Continuous-flow synthesis of [11C]raclopride, a positron emission tomography radiotracer, on a microfluidic chip

Article abstract of DOI:10.1139/cjc-2012-0331

¹¹C-labelled radiotracers such as [¹¹C]raclopride are produced in a process that can take between 45 and 60 min to complete. These conventional approaches can consume upwards of 75% of the ¹¹C (t _1/2 = 20 min) due to radioactive decay alone, even more if synthesis losses are considered. To compensate, a large starting quantity of radioactive precursors such as [¹¹C]methyl iodide is required to produce an adequate amount of the tracer for injection. In this investigation, a continuous-flow microchip is explored for the purpose of synthesizing ¹¹C radiotracers in a shorter time by exploiting the favorable reaction kinetics of using smaller reaction volumes. To enhance the mixing of reagents within the microchannel, a micromixer loop design was used in fabricating various polydimethylsiloxane chip styles. With a loop design implemented in an abacus-style chip for the production of nonradioactive raclopride, shorter reaction times, reduced precursor use, and improved yields were possible when compared with the use of a simple serpentine design (no loop-style chip). However, when performing the equivalent radiochemical reaction, the results were not as favorable. Using the loop design in a full loop-style chip, parameters such as premixing the reagents, reducing flow rate, and varying reagent concentrations were explored to improve the yields of [ ¹¹C]raclopride (in terms of relative radioactivity) formed. The full loop chip design produced the best results, and future work will see the polydimethylsiloxane prototype chip design translated into a glass chip for further optimization.

Full text of DOI:10.1139/cjc-2012-0331

326
ARTICLE
Continuous-flow synthesis of [11C]raclopride, a positron emission
tomography radiotracer, on a microfluidic chip
Samar Haroun, Zahra Sanei, Salma Jivan, Paul Schaffer, Thomas J. Ruth, and Paul C.H. Li
Abstract: 11C-labelled radiotracers such as [11C]raclopride are produced in a process that can take between 45 and 60 min to
complete. These conventional approaches can consume upwards of 75% of the 11C (t1/2 = 20 min) due to radioactive decay alone,
even more if synthesis losses are considered. To compensate, a large starting quantity of radioactive precursors such as
[11C]methyl iodide is required to produce an adequate amount of the tracer for injection. In this investigation, a continuous-flow
microchip is explored for the purpose of synthesizing 11C radiotracers in a shorter time by exploiting the favorable reaction
kinetics of using smaller reaction volumes. To enhance the mixing of reagents within the microchannel, a micromixer “loop”
design was used in fabricating various polydimethylsiloxane chip styles. With a loop design implemented in an abacus-style chip
for the production of nonradioactive raclopride, shorter reaction times, reduced precursor use, and improved yields were
possible when compared with the use of a simple serpentine design (no loop-style chip). However, when performing the
equivalent radiochemical reaction, the results were not as favorable. Using the loop design in a full loop-style chip, parameters
such as premixing the reagents, reducing flow rate, and varying reagent concentrations were explored to improve the yields of
[11C]raclopride (in terms of relative radioactivity) formed. The full loop chip design produced the best results, and future work
will see the polydimethylsiloxane prototype chip design translated into a glass chip for further optimization.
Key words: microfluidic reactor, positron emission tomography (PET) imaging, radiotracer, raclopride, polydimethylsiloxane
(PDMS) chip.
Résumé : Les traceurs radioactifs marqués au 11C, tel le [11C]raclopride, sont produits selon un procédé pouvant durer de 45 a`
60 min. Il arrive que ces méthodes conventionnelles consomment jusqu'a` 75 % du 11C (t1/2 = 20 min.) rien que par désintégration
radioactive, et davantage si l'on tient compte des pertes de synthèse. Pour compenser ces pertes, on doit utiliser une grande
quantité initiale de précurseurs radioactifs, comme l'iodure de [11C]méthyle, afin de produire une quantité adéquate de traceur
pour l'injection. La présente étude porte sur l'emploi d'un microréacteur a` flux continu pour synthétiser plus rapidement des
traceurs radioactifs marqués au [11C] en tirant parti de la cinétique de réaction favorable due a` l'utilisation de plus petits volumes
de réaction. Pour améliorer le mélange des réactifs dans les microcanaux, on a utilisé un dessin de micromélangeur a` « boucles »
pour fabriquer divers styles de puce en polydiméthylsiloxane. L'utilisation du dessin a` boucles dans une puce de style abaque
pour produire du raclopride non radioactif a permis de réduire les temps de réaction ainsi que la consommation de précurseurs
et d'augmenter le rendement comparativement a` l'utilisation d'un simple dessin en serpentin (puce de style sans boucles).
Cependant, lors de l'exécution de la réaction radiochimique équivalente, les résultats n'ont pas été aussi encourageants. En se
servant du dessin a` boucles dans une puce de style entièrement en boucles, on a examiné des paramètres tels que le prémélange
des réactifs, la réduction du débit et diverses concentrations de réactifs en vue d'améliorer le rendement en [11C]raclopride
(mesuré par la radioactivité relative). La puce entièrement en boucles est celle qui a donné les meilleurs résultats, et dans le cadre
de futurs travaux, le prototype en polydiméthylsiloxane sera converti en une puce en verre afin de poursuivre l'optimisation.
[Traduit par la Rédaction]
Mots-clés : réacteur microfluidique, tomographie par émission de positrons (TEP), traceur radioactif, raclopride, puce en polydi-
méthylsiloxane (PDMS).
in neurology for amyloid plaque imaging, and L-[methyl-11C]-
Introduction
methionine ([11C]MET) is used in oncology for tumor imaging.1–3
Positron emission tomography (PET) is a sensitive technique
Currently, the challenge with C-11 radiotracer production is in
used for the in vivo monitoring of biochemical and physiological
developing a rapid, efficient, and safe radiolabelling platform.2
processes. This nuclear imaging technique uses short-lived posi-
tron emitting isotopes such as carbon-11 (11C) (t1/2 = 20 min) in
synthesized radiotracers that are administered to patients for
research or diagnostic purposes. PET has demonstrated uses in
various areas of medicine, e.g., [11C]raclopride ([11C]Rac) is used
in psychiatry for dopamine receptor imaging, [11C]meta-
hydroxyephedrine is used in cardiology for adrenergic receptor
imaging, 11C-labelled Pitsburgh compound B ([11C]PiB) is used
Since the radiolabelled precursors are short-lived, any production
processes completed within 2–3 half-lives can result in over 75%–
88% radioactivity loss through radioactive decay.4 Therefore, large
amounts of initial activities are needed to compensate for the loss
in decay and excess amounts of expensive precursor (102–106 mol
difference between the precursor and 11C-labelled methyl iodide
([11C]MeI)) are used to accelerate the radiolabelling process.5 Re-
Received 23 August 2012. Accepted 31 December 2012.
S. Haroun, Z. Sanei, and P.C.H. Li. Department of Chemistry, Simon Fraser University, Burnaby, BC V5A 1S6, Canada.
S. Jivan. Nuclear Medicine Division, TRIUMF, Vancouver, BC V6T 2A3, Canada.
P. Schaffer and T.J. Ruth. Department of Chemistry, Simon Fraser University, Burnaby, BC V5A 1S6, Canada; Nuclear Medicine Division, TRIUMF, Vancouver, BC V6T 2A3,
Canada.
Corresponding author: Paul C.H. Li (e-mail: paulli@sfu.ca).
Can. J. Chem. 91: 326–332 (2013) dx.doi.org/10.1139/cjc-2012-0331
Published at www.nrcresearchpress.com/cjc on 17 January 2013.

Haroun et al.
327
Scheme 1. [11C]Rac synthesis using [11C]MeI.5,31,33
CH3
CH3
OH
O
OH
Cl
O
O
Cl
Cl
N
N
11CH3I
NH
NH
DMSO/NaOH
OH
11CH3
Cl
cently, microfluidic reactors have shown improved synthesis per-
formance and product purities over conventional processes.6,8–11
This is attributed to the enhanced mass and heat transfer because
of the increased surface area to volume ratio in the microchannels
(up to 50 000 m2/m3 compared with conventional laboratory reac-
tors at 1000 m2/m3).4,7,12–21 The microfluidic reactor geometry
facilitates the optimization of reaction conditions such as temper-
ature, flow rate, and residence time.17–21 For example, Ratner et al.
used a simple serpentine microfluidic reactor to explore a glyco-
sylation reaction where the product stereochemistry is dependent
on temperature, solvent choice, stoichiometry, and concentra-
tion.12,22 Using the microfluidic reactor, they were able to reduce
reagent quantities compared with conventional methods and ef-
ficiently handle microlitres of volume. Also, using a single batch
of reagents usually used for one conventional experiment, they
were able to carry out up to 44 experiments at varied experimen-
tal conditions. Furthermore, maximum conversions were
achieved within a short residence time (26–213 s).22 With an en-
closed system and controlled reaction conditions, they were also
able to retrieve data that improved their understanding of the
reaction kinetics.12,22
The development of a quick process using microfluidics for
PET radiotracer production has attracted a great deal of atten-
tion.12,23–28 Miniaturization of PET radiolabelling processes has
the potential of reducing probe precursor amounts as well as
providing better fluid handling and control over reaction con-
ditions.4,6,29 In addition, the small chip dimensions can reduce
workspace and facilitate safer and more localized shielding re-
quirements.4,7,11,23 11C radiolabelling examples in the literature
have shown reaction times and radiochemical yields (RCY) that
are similar to the conventional methods. For example, Lu et. al.
were able to radiolabel a carboxylic ester using [11C]MeI on a
continuous-flow T-shaped microchip that is 200 ␮m wide and
60 ␮m deep with a total volume capacity of 0.2 ␮L. The precursors
3-(3-pyridinyl)propionic acid and [11C]MeI were both introduced
through separate inlets, and at a 1 ␮L/min flow rate, they were able
to achieve RCYs as high as 88%.14 This was completed using re-
duced reagent amounts in a residence time of 12 s and total pro-
cessing time of 10 min at room temperature. However, to the best
of our knowledge, there have been no investigations of [11C]Rac
synthesis on a microfluidic chip.
