N-(4-(di-tert-butyl[18F]fluorosilyl)benzyl)-2-hydroxy-N,N- dimethylethylammonium bromide ([18F]SiFAN+Br-): A novel lead compound for the development of hydrophilic SiFA-based prosthetic groups for 18F

Article abstract of DOI:10.1016/j.jfluchem.2010.10.008

¹⁸F-labeled compounds play a major role in the development of new in vivo imaging agents for Positron Emission Tomography (PET), a non invasive imaging modality depicting the biodistribution of radioactive compounds in humans. Recently we repor

Full text of DOI:10.1016/j.jfluchem.2010.10.008

Journal of Fluorine Chemistry 132 (2011) 27–34
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
N-(4-(di-tert-butyl[18F]fluorosilyl)benzyl)-2-hydroxy-N,
N-dimethylethylammonium bromide ([18F]SiFAN+Brꢀ): A novel lead compound
for the development of hydrophilic SiFA-based prosthetic groups for 18F-labeling
Alexey P. Kostikov a, Liuba Iovkova b, Joshua Chin a, Esther Schirrmacher a, Bjo¨rn Wa¨ngler c,
Carmen Wa¨ngler c, Klaus Jurkschat b, Gonzalo Cosa d, Ralf Schirrmacher a,
*
a McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, 3801 University Street, Montreal, Quebec H3A3B4, Canada
b
¨
Lehrstuhl fu¨r Anorganische Chemie, Technische Universitat Dortmund, Germany
c Department of Nuclear Medicine, Ludwig-Maximillians-University, Munich, Germany
d Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, QC H3A 2K6, Canada
A R T I C L E I N F O
A B S T R A C T
18F-labeled compounds play a major role in the development of new in vivo imaging agents for Positron
Emission Tomography (PET), a non invasive imaging modality depicting the biodistribution of
radioactive compounds in humans. Recently we reported a new method for the introduction of fluorine-
18 into a range of organic molecules exploiting the very fast 18F–19F isotope exchange of fluorosilanes
(termed SiFA compounds). Here, we wish to report the labeling of the first charged SiFA molecule N-(4-
(di-tert-butylfluorosilyl)benzyl)-2-hydroxy-N,N-dimethylethylammonium bromide (SiFAN+Brꢀ) serv-
ing as a lead compound in the development of SiFA-based prosthetic groups of reduced lipophilicity for
biomolecule labeling. Mild conditions for synthesis of [18F]SiFAN+Brꢀ and an easy purification procedure
Article history:
Received 28 September 2010
Received in revised form 26 October 2010
Accepted 27 October 2010
Available online 3 November 2010
Keywords:
Isotopic exchange
Silicon–fluoride-acceptor (SiFA)
PET
using simple C-18 solid phase cartridge have been developed yielding the
[
18F]SiFAN+Brꢀ in
radiochemical yields of 34% (non-decay corrected) within 40 min. A series of kinetic experiments
were performed that show high isotopic exchange rate constants. Low activation energy (15.7 kcal/mol)
and a large preexponential factor (7.9 ꢁ 1013 Mꢀ1 sꢀ1) were calculated for the isotopic exchange reaction
from a corresponding Arrhenius plot. For comparison, the 18F-fluorination of ethyleneglycol-di-p-
tosylate via the formation of a carbon–18F bond showed a 1.3 kcal/mol higher activation energy and a
much lower preexponential factor of 2.9 ꢁ 109 Mꢀ1 sꢀ1. Moderate hydrophilicity (log D = 0.44), stability
Radiotracer
Fluorine-18
in aqueous media at pH up to 7.4 and a high specific activity of [18F]SiFAN+Brꢀ (SA = 20.4 GBq/
m
mol,
mol) make this charged SiFA compound useful for the development of novel SiFA-based 18F-
labeling synthons.
0.55 Ci/
m
ß 2010 Elsevier B.V. All rights reserved.
1. Introduction
overall radiochemical yields (RCYs) of the desired product. Novel
mild methods for the labeling of small as well as larger
biomolecules are thus highly sought after. The radiochemistry of
PET radioisotopes and the particular chemistry of 18F and its role in
PET tracer development have been the subject of many reviews [2].
Most notably, the 18F-labeling of more complex molecules such as
peptides and proteins still predominantly depends on prosthetic
18F-labeling agents obtained via nucleophilic substitution employ-
ing the 18Fꢀ anion. The prosthetic group (‘‘18F-labeling synthon’’) is
synthesized in advance and reacted with the biomolecule intended
for PET imaging. Numerous 18F-labeled precursors have been
reported in the literature predominantly utilizing click chemistry
type reactions for the efficient conjugation to peptide or protein
based PET imaging agents [2 g]. Recently, direct one-step 18F-
labeling procedures for peptides taking advantage of either
activated trimethylammonium leaving groups directly coupled
to the N-terminus of the peptide or nucleophilic ring-opening of
Fluorine-18 is among the most widely used radioisotopes
for Positron Emission Tomography (PET) due to its relatively long
half-life (110 min) and low positron energy (635 keV) which
allows for high-resolution imaging. Although many tracers have
been synthesized and successfully used as PET imaging agents,
only very few, most notably 2-[18F]fluoro-2-deoxy-
D-glucose
([18F]FDG), have found widespread application in nuclear medicine
[1]. This is mainly due to the fact that the nucleophilic reaction
between azeotropically ‘‘dried’’ 18Fꢀ and leaving group bearing
precursors often requires harsh reaction conditions and thus time-
consuming purification procedures, ultimately reducing the
* Corresponding author. Tel.: +1 514 398 1624; fax: +1 514 398 8195.
