstly, we explored the effect of varying benzylamine concentration (1%, 5% and
10%) in the presence of a fixed content of DBU (10% v/v) (Figure 2). A solution of 10%
DBU in MeCN without adding benzylamine (Figure 2) showed good Tsol (62.7 ± 2.3%), but
very low Tloop (2.1 ± 0.1%). While the solution can trap [11C]CO2 reasonably well, it is poorly
retained in the loop. Any solutions to be used within this loop trapping/fixation methodology
must therefore exhibit high Tloop. Increasing benzylamine content up to 10%, both Tsol and
Tloop increased. Notably, adding 1%, 5% and 10% benzylamine, Tsol showed near-quantitative
trapping of total [11C]CO2 (93.8 ± 2.5%, 97.6 ± 0.8% and 97.5 ± 0.4%, respectively).
However despite the comparable Tsol seen for these solutions, there was a marked difference
in Tloop values (Figure 2). Using 1% and 5% benzylamine content showed a two-fold
increase of Tloop (10.9 ± 4.4% and 24.4 ± 8.0% respectively), however increasing further the
content of benzylamine up to 10% Tloop did not significantly improve (27.0 ± 3.2%).
These results show that – chemically – these mixed benzylamine/DBU solutions, are
exquisitely fine [11C]CO2 trapping/fixation solutions, even at relatively low concentrations of
benzylamine (1%, Tsol > 90%). However for their usage in loop trapping/fixation, only the
solutions with high Tloop values are suitable (10% benzylamine, Tsol = 97%, Tloop = 27%). We
suspected that the increased benzylamine content of the solutions with the highest Tloop values
resulted in the increased viscosity of these solutions, replacing less-viscous MeCN (0.343
mPa.s)30 with more-viscous benzylamine (1.492 mPa.s),31 which therefore gave an increased
retention of the solution on the walls of the loop.
This article is protected by copyright. All rights reserved.

Since we observed no significant improvement of Tloop using 5% or 10% of
benzylamine we decided to fix the content of benzylamine to 10% in the next experiments.
We then explored the effect of varying DBU concentration in our solutions (Figure 3). 10%
benzylamine in MeCN (no DBU added) showed a reasonable Tsol (44.1 ± 15.4%), but very
low Tloop (1.6 ± 1.1%). These results are very similar to those seen for a solution of 10%
DBU in MeCN (Figure 2), indeed while both solutions (10% DBU in MeCN or 10%
benzylamine in MeCN) can trap [11C]CO2 fairly well, their poor loop-retention hampers the
Tloop. Adding DBU at 10%, 50% and 90% led to near-quantitative Tsol (97.5 ± 0.4%, 96.5
±1.8%, and 89.9 ± 4.0%, respectively), again demonstrating that these are highly powerful
[11C]CO2 trapping/fixation solutions (Figure 3). However Tloop increased by adding 10%,
50%, and 90% DBU (27.0 ± 3.2%, 26.8 ± 15.6%, and 41.8 ± 7.1, respectively). The highest
Tloop value (42%) for this three-component trapping setup was obtained using a MeCN-free
system containing 90% DBU and 10% benzylamine (v/v) mixture. This again supports the
suggestion that solution viscosity dictates loop-retention and therefore affects Tloop, since we
are increasing the content of more-viscous DBU (11.76 mPa.s)32 and decreasing the content
of less-viscous MeCN (0.343 mPa.s).31
This section of work demonstrated the successful trapping and fixation of [11C]CO2
using a mixture of benzylamine/DBU solutions coating the walls of a 150 µL ETFE loop.
The modular three-component experimental setup was easy to set-up as well as to disconnect
for radioactivity measurements, and so allows the rapid screening of a number of different
trapping solutions. We next wanted to explore the potential integration of this in-loop
[11C]CO2 trapping into more complex flow syntheses of 11C-labeled compounds. To
investigate this, we performed a proof-of-concept flow synthesis of N,N’-[11C]dibenzylurea
using a more complex setup, and using the optimal trapping solution developed above: 90%
DBU and 10% benzylamine (v/v) mixture.
In-loop flow radiosynthesis of N,N’-[11C]dibenzylurea
The apparatus setup for the flow-synthesis section of this work was based on the simple
trapping apparatus described above. The [11C]CO2 trapping/fixation in-loop was initially
performed in a similar manner, before passing a solution of Mitsunobu reagents (PBu3 and
DBAD in MeCN) through the trapping loop, through a second 150 µL reaction loop (to
ensure adequate mixing) and into a product vial (Figure 4). The [11C]CO2 is trapped initially
as an N-[11C]benzyl carbamate intermediate, which is converted by the Mitsunobu reagents to
a highly reactive 11C-isocyanate intermediate. This then undergoes attack from another
molecule of benzylamine, to form the desired N,N’-[11C]dibenzylurea product (Figure 4).
The setup (Figure 4) includes two switching valves (V1 and V2) on an E&Z Modular
Lab automated synthesis system. The trapping loop is connected to these two valves, the inlet
is connected to V1, and the outlet is connected to V2. V1 can be switched to either the
cyclotron outlet (for helium flush and [11C]CO2 delivery), or to a syringe containing the
Mitsunobu reagents. V2 can divert the flow towards either the waste vial or through the
reaction loop and into the product vial (see materials and methods section for step-by-step
procedure). To ascertain the quantity of [11C]CO2 delivered through the system per-
experiment, the system was washed with MeCN and the radioactivity of all the washings and
components (lines, fittings and needles) was measured.
This article is protected by copyright. All rights reserved.

An additional feature of this setup is the 1 mL H2O added as a quenching solution in
the final product vial. This ensures that any 11C-urea derivatives detected in the crude HPLC
are products formed exclusively in the loop. Without this simple addition, the reaction could
feasibly be simply occurring in the product vial that receives the reaction mixture, during the
time taken to measure product radioactivity and prepare the sample for HPLC analysis. The
active Mitsunobu intermediate formed is a Morrison-Brunn-Huisgen betaine, which will react
with protic substrates: alcohols, amines and carboxylic acids.33–37 Our own non-radioactive
experiments have confirmed that this betaine reacts with water forming the tri-n-
butylphosphine oxide (PBu3O) and di-tert-butyl hydrazodiformate (DBAD-H2). Therefore in
the 11C-urea synthesis, any remaining unreacted urea-forming betaine will react with the
excess water (in the product vial) upon leaving the reaction loop, ensuring that the crude-
HPLC is representative of the 11C-labeled species formed in-loop, not in-vial. In addition this
serves to pre-dilute the crude product ready for HPLC injection, which helps to streamline the
synthetic process.
One notable observation during these 11C-urea syntheses was that we saw a
significantly increased Tloop (78.3 ± 3.6%) compared to that seen within the trapping
experiments (41.8 ± 7.1%), indicating an increase in loop-retention of our optimised trapping
solution. This variance in Tloop value might be due to the difference between the two setups;
in the flow synthesis of N,N’-[11C]dibenzylurea the loops are routed via the E&Z switching
valves (V1 and V2). We speculate that the use of these valves increases the backpressure in
the system, and correspondingly increases the loop-retention of the trapping solution.
Due to the demanding time-constraints placed upon carbon-11 radiochemistry in the
development of this method we attempted to minimise all process-times, avoiding any
product losses due to radioactive decay. Since we avoided pre-trapping or concentration of
[11C]CO2 (instead delivered diluted in the helium carrier gas), and the rate of delivery was not
slowed from the cyclotron’s 70 mL/min, the [11C]CO2 is delivered through the system and
trapped in the loop within 105 sec of the end-of-bombardment (EOB). V1 and V2 are
instantly switched and the trapping loop is filled with Mitsunobu reagents within 30 sec
(Figure 4). V1 is then switched back to a helium flush and the reagents are pushed through
the reaction loop and into the product vial at 70 mL/min. Therefore, the process is complete
within just 3 min from EOB, and 1 min from end-of-delivery (EOD).
In the synthesis of our model substrate, N,N’-[11C]dibenzylurea, we had to consider
the concentration and stoichiometry of our Mitsunobu reagents added to the 150 µL trapping
loop. Based on the assumption of 7-8 µL trapping solution retention (based on preliminary
non-radioactive flushing experiments filling the loop with 75 µL of solution) and considering
11
the optimal conditions found in our previous radiosynthesis of C-symmetrical ureas,27 the
reagent concentration was selected to ensure a 2:1 stoichiometric ratio of Mitsunobu reagents
(PBu3/DBAD) to benzylamine.