Experimental
General
Rac was purchased from ABX GmBH (Radeberg, Germany) and
the free base DMR precursor was purchased from JML Biopharm
Inc. (Vancouver, British Columbia). Both reagents were stored at
−20 °C until use. The SU-8 50 photoresist and SU-8 developer solu-
tion used in the microchip molding master fabrication were pur-
chased from MicroChem (Newton, Massachusetts). Sylgard 184
crosslinking silicone prepolymer and the curing agent used to
make the polydimethylsiloxane (PDMS) chips were acquired from
Dow Corning (Midland, Texas). HPLC-grade acetonitrile (MeCN)
was purchased from Caledon (Georgetown, Ontario). Extra dry
sure-sealed bottle solvents DMSO and MeCN used as reaction sol-
vents were purchased from Sigma Aldrich (Oakville, Ontario).
Fabrication of polymer–glass microfluidic chip
A micromixer design, initially explored by Garstecki et al.35 for
the purpose of mixing samples, was adopted for the radioactive
synthesis of [11C]Rac in this work. It was anticipated that the mi-
cromixer loop design would induce more mixing as compared
with a serpentine or a T-shaped microfluidic reactor chip, previ-
ously used for radiolabelling.6,14,28,29,36–38 Three microchip designs
(shown in Fig. 1) were made: no loop (or serpentine, no micromixer
loop designs were incorporated), abacus (a combination of micro-
mixer loop design and serpentine design was incorporated in the
microchannel), and full loop (the micromixer loop design is incorpo-
rated throughout all of the microchannels). The split and recombine
design of the microreactor loop design is expected to induce passive
mixing through splitting and recombining the fluid streams. There-
fore, the full loop chip that has more microreactor loops is expected
to have more efficient mixing and high yield. PDMS was used to
prototype the microreactor design using a single-layer mold to fab-
ricate the chip.39–40 Once the PDMS mold (1.2 mm thick) was created,
it was bonded permanently to a glass wafer (0.6 mm) using O2 plasma
treatment. The dimensions of the PDMS–glass microchips were 45
mm (length) × 15 mm (height) × 1.8 mm (thickness) with a channel
depth of 100 ␮m. Along with its fast prototyping time, the polymer
(PDMS) material was used due to its low cost of fabrication and its
compatibility with the solvents (DMSO and MeCN) used in the
[11C]Rac synthetic process.
In this work, we investigate the radioactive synthesis of a 11C
radiotracer, [11C]Rac. This is an important compound for brain
research due to its high binding affinity and selectivity as a dopa-
mine D2/3 receptor antagonist.7 Currently, [11C]Rac is synthesized
in a methylation reaction using [11C]MeI (see Scheme 1).14 The
reaction consumes up to 1.7 mg of expensive precursor (desmethyl
raclopride (DMR)) with a total production time of up to 45 min,
limiting the commercial availability of [11C]Rac.5,30–34 Our goal is
to explore the use of a continuous-flow microreactor chip for the
rapid, efficient, and safe radiosynthesis of [11C]Rac. A continuous-
flow microfluidic reactor system involves the continuous pump-
ing of fluids into the microchannels and collection of product
from the outlets (plug-flow).10 This type of microreactor is found
to shorten the reaction time, reduce the precursor consumption,
and allow for safe and remote operation of a disposable microre-
actor chip.
The microchip was mounted in an in-house built Al microchip
holder for interfacing to the fluid delivery system. The fluidic
connections were made using polyetheretherketone (PEEK) (Fisher
Scientific, Ottawa, Ontario) tubing (outside diameter = 1.6 mm) with
an inlet tubing inside diameter of 175 ␮m and an outlet tubing
inside diameter of 400 ␮m. The larger outlet tubing inside diam-
eter was used to avoid pressure buildup (see Fig. 2).6
Microfluidic nonradioactive synthesis
Two solutions were made for Rac synthesis: (1) 0.8 mg of DMR
(2.4 ␮mol), unless otherwise noted, was dissolved in 400 ␮L of
DMSO and mixed with 7 ␮L of 5 mol/L NaOH solution and (2)
120 ␮mol of MeI added to HPLC-grade MeCN (750 ␮L). Both solutions
were loaded on two separate syringes and the experiment was
carried out using a setup similar to the one shown in Fig. 2. A flow
rate of 8 ␮L/min on each pump (16 ␮L/min total for both) was used.
Published by NRC Research Press

328
Can. J. Chem. Vol. 91, 2013
Fig. 1. Images of the microchips of three micromixer designs
(45 mm × 15 mm): (a) no loop chip (total channel length = 272 mm,
volume = 2.7 ␮L), (b) abacus chip (length = 437 mm, volume = 4.4 ␮L),
and (c) full loop chip (length = 356 mm, volume = 3.6 ␮L). All
channels (100 ␮m wide and 100 ␮m deep) were filled with blue dyed
solutions.
product sample of 40 ␮L (or 80 ␮L) could be quenched immedi-
ately after collection.
HPLC product analysis was carried out on a UV-RAD Waters
HPLC instrument (Waters 200E; Waters Limited, Mississauga,
Ontario).
Results and discussion
No Loop
(a)
Rac nonradioactive synthesis
Initially, the nonradioactive synthesis was explored on two pro-
totypes of the PDMS–glass microchips (i.e., no loop and abacus).
Although glass chips are the preference for organic synthesis, we
used PDMS chips in this work because of the advantage of fast
chip prototyping. The plan is to translate the best design onto a
glass chip for future work.
After the microchip synthesis, product mixtures were analyzed
using HPLC. As shown in Fig. 3, the peaks of Rac, DMR, and the IS
are well separated. The IS was used to determine the peak area
ratios of Rac/IS, and these together with the amount of the limit-
ing reagent DMR were used to calculate the Rac yield. As shown in
Table 1, at the same temperature of say 82 °C, the abacus chip
produced a yield of 5.8%, which was higher than that of the no
loop chip. It was thus confirmed that the microchip that incorpo-
rated the micromixer loop design gave higher yields compared
with the serpentine or no loop microchip reported in the liter-
ature. It is believed that the micromixer loop design induces
mixing through the split–recombine path of the reagent streams.
Furthermore, the continuous-flow microchip allowed us to ex-
plore the effect of temperature through multiple runs using a
single batch of DMR and MeI, indicating that the Rac yield in-
creased with temperature.
The microchip Rac synthesis process was also compared with
the conventional technique. As shown in Table 2, an experiment
on the abacus microchip showed a Rac yield of 5.8%, which is
comparable with that of the conventional method, conducted at a
slightly lower temperature using a previously reported proce-
dure.5,41 More importantly, the microchip method was achieved
in a shorter time (one third), with a reduced amount of MeI (ϳ1/20)
needed.
Abacus
(b)
Full Loop
(c)
Microchip radioactive synthesis of [11C]Rac
The microchip with the microchip holder was placed on a hot
plate, with the temperature monitored using a calibrated thermo-
couple. During collection of product with a volume of 40 ␮L (or
80 ␮L, depending on the collection time), it was immediately
quenched using 100 ␮L (or 200 ␮L) of 0.1 mol/L ammonium for-
mate (HC(O)ONH4) solution. Multiple runs at various tempera-
tures were carried out.
The product samples were then analyzed using HPLC (Agilent
1100; Agilent Technologies Canada Inc., Mississauga, Ontario). The
nonradioactive Rac yield was then quantified using an internal
standard (IS), 2,4-dihydroxybenzoic acid.
After investigating the nonradioactive synthesis on the micro-
chip, the abacus chip, but not the no loop chip, was selected to
conduct radioactive synthesis. Figure 4 shows the UV-HPLC and
RAD-HPLC results. On the lower trace, a Rac cold (12C) mass peak
was not observed, indicating that the concentrations were below
detectable levels in each product sample. Meanwhile, on the in-
verted RAD-HPLC trace, the corresponding radiolabelled products
[11C]Rac and unreacted [11C]MeI were observed.
As substantial 11C side products were formed and the yield (now
based on the limiting reagent [11C]MeI) was not optimized, we
adopted the relative activity of [11C]Rac (as defined in eq. 1) for
comparison purposes.
Microfluidic radioactive synthesis
The [11C]Rac relative activity equation accounts for 11C activity
used to make [11C]Rac from [11C]MeI:
The radioactive synthesis of [11C]Rac was conducted in the setup
shown in Fig. 2. In this case, an in-house made aluminum rotating
collection vial holder, capable of holding 12 vials, was used to
facilitate vial switching during consecutive collections and to de-
crease radiation exposure due to reduced contact with vials by the
user. DMR (0.8 mg or 2.4 ␮mol, unless otherwise noted) was dis-
solved in 250 ␮L of extra dry DMSO in a reaction vial and mixed
with 7 ␮L of 5 mol/L NaOH. On the other hand, 50–80 mCi of
[11C]MeI, which varied by the different amounts of [11C]CH4 sup-
plied by the cyclotron, was trapped in 250 ␮L of extra dry MeCN.
With a specific activity of 10 Ci/␮mol, the radioactivity of [11C]MeI
translated to 5–8 nmol.5 For each batch of DMR, multiple runs
were conducted using various flow rates and collection times. A
100 ␮L (or 200 ␮L) amount of 0.1 mol/L ammonium formate
(HC(O)ONH4) solution was added to each collection vial so that the
[1]
[11C]Rac relative activity (%)
[11C]Rac RAD-HPLC peak area
ϭ
× 100
Total RAD-HPLC peak areas
As shown in Table 3, three consecutive experiments were con-
ducted at a flow rate of 10 ␮L/min (with a residence time of 13 s).
The relative activities of [11C]Rac in the three runs are quite repro-
ducible, giving rise to 2.0%. This low value is in part due to the
underutilization of [11C]MeI, resulting in substantial relative activ-
ity of unreacted [11C]MeI. Nonradioactive MeI was, however, unde-
tected in UV-HPLC due to the low UV absorbance of MeI.
Published by NRC Research Press

Haroun et al.
329
Fig. 2. Microchip synthesis reaction setup in which two syringe pumps are used to deliver the precursor solution (DMR in DMSO) and [11C]MeI
solution ([11C]MeI in MeCN).
Syringe
pumps
Chip in Al
[11C]MeI/MeCN
holder
Collection
vial holder
Precursor
solution
Hot Plate
Fig. 3. UV-HPLC chromatogram of a standard mixture containing 2,4-dihydroxybenzoic acid (IS) – DMR – Rac standard solution. The IS, DMR,
and Rac concentrations used were 130, 523, and 480 ␮mol/L, respectively. The HPLC conditions employed a C18 column (Zorbax, 4.6 mm × 150 mm,
Agilent) and a UV detector (␭ = 240 nm). The mobile phase composition for analysis was 5 mmol/L sodium dihydrogen phosphate buffer (pH ϳ 3.0)
and MeCN (70:30).
DMSO
DMR
Rac
IS
Time (min)
Table 1. Rac nonradioactive synthesis results on
the no loop and abacus microchip designs using
2.4–3.9 ␮mol of DMR and 7.5 ␮L (120 ␮mol) of MeI
in 750 ␮L of MeCN at 62–92 °C.
Table 2. Rac reaction conditions and results com-
parison in the conventional and microchip nonra-
dioactive synthesis.
Abacus
Rac yield (%)
Parameter
Conventional microchip
DMR (␮mol)
MeI (␮mol)
Reaction time (min)
Temperature (°C)
Rac yield (%)
aMultiple runs (n = 2) for this conventional process
showed a 4.4% 0.2% yield.
4.2
120
15
0.2
6.4
5
82
5.8
Temperature (°C)
No loopa
Abacusb
62
72
82
92
2.8 0.7c
3.3 0.1c
5.8
0.7
1.3
2.2
78
4.6a
a3.9 ␮mol of DMR was used.
b2.4 ␮mol of DMR was used.
cErrors represent standard deviation of two or three
experimental results. In other cases, errors are <0.7% ac-
cording to the calibration curve.
2 ␮L/min increased the residence time from 13 to 66 s and the
relative activity of [11C]Rac from 2.0% to 5.1%.
Although the relative radioactivity of unreacted MeI decreased
from 84.0% to 51.8%, that of the undesirable 11C side products
increased from 14.0% to 43.1%. This suggested that the increased
residence time also enhanced side product formation, possibly
through other methylation processes or decomposition of the
radiolabelled products.25 The enhanced side product formation
In an attempt to improve the yield of [11C]Rac, we decreased the
rate of fluid delivery into the microchannels or the liquid flow
rate. It was believed that the decrease in flow rate increased the
residence time, leading to more methylation products and less
unreacted [11C]MeI. For instance, a decrease in flow rate from 10 to
Published by NRC Research Press

330
Can. J. Chem. Vol. 91, 2013
Fig. 4. UV-HPLC and RAD-HPLC for [11C]Rac synthesis on the abacus microchip at 82 °C and a flow rate of 1 ␮L/min in flow mix mode. The
HPLC conditions are a C8 column (Zorbax-SD, 3.5 mm × 150 mm, Agilent) and UV detection at ␭ = 254 nm and a radioactivity detector (NaI(Tl)/
PMT, FC-6106, photodiode). The mobile phase consisted of 5 mmol/L potassium dihydrogen phosphate buffer (pH ϳ 2.8) and MeCN (63:37).
[11C]Rac
UV-HPLC
RAD-HPLC
[11C]MeI
6.7%
42.2%
11C – side
products
No Rac
DMR
1 AU
Peak
0
1
2
3
4
5
Time (min)
Note: The solvent DMSO was detected at ~1.8 min
Table 3. [11C]Rac synthesis results using the aba-
cus microchip where 0.8–1.0 ␮mol of DMR and
[11C]MeI are reacted at 82 °C at varied flow rates.
and [11C]MeI in the microchannel. Instead of the reactant mol-
ecules interacting for the first time at the T-junction, they are
already interacting as they are introduced into the microchan-
nel. Premixing the reagents may also prevent [11C]MeI from
being too concentrated, which can prevent it from reacting
with the chip substrate rather than with DMR.
Flow rate Calculated
[11C]Rac relative
residence time (s)a radioactivity (%)
(␮L/min)
10
13
2.0
2.0
2.0
4.0
5.1
6.7
Another factor explored was the microchip design; we per-
formed synthesis on the full loop chip in addition to the abacus
chip. Results in Table 4 confirmed increased efficiency when pre-
mixing the reagents for the [11C]Rac synthesis using the full loop
microchip (25.6%) over the abacus design (16.0%) at a flow rate of
2 ␮L/min. This improvement was attributed to a higher ratio of
micromixer loops to channel length (loop number/length in
millimetres) in the full loop microchip (0.26 mm–1) as compared
with the abacus microchip design (0.18/mm). The higher relative
activity obtained with the full loop chip is attributed to the better
mixing and hence better utilization of [11C]MeI, as quantified by
the [11C]Rac conversion defined by eq. 2. This parameter deter-
mines the efficiency of converting [11C]MeI into 11C compounds,
not 11C side products, and disregards the unreacted [11C]MeI, i.e., a
higher conversion indicates less side products formed. The high-
est conversion of 25.6% 0.6% was observed when the reagents
were flowed in at 2 ␮L/min in the full loop chip. However, a con-
version of 25.6% meant a side product level of 74.4%, most likely
due to the radioactivity loss through the use of the PDMS polymer
as the microchip substrate (vide infra).
5
2
1
26
66
131
aResidence time is calculated using the flow rate and
the microchannel length.
indicates that the microchip flow mixing might be efficient for
other fast and simple methylation syntheses, while not perform-
ing efficiently with the kinetically slower formation of [11C]Rac,
resulting in the low relative activity of [11C]Rac.
For further investigation of the residence time effects, the reac-
tant solutions (DMR and [11C]MeI) were premixed before being
introduced into the microchannels. This premix mode showed an
increase in the [11C]Rac relative activity with reduced levels of
unreacted MeI in comparison with the flow mix mode. As shown
in Table 4, at a flow rate of 10 ␮L/min, the [11C]Rac relative activi-
ties increased by a factor of ϳ5 in the premix mode as compared
with the flow mix mode. The highest relative radioactivity of 16%
was observed when the reagents were flowed in at 2 ␮L/min.
Premixing the reactants increases the interaction time of DMR
Rac conversion accounts for the 11C side products produced
during the methylation process:
[11C]Rac RAD-HPLC peak area
Total RAD-HPLC peak areas Ϫ unreacted [11C]MeI RAD-HPLC peak areas
[2]
[11C]Rac conversion (%) ϭ
× 100
From the relative activity of [11C]Rac, we were able to calculate
amounts (ϳ1/10 [11C]MeI and ϳ1/20 DMR) as used in the conven-
tional method. Although the specific activity cannot be evaluated
due to the undetected levels of Rac cold mass, it is noted that the
final activity is comparable with doses used in some [11C]Rac ani-
mal studies in the literature.42–44 For example, the study dis-
played in one article uses doses of 20 MBq (or 500 ␮Ci) of [11C]Rac
the absolute activity of [11C]Rac (see Table 4). After starting with
7.48 mCi of [11C]MeI, the actual [11C]Rac activity was as high as
164 ␮Ci (411 ␮Ci after decay correction) for the abacus chip's pre-
mixed experiment at a flow rate of 10 ␮L/min. This absolute activ-
ity was obtained at a short time (3 min) and lower precursor
Published by NRC Research Press

Haroun et al.
331
Table 4. [11C]Rac conversion and absolute product radioactivity at varied flow rates using the abacus
and full loop microchip in the premix mode at 82 °C.
Flow rate Calculated
Microchip (␮L/min) residence time (s) radioactivity
[11C]Rac relative [11C]Rac
[11C]Rac absolute
conversion (%) activity (␮Ci)a
Abacus
10
5
2
13
26
66
53
10.0
11.2
16.0
25.6 0.6
20.5
16.8
16.9
25.6 0.6
411
386
330
441 138
Full loopb
2
aDecay-corrected.
bErrors for the full loop microchip represent a standard deviation of 4 (n = 4) experimental results.
activity to study the effect of amphetamine on the binding affinity
of dopamine D2 receptors over a 90 min period.42
References
(1) Available from http://www.gehealthcare.com/ [accessed 18 September 2010].
(2) Miller, P. W.; Long, N. J.; Vilar, R.; Gee, A. D. Angew. Chem. Int. Ed. 2008, 47,
8998. doi:10.1002/anie.200800222.
From the absolute radioactivity of [11C]Rac and the initial radio-
activity of [11C]MeI, both decay-corrected, we can determine the
RCY. In the case of the full loop chip at a flow of 2 ␮L/min, the RCY
was found to be 8.9% 2.0%. Our RCY value is lower than the
values of 17%–45% reported by conventional production meth-
ods,5,45 which can be partially attributed to the use of PDMS as the
chip substrate (see Supplementary material for activity loss mea-
surement). Since the PDMS microchip is a prototype, there is
room for improvement. Nevertheless, the finding that the full
loop chip is a better design is an encouraging result that will lead
us to future studies that involve the fabrication of the glass chip
with the full loop design for evaluation of the microchip synthesis
method.46
(3) Lang, T.; Hernandez-Pampaloni, M.; Hawkins, R.; Seo, Y.; Sayre, G.;
Harnish, R.; Streeper, T.; Saeed, I.; Schreck, C.; Dannoon, S.; Slater, J.;
Blecha, J.; VanBrocklin, H. Osteoporos. Int. 2011, 22, 59.
(4) Miller, P. W. J. Chem. Technol. Biotechnol. 2009, 84, 309. doi:10.1002/jctb.2061.
(5) Adam, M. J.; Jivan, S.; Huser, J. M.; Lu, J. Radiochim. Acta 2000, 88, 207.
doi:10.1524/ract.2000.88.3-4.207.
(6) Elizarov, A. M. Lab Chip 2009, 9, 1326. doi:10.1039/b820299k.
(7) Kegeles, L. S.; Abi-Dargham, A.; Frankle, G.; Gil, R.; Cooper, T. B.;
Slifstein, M.; Hwang, D. R.; Huang, Y. Y.; Haber, S. N.; Laruelle, M. Arch. Gen.
Psychiatry 2010, 67, 231. doi:10.1001/archgenpsychiatry.2010.10.
(8) deMello, A. J. Nature 2006, 442, 394. doi:10.1038/nature05062.
(9) Watts, P.; Wiles, C. Org. Biomol. Chem. 2007, 5, 727. doi:10.1039/b617327f.
(10) Wirth, T. Microreactors in organic synthesis and catalysis; Wiley-VCH: Wein-
heim, 2008.
(11) Wheeler, T. D.; Zeng, D. X.; Desai, A. V.; Onal, B.; Reichert, D. E.;
Kenis, P. J. A. Lab Chip 2010, 10, 3387. doi:10.1039/c0lc00162g.
(12) Hartman, R. L.; Jensen, K. F. Lab Chip 2009, 9, 2495. doi:10.1039/b906343a.
(13) Abdelgawad, M.; Wheeler, A. R. Transducers '07 & Eurosensors Xxi, Digest of
Technical Papers, Vols. 1 and 2; U913, 2007.
(14) Lu, S. Y.; Watts, P.; Chin, F. T.; Hong, J.; Musachio, J. L.; Briard, E.;
Pike, V. W. Lab Chip 2004, 4, 523. doi:10.1039/b407938h.
(15) Cherlo, S. K. R.; Sreenath, K.; Pushpavanam, S. Ind. Eng. Chem. Res. 2009, 48,
8678. doi:10.1021/ie900306j.
(16) Bruus, H. Theoretical microfluidics; Oxford University Press: Oxford, New
York, 2008.
(17) Haswell, S. J.; O'Sullivan, B.; Styring, P. Lab Chip 2001, 1, 164. doi:10.1039/
b104035a.
(18) Wilson, N. G.; McCreedy, T. Chem. Commun. 2000, 733. doi:10.1039/B001705L.
(19) Lu, H.; Schmidt, M. A.; Jensen, K. F. Lab Chip 2001, 1, 22. doi:10.1039/
b104037p.
(20) Ueno, M.; Hisamoto, H.; Kitamori, T.; Kobayashi, S. Chem. Commun. 2003,
936. doi:10.1039/B301638B.
(21) Hisamoto, H.; Saito, T.; Tokeshi, M.; Hibara, A.; Kitamori, T. Chem. Commun.
2001, 2662. doi:10.1039/B106494K.
(22) Ratner, D. M.; Murphy, E. R.; Jhunjhunwala, M.; Snyder, D. A.; Jensen, K. F.;
Seeberger, P. H. Chem. Commun. 2005, 578. doi:10.1039/B414503H.
(23) Fletcher, P. D. I.; Haswell, S. J.; Pombo-Villar, E.; Warrington, B. H.; Watts, P.;
Wong, S. Y. F.; Zhang, X. L. Tetrahedron 2002, 58, 4735. doi:10.1016/S0040-
4020(02)00432-5.
(24) Lu S. Y.; Pike V. W. Ernst Schering Res Found Workshop, 271, 2007.
(25) Pascali, G.; Mazzone, G.; Saccomanni, G.; Manera, C.; Salvadori, P. A. Nucl.
Med. Biol. 2010, 37, 547. doi:10.1016/j.nucmedbio.2010.03.006.
(26) Miller, P. W.; Audrain, H.; Bender, D.; de Mello, A. J.; Gee, A. D.; Long, N. J.;
Vilar, R. Chem. Eur. J. 2011, 17, 460. doi:10.1002/chem.201002644.
(27) Pascali, G.; Nannavecchia, G.; Pitzianti, S.; Salvadori, P. A. Nucl. Med. and Biol.
2011, 38, 637. doi:10.1016/j.nucmedbio.2011.01.005.
(28) Wang, M. W.; Lin, W. Y.; Liu, K.; Masterman-Smith, M.; Shen, C. K. F. Mol.
Imaging 2010, 9, 175.
(29) Brady, F.; Luthra, S. K.; Gillies, J. M.; Jeffery, N. T.; Imaging Research Solu-
tions Ltd., 2003.
Conclusion
We have demonstrated that the use of a microchip maintains
advantages through improved yields and shorter reaction times
for the production of nonradioactive as well as [11C]raclopride.
Microfluidic synthesis has improved the nonradioactive synthesis
results in comparison with the conventional process by reducing
the reaction time by ϳ33% and required ϳ5% of the MeI used in
conventional preparation. The abacus microchip was found to be
a better design than the no loop or serpentine chip to provide a
better yield.
For the [11C]Rac synthesis, the full loop chip produced better
results, demonstrating higher [11C]Rac relative radioactivity and
higher conversion of [11C]MeI to [11C]Rac than the abacus micro-
chip. To evaluate the microchip method further, the reactants
were premixed and it was found that the full loop chip still pro-
duced better results than the abacus chip, reinforcing the impor-
tance of a micromixer design for this process. We have also
improved the [11C]Rac synthesis by reducing the flow rate to
2 ␮L/min. Although we have demonstrated the advantages of
lower precursor consumption and safer operation, this modifica-
tion would not be the most efficient solution on a production
scale, since a lower flow rate will decrease production speed. The
fabrication of a glass chip with the micromixer design is under-
way for the full investigation and optimization of the microchip
radiosynthesis of [11C]Rac and other labelled compounds. In addi-
tion, the behavior of reagents in the microchannels is further
explored in a computational fluid dynamics study.46
(30) Ehmann, W. D.; Vance, D. E. Radiochemistry and nuclear methods of analysis;
Wiley: New York, 1991.
(31) Fei, X. S.; Mock, B. H.; DeGrado, T. R.; Wang, J. Q.; Glick-Wilson, B. E.;
Sullivan, M. L.; Hutchins, G. D.; Zheng, Q. H. Synth. Commun. 2004, 34, 1897.
doi:10.1081/SCC-120034174.
(32) Langer, O.; Nagren, K.; Dolle, F.; Lundkvist, C.; Sandell, J.; Swahn, C. G.;
Vaufrey, F.; Crouzel, C.; Maziere, B.; Halldin, C. J. Labelled Comp. Radiopharm.
1999, 42, 1183. doi:10.1002/(SICI)1099-1344(199912)42:12<1183::AID-JLCR
274>3.0.CO;2-Z.
(33) Welch, M. J.; Redvanly, C. S. Handbook of radiopharmaceuticals: radiochemistry
and applications; J. Wiley: New York, 2003.
Supplementary material
Supplementary data are available with the article through the
journal Web site at http://nrcresearchpress.com/doi/suppl/10.1139/
cjc-2012-0331.
Acknowledgements
We are grateful to the Natural Sciences and Engineering Re-
search Council of Canada for financial support. We acknowledge
the Simon Fraser University technical support of Ken Van Wieren
(collection vial holder) and Frank Haftbaradaran (HPLC). We also
thank the TRIUMF Nuclear Medicine Division for all their support
including access to the TR13 cyclotron.
(34) Cheung, M. K.; Ho, C. L. Appl. Radiat. Isot. 2009, 67, 581. doi:10.1016/j.apradiso.
2008.08.018.
(35) Garstecki, P.; Fuerstman, M. J.; Fischbach, M. A.; Sia, S. K.; Whitesides,
G. M. Lab Chip 2006, 6, 207. doi:10.1039/b510843h.
Published by NRC Research Press

332
Can. J. Chem. Vol. 91, 2013
(36) Lee, C. C.; Sui, G. D.; Elizarov, A.; Shu, C. Y. J.; Shin, Y. S.; Dooley, A. N.;
Huang, J.; Daridon, A.; Wyatt, P.; Stout, D.; Kolb, H. C.; Witte, O. N.;
Satyamurthy, N.; Heath, J. R.; Phelps, M. E.; Quake, S. R.; Tseng, H. R. Science
2005, 310, 1793. doi:10.1126/science.1118919.
(37) Gillies, J. M.; Prenant, C.; Chimon, G. N.; Smethurst, G. J.; Dekker, B. A.;
Zweit, J. Appl. Radiat. Isot. 2006, 64, 333. doi:10.1016/j.apradiso.2005.08.009.
(38) Gillies, J. M.; Prenant, C.; Chimon, G. N.; Smethurst, G. J.; Perrie, W.;
Hambletta, I.; Dekker, B.; Zweit, J. Appl. Radiat. Isot. 2006, 64, 325. doi:10.
1016/j.apradiso.2005.08.007.
(42) Kilbourn, M. R.; Domino, E. F. Eur. J. Pharmacol. 2011, 654, 254. doi:10.1016/j.
ejphar.2011.01.008.
(43) Pedersen, K.; Simonsen, M.; Ostergaard, S. D.; Munk, O. L.; Rosa-Neto, P.;
Olsen, A. K.; Jensen, S. B.; Moller, A.; Cumming, P. NeuroImage 2007, 35, 38.
doi:10.1016/j.neuroimage.2006.11.038.
(44) Hoekzema, E.; Herance, R.; Rojas, S.; Pareto, D.; Abad, S.; Jimenez, X.;
Figueiras, F. P.; Popota, F.; Ruiz, A.; Torrent, E.; Soriano, F. J. F.; Rocha, M.;
Rovira, M.; Victor, V. M.; Gispert, J. D. Neuroscience 2010, 171, 1283. doi:10.
1016/j.neuroscience.2010.10.012.
(39) Hessel, V.; Lowe, H.; Schonfeld, F. Chem. Eng. Sci. 2005, 60, 2479. doi:10.1016/
j.ces.2004.11.033.
(40) Nguyen, N. T.; Wu, Z. G. J. Micromech. Microeng. 2005, 15, R1. doi:10.1088/0960-
1317/15/2/R01.
(45) Ishiwata, K.; Ishii, S.; Senda, M. Ann. Nucl. Med. 1999, 13, 195. doi:10.1007/
BF03164862.
(46) Haroun, S.; Wang, L.; Ruth, T. J.; Li, P. C. H. Chem. Eng. Process. Submitted 2012.
(41) Wilson, A. A.; Garcia, A.; Jin, L.; Houle, S. Nucl. Med. Biol. 2000, 27, 529.
doi:10.1016/S0969-8051(00)00132-3.
Published by NRC Research Press