E-mail address: ralf.schirrmacher@mcgill.ca (R. Schirrmacher).
0022-1139/$ – see front matter ß 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2010.10.008

[()TD$FIG]
28
A.P. Kostikov et al. / Journal of Fluorine Chemistry 132 (2011) 27–34
Scheme 1. Synthesis of SiFAN+Brꢀ starting from SiFA–benzyl alcohol (1) which was converted into the corresponding bromo compound 2 and reacted with N,N-dimethyl
ethanolamine (DMAE) to yield SiFAN+Brꢀ (3).
activated aziridines have proven to be feasible [3]. Although these
approaches yield the 18F-labeled peptides in one-step in high RCYs
the final purification still required an HPLC.
2. Results and discussion
2.1. Synthetic aspects and molecular structure of the SiFAN+Brꢀ (3)
Besides the common formation of a carbon–18F bond by
nucleophilic substitution, only a few groups have investigated the
potential of ‘‘non-conventional’’ phosphorous–, boron–, 18F–
aluminium chelates and silicon–18F-radiochemistry for the 18F-
labeling of organic compounds [4]. Our group has introduced a
new 18F-labeling method based on the 18F–19F isotope exchange
reaction using nanomolar amounts of para substituted di-tert-
butylphenylfluorosilanes (SiFA-compounds; coined after sili-
con_fluoride_acceptor) at room temperature [5]. Isotopic ex-
change reactions involving a carbon–fluorine bond have already
been applied in 18F-labeling chemistry and it could be demon-
strated that only products of low specific activity (SA) can be
obtained hampering their use in PET tracer development [6]. In
contrast to this, SiFA compounds can be obtained in high RCYs as
well as high SA because of the efficiency of the isotopic exchange
The preparation of the SiFAN+Brꢀ (3) started from the SiFA–
benzyl alcohol (1), the synthesis of which is described elsewhere
[5d]. The reaction of the alcohol 1 with tetrabromomethane [8]
gave the corresponding benzylbromide derivative (2) in good yield.
The latter was reacted with an excess of N,N-dimethylaminoetha-
nol at 80 8C in a closed vessel (Scheme 1). After evaporation of the
excess ethanolamine, compound 3 was obtained as a white and
rather hygroscopic solid (m.p. 148 8C) soluble in common polar
solvents.
Single crystals of compound 3 suitable for X-ray diffraction
analysis were obtained by recrystallization from n-hexane/
chloroform. The molecular structure of 3 is shown in Fig. 1,
selected geometric parameters are collected in Table 1.
Compound 3 crystallizes monoclinically with four molecules in
the unit cell. The silicon atom is four-coordinate and shows a
distorted tetrahedral configuration. The largest deviations from
the tetrahedral angles are found for C(11)–Si–C(15) (119.41(14)8)
reaction.
A great variety of SiFA compounds differing in
substitution pattern at the benzene structure have been evaluated
towards their amenability for isotopic exchange. Among those,
p-(di-t-butyl[18F]fluorosilyl)benzaldehyde (SiFA–aldehyde) [5b]
and p-(di-t-butyl[18F]fluorosilyl)benzene thiol (SiFA–SH) [5c]
have been successfully applied as synthons for the labeling of
peptides and proteins, respectively. Despite being successful
labeling building blocks for proteins, yielding suitable in vivo
images, a major drawback of these previously reported SiFA
compounds for the labeling of peptides is their extremely high
lipophilicity, accounting for a high accumulation of the radiotrac-
er in the liver and unspecific binding to the target tissue. These
findings were corroborated by Hoehne et al. who have indepen-
˚
and F–Si–C(1) (103.95(11)8). The Si–F distance is 1.6051(16) A and
falls into the range of the distances of other SiFA-compounds [5]. A
noteworthy feature is the O(1)–Hꢂ ꢂ ꢂBr(1) hydrogen bridge (O(1)–
˚
˚
H(1) 0.71(3) A, H(1)ꢂ ꢂ ꢂBr(1) 2.52(3) A, O(1)ꢂ ꢂ ꢂBr(1) 3.22(2), \ O(1)–
H(1)–Br(1) 172(4)8) that links the cationic and the anionic part of
the compound.
2.2. Optimization of the reaction conditions for the isotopic exchange
Based on our previous experience, [18F]SiFA building blocks
display a high stability at pH ꢃ 7.4 found under physiological
conditions, but undergo relatively rapid hydrolysis at higher pH.
When performing the isotope exchange reaction in dipolar aprotic
solvents such as MeCN (Scheme 2), the base used for the azeotropic
drying of 18Fꢀ as well as its concentration during the isotope
exchange reaction appears to have a crucial influence on 18F
incorporation yields despite the absence of water leading to the
formation of hydroxide anions. Thus, in order to optimize the
incorporation yield of 18Fꢀ into 3, the following four parameters
were varied: (a) the nature of the base (K2CO3 or K2(COO)2), (b) the
concentration of the base, (c) the amount of 3 and (d) the reaction
time of the isotopic exchange. For radio-tracer production, the
dently developed
a no-carrier-added silicon-based labeling
technique using Si–F building blocks of similar structure [7]. It
became apparent that the steric hindrance imparted by the tert-
butyl groups of the SiFA building block is mandatory for
maintaining a high stability of the Si–18F bond towards hydrolysis
[5a]. At the same time, these bulky groups are responsible for the
high lipophilicity of SiFA complicating their use as universally
applicable 18F-labeling synthons especially for compounds of
lower molecular weight.
In order to reduce the overall lipophilicity of the SiFA-building
block, and to find a chemical lead compound for the development
of novel hydrophilic SiFA-based prosthetic groups, we introduced a
positive chemical charge in the form of a tertiary ammonium salt
into the SiFA molecule and investigated whether the isotope
exchange reaction with 18Fꢀ still occurs. We synthesized N-(4-(di-
tert-butylfluorosilyl)benzyl)-2-hydroxy-N,N-dimethylethylam-
monium bromide (SiFAN+Brꢀ), optimized the conditions for
introducing 18Fꢀ via isotopic exchange, measured the rate
constants of the isotopic exchange at different temperatures and
calculated the preexponential factor and activation energy for the
18F–19F exchange from the corresponding Arrhenius plot. To
further characterize SiFAN+Brꢀ, we determined lipophilicity,
stability in aqueous media at different pH and specific activity
(SA), setting a compelling foundation for the future design of
charged SiFA-based building blocks for the 18F-labeling of
biomolecules such as peptides and proteins.
whole batch of dried 18Fꢀ in 0.5–1 mL MeCN, containing
mmolar
amounts of a base (usually K2CO3) and Kryptofix 2.2.2 is normally
used for the 18F-radiolabeling reaction. The results show that the
yield of the 3/18Fꢀ reaction in MeCN significantly drops when
higher amounts of potassium carbonate were used instead of
potassium oxalate as a base at a constant amount of the starting
material (Fig. 2). Using 10
between 0.2 and 0.7 mL of 18Fꢀ (20–50 mCi) from a stock solution
of K2.2.2–K2C2O4 (27 mol) in MeCN (2 mL), an almost constant
RCY of >90% of [18F]3 could be maintained whereas using K2.2.2–
K2CO3 stock-volumes of the same concentration (13.5 mol/mL) of
mg (24 nmol) of 3 and different volumes
m
m
>0.4 mL led to a constant decrease of RCYs (Fig. 2). When 0.7 mL of
the stock solution was used, RCYs of only 10% could be obtained
with K2CO3 in contrast to 90% when using potassium oxalate.

[
A.P. Kostikov et al. / Journal of Fluorine Chemistry 132 (2011) 27–34
29
Fig. 1. Molecular structure of 3 measured at 293 K. Atomic displacement parameters are drawn at 30% probability level.
Moreover, using oxalate salt allowed for reducing the amount
of the starting material 3 without compromising the RCY of
the labeled compound (Fig. 3). In order to maintain RCYs of >50%
reaction mixture was diluted with water and passed through a C18
cartridge, [18F]3 was quantitatively retained on the solid phase
material whereas the 18Fꢀ and the Kryptofix 2.2.2. passed through.
Interestingly, none of the commonly used organic solvents such as
MeOH, MeCN or THF eluted [18F]3 from the C18 cartridge. However,
(
18Fꢀ incorporation), the amount of the precursor 3 can be kept at
as low as 2
mg (5 nmol). The maximum achievable RCY was always
reached within 3–5 min upon mixing of the reagents. The
azeotropic 18F-drying process using oxalate was optimized in
order to minimize losses of 18Fꢀ (<10%) during the evaporation
steps.
[
18F]3 was quantitativelyrecoveredusingamixtureofMeCN/0.01 M
H3PO4 (4:1). The addition of H3PO4 enhances the elution strength of
the solvent to efficiently elute charged [18F]3 from the solid phase
cartridge. After gentle co-evaporation of the H3PO4/MeOH with
ethanol, the residue in the flask appeared to be dry. For the
formulation of an injectable solution, H3PO4 poses no problem since
sterile phosphate buffer is added for pH adjustment at the end of
synthesis. The radiochemical purity was >99%, as determined by
both TLC and HPLC (Fig. 4). In a typical experiment using a starting
activity of 1.85 GBq (50 mCi) of 18Fꢀ, [18F]3 was obtained in RCYs of
34% (non decay-corrected) after 40 min total preparation time
without using HPLC purification.
2.3. Purification of [18F]SiFAN+Brꢀ ([18F]3)
Next, we developed conditions for the purification and final
formulation of [18F]3. One of the main advantages of [18F]SiFA
compounds is their relatively simple purification without HPLC use.
Due to very mild reaction conditions no radioactive side-products
are usually observed, besides unreacted 18Fꢀ. As a result of their very
different lipophilicities, the separation of [18F]3 from 18Fꢀ could be
achieved by simple solid phase cartridge extraction. When the crude
2.4. Specific activity of [18F]3
Previously we calculated the SAs of the 18F-labeled SiFA
compounds simply by dividing their radioactivity amount (Ci)
by the used amount of SiFA compound (m
mol). In case of [18F]3, we
Table 1
˚
Selected bond lengths [A] and bond angles [8] for compound 3.
˚
decided to compare the calculated SA with the experimentally
determined SA, measured by HPLC chromatography to prove that a
Bond lenghts [A]
Bond angles [8]
Si(1)–F(1)
Si(1)–C(1)
Si(1)–C(11)
Si(1)–C(15)
1.6051(16)
1.862(3)
1.875(3)
1.877(3)
F(1)–Si(1)–C(1)
F(1)–Si(1)–C(11)
F(1)–Si(1)–C(15)
C(1)–Si(1)–C(11)
C(1)–Si(1)–C(15)
C(11)–Si(1)–C(15)
103.95(11)
104.92(12)
105.26(12)
112.55(13)
109.22(13)
119.41(14)
simple calculation is sufficient. An aliquot (40
from the reaction of 15
g (36 nmol) 3 and 1.85 GBq (50 mCi) 18Fꢀ
m
L) of purified [18F]3
m
that gave the product in 34% yield was injected into an HPLC
system (cf. Section 3). The amount of carrier was determined from
the UV calibration curve. The SA of [18F]3 was precisely calculated
[()TD$FIG]
Scheme 2. Radioactive labeling of 3 via 18F–19F isotopic exchange.

[
A.P. Kostikov et al. / Journal of Fluorine Chemistry 132 (2011) 27–34
I
Fig. 2. Dependence of the incorporation yield of the 18Fꢀ on the concentration and
the kind of the base using a constant amount of the starting material (10 g,
m
24 nmol). Using K2(COO)2 the amount of added base has just a minor influence on
the RCYs of the isotopic exchange.
knowing the total volume of the product (4.8 mL), its radioactivity
(13.3 mCi) and the concentration of non-radioactive 3 (15
36 nmol) in this aliquot. The experimental value of 0.52 Ci/
closely matched the calculated specific activity of 0.47 Ci/
Higher specific activities could be easily achievable by increasing
the initial activity of 18Fꢀ and reduction of the starting material 3,
which was limited in our case by radiation safety regulations (of
manual manipulations).
m
g,
mol
mol.
m
m
Fig. 4. HPLC chromatograms of solid phase purified SiFAN+Brꢀ [reaction: 18Fꢀ
(1.85 GBq (50 mCi)), K2.2.2/K2(COO)2 (13.5 mol) in 1 mL MeCN, rt, total synthesis
m
time: 40 min]: (A) radio-HPLC chromatogram, (B) corresponding UV chromatogram.
2.5. Lipophilicity of [18F]3
2.6. Hydrolytic stability of [18F]3
A serious drawback of the SiFA building block comprising two t-
butyl groups and one benzene moiety is its high lipophilicity. A
typical log D value, as determined for 4-(di-tert-butylfluorosilyl)-
benzaldehyde (SiFA–CHO), is 3.6. So far, this has significantly
limited possible applications of SiFA compounds for in vivo PET
imaging. The lipophilicity is less detrimental regarding the labeling
of proteins where the SiFA-based labeling synthon imparts only a
negligible change onto the protein. [18F]3, despite having bulky t-
butyl and phenyl substituents is a positively charged molecule,
containing a hydrophilic quaternary ammonium cation. In order to
determine the lipophilicity of [18F]3, we measured the partition of
purified [18F]3 between aqueous phosphate buffer (pH 7.4) and
octanol-1. The positive charge reduced the overall lipophilicity by a
factor of 8 yielding a log D value of 0.44.
As expected and in congruence with our previous findings,
[
18F]3 was completely stable in acidic media – there is no trace of
18Fꢀ ions even after 2 h of incubation in aqueous buffer solution at
pH 4. In neutral and very slightly basic media (pH 7.0 and 7.4) some
hydrolysis was observed but radiochemical impurities did not
exceed 5% of the total radioactivity after 2 h. Most importantly,
[
18F]3 was found to be very stable in human blood serum upon
incubation at 37 8C for 2 h (18
frame of most in vivo PET studies. Increasing hydrolysis over
time occurred in stronger basic media (pH 10) – half of [18F]3
F
ꢀ ꢃ 5%) which is the typical time
was decomposed after
a 2 h incubation time. Kinetic data
of the hydrolysis are represented in Fig. 5 as a concentration of
free 18Fꢀ anion versus time at pH 10. The hydrolysis pseudo-first
order rate constant was calculated to be k0 = 1.55 ꢁ 10ꢀ4 sꢀ1
[)(TD$FIG]
,
[()TD$FIG]
Fig. 3. Dependence of the incorporation yield of the 18Fꢀ on the concentration of the
SiFAN+Brꢀ at a constant concentration of the base (2.7 ꢁ 10ꢀ3 M; 0.2 mL from a
stock solution (13.5 mmol/L)): (1) carbonate (blue line); (2) oxalate (red line). (For
interpretation of the references to color in this figure legend, the reader is referred
to the web version of the article.)
Fig. 5. Hydrolysis of [18F]SiFAN+ (ca. 37 MBq (1 mCi) in 1 mL of borate buffer) at pH
10. The appearance of 18Fꢀ in the solution as a result of hydrolysis is plotted against
time.

[
A.P. Kostikov et al. / Journal of Fluorin
]
e Chemistry 132 (2011) 27–34
31
Fig. 7. Energy diagram of the 18F–19F isotopic exchange reaction (
D
G ꢄ 0 because
isotope effect is negligibly small). The Ea of the isotopic exchange reaction was
Fig. 6. Kinetic curves reflecting the disappearance of 18Fꢀ during the isotopic
determined to be 15.7 kcal/mol.
exchange reaction in MeCN at different temperatures over time for the
determination of the initial rate constants [3: 20
18Fꢀ: 3–5 mCi; VMeCN = 2 mL].
mg (47 nmol); starting activity
added total radioactivity amount of 18Fꢀ for each experiment was
between 111 and 185 MBq (3–5 mCi). The isotopic exchange rate
at each individual temperature was monitored by radio-TLC. Fig. 6
shows the resulting kinetic curves of the 18Fꢀ consumption fitted
to an exponential regression, since isotopic exchange is assumed to
be a pseudo-first order SN2 reaction.
which corresponds to second order rate constant of k = k0/
[OHꢀ] = 1.55 L molꢀ1 sꢀ1 ([OHꢀ] = 1 ꢁ 10ꢀ4 mol Lꢀ1).
2.7. Rate constants and activation energy of the isotope exchange
reaction
Pseudo-first order initial rate constants at different
temperatures 0were calculated from the exponential fit
Previously we reported that the high rates of fluorine exchange
reaction at the silicon atom are due to the low energy barrier of
formation of the penta-coordinated silicon transition state, as
confirmed previously by DFT calculations [5]. However, neither
isotopic exchange reaction rate constants nor the corresponding
activation energy has been previously determined experimentally.
In general, these important parameters are very rarely calculated
in PET radiochemistry and activation energies of very few 18F-
fluorination reactions have been reported so far [6,9]. In order to
determine the Arrhenius preexponential factor and activation
energy for the 18F–19F isotopic exchange at the silicon atom of
compound 3 we decided to carry out a series of experiments at
different temperatures. We found experimentally that the optimal
temperature range to monitor the labeling of 3 is between ꢀ20 8C
and 0 8C, since the reaction is too fast at higher temperatures and
too slow at lower temperatures. Concentrations of unlabeled 3 and
Kryptofix 2.2.2/C2O42ꢀ/K+ complex in MeCN solutions were kept
constant at 2.38 ꢁ 10ꢀ4 M and 6.7 ꢁ 10ꢀ3 M, respectively. The
equation
(k273 ¼ 4:70 ꢁ 10ꢀ3 sꢀ1
;
k20 68 ¼ 3:51 ꢁ 10ꢀ3 sꢀ1
;
k0263 ¼ 2:01 ꢁ 10ꢀ3
s
ꢀ1; k20 58 ¼ 8:84 ꢁ 10ꢀ4
s
ꢀ1) and then divided
by concentration of 3 to determine the actual second-order rate
constants
as
follows:
k273 = 19.75 L molꢀ1 sꢀ1
;
k268 =
14.75 L molꢀ1 sꢀ1; k263 = 8.45 L molꢀ1 sꢀ1; k258 = 3.71 L molꢀ1 sꢀ1
.
These rate constants were used to create an Arrhenius plot (ln k
versus Tꢀ1) to calculate the activation energy (Ea) of the isotope
exchange reaction, which was found to be 65.6 kJ molꢀ1 (15.7 kcal/
mol) (Figs. 7 and 8a). A value of 7.6 ꢁ 1013
M
ꢀ1 sꢀ1 was determined
for the pre-exponential factor (A). In a comparative study, we
determined the activation energy and preexponential values for
the 18F-fluorination of ethyleneglycol-di-p-tosylate, an example of
conventional C-18F bond formation. This 18F-fluorination yields
one of the most frequently used secondary labeling precursors in
radiochemistry [10]. To the best of our knowledge, the kinetic
parameters of this benchmark nucleophilic fluorination reaction
have never been reported. Rate constants for the fluorination at
60–80 8C with 5 8C intervals were calculated in the same fashion as
[()TD$FIG]
Fig. 8. Arrhenius plots of ln k against 1/T for (A) the isotopic exchange reaction of 3 with 18Fꢀ [3: 20
m
g (35.7 nmol); starting activity 18Fꢀ: 110–185 MBq (3–5 mCi);
VMeCN = 1 mL, T = ꢀ20 to 0 8C] and (B) the reaction of ethyleneglycol-di-p-tosylate with 18Fꢀ [ethyleneglycol-di-p-tosylate (10 mg, 27
m
mol); starting activity 18Fꢀ: 110–
185 MBq (3–5 mCi); VMeCN = 1 mL, T = 60–80 8C]. An activation energy of 65.6 kJ molꢀ1 (15.7 kcal/mol) and 71.5 kJ molꢀ1 (17 kcal/mol) was calculated from plotting ln k
against Tꢀ1 for the isotopic exchange and the fluorination of the tosylate, respectively.

32
A.P. Kostikov et al. / Journal of Fluorine Chemistry 132 (2011) 27–34
described above and the ln k versus Tꢀ1 dependence was plotted.
The activation energy for this nucleophilic fluorination was
found to be 71.5 kJ molꢀ1 (17 kcal/mol) (Fig. 8b) which is
1.3 kcal/mol higher than the isotope exchange reaction. The
preexponential factor we determined for this reaction was
2.9 ꢁ 109 Mꢀ1 sꢀ1. The large difference in reactivity for the
isotope exchange reaction versus the 18F-fluorination of
ethyleneglycol-di-p-tosylate can thus be ascribed to a difference
carrier C atom for non-methyl and 1.5 times Ueq of the carrier C
atom for methyl groups. Atomic scattering factors for neutral
atoms and real and imaginary dispersion terms were taken from
International Tables for X-ray Crystallography [13]. The figure
was created by SHELXTL [14]. Crystallographic data for the
structure reported in this paper have been deposited by the
Cambridge Crystallographic Data Centre as supplementary
material publication no. CCDC-790848. Copy of the data can
be obtained free of charge on application to CCDC, 12 Union
Road, Cambridge, CB2 1EZ, UK (fax: +44 1223 336033, e mail:
deposit@ccdc.cam.ac.uk).
in the preexponential factor, over
4 orders of magnitude
larger for the isotope exchange reaction. The smaller (ca.
1.3 kcal) activation energy for the isotope exchange reaction
over the 18F-fluorination of ethyleneglycol-di-p-tosylate would
also account for the larger rate constant observed for the former
over the latter.
3.3. Synthesis of the [4-(di-tert-butylfluorsilyl)benzyl]-(2-
hydroxyethyl)-dimethylammonium bromide, SiFAN+Brꢀ (3)
3. Experimental
3.3.1. (4-Bromomethylphenyl)-di-tert-butyl-fluorosilane (2)
To a 0 8C cooled solution of the SiFA–benzyl alcohol 1 (3.08 g,
11.5 mmol) and tetrabromomethane (4.18 g, 12.6 mmol,
1.1 equiv.) in 100 mL dichloromethane triphenylphosphine
3.1. General procedures
Highly enriched [18O]water (>97%) was purchased from
Cambridge Isotopes Laboratories. All other commercially avail-
able chemicals such as Kryptofix 2.2.21, ethyleneglycol-di-p-
tosylate, acetonitrile, potassium carbonate and potassium oxalate
were of the highest available purity purchased from Sigma–
Aldrich. The labeling precursor N-(4-(di-tert-butylfluorosilyl)-
(3.30 g, 12.6 mmol, 1.1 equiv.) was added over a period of
30 min in small portions [8]. The solution was stirred for 2 h at
room temperature and turned yellow during this time. The solvent
was removed in vacuo and the residue was washed with cold n-
hexane (150 mL). The white precipitate was removed by filtration,
the clear solution was concentrated in vacuo and purified by
column chromatography (SiO2, 100% pentane). Compound 2 was
isolated as a colorless oil (3.06 g, 9.2 mmol, 80%), which solidified
on standing (m.p. 52 8C).
benzyl)-2-hydroxy-N,N-dimethylethylammonium
bromide
(SiFAN+Brꢀ, 3) was synthesized as described below. Radio TLCs
(silicagel-60 plates) were monitored using an Instant Imager
(Packard). Reactions at low temperatures were carried out using a
thermostat (FTS Systems). Analytical HPLC was performed on an
1H NMR (400.13 MHz, CDCl3):
d
7.58 (d, 3J(1H–1H) = 8 Hz, 2H,
Hm), 7.40 (d, 3J(1H–1H) = 8 Hz, 2H, Ho), 4.49 (s, 2H, CH2), 1.05 (s,
Agilent Technologies 1200 system equipped with
a
Gabi
18H, Si(C(CH3)3)2).
radioactivity detector (Raytest). Kinetic experiments were per-
formed using a FTS systems cryostat and the curves were plotted
and fitted using Sigma Plot 11.0 software. All solvents used for the
syntheses of 1–3 were purified by distillation under argon
atmosphere from appropriate drying agents. The NMR experi-
ments were carried out with Bruker DRX 400, Bruker DRX 300 and
13C{1H}
NMR
(100.63 MHz,
CDCl3):
d
138.8
(d,
3J(13C–19F) = 14 Hz, Cp), 134.4 (d, 3J(13C–19F) = 4 Hz, Cm), 128.1 (s,
Co) 125.9 (s, Ci), 27.2 (s, Si(C(CH3)3)2), 20.2 (d, 2J(13C–19F) = 12 Hz,
Si(C(CH3)3)2).
19F NMR (282.38 MHz, CDCl3):
29Si{1H}
NMR (59.63 MHz,
1J(29Si–19F) = 299 Hz).
d
188.9 (s, 1J(19F–29Si) = 299 Hz).
CDCl3): 14.3 (d,
d
Varian Mercury 200 spectrometers. The chemical shifts dare given
in ppm and are referenced to the solvent peaks with the usual
values calibrated against tetramethylsilane (1H, 13C, 29Si) and
CFCl3 (19F). The high resolution mass spectra were obtained with
LTQ Orbitrap mass spectrometer (Thermo Electron) using
acetonitrile as mobile phase. The acetonitrile solutions were
injected via a TriPlus Autosampler onto a DFS-system (Perfluor-
kerosene as reference), connected with a Trace GC Ultra 2000
Elemental analyses: calculated (%) for C15H24FSiBr (331.34 g/
mol) C 54.4, H 7.3, found C 54.4, H 7.6.
HR-GC–EI-MS: calculated (m/z) for C15H24F28Si79Br+ 330.0815,
found 330.0814.
3.3.2. SiFAN+Brꢀ (3)
A mixture of the SiFA–benzyl bromide 2 (0.57 g, 1.72 mmol)
and N,N-dimethyl ethanolamine (0.78 g, 1 mL, 8.81 mmol) was
heated for 24 h at 80 8C in a glass vessel with a Teflon valve (F.
Young). The excess of ethanolamine was removed in vacuo and the
ammonium salt was re-crystallized from CHCl3/n-hexane. Com-
pound 3 (0.41 g, 0.97 mmol, 56%) was obtained as colorless crystals
(m.p. 148 8C).
system (column: DB-5MS (25 m, 0.25 mm ID, film 0.1
mm)). FT
infrared spectra were recorded using a Bruker IFS28 spectrometer.
Elemental analyses were performed on
analyser.
a LECO-CHNS-932
3.2. X-ray crystallographic procedures
1H NMR (400.13 MHz, CDCl3):
d 7.65 (s, 4H, Harom), 5.12 (t,
Crystals of compounds 3 (Fig. 1) suitable for single-crystal X-ray
diffraction analyses were grown by re-crystallization from CHCl3/
n-hexane. Intensity data were collected with a Xcalibur2 CCD
diffractometer (Oxford diffraction) with graphite monochromated
3J(1H–1H) = 5 Hz, 1H, OH), 4.92 (s, 2H, PhCH2), 4.22 (m, 2H, NCH2),
3.81 (m, 2H, CH2OH), 3.32 (s, 6H, NCH3), 0.99 (s, 18H, Si(C(CH3)3)2).
13C{1H}
NMR
(100.63 MHz,
CDCl3):
d
137.4
(d,
3J(13C–19F) = 14 Hz, Cp), 134.6 (d, 3J(13C–19F) = 4 Hz, Cm), 132.3 (s,
Co), 128.3 (s, Ci), 68.6 (s, PhCH2), 65.9 (s, NCH2), 55.9 (s, CH2OH),
50.8 (s, NCH3), 27.2 (s, Si(C(CH3)3)2), 20.1 (d, 2J(13C–19F) = 12 Hz,
Si(C(CH3)3)2).
MoK
whole sphere of the reciprocal space with 8 sets at different
angles and 485 frames by -rotation (
= 18) at 2 ꢁ 40 s per
a radiation at 173 K. The data collection covered almost the
k
v
D/v
frame. Crystal decay was monitored by repeating the initial frames
at the end of the data collection. After analysis of the duplicate
reflections, there was no indication of any decay. The structure was
solved by direct methods (SHELXS97) [11]. Refinement applied
full-matrix least-squares methods (SHELXL97) [12]. All H atoms
were located in the difference Fourier map and their positions were
isotropically refined with Uiso constrained at 1.2 times Ueq of the
19F NMR (282.38 MHz, CDCl3):
29Si{1H}
NMR (59.63 MHz,
1J(29Si–19F) = 299 Hz).
Elemental analyses: calculated (%) for
d
187.8 (s, 1J(19F–29Si) = 299 Hz).
CDCl3): 13.7 (d,
d
C
19H35FNOSiBr
(420.48 g/mol) C 54.3, H 8.4, N 3.3, found C 52.3, H 8.5, N 3.4.
HR-LC–ESI-MS: calculated (m/z) for C19H35FNO28Si+ 340.2466,
found 340.2466.

A.P. Kostikov et al. / Journal of Fluorine Chemistry 132 (2011) 27–34
33
3.4. Radiochemistry
was ꢅ50% as determined by TLC) then quenched with water
(20 mL), and purified on a C18 cartridge as described above. [18F]3
(492–629 MBq (13.3–17 mCi), radiochemical purity >99%) was
dissolved in EtOH (0.5 mL) and sterile phosphate buffer (4.5 mL).
3.4.1. Azeotropic drying of 18Fꢀ
No-carrier-added (nca) 18Fꢀ was generated by the bombard-
ment of an enriched [18O]water target (>97%, 2.2 mL) with a
proton beam (35
m
A) using an IBA cyclotron (Cyclon 18/9) for
3.4.4. Determination of specific activity (SA) of [18F]3
10 min. The resulting aqueous [18F]Fꢀ/H2[18O]O (3.7–7.4 GBq;
100–200 mCi) was transferred to a hot cell and passed through a
preconditioned (10 mL 0.5 M K2CO3 followed by 10 mL deionized
water) QMA SepPak Light cartridge (Waters). The H2[18O]O was
collected and the fixed 18Fꢀ was eluted from the QMA cartridge
with a solution of Kryptofix 2.2.21 (10 mg), either 0.25 M K2CO3
To determine the SA of [18F]3, a small aliquot (40
m
L) of the
aforementioned solution (cf. Section 3.4.3) was analyzed by means
of radio-HPLC using a Prodigy ODS (5
m
m, 250 mm ꢁ 4.6 mm)
column (Phenomenex) and acetonitrile/20 mM KH2PO4 (55:45) as
a mobile phase (flow rate: 1 mL/min). A calibration curve (UV
detection at 230 nm) to calculate the amount of 3 was created by
injecting solutions of unlabeled 3 of different concentrations (1, 2,
(60
mL) or 0.25 M K2C2O4 (60 mL) and MeCN (1 mL). The solvents
were removed under a gentle vacuum of ca. 500 mbar and a sweep
stream of argon at 90 8C. The remaining traces of water were
removed by azeotropic co-evaporation with MeCN (1 mL) twice.
After all the solvents were visibly removed the argon flow was
stopped, the vacuum was set to maximum (ca. 10 mbar) and the
final drying was continued for 2 min.
5 and 10 mg/mL). The retention time of 3 at a flow rate of 0.8 mL/
min was 8.8 min (Fig. 4).
3.4.5. Determination of the lipophilicity of [18F]3
In order to determine the log D value, the dried [18F]3 (ca.
37 MBq (1 mCi)) was dissolved in phosphate buffer (pH 7.4) and an
aliquot of this solution (0.5 mL) was placed into an Eppendorf tube.
Octanol-1 (0.5 mL) was added and the mixture was vigorously
shaken using a Vortex apparatus for 5 min. After centrifugation
(14,500 rpm, 3 min), the aqueous and organic phases were
carefully separated and their radioactivity content measured in
a calibrated curimeter. The experiment was done in triplicate and
the ratio of the total radioactivities of the two phases was used to
calculate the log D. In addition to this procedure and to corroborate
3.4.2. 18F-labeling of 3 via isotopic exchange at room temperature
In order to optimize the reaction conditions, the dried [18F]Fꢀ/
Kryptofix 2.2.21/K+ complex was re-dissolved in 1 mL of acetoni-
trile. Aliquots of this solution were mixed at room temperature
with aliquots of a stock solution of precursor 3 ((c = 1
mg/mL),
investigated amount of 3: 1–800 g) in anhydrous MeCN in
m
Eppendorf vials (1.5 mL). Each reaction mixture was kept at room
temperature and was analyzed by TLC (CH2Cl2/MeOH/AcOH 7:2:1,
Rf = 0.5) at different time points (3–50 min) and the 18Fꢀ
incorporation yield was determined as percentage of the
radioactivity of the spot corresponding to [18F]3 to the total
radioactivity of the TLC plate (only two spots were detectable:
the results, equal volumes (1
mL) from each phase were spotted on
a TLC plate and the ratio of the activities was determined with an
Instant Imager (Packard Canberra). The calculated log D values
using these two different methods matched and were averaged
over three experiments for each method.
[
18F]3 + unreacted 18Fꢀ). To separate the labeled product from
unreacted 18Fꢀ, the crude reaction mixture was diluted with water
(10ꢁ the volume of the crude reaction mixture) and the resulting
solution was passed through a preconditioned (10 mL MeOH
followed by 10 mL of water) C-18 light cartridge (SepPak, Waters).
The cartridge was washed with 10 mL water to remove all traces of
residual 18Fꢀ and Kryptofix 2.2.21. The desired [18F]3 was nearly
quantitatively (>97%) eluted from the cartridge with MeOH/
0.01 M H3PO4 (2 mL, 4:1). Both aqueous and organic eluates were
analyzed by means of radio-TLC using the conditions reported
above to confirm the quantitative separation of unreacted 18Fꢀ
anion and [18F]3. The product was collected in a round-bottom
flask and the solvent was removed by evaporation using a water
bath at gentle heat (ꢃ70 8C) and a vacuum of 20 mbar. Higher
temperatures led to a decomposition of [18F]3. Ethanol (100%,
2 mL) was then added and the solution was re-evaporated in order
to remove all traces of methanol and H3PO4. Finally, [18F]3 was re-
dissolved in ethanol (0.3 mL) and sterile phosphate buffer (2.7 mL)
for the final formulation. It could be shown that the compound did
not decompose under these conditions mimicking an injectable
solution for animal or human PET experiments.
3.4.6. Stability of [18F]3 towards hydrolysis at different pH
[
18F]3 was synthesized and purified as described above and
then re-dissolved in aqueous methanol (10%, 2 mL). Equal aliquots
(300 L, ca. 1 mCi) of the resulting solution were diluted with
standard commercially available aqueous buffers of known pH
(700 L, pH 4, 7, 7.4 and 10) at room temperature and with human
m
m
blood serum (1 mL, pH 7.4) at 37 8C. Each solution was analyzed by
radio-TLC for up to 4 h by taking small aliquots at different time
points (20–240 min) to determine the concentration of unbound
18Fꢀ. For the experiment at pH 10, the amount of unbound 18Fꢀ was
plotted versus time to calculate a hydrolysis rate constant.
3.4.7. Rate constants of isotopic exchange and Arrhenius plot
3.4.7.1. [18F]3. To determine the rate constants and activation
energy (Ea), the 18F–19F isotope exchange reaction between 3 and
18Fꢀ was monitored at four different temperatures (0 8C, ꢀ5 8C,
ꢀ10 8C, and ꢀ15 8C) using equal amounts of the precursor (20
mg,
47 nmol) and 18Fꢀ/Kryptofix 2.2.21/oxalate/K+ complex (3–5 mCi).
Each reaction was monitored by radio-TLC at several time points
up to 40 min. The radioactivity amount corresponding to 18Fꢀ was
plotted versus time to calculate the isotope exchange pseudo-first
order rate constants at different temperatures. The resulting data
were used to create an Arrhenius plot (ln k versus 1/T) to calculate
the corresponding preexponential factor and the activation energy
of the isotope exchange from the slope of the plot.
3.4.3. Synthesis of [18F]3 at a higher MBq (mCi) level
We applied the optimized isotopic exchange reaction condi-
tions to synthesize a large radioactivity amount of [18F]3 and
calculate the overall preparative radiochemical yield of the
product. 1.85 GBq (50 mCi) 18Fꢀ was fixed on the QMA cartridge
and eluted with a mixture of Kryptofix 2.2.2 (5 mg, 13.3
mmol) and
3.4.7.2. Nucleophilic fluorination of ethyleneglycol-di-p-tosylate. A-
K2C2O4 (6.7 mol) in acetonitrile/water (95:5, V = 2 mL). The
m
liquots of ethyleneglycol-di-p-tosylate solution in MeCN (2.7
in 100
L) were added to the stock solution of the [18F]Fꢀ/
Kryptofix 2.2.21/oxalate/K+ (100
L 3–5 mCi) complex in Eppen-
mmol
resulting complex was dried as described above (a slightly lower
vacuum was applied to keep losses of 18Fꢀ during the drying below
m
m
10%) and 3 (15
(1 g/ L) in anhydrous MeCN (1 mL). The reaction mixture was
kept at room temperature for 5 min (incorporation yield of 18Fꢀ
mg, 35.7 nmol) was added from a stock solution
dorf tubes and reaction mixtures were heated up to 60–80 8C in
5 8C increments and monitored by radio-TLC at certain time-
m
m

34
A.P. Kostikov et al. / Journal of Fluorine Chemistry 132 (2011) 27–34
[2] (a) D. Le Bars, J. Fluorine Chem. 127 (2006) 1488–1493;
(b) R. Schirrmacher, C. Wa¨ngler, E. Schirrmacher, Mini Rev. Org. Chem. 4 (2007)
317–329;
points. The rate constants were determined in the same fashion as
described above for the [18F]3 and were used to create an Arrhenius
plot to calculate the preