Evaluation of tetraethylammonium bicarbonate as a phase-transfer agent in the formation of [18F]fluoroarenes

Article abstract of DOI:10.1016/j.jfluchem.2012.07.015

Tetraethylammonium bicarbonate, Et₄N·HCO₃, was found to be an efficient phase-transfer agent, in a microreactor, for the production of [¹⁸F]fluoroarenes from a range of precursors (diaryliodonium salt, nitroarene and [

Full text of DOI:10.1016/j.jfluchem.2012.07.015

Journal of Fluorine Chemistry 143 (2012) 231–237
Contents lists available at SciVerse ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Evaluation of tetraethylammonium bicarbonate as a phase-transfer agent in the
formation of [18F]fluoroarenes
*
Christopher D. Reed, Guillaume G. Launay, Michael A. Carroll
School of Chemistry, Bedson Building, Newcastle University, Newcastle upon Tyne NE1 7RU, UK
A R T I C L E I N F O
A B S T R A C T
Article history:
Tetraethylammonium bicarbonate, Et4NꢁHCO3, was found to be an efficient phase-transfer agent, in a
microreactor, for the production of [18F]fluoroarenes from a range of precursors (diaryliodonium salt,
nitroarene and [19F]fluoroarene). This study has established Et4NꢁHCO3 as a practical alternative to the
conventional phase-transfer system – Kryptofix1 222/K2CO3 – as the radiolabelled products were
generated in comparable radiochemical yields. The use of Et4NꢁHCO3 also dramatically increased
productivity by eliminating the frequent blockages of the microreactor experienced with the traditional
system.
Received 17 April 2012
Received in revised form 19 July 2012
Accepted 24 July 2012
Available online 15 August 2012
Keywords:
Radiochemistry
Microfluidics
Fluorinated aromatics
Fluorine-18
ß 2012 Elsevier B.V. All rights reserved.
Fluorination
1. Introduction
the routine production of PET imaging agents. A critical advantage
of microfluidic systems, in PET radiochemistry, when compared to
Positron-emission tomography (PET) is an imaging technique
for the absolute measurement, in vivo, of positron emitters,
enabling their pharmacokinetics and biodistribution to be
elucidated by non-invasive means and is rapidly becoming an
integral component of numerous aspects of the drug discovery/
development process [1]. The associated increase in demand for
PET radiopharmaceuticals has prompted the need to overcome the
complexities involved in the production of radiopharmaceuticals
bearing short-lived radioisotopes, such as fluorine-18
(t1/2 = 109.7 min) and carbon-11 (t1/2 = 20 min). The development
of efficient, automated processes is critical to the realisation of this
goal.
Microfluidic technology has recently emerged as an invaluable
tool in the development of PET radiochemical methods allowing
the quantities of precursors/reagents/solvents to be significantly
reduced facilitating timely isolation of the final radiolabelled
product [2]. This methodology, coupled with the strict control of
the reaction parameters, also has the potential to enhance
selectivity for the desired transformation over by-product
generation, improve radiochemical yields (RCYs) and thus increase
the purity of the target radiopharmaceutical. In addition, the high
degree of reproducibility inherent to this approach is essential for
conventional batch platforms, is that they enable multiple
reactions to be conducted from a single batch of radioisotope
allowing a degree of process optimisation not typically associated
with short-lived radioisotopes.
To expand the range of PET imaging agents available to our drug
discovery and development programmes we are exploring the
many benefits that microfluidic production could convey. The
focus of the initial studies was the direct translation of
conventional batch radiofluorination protocols to the microfluidic
system however this resulted in frequent blockages of the
microreactor [3]. In the reaction types investigated this was
attributed to the use of the traditional phase-transfer system,
Kryptofix1 222/K2CO3, and we now wish to report our evaluation
of an alternative system which has eliminated the problem and
dramatically improved productivity.
2. Results and discussion
Fluorine-18 is often referred to as the ‘radioisotope of choice’ in
PET as decay is almost exclusively by positron emission (b+:EC,
97:3), there are no specific issues arising from the decay product –
oxygen-18, the positron energy (0.635 MeV) is one of the lowest of
the commonly used positron-emitters resulting in a limited
positron range (ꢀ2 mm in water), which in turn results in the
highest image quality as this distance is comparable to the highest
spatial resolution possible with modern PET imaging systems [4].
* Corresponding author. Tel.: +44 0191 222 7074.
E-mail address: m.a.carroll@ncl.ac.uk (M.A. Carroll).
0022-1139/$ – see front matter ß 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jfluchem.2012.07.015

232
C.D. Reed et al. / Journal of Fluorine Chemistry 143 (2012) 231–237
O
OEt
O
OEt
readily prepared from a single experimental set-up once the
critical parameters are established. To demonstrate the translation
of our batch radiofluorination protocol to the microfluidic platform
(Advion NanoTek Microfluidic System [8]) we decided to utilise a
simple diaryliodonium salt as a model study [6e] (Scheme 1) prior
to the realisation of more complex targets.
The diaryliodonium salt needed for the model study was
prepared, via the aryltributylstannane, as a white crystalline solid
as shown in Scheme 2. We [6c,6f,6g] and others [7a,7c,7d,7f,7k]
have used DMF as the solvent for the fluorination of diaryliodo-
nium salts and therefore this was selected to perform the reactions
in this study. The conventional phase-transfer system for radio-
fluorinations was used (Kryptofix1 222/K2CO3 as a solution in
MeCN:H2O; 9:1, v/v) [9] and the diaryliodonium salt prepared as a
solution in DMF (5–10 mg/mL).
[
18F]Fluoride
1
I
18F
18F]2
CF3COO
S
[
Scheme 1. Formation of ethyl 4-[18F]fluorobenzoate [18F]2 from a diaryliodonium
salt.
O
OEt
O
OEt
b
a
1
The initial phase of this study used the microreactor setup as
described (Fig. 1) and was characterised by frequent blocked
microreactors which we attributed to the concentration of the
phase-transfer system employed as the diaryliodonium salt and
the reaction products are all highly soluble in DMF. Reducing the
concentration of the Kryptofix1 222/K2CO3 greatly improved the
situation [10] but the level of blocked reactors remained a concern
and an alternative phase-transfer system was needed.
Tetraalkylammonium salts with their enhanced solubility in
organic solvents have been widely used as an alternative to the
Kryptofix1 222/K2CO3 system with Bu4NꢁHCO3 being the most
common. This is usually prepared, with very limited characterisa-
tion, by treating an aqueous solution of the Bu4NꢁOH with carbon
dioxide [11]. Following elution of the [18F]fluoride from the anion
exchange resin (QMA) with the phase-transfer system the next
step is usually the generation of anhydrous [18F]fluoride via
azeotropic removal of the water using multiple distillations with
acetonitrile. Given the propensity for tetraalkylammonium salts,
and in particular the tetrabutylammonium salts, to form clathrates
I
SnBu3
3
4
Scheme 2. (a) (PPh3)4Pd, toluene, 110 8C, 59% and (b) TFA, diacetoxyiodo-2-
thiophene 5, DCM, ꢂ30 8C to RT, 82%.
The half-life of fluorine-18 is also sufficient to allow both the
distribution of imaging agents from a central cyclotron production
facility and the acquisition of scan data to take place over several
hours. As a result there is substantial interest in newmethods forthe
incorporation of fluorine-18 into diverse bioactive structures [5].
To this end, we [6] and others [7] have a particular interest in the
formationof [18F]fluoroarenes usingdiaryliodonium salt precursors,
as this versatile synthetic method has a number of distinct
advantages over conventional nucleophilic aromatic substitution.
It places little or no restriction on the nature of the target aromatic
ringoritssubstituents, for example, both electron-richand electron-
deficient [18F]fluoroarenes may be prepared. To date, this technolo-
gy also provides the only practical method for the production of
18F]fluoroarene derivatives not accessible using conventional
synthetic approaches [6g,7a,7b,7f,7j], is compatible with hetero-
aromatic ring systems [6d] and reactive functionality [6f].
The surprising consistency in the radiofluorination conditions
employed, coupled with an excellent level of substrate tolerance,
suggests that multiple, yet diverse [18F]fluoroarenes may be
[12] the generation of anhydrous
significant challenge.
[
18F]Bu4NꢁF represents
a
[
[
18F]Fluoroalkanes [13], 16-[18F]fluorohexadecanoic acid [14]
and [18F]FDG [15] have all been prepared successfully using the
related [18F]Et4NꢁF. This reagent had been prepared from Et4NꢁOH,
either by direct treatment with [18F]fluoride or, via [18F]Me3SiF to
generate an anhydrous reagent, however, given the complexity of
Waste
DMF
Waste
DMF
Auto-injector
Waste
Waste
6
1
Loop
plug
5
2
3
4
HPLC column
a
b
a
b
c
d
hg P3
e
f
c
d
hg P1
plug
plug
e
f
HPLC pump
Waste
Precursor in DMF
Loop 3
Loop 1
Micro-reactor
100 µm x 4 m
Concentrator
[
18F]Fluoride Transfer Line
e
d
c
[
18O]H2O recovery
C
f
a
b
Acetonitrile
PTS-1/PTS-2
0.010" ID PEEK (254 µm)
0.020" ID PEEK (508 µm)
0.030" ID PEEK (762 µm)
Check Valve
QMA cartridge
vent
Target Water
N2 Flow
Back Pressure Regulator (500 psi)
Pressure Transducer
Drying
Vial
Fig. 1. Set-up of the Advion NanoTek Microfluidic System for radiofluorination.

C.D. Reed et al. / Journal of Fluorine Chemistry 143 (2012) 231–237
233
due to the shorter alkyl chains making anhydrous [18F]Et4NꢁF a
more realistic proposition. Given these favourable characteristics
of Et4NꢁHCO3 over Bu4NꢁHCO3 we decided to evaluate this material
as a phase-transfer agent in the microfluidic radiofluorinations.
The first assessment was to determine how efficient Et4NꢁHCO3
was at eluting the [18F]fluoride from the anion exchange resin
(QMA) after it had been adsorbed from the cyclotron produced
[
[
18O]H2O. The amount of activity (5–100 mCi) and volume of
18O]H2O (1–5 mL) differed between batches, however, these
broad variations had no observed effect in elution efficiency with
Et4NꢁHCO3 able to elute > 99% (n = 10) of the radioactivity from the
QMA cartridge which is consistent with the performance of the
Kryptofix1 222/K2CO3 system [18].
With the efficient formation of [18F]Et4NꢁF established the next
step was to compare the Kryptofix1 222/K2CO3 system (PTS-1)
with the Et4NꢁHCO3 system (PTS-2) in the formation of [18F]2
(Scheme 1).
For this study the precursor solution (5–20
mL) and the phase-
transfer system (PTS-1 or PTS-2, 10 L) were infused into the
m
microreactor at the same time. It was noted that the concentration
of the diaryliodonium salt used (5–10 mg/mL) had little effect on
the outcome of the reaction [19] however the other parameters;
flow-rate, temperature and stoichiometry (P1:P3); did influence
the result obtained. It should also be noted that the actual
radiochemical yield obtained was found to vary somewhat
between different batches of [18F]fluoride, however the trend
observed remained the same (e.g. for Method B: increasing RCY
with temperature, decreased RCY with increased flow rate) [20].
We found that longer residence times were beneficial (Fig. 2a)
when using PTS-2 however this effect was not as pronounced
when PTS-1 was employed. Using optimum flow-rates both phase-
transfer systems displayed improved performance with increasing
temperature (Fig. 2b) and a stoichiometry of 1:1 was clearly found
to be superior for PTS-2 whereas PTS-1 was much more tolerant in
this regard.
These promising results demonstrated that Et4NꢁHCO3 system
(PTS-2) was a practical alternative to the Kryptofix1 222/K2CO3
system (PTS-1) for this type of transformation and it was
encouraging that blocked microreactors no longer occurred.
To demonstrate the general utility of PTS-2 for radiofluorina-
tion we also reviewed the 19F/18F isotopic exchange reaction [21]
which may provide rapid access to preliminary PET imaging
studies as the fluorine-19 derivative is often available directly from
drug discovery programmes [22] (Scheme 3).
Interestingly, for the isotope exchange reaction the effect of
flow-rate (Fig. 3a) was found to be opposite to that observed when
using the diaryliodonium salt precursor (Fig. 2a) with the longer
residence times benefiting PTS-1. However increasing the micro-
reactor temperature had a similar effect with the best radiochemi-
cal yields being obtained at 190 8C (Fig. 3b). For PTS-1 the effect of
the P1:P3 ratio was similar for both reaction types (Figs. 2c and 3c)
however the preference for the 1:1 stoichiometry for PTS-2 was
less pronounced (Fig. 3c).
Conventional nucleophilic aromatic substitution is a very
common approach to the formation of electron-deficient
Fig. 2. (a) Method A: PTS-1, 190 8C, 1 (5 mg/mL), P1:P3 1:1; Method B: PTS-2,
190 8C, 1 (10 mg/mL), P1:P3 1:1; (b) Method A: PTS-1, flow-rate (30 L/min), 1
(5 mg/mL), P1:P3 1:1; Method B: PTS-2, flow-rate (10 L/min), 1 (10 mg/mL), P1:P3
1:1; (c) Method A: PTS-1, flow-rate (25 L/min), 190 8C, 1 (5 mg/mL); Method B:
PTS-2, flow-rate (10
L/min), 190 8C, 1 (10 mg/mL). All result are n ꢃ 3.
m
m
m
m
O
OEt
O
OEt
[
18F]Fluoride
these production methods we proposed that access to [18F]Et4NꢁF
may be more readily achieved by the use of tetraethylammonium
bicarbonate (Et4NꢁHCO3) as the phase-transfer agent. Unlike the
tetrabutyl-derivative Et4NꢁHCO3 is readily available as a crystalline
solid [16] facilitating drying and characterisation yet retaining the
required solubility in anhydrous organic solvents. In addition
hydration of the tetraethylammonium cation is much reduced [17]
19F
19F]2
18F
18F]2
[
[
Scheme 3. Formation of ethyl 4-[18F]fluorobenzoate [18F]2 from [19F]2.

234
C.D. Reed et al. / Journal of Fluorine Chemistry 143 (2012) 231–237
Fig. 4. (a) Method A: PTS-1, 190 8C, 6 (10 mg/mL), P1:P3 1:1; Method B: PTS-2,
190 8C, 6 (10 mg/mL), P1:P3 1:1; (b) Method A: PTS-1, flow-rate (15 L/min), 6
(5 mg/mL), P1:P3 1:1; Method B: PTS-2, flow-rate (20 L/min), 6 (10 mg/mL), P1:P3
1:1; (c) Method A: PTS-1, flow-rate (10 L/min), 190 8C, 6 (5 mg/mL); Method B:
PTS-2, flow-rate (20
L/min), 190 8C, 6 (10 mg/mL). All results are n ꢃ 3.
Fig. 3. (a) Method A: PTS-1, 190 8C, [19F]2 (10 mg/mL), P1:P3 1:1; Method B: PTS-2,
190 8C, [19F]2 (10 mg/mL), P1:P3 1:1; (b) Method A: PTS-1, flow-rate (10
L/min),
19F]2 (5 mg/mL), P1:P3 1:1; Method B: PTS-2, flow-rate (20 L/min), [19F]2
(10 mg/mL), P1:P3 1:1; (c) Method A: PTS-1, flow-rate (10
L/min), 190 8C, [19F]2
L/min), 190 8C, [19F]2 (10 mg/mL). All
m
m
m
[
m
m
m
m
(5 mg/mL); Method B: PTS-2, flow-rate (20
results are n ꢃ 3.
m
[
18F]fluoroarenes. It is employed extensively where an electron-
Again, it was noticed that longer residence times gave improved
radiochemical yields but the effect was more pronounced when
PTS-1 was employed (Fig. 4a). Conducting the reaction at higher
temperatures resulted in excellent incorporation of the fluorine-18
for both phase-transfer systems (Fig. 4b). The optimum stoichi-
ometry peaked at 1:1.5 when using PTS-1 and 1:1 when using PTS-
withdrawing group exists in the 4-position relative to the site for
the introduction of fluorine-18, for example in the production of
[
result we also decided to evaluate PTS-2 for this widespread
method of fluorine-18 introduction (see Scheme 4).
18F]altanserin [23], [18F]MPPF [24], and [18F]setoperone [25]. As a

C.D. Reed et al. / Journal of Fluorine Chemistry 143 (2012) 231–237
235
O
OEt
O
OEt
chromatography (SiO2, hexane:diethyl ether 40:1) to yield the
product as a colourless oil (3.94 g, 8.97 mmol, 59%). Rf = 0.87
(petrol:EtOAc, 9:1); Anal. Calcd. for C21H36O2Sn: C, 57.43; H, 8.26.
[
18F]Fluoride
Found: C, 57.57; H, 8.35; IR (neat)
n 2925, 1719, 1273; dH
(400 MHz; CDCl3) 7.95 (2H, d, H2/H6, J 7 Hz), 7.54 (2H, d, H3/H5, J
7 Hz), 4.36 (2H, q, OCH2, J 7 Hz), 1.58–1.44 (6H, m, SnCH2CH2), 1.38
(3H, t, OCH2CH3, J 7 Hz), 1.31 (6H, m, CH2CH3), 1.17–0.97 (6H, m,
SnCH2), 0.87 (9H, CH2CH3, t, J 7 Hz); dC (101 MHz; CDCl3) 166.9
(CO), 149.3 (C4), 136.4 (C2/C6), 130.1 (C1), 128.5 (C3/C5), 60.7
(OCH2), 29.1 (SnCH2CH2CH2CH3), 27.4 (SnCH2CH2CH2CH3), 14.4
(OCH2CH3), 13.7 (SnCH2CH2CH2CH3), 9.7 (SnCH2CH2CH2CH3) dSn
(149 MHz; CDCl3) ꢂ38.5; m/z (ESI) 463 (M++Na, 13%), 301 (93)
(found: M++Na, 463.1630. C21H36O2116Sn requires 463.1633).
NO2
18F
18F]2
6
[
Scheme 4. Formation of ethyl 4-[18F]fluorobenzoate [18F]2 from the relevant
nitroarene.
2 (Fig. 4c) with the difference in the absolute radiochemical yields
being attributed to inter-batch variation of the starting [18F]fluo-
ride due to the similarities observed during the evaluation of the
other reaction parameters [20].
4.2. Diacetoxyiodo-2-thiophene (5)
Sodium perborate tetrahydrate (33.85 g, 220 mmol) was added
to a stirred solution of 2-iodothiophene (2.21 mL, 20 mmol) in
glacial acetic acid (45 mL) in small portions over 30 min. The
mixture was then heated to 50 8C, stirred for 5 h and then allowed
to cool to RT. The mixture was extracted with DCM (3 ꢄ 100 mL),
the organic fractions were combined, washed with water
(3 ꢄ 100 mL) and the organic layer concentrated in vacuo to give
the crude product (2.8 g, 8.53 mmol, 42%); Rf = 0.17 (DCM:MeOH,
96:4); m.p. = 118–121 8C (decomp) (from DCM–ether–petrol) (lit.
[28] 120–122 8C); Anal. Calcd. for C8H9IO4S: C, 29.28; H, 2.7. Found
3. Conclusion
In summary we have demonstrated, that Et4NꢁHCO3 is a viable
alternative to the traditional phase transfer system, Kryptofix1
222/K2CO3, for the production of [18F]fluoroarenes using a key
range of synthetic methods. Of particular benefit, in a microreactor,
was the performance of Et4NꢁHCO3 where the occurrence of
blocked reactors experienced with the conventional phase-
transfer system was eliminated thus dramatically increasing
productivity. In addition the Advion NanoTek Microfluidic System
provided an efficient, versatile and convenient methodology for
the rapid optimisation of radiofluorination reactions.
C, 29.18; H, 2.62; IR (neat)
7.77 (1H, dd, H5, J 4, 1 Hz), 7.63 (1H, dd, H4, J 5, 1 Hz,), 7.12 (1H, dd,
H3, J 5, 4 Hz), 2.01 (6H, s, COCH3); C (101 MHz; CDCl3) 177.1 (CO),
n 3080, 1634, 1267; dH (400 MHz; CDCl3)
d
139.4 (C3), 134.9 (C5), 128.7 (C4), 106.3 (C2), 20.4 ((OCOMe)2); m/z
(ESI) 269 (M+, 100%), 227 (27), 210 (98) (found: M+, 268.9130.
C6H6IO2S requires 268.9128).
4. Experimental
Reactions requiring anhydrous conditions were performed using
oven-dried glassware and conducted under a positive pressure of
nitrogen. Ethyl 4-nitrobenzoate (98%) and ethyl 4-fluorobenzoate
(99%) were purchased from Alfa Aesar (UK). Potassium carbonate
(99%), Kryptofix1 222 (4,7,13,16,21,24-hexaoxa-1,10-diazabicy-
clo[8.8.8]hexacosane; 98%), and tetraethylammonium bicarbonate
(ꢃ95%) were purchased from Sigma–Aldrich (UK) and used as
received. For radiofluorinations, acetonitrile (>99%) and dimethyl-
formamide (>99%) were purchased from Fisher Chemicals (UK) and
pre-driedonmolecularsieves(24 hon4 AMSthenstoredon3 AMS)
before being used, other anhydrous solvents were prepared in
accordance with standard protocols. Infrared spectra were
recorded on a Varian 800 FT-IR Scimitar Series spectrometer with
internalcalibration. 1H,13C, 19F, 119SnNMR spectrawererecordedon
a Jeol ECS 400 MHz spectrometer with residual protic solvent as an
internal reference for the 1H and 13C spectra, CFCl3 and Me4Sn as
external references for the 19F and 119Sn spectra respectively.
Elemental analyses were carried out at London Metropolitan
University and are reported as the average of two runs. Mass
spectra and accurate masses were recorded at the EPSRC
Mass Spectrometry Service, Swansea. It should be noted that
hypervalent iodine compounds are known to exist as mixtures (incl.
dimers/trimers) and these are evident in the MS data, only the data
for the monomer is reported. Melting points were recorded on a
Gallenkamp MF-370 melting point apparatus and are uncorrected.
Caution: hypervalent iodine compounds are potentially explosive
and should be handled taking appropriate precautions [26].
4.3. 4-Ethoxycarbonylphenyl(thiophen-2-yl)iodonium
trifluoroacetate (1)
Trifluoroacetic acid (0.74 mL, 10 mmol) was added dropwise, at
ꢂ30 8C and under N2, to a solution of 5 (1.96 g, 6 mmol) in DCM
(30 mL). After 30 min of stirring at ꢂ30 8C, the solution was
warmed to RT and stirred for 1 h when the solution was re-cooled
to ꢂ30 8C and 4 (1.72 mL, 5 mmol) was added. The resulting
mixture was allowed to warm to RT overnight and the solvent
removed in vacuo to give the crude product. Crystallisation gave 1
as a white crystalline solid (1.47 g, 4.1 mmol, 82%); Rf = 0.31
(DCM:MeOH, 93:7); m.p. 136–138 8C (decomp.) (from DCM–
ether–petrol); Anal. Calcd. for C15H12F3IO4S: C, 38.15; H, 2.56.
˚
˚
Found C, 38.21; H,2.61; IR (neat)
n 3087, 1704, 1651; dH (400 MHz,
d6-DMSO): 8.32 (2H, d, J 9 Hz), 8.05 (1H, dd, J 4, 1 Hz), 7.97 (2H, d, J
9 Hz), 7.94 (1H, dd, J 5, 1 Hz), 7.14 (1H, dd, J 5, 4 Hz), 4.28 (2H, q, J
7 Hz), 1.26 (3H, t, J 7 Hz); dC (101 MHz, d6-DMSO) 165.1 (C55O),
141.2 (C50), 138.0 (C40), 135.4 (C3/C5), 133.2 (C1), 132.3 (C2/C6),
130.2 (C30), 124.7 (C4), 101.9 (C20), 62.0 (OCH2), 14.56 (Me); dF
(376 MHz, d6-DMSO) ꢂ73.4 (COCF3). m/z (ESI) 358 (M+, 100%)
(found: M+, 358.9588. C13H12IO2S requires 358.9597).
4.4. Radiochemistry
4.4.1. Operation of the Advion NanoTek Microfluidic System
The radiosynthetic work, including the preparation and drying
of the [18F]fluoride was performed on the Advion NanoTek
Microfluidic System [8]. This is a modular-based microfluidic
system designed for the synthesis of PET imaging agents. The
system consists of a base module (BM), concentrator module (CM),
and reactor module (RM), all are operated and controlled by the
NanoTek 1.4 software. The BM consists of two reagent cartridges
(P1 and P2) used to dispense metered amounts of reactants to the
4.1. Ethyl 4-(tri-butylstannyl)benzoate (4) [27]
Pd(PPh3)4 (120 mg, 0.01 mmol) was added to a solution of ethyl
4-iodobenzoate (4.14 g, 15 mmol) and bis(tributyltin) (17.41 g,
30 mmol) in anhydrous toluene (100 mL). The solution was heated
at reflux for 24 h under nitrogen. The solvent was then removed in
vacuo providing the crude product which was purified by column

236
C.D. Reed et al. / Journal of Fluorine Chemistry 143 (2012) 231–237
microreactors for radiolabelling reactions. These cartridges com-
prise of a high-pressure syringe pump connected to an eight-way
bridged valve supporting a looped-reservoir from which reagents
are dispensed towards the microreactor. The RM consists of the
isotope reagent (P3) and microreactor cartridges (upto four), as
well as an eight-way distribution valve (DV) used to direct the
reaction bolus to the user-desired location. P3 receives the solution
of [18F]fluoride (PTS-1 or PTS-2) from the CM module where it is
prepared from a volume of cyclotron produced [18F]fluoride in
volume of system solvent at a preset flow rate. The solutions
initially mix at the entrance to the microreactor. User-operated
computer software determines the reaction parameters – temper-
ature, time and stoichiometry (P1:P3). Following completion of the
reaction, P3 sweeps the crude mixture to the DV where it is
directed towards the electronic injector and radio-HPLC system
allowing the radiochemical yield of the process to be determined.
At the start and end of each series of experiments the BM, RM,
microreactor and associated transfer lines were cleaned using DMF
and the CM system with acetonitrile.
[
18O]H2O. The CM module consists of a low-pressure, six-way valve
reagent cartridge (CM1) and vessel chamber capable of rapid
heating and evaporation of solvent. It is possible to subject the CM
vessel to a combination of reduced pressure and a positive flow of
nitrogen gas allowing the [18F]fluoride to be dried via several
azeotropic distillations with MeCN. The apparatus was setup as
described in Fig. 1.
4.4.4. Radioanalytical HPLC methods
HPLC methods were carried out using an Agilent 1200 HPLC
system equipped with a UV absorbance detector (lmax 254 nm)
and a radioactivity detector (LabLogic Flow Count). The [18F]fluor-
oarenes were not isolated and the radiochemical yield (RCY)
reported relates to the amount of radioactivity of the product
relative to the total radioactivity detected by radio-HPLC on
analysis of the reaction mixture.
Throughout this work, the crude reaction mixture was swept
into an electronic injection valve (Smartline Valve Drive: Knauer,
Germany) for direct injection onto the radio-HPLC system. The
enclosed fluid paths of this tandem synthetic and analytical system
ensured that the results obtained were representative of the crude
reaction mixture as potential volatile products could not be lost
[29]. Operation of this valve and the initialisation of the HPLC
analysis was also controlled by the NanoTek software. This
software may be operated in several modes; in ‘Discovery Mode’
user-directed macros automate all procedures involved in the
synthetic process, thereby allowing simple and rapid throughput
of multiple reactions. These procedures can also be divided into
their constituent parts and controlled separately within ‘Manual
Mode’ which, therefore, allows for finer control of the system. In
‘Sequence Mode’ detailed macros can be written, manipulated and
combined to enable extensive and complex processes to be
performed with minimal user intervention. Throughout this study,
user-control fluctuated between the three modes, for optimal
system management.