Using the in-loop flow radiosynthesis setup: the radio-HPLC of the crude solution
showed a radiochemical purity (RCP) of 82.6 ± 3.3%. This coupled with the overall Tloop of
78.3 ± 3.6%, led to a decay-corrected non-isolated radiochemical yield (RCY) of 72.3 ± 5.1%
(n=3) for the synthesis of N,N’-[11C]dibenzylurea, in 3 min from EOB, with a molar
radioactivity of 0.72 ± 0.17 GBq/µmol (for 300-350 MBq initial [11C]CO2; other work in the
group has demonstrated that this molar radioactivity would be expected to increase towards
This article is protected by copyright. All rights reserved.

60-70 GBq/µmol for an initial 30 GBq [11C]CO2 production,28 which is consistent with the
molar radioactivities obtained for other clinical 11C-labelled radiotracers within our
institution).38 This therefore presents a reproducible, rapid, and high-yielding synthesis of a
carbon-11 labeled urea product directly from [11C]CO2.
The radiochemical yields for the synthesis of N,N’-[11C]dibenzylurea using this in-
loop method (72% decay-corrected) are slightly lower than those for the traditional in-vial
method (82% decay-corrected), due to a lower [11C]CO2 trapping efficiency (78% versus
96%), but with a comparable RCP (83% vs 85%, by crude radio-HPLC).27 However this in-
loop method: uses smaller quantities of reagents compared to the in-vial method (75 µL
versus 400 µL), which should simplify purification; will minimise most of the transfer losses
associated with in-vial synthesis; and the use of cheap, disposable, ETFE loops means that
this method is particularly well suited to GMP production. These factors combined therefore
mean that this in-loop method presents an appealing alternative to in-vial [11C]CO2 fixation
reactions.
Conclusions
In this work we demonstrated that cyclotron-produced [11C]CO2 can be fixed in a low-
volume (150 µL) ETFE loop. We showed that while amine/DBU solutions are chemically
efficient [11C]CO2 fixation agents in-vial, both the chemical and physical properties
(primarily viscosity) of these solutions determine their degree of loop trapping efficiency
(Tloop). This optimised direct-from-cyclotron [11C]CO2 fixation methodology avoids the need
for cryogenic pre-concentration commonly used in carbon-11 procedures. This setup was
implemented in a proof-of-concept in-loop flow radiosynthesis of N,N’-[11C]dibenzylurea by
passing Mitsunobu reagents through the loop containing a trapped N-[11C]benzyl carbamate
intermediate and benzylamine. N,N’-[11C]dibenzylurea was obtained with high non-isolated
radiochemical yields (~ 72%), comparable to those previously reported in-vial (~ 82%),
within 3 min from EOB (recently presented in abstract form at the International Symposium
on Radiopharmaceutical Sciences, ISRS, Dresden, 14-19 May 2017).39 This novel
methodology has demonstrated the potential for direct-from-cyclotron in-loop [11C]CO2
fixation and has demonstrated that this can be employed as part of a more complex synthesis
(eg. amides, carbamates).28 The speed of the reaction (3 min from EOB) and the
cheap/disposable ETFE tubing setup (ideal for GMP production) mean that this method
should be suitable for further applications in the direct trapping/fixation reactions of
[11C]CO2. We anticipate that this new methodology will facilitate a more widespread uptake
and application of these powerful reactions, for the radiolabeling of a diverse array of
structures directly from [11C]CO2.
Experimental
Materials and general methods
Anhydrous acetonitrile (MeCN, 99.8%), ascarite, benzylamine (99%), di-tert-butyl-
azodicarboxylate (DBAD, 98%), triethylamine (Et3N, ≥ 99.5%) and tri-n-butyl phosphine
(PBu3, 99%) were purchased from Sigma-Aldrich. Ethyl acetate (EtOAc, ≥ 99.5%) was
purchased from Fisher Scientific. Anhydrous magnesium sulfate (MgSO4, 98%) was
This article is protected by copyright. All rights reserved.

purchased from Fluka. 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU, 99%) and benzyl
isocyanate (98%) were purchased from Alfa Aesar. Carbon dioxide (CO2) was purchased
from BOC Gases.
ETFE tubing (1/16” O.D. x 0.75 mm I.D., 25 m/pkg) was obtained from VICI Jours.
The ascarite traps were constructed from empty SPE-ED cartridges, obtained from Biosys
Solutions Ltd: Fritted Empty MiniSPE-ED Cartridges, part # 2447. All fluidic connections
were obtained from Upchurch Scientific, the product codes are as follows: fingertight
flangeless fitting short, PEEK, XP-235X; female to male quick-connect Luer adapter, P-675-
01.
[11C]CO2 was produced using a Siemens RDS112 cyclotron in a 14N(p,α)11C reaction,
by the 11 MeV proton bombardment of nitrogen (+ 1% O2) gas. The cyclotron produced
[11C]CO2 was transferred in a stream of helium gas at 70 mL/min directly into a switching
valve of an E&Z Modular Lab automated synthesis unit. Unless otherwise specified, all
radioactive experiments used a 5µA bombardment for 1 min, giving on average 300-350
MBq [11C]CO2 at EOB. RCYs reported are calculated as a percentage of the total
radioactivity delivered from the cyclotron. Unless otherwise stated, all experiments were
repeated three times (n = 3).
1H and 13C{1H} NMR spectra were recorded on a Bruker Avance DRX 400 MHz
spectrometer at 294 K. Chemical shifts are reported in parts per million (ppm) relative to the
residual solvent proton impurities (1H) or the residual solvent carbon impurities (13C), as
internal standards.
HPLC analysis was performed on an Agilent 1200 system, with a variable wavelength
UV detector and a LabLogic Flow-RAM β+ detector equipped in series. Analytical reverse-
phase column: Agilent XDB-C18, 5 µm, 4.6 x 150 mm. Gradient used: 95% H2O, 5%
MeCN; to 5% H2O, 95% MeCN; over 9 min. Identity of radioactive products was confirmed
by co-elution with the non-radioactive standard compounds. HPLC was used to determine
molar radioactivities, by reference to a variable dilution calibration curve. Previous
experiments within our lab have shown that scaling reactions to higher starting
radioactivities, with all other factors kept constant, leads to correspondingly higher molar
radioactivities. From these results we assume that a 100-fold increase from approx. 300 MBq
(preliminary experiments) to 30 GBq initial [11C]CO2 (clinical production) will give a
roughly 100-fold increase in molar radioactivity.
Mitsunobu solutions were prepared by dissolving DBAD (21.1 mg, 91.6 µmol, 3 eq.)
in 1 mL MeCN (anhydrous). PBu3 (22.9 µL, 91.6 µmol, 3 eq.) was added and the mixture
was briefly shaken. A colour change from pale yellow-to-colourless was observed on
successful formation of the active Mitsunobu intermediate. Trapping solutions were prepared
as 1 mL solutions by diluting the corresponding percentages (v/v) of DBU and benzylamine
in MeCN. All trapping (benzylamine/DBU/MeCN) and Mitsunobu (PBu3/DBAD/MeCN)
solutions were prepared under an inert argon atmosphere, using anhydrous MeCN.
Three-component [11C]CO2 trapping apparatus: design and setup
The trapping apparatus setup is shown in Figure 1. One end of a 35 cm length of ETFE
tubing (trapping loop, 1/16” O.D., 0.75mm I.D., 150 µL volume) was fitted with a
fingertight screw fitting, and the other end was cleanly cut at a 45º taper. The loop was tightly
This article is protected by copyright. All rights reserved.

coiled and placed inside a 10 mL glass vial for ease of handling. The tapered end was inserted
through the rubber septum of a sealed crimped 10 mL waste vial. This vial was vented via a
needle through an ascarite trap into a gas waste bag. The loop was half-filled with 75 µL
trapping solution, using a 100 µL syringe, and connected via a luer slip fitting to the
[11C]CO2 outlet line from the E&Z Modular Lab. Providing a simple, modular and easy to
disconnect setup which traps all [11C]CO2 passed through.
11C-Urea derivative flow synthesis apparatus: design and setup
The 11C-urea synthesis apparatus builds upon the trapping apparatus proof-of-concept, and is
shown in Figure 4. The cyclotron [11C]CO2 outlet line was connected to switching valve 1
(V1) as was a syringe containing Mitsunobu reagents (PBu3 and DBAD in MeCN). Both ends
of a 35 cm length of ETFE tubing (trapping loop, 1/16” O.D., 0.75mm I.D., 150 µL volume)
were fitted with fingertight screw fittings, and the loop was half filled with 75 µL trapping
solution. One end was attached to the outlet of V1 and the other to the inlet of switching
valve 2 (V2). To one outlet of V2, a short length of ETFE tubing was connected to a waste
vial, vented via an ascarite trap. To the other outlet is connected a second 35 cm length of
coiled ETFE tubing (reaction loop, 1/16” O.D., 0.75mm I.D., 150 µL) running into a sealed
crimped product vial (containing 1 mL water), vented via an ascarite trap.
Generic procedure: [11C]CO2 trapping
The preloaded loop was connected to the [11C]CO2 delivery line from the cyclotron and
helium was flushed through the system for 3 min at 70 mL/min, leaving a just a small,
residual amount of trapping solution coating the walls of the trapping loop, and flushing the
bulk of the trapping solution into the waste vial. The cyclotron produced [11C]CO2 was then
directly delivered diluted in carrier helium gas (without prior trapping and concentration) at
70 mL/min into the E&Z Modular Lab and subsequently through the three-component
trapping apparatus. Since all [11C]CO2 is trapped within either the trapping loop, the waste
vial, or the ascarite trap, these three components are quickly separated and their
radioactivities measured within a Capintec® dose calibrator. Comparison of the distribution of
radioactivity within this apparatus allowed calculation of different solvent trapping
efficiencies.
Generic procedure: 11C-urea derivative synthesis
1) Helium was flushed through the preloaded trapping loop to the waste vial for 3 min at 70
mL/min. 2) Cyclotron produced [11C]CO2 was then directly delivered, diluted in carrier
helium gas (without prior trapping and concentration) at 70 mL/min, through the trapping
loop, waste vial and ascarite trap. 3) V1 and V2 are switched and the trapping loop was filled
with 150 µL Mitsunobu solution (PBu3/DBAD in MeCN). 4) V1 was switched and a 70
mL/min helium flush from the cyclotron transferred the contents of the trapping loop through
the reaction loop and into the product vial (containing 1 mL water as a quench). 5) The crude
products were analysed by radio-HPLC and the system was washed with MeCN and all
washings and component radioactivities were measured to determine the total [11C]CO2
radioactivity delivered from the cyclotron.
This article is protected by copyright. All rights reserved.

Synthesis of N,N’-Dibenzylurea
Benzylamine (1.2 mmol, 131 µL, 3 eq.) was dissolved in 2 mL EtOAc with stirring at room
temperature. To this was added Et3N (1.2 mmol, 167 µL, 3 eq.) and benzyl isocyanate (400
µmol, 49.4 µL, 1 eq.). Solution was stirred for 2 hours under ambient conditions. 1M HCl
was added and the solution was extracted three times with EtOAc. The combined organic
extracts were washed with brine and dried over MgSO4. The solvent was removed under
reduced pressure to yield a white solid (83.5 mg, 87% yield); 1H NMR (400 MHz, CDCl3) δ
7.25-7.1 (m, 10H, CH), 4.75 (br s, 2H, NH), 4.26 (d, 4H, CH2); 13C NMR (100 MHz, CDCl3)
δ 158.1 (CO), 139.0 (C), 128.6 (CH), 127.4 (CH), 127.3 (CH), 44.5 (CH2).
Acknowledgments
This work was supported by the Medical Research Council (MRC, MR/K022733/1) and the
King’s College London and Imperial College London EPSRC Centre for Doctoral Training in
Medical Imaging (EP/L015226/1). The authors acknowledge financial support from the
Department of Health via the National Institute for Health Research (NIHR) comprehensive
Biomedical Research Centre award to Guy’s & St Thomas’ NHS Foundation Trust in
partnership with King’s College London and King’s College Hospital NHS Foundation Trust
and the Centre of Excellence in Medical Engineering funded by the Wellcome Trust and
EPSRC under grant number WT088641/Z/09/Z.
References
1.
2.
3.
4.
5.
6.
7.
Miller PW, Long NJ, Vilar R, Gee AD. Synthesis of 11C, 18F, 15O, and 13N Radiolabels for Positron Emission
Tomography. Angew Chemie Int Ed. 2008;47(47):8998-9033. doi:10.1002/anie.200800222.
Antoni G. Development of carbon-11 labelled PET tracers-radiochemical and technological challenges in a historic
perspective. J Labelled Comp Radiopharm. 2015;58(3):65-72. doi:10.1002/jlcr.3258.
Tu Z, Mach RH. C-11 radiochemistry in cancer imaging applications. Curr Top Med Chem (Sharjah, United Arab
Emirates). 2010;10(11):1060-1095. doi:10.2174/156802610791384261.
Ametamey SM, Honer M, Schubiger PA. Molecular imaging with PET. Chem Rev. 2008;108(5):1501-1516.
doi:10.1021/cr0782426.
Piel M, Vernaleken I, Rösch F. Positron emission tomography in CNS drug discovery and drug monitoring. J Med
Chem. 2014;57(22):9232-9258. doi:10.1021/jm5001858.
Rotstein BH, Liang SH, Holland JP, et al. 11CO2 fixation: a renaissance in PET radiochemistry. Chem Commun.
2013;49(50):5621. doi:10.1039/c3cc42236d.
Pike VW, Eakins MN, Allan RM, Selwyn AP. Preparation of carbon-11 labelled acetate and palmitic acid for the
study of myocardial metabolism by emission-computerised axial tomography. J Radioanal Chem. 1981;64(1-
2):291-297. doi:10.1007/BF02518360.
8.
9.
Kawashima H, Yajima K, Kuge Y, Hashimoto N, Miyake Y. Synthesis of [1-11C]-2-octynoic acid, [1-11C]-2-
decynoic acid and [1-11C]-3-(R,S)-methyloctanoic acid as potential markers for PET studies of fatty acid
metabolism. J Label Compd Radiopharm. 1997;39(3):181-193. doi:10.1002/(SICI)1099-
1344(199703)39:3<181::AID-JLCR962>3.0.CO;2-W.
Runkle AC, Shao X, Tluczek LJM, Henderson BD, Hockley BG, Scott PJH. Automated production of [11C]acetate
and [11C]palmitate using a modified GE Tracerlab FXC-Pro. Appl Radiat Isot. 2011;69(4):691-698.
doi:10.1016/j.apradiso.2011.01.002.
10.
11.
Davenport RJ, Dowsett K, Pike VW. A simple technique for the automated production of no-carrier-added [1-
11C]acetate. Appl Radiat Isot. 1997;48(8):1117-1120. doi:10.1016/S0969-8043(97)00118-8.
Rami-Mark C, Ungersboeck J, Haeusler D, et al. Reliable set-up for in-loop 11C-carboxylations using Grignard
This article is protected by copyright. All rights reserved.

reactions for the preparation of [carbonyl-11C]WAY-100635 and [11C]-(+)-PHNO. Appl Radiat Isot. 2013;82:75-80.
doi:10.1016/j.apradiso.2013.07.023.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
Wilson AA, Garcia A, Jin L, Houle S. Radiotracer synthesis from [11C]-iodomethane: A remarkably simple captive
solvent method. Nucl Med Biol. 2000;27(6):529-532. doi:10.1016/S0969-8051(00)00132-3.
Wilson AA, Garcia A, Houle S, Vasdev N. Utility of commercial radiosynthetic modules in captive solvent [11C]-
methylation reactions. J Label Compd Radiopharm. 2009;52(11):490-492. doi:10.1002/jlcr.1618.
Iwata R, Pascali C, Bogni A, Miyake Y, Yanai K, Ido T. A simple loop method for the automated preparation of
[11C]raclopride from [11C]methyl triflate. Appl Radiat Isot. 2001;55(1):17-22. doi:10.1016/S0969-8043(00)00368-7.
Shao X, Kilbourn MR. A simple modification of GE tracerlab FX C Pro for rapid sequential preparation of
[11C]carfentanil and [11C]raclopride. Appl Radiat Isot. 2009;67(4):602-605. doi:10.1016/j.apradiso.2008.12.013.
Shao X, Schnau PL, Fawaz M V., Scott PJH. Enhanced radiosyntheses of [11C]raclopride and [11C]DASB using
ethanolic loop chemistry. Nucl Med Biol. 2013;40(1):109-116. doi:10.1016/j.nucmedbio.2012.09.008.
Rotstein BH, Liang SH, Placzek MS, et al. 11C=O bonds made easily for positron emission tomography
radiopharmaceuticals. Chem Soc Rev. 2016;45(17):4708-4726. doi:10.1039/C6CS00310A.
Jessop PG, Heldebrant DJ, Li X, Eckert CA, Liotta CL. Green chemistry: Reversible nonpolar-to-polar solvent.
Nature. 2005;436(7054):1102-1102. doi:10.1038/4361102a.
Heldebrant DJ, Yonker CR, Jessop PG, Phan L. Organic liquid CO2 capture agents with high gravimetric CO2
capacity. Energy Environ Sci. 2008;124(July):487-493. doi:10.1039/b809533g.
Hooker JM, Reibel AT, Hill SM, Schueller MJ, Fowler JS. One-Pot, Direct Incorporation of [11C]CO2 into
Carbamates. Angew Chemie Int Ed. 2009;48(19):3482-3485. doi:10.1002/anie.200900112.
Wilson AA, Garcia A, Houle S, Vasdev N. Direct fixation of [11C]-CO2 by amines: formation of [11C-carbonyl]-
methylcarbamates. Org Biomol Chem. 2010;8(2):428-432. doi:10.1039/B916419G.
Heldebrant DJ, Jessop PG, Thomas CA, Eckert CA, Liotta CL. The reaction of 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) with carbon dioxide. J Org Chem. 2005;70(13):5335-5338. doi:10.1021/jo0503759.
Nicholls R, Kaufhold S, Nguyen BN, et al. Observation of guanidine–carbon dioxide complexation in solution and
its role in the reaction of carbon dioxide and propargylamines. Catal Sci Technol. 2014;4(10):3458-3462.
doi:10.1039/C4CY00480A.
24.
25.
26.
27.
Vasdev N, Sadovski O, Garcia A, et al. Radiosynthesis of [11C]SL25.1188 via [11C]CO2 fixation for imaging
monoamine oxidase B. J Label Compd Radiopharm. 2011;54(10):678-680. doi:10.1002/jlcr.1908.
Wilson AA, Garcia A, Houle S, Sadovski O, Vasdev N. Synthesis and Application of Isocyanates Radiolabeled
with Carbon-11. Chem - A Eur J. 2011;17(1):259-264. doi:10.1002/chem.201002345.
Haji Dheere AK, Yusuf N, Gee A. Rapid and efficient synthesis of [11C]ureas via the incorporation of [11C]CO2 into
aliphatic and aromatic amines. Chem Commun (Camb). 2013;49(74):8193-8195. doi:10.1039/c3cc44046j.
Dheere A, Bongarzone S, Taddei C, Yan R, Gee A. Synthesis of 11C-Labelled Symmetrical Ureas via the Rapid
Incorporation of [11C]CO2 into Aliphatic and Aromatic Amines. Synlett. 2015;26(16):2257-2260. doi:10.1055/s-
0034-1381055.
28.
29.
30.
Bongarzone S, Runser A, Taddei C, et al. From [11C]CO2 to [11C]amides: a rapid one-pot synthesis via the
Mitsunobu reaction. Chem Commun. 2017;53(38):5334-5337. doi:10.1039/C7CC01407D.
Mossine A V., Brooks AF, Jackson IM, et al. Synthesis of Diverse 11C-Labeled PET Radiotracers via Direct
Incorporation of [11C]CO2. Bioconjug Chem. 2016;27(5):1382-1389. doi:10.1021/acs.bioconjchem.6b00163.
CHERIC. CHERIC: Chemical Engineering and Materials Research Information Center, Pure Component
Properties, Acetonitrile. https://www.cheric.org/research/kdb/hcprop/showcoef.php?cmpid=1386&prop=VSL.
Accessed June 14, 2017.
31.
32.
Weng WL, Chang LT, Shiah IM. Viscosities and densities for binary mixtures of benzylamine with 1-pentanol, 2-
pentanol, 2-methyl-1-butanol, 2-methyl-2-butanol, 3-methyl-1-butanol, and 3-methyl-2-butanol. J Chem Eng Data.
1999;44(5):994-997. doi:10.1021/je990031d.
Fisher Scientific. DBU MSDS, Fisher Scientific.
https://www.fishersci.com/shop/msdsproxy?productName=AC160610025&productDescription=1+8-
DIAZABICYCLO+5.4.0+U+2.5KG&catNo=AC160610025&vendorId=VN00032119&storeId=10652. Published
2014. Accessed June 14, 2017.
This article is protected by copyright. All rights reserved.

33.
34.
Brunn E, Huisgen R. Structure and Reactivity of the Betaine Derived from Triphenylphosphine and Dimethyl
Azodicarboxylate. Angew Chemie Int Ed English. 1969;8(7):513-515. doi:10.1002/anie.196905131.
Satish Kumar N, Praveen Kumar K, Pavan Kumar KVP, Kommana P, Vittal JJ, Kumara Swamy KC. Diverse
Modes of Reactivity of Dialkyl Azodicarboxylates with P(III) Compounds: Synthesis, Structure, and Reactivity of
Products Other than the Morrison-Brunn-Huisgen Intermediate in A Mitsunobu-Type Reaction. J Org Chem.
2004;69(6):1880-1889. doi:10.1021/jo035634d.
35.
36.
37.
38.
39.
But TYS, Toy PH. The Mitsunobu reaction: Origin, mechanism, improvements, and applications. Chem - An Asian
J. 2007;2(11):1340-1355. doi:10.1002/asia.200700182.
Swamy KCK, Kumar NNB, Balaraman E, Kumar KVPP. Mitsunobu and Related Reactions: Advances and
Applications. Chem Rev. 2009;109(6):2551-2651. doi:10.1021/cr800278z.
Fletcher S. The Mitsunobu reaction in the 21 st century. Org Chem Front. 2015;2(6):739-752.
doi:10.1039/C5QO00016E.
One example of a clinical radiotracer produced at our institution is [11C]methionine, which is obtained by 11C-
methylation using [11C]CH3I produced via the gas-phase method.
Downey J, Bongarzone S, Hader S, Gee AD. Oral Presentations (O 066). In: Journal of Labelled Compounds and
Radiopharmaceuticals. Vol 60. ; 2017:S88-S89. doi:10.1002/jlcr.3507.
This article is protected by copyright. All rights reserved.

Figure 1. Representative three-component [11C]CO2 trapping: trapping loop, vial and [11C]CO2 trapping cartridge (ascarite).
This article is protected by copyright. All rights reserved.

Effect of varying % benzylamine with fixed 10 % DBU (v/v) in MeCN
97.6
97.5
93.8
100
80
60
40
20
0
Loop-trapping
Total-solution trapping
62.7
24.4
27.0
10.9
2.1
0
1
5
10
% Benzylamine (v/v)
Figure 2. Effect of varying % benzylamine with fixed 10 % DBU (v/v) in MeCN.
This article is protected by copyright. All rights reserved.

Effect of varying % DBU with fixed 10% benzylamine (v/v) in MeCN
97.5
96.5
100
80
60
40
20
0
89.9
Loop-trapping
Total-solution trapping
44.1
41.8
26.8
27.0
1.6
0
10
50
90
% DBU (v/v)
Figure 3. Effect of varying % DBU with fixed 10% benzylamine (v/v) in MeCN.
This article is protected by copyright. All rights reserved.

Figure 4. Schematic of in-loop flow radiosynthetic setup showing the [11C]CO2 trapping loop as well as the additional
reaction loop to allow Mitsunobu reaction to form N,N’-[11C]dibenzylurea.
This article is protected by copyright. All rights reserved.