4.4.4.1. Ethyl 4-[18F]fluorobenzoate [18F]2 from 1. The radioactive
product was separated on a PolymerX 5
mm RP-1 100 column
(Phenomenex, USA) eluting at 1.1 mL/min with EtOH–water.
Mobile phase composition started at 25% EtOH which increased
linearly to 100% at EtOH over 7 min which continued at this
composition for a further 2 min. [18F]2 tR = 7 min. Other possible
components: thiophene tR = 6.5 min (not observed), ethylbenzoate
tR = 7.5 min (not observed), 2-iodothiophene tR = 8.5 min, ethyl 4-
iodobenzoate 3 tR = 9 min.
4.4.4.2. Ethyl 4-[18F]fluorobenzoate [18F]2 from [19F]2. The radio-
active product was detected on a PolymerX 5
(Phenomenex, USA) eluting at 1.1 mL/min with EtOH–water.
Mobile phase composition started at 60% EtOH which increased
linearly to 80% at EtOH over 6 min. [18F]2 tR = 4.5 min.
mm RP-1 100 column
4.4.2. Preparation of [18F]KF/Kryptofix1 222 or [18F]Et4NꢁF
For reactions conducted using the Advion NanoTek Microfluidic
System [8] no-carrier-added [18F]fluoride (5–100 mCi) in [18O]H2O
4.4.4.3. Ethyl 4-[18F]fluorobenzoate [18F]2 from 6. The radioactive
product was separated on a PolymerX 5
mm RP-1 100 column
(Phenomenex, USA) eluting at 1.1 mL/min with EtOH–water.
Mobile phase composition started at 40% EtOH which increased
linearly to 60% at EtOH over 2 min and then to 80% at 5 min which
continued at this composition for a further 7 min. [
18F]2
tR = 6.2 min. Other possible components: ethyl 4-nitrobenzoate 6
tR = 6.6 min.
(1–5 mL) [30] was adsorbed onto
exchange resin (QMA: Waters Sep-Pak1 Light Accell Plus) before
being released using a solution of either K2CO3 (1.18 mg; 8 mol)
and Kryptofix1 222 (6.75 mg; 18
mol) in MeCN/H2O (9:1 (v/v)
450 mol) in MeCN/H2O
L) (PTS-1); or Et4NꢁHCO3 (3.4 mg; 17
(9:1 (v/v) 450 L) (PTS-2) and into a V-vial (2 mL, Wheaton). The
solutions were dried by two successive azeotropic evaporations
a pre-conditioned anion
m
m
m
m
m
Acknowledgements
using MeCN (450
[
mL) at 100 8C and under a positive flow of N2. The
18F]KF/Kryptofix1 222 or [18F]Et4NꢁF was then dissolved in DMF
This research was supported in part (MAC, CDR) by a Cancer
Research UK/EPSRC/MRC/NIHR Cancer Imaging Programme (Ref.
C29821/A10348), the EPSRC (GGL, Ref. EP/H029753/1) and New-
castle University. We also thank the EPSRC for provision of the
Swansea Mass Spectrometry Service. We would also like to thank
Advion Biosciences and LabLogic UK for their continuing support
and Erigal Ltd. for their supply of fluorine-18.
(450
mL) and loaded into the storage loop (P3: 401 mL) on the
microfluidic system.
4.4.3. Radiofluorination of 1, [19F]2 and 6
Dry [18F]fluoride (5–100 mCi, 450
mL DMF) and precursor
solutions (1, [19F]2 or 6 in DMF at 10 mg/mL) were loaded into
their respective storage loops (loop 3 and loop 1 respectively).
Capillaries that lead from P1 and P3 to the microreactor were
primed with their respective reagents. Prior to recording data
several priming reaction runs were performed to confirm the
integrity of the synthetic platform and the in-line analytical HPLC
References
[1] (a) J.S. Fowler, A.P. Wolf, Accounts of Chemical Research 30 (1997) 181–188;
(b) M.E. Phelps, Proceedings of the National Academy of Sciences of the United
States of America 97 (2000) 9226–9233;
(c) S.M. Ametamey, M. Honer, P.A. Schubiger, Chemical Reviews 108 (2008)
1501–1516;
(d) Z. Li, P.S. Conti, Advanced Drug Delivery Reviews 62 (2010) 1031–1051;
(e) S.L. Pimlott, A. Sutherland, Chemical Society Reviews 40 (2011) 149–162.
[2] For examples see:
system. A set volume of either PTS-1 or PTS-2 (10
variable volume of the precursor (5–20 L: stoichiometry assess-
ment) were dispensed into the pre-heated microreactor (130–
190 8C) [31] at pre-determined flow rates (10–30 L/min). This is
achieved by the syringes of P1 and P3, simultaneously dispensing
precursor and [18F]fluoride solutions (using either PTS-1 or PTS-2)
from their respective loops, into the microreactor using a fixed
mL) and a
m
m
(a) S. Lu, V. Pike, in: P.A. Schubiger, L. Lehmann, M. Friebe (Eds.), PET Chemistry,
The Driving Force in Molecular Imaging, Springer, Berlin, 2007, pp. 271–287;
(b) G. Pascali, G. Mazzone, G. Saccomanni, C. Manera, P.A. Salvadori, Nuclear
Medicine and Biology 37 (2010) 547–555;

C.D. Reed et al. / Journal of Fluorine Chemistry 143 (2012) 231–237
237
(c) R. Bejot, A.M. Elizarov, E. Ball, J. Zhang, R. Miraghaie, H.C. Kolb, V. Gouverneur,
Journal of Labelled Compounds and Radiopharmaceuticals 54 (2011) 117–122;
(d) P.W. Miller, Journal of Chemical Technology and Biotechnology 84 (2009)
309–315;
[10] During the course of this study the radiofluorination of diaryliodonium salts
has been reported using the Advion NanoTek microfluidic system using
Kryptofix1 222/K2CO3 at concentrations similar to PTS-1. See references
[7c,7h,7k].
(e) H. Audrain, Angewandte Chemie International Edition 46 (2007) 1772–1775.
[3] For examples of blockages in microreactors see:
[11] P.A. Culbert, M.J. Adam, E.T. Hurtado, J.M.A. Huser, S. Jivan, J. Lu, T.J. Ruth, S.K.
Zeisler, Applied Radiation and Isotopes 46 (1995) 887–891.
[12] (a) G.A. Jeffrey, R.K. McMullan, in: F.A. Cotton (Ed.), Progress in Inorganic
Chemistry, vol. 8, John Wiley & Sons, Inc., Hoboken, NJ, USA, 1967, pp. 43–108;
(b) L.S. Aladko, Y.A. Dyadin, T.V. Rodionova, I.S. Terekhova, Journal of Molecular
Liquids 106 (2003) 229–238.
(a) C. Amador, A. Gavriilidis, P. Angeli, Chemical Engineering Journal 101 (2004)
379–390;
(b) Y. Tanaka, O. Tonomura, K. Isozaki, S. Hasebe, Chemical Engineering Journal
167 (2011) 483–489.
[4] S.E. Snyder, M.R. Kilbourn, in: M.J. Welch, C.S. Redvanly (Eds.), Handbook of
[13] S.J. Gatley, R.D. Hichwa, W.J. Shaughnessy, R.J. Nickles, The International Journal
of Applied Radiation and Isotopes 32 (1981) 211–214.
[14] A.L. Bosch, T.R. Degrado, S.J. Gatley, International Journal of Radiation Applica-
tions and Instrumentation – Part A: Applied Radiation and Isotopes 37 (1986)
305–308.
Radiopharmaceuticals: Radiochemistry and Applications, John Wiley
Ltd., Chichester, UK, 2003, pp. 195–227.
& Sons
[5] For examples see:
(a) L.Cai,S.Lu, V.W. Pike, European JournalofOrganic Chemistry(2008)2853–2873;
(b) P.W. Miller, N.J. Long, R. Vilar, A.D. Gee, Angewandte Chemie International
Edition 47 (2008) 8998–9033;
(c) L. Hoigebazar, J.M. Jeong, J.-Y. Lee, D. Shetty, B.Y. Yang, Y.-S. Lee, D.S. Lee, J.-K.
Chung, M.C. Lee, Journal of Medicinal Chemistry 55 (2012) 3155–3162;
(c) E. Lee, A.S. Kamlet, D.C. Powers, C.N. Neumann, G.B. Boursalian, T. Furuya, D.C.
Choi, J.M. Hooker, T. Ritter, Science 334 (2011) 639–642;
(d) H. Teare, E.G. Robins, A. Kirjavainen, S. Forsback, G. Sandford, O. Solin, S.K. Luthra,
V. Gouverneur, Angewandte Chemie International Edition 49 (2010) 6821–6824;
(e) B. Wa¨ngler, G. Quandt, L. Iovkova, E. Schirrmacher, C. Wa¨ngler, G. Boening, M.
Hacker, M. Schmoeckel, K. Jurkschat, P. Bartenstein, R. Schirrmacher, Bioconjugate
Chemistry 20 (2009) 317–321.
[15] L.G. Hutchins, A.L. Bosch, M.S. Rosenthal, R.J. Nickles, S.J. Gatley, The International
Journal of Applied Radiation and Isotopes 36 (1985) 375–378.
[16] Tetraethylammonium bicarbonate is commercially available: Sigma-Aldrich, cat.
No. 11268 (CAS: 17351-61-0). http://www.sigmaaldrich.com.
[17] (a) T.C.W. Mak, Journal of Inclusion Phenomena 3 (1985) 347–354;
(b) T.C.W. Mak, H.J. Bruins Slot, P.T. Beurskens, Journal of Inclusion Phenomena 4
(1986) 295–302.
[18] The efficiency of the elution of radioactivity from the QMA cartridge took into
account potential losses in the apparatus/process. The initial level of adsorption of
the [18F]fluoride from the cyclotron produced [18O]H2O onto the QMA cartridge
was determined by analysis of the recovered [18O]H2O and the radioactivity
remaining in the original supply vial. The result was also corrected in light of any
residual radioactivity on the QMA cartridge.
[6] (a) S. Martı´n-Santamarı´a, M.A. Carroll, C.M. Carroll, C.D. Carter, V.W. Pike, H.S.
Rzepa, D.A. Widdowson, Chemical Communications (2000) 649–650;
(b) M.A. Carroll, L.M. Kamara, D.A. Widdowson, V.W. Pike, Journal of Labelled
Compounds and Radiopharmaceuticals 48 (2005) 519–520;
(c) M.A. Carroll, J. Nairne, G. Smith, D.A. Widdowson, Journal of Fluorine Chem-
istry 128 (2007) 127–132;
[19] Precursor concentrations of 5–10 mg/mL are suitable for a range of diaryliodo-
nium precursors, we have observed precipitation of some precursors at concen-
tration ꢃ30 mg/mL.
[20] The variation between batches of [18F]fluoride has been noted by other research-
ers, for example see: G. Pascali, L. D’Antonio, P. Bovone, P. Gerundini, T. August,
Nuclear Medicine and Biology, 36 (2009) 569–574.
(d) M.A. Carroll, J. Nairne, J.L. Woodcraft, Journal of Labelled Compounds and
Radiopharmaceuticals 50 (2007) 452–454;
(e) R. Yan, M.A. Carroll, Journal of Labelled Compounds and Radiopharmaceu-
ticals 51 (2008) 259–260;
[21] E. Blom, F. Karimi, B. La˚ ngstro¨m, Journal of Labelled Compounds and Radio-
pharmaceuticals 52 (2009) 504–511.
(f) M. Carroll, R. Yan, F. Aigbirhio, D. Soloviev, L. Brichard, Journal of Nuclear
Medicine 49 (2008) 298P;
(g) M. Carroll, R. Yan, F. Aigbirhio, D. Soloviev, L. Brichard, Journal of Nuclear
Medicine 49 (2008) 303P.
[22] For examples of fluorine in medicinal chemistry see:
(a) I. Ojima, Fluorine in Medicinal Chemistry and Chemical Biology, John Wiley &
Sons Ltd., Chichester, UK, 2009;
(b) S. Purser, P.R. Moore, S. Swallow, V. Gouverneur, Chemical Society Reviews 37
(2008) 320–330;
[7] For selected examples see:
(a) T.L. Ross, J. Ermert, C. Hocke, H.H. Coenen, Journal of the American Chemical
Society 129 (2007) 8018–8025;
(b) M.-R. Zhang, K. Kumata, K. Suzuki, Tetrahedron Letters 48 (2007) 8632–8635;
(c) J.-H. Chun, S. Lu, Y.-S. Lee, V.W. Pike, Journal of Organic Chemistry 75 (2010)
3332–3338;
(c) D. O’Hagan, Journal of Fluorine Chemistry 131 (2010) 1071–1081;
(d) W.K. Hagmann, Journal of Medicinal Chemistry 51 (2008) 4359–4369.
[23] C. Lemaire, R. Cantineau, M. Guillaume, A. Plenevaux, L. Christiaens, Journal of
Nuclear Medicine 32 (1991) 2266–2272.
[24] D. Le Bars, C. Lemaire, N. Ginovart, A. Plenevaux, J. Aerts, C. Brihaye, W. Hassoun, V.
Leviel, P. Mekhsian, D. Weissmann, J.F. Pujol, A. Luxen, D. Comar, Nuclear Medi-
cine and Biology 25 (1998) 343–350.
(d) S. Telu, J.-H. Chun, F.G. Sime´on, S. Lu, V.W. Pike, Organic & Biomolecular
Chemistry 9 (2011) 6629–6638;
(e) F. Basuli, H. Wu, G.L. Griffiths, Journal of Labelled Compounds and Radio-
pharmaceuticals 54 (2011) 224–228;
[25] J. Blin, S. Pappata, M. Kiyosawa, C. Crouzel, J.C. Baron, European Journal of
Pharmacology 147 (1988) 73–82.
(f) T.L. Ross, J. Ermert, H.H. Coenen, Molecules 16 (2011) 7621–7626;
(g) B. Wang, R.L. Cerny, S. Uppaluri, J.J. Kempinger, S.G. DiMagno, Journal of
Fluorine Chemistry 131 (2010) 1113–1121;
(h) J.-H. Chun, V.W. Pike, Journal of Labelled Compounds and Radiopharmaceu-
ticals 54 (2011) S482;
(i) D. Zimmerman, T.L. Ross, Journal of Labelled Compounds and Radiopharma-
ceuticals 54 (2011) S488;
(j) K.D. Neumann, A. Vavere, S. Snyder, S.G. DiMagno, Journal of Labelled Com-
pounds and Radiopharmaceuticals 54 (2011) S492;
[26] (a) D.B. Dess, J.C. Martin, Journal of the American Chemical Society 113 (1991)
7277–7287;
(b) H. Tohma, S. Takizawa, T. Maegawa, Y. Kita, Angewandte Chemie Interna-
tional Edition 39 (2000) 1306–1308.
[27] P. Tang, T. Furuya, T. Ritter, Journal of the American Chemical Society 132 (2010)
12150–21254.
[28] H. Togo, T. Nabana, K. Yamaguchi, Journal of Organic Chemistry 65 (2000) 8391–
8394.
[29] For example fluorobenzene and 2-fluorothiophene have boiling points lower than
the reaction temperature and may therefore be lost from an open system, see:
M.A. Carroll, C. Jones, S.-L. Tang, Journal of Labelled Compounds and Radio-
pharmaceuticals, 50 (2007) 450–451.
(k) G. Pascali, S. Pitzianti, G. Saccomani, S. Del Carlo, C. Manera, M. Macchia, P.A.
Salvadori,Journalof LabelledCompoundsandRadiopharmaceuticals54(2011) S502.
[8] Advion Biosciences Ltd., Ithaca, New York. http://www.advion.com/biosystems/
nanotek/nanotek-positron-emission-tomography.php.
[30] [18F]Fluoride in [18O]H2O was supplied by Erigal Limited. http://www.erigal.co.uk.
[31] This temperature range was selected as it was found that limited reaction was
evident at the lower temperatures and the maximum operating temperature of
the Advion NanoTek Microfluidic System was 200 8C.
[9] Advion NanoTek LF 07-08 User Manual; Appendix A. [18F]Fluoride ion prepara-
tion; Kryptofix1 222 (30 mg/mL) and K2CO3 (5.5 mg/mL) in MeCN/H2O
(9:1, v/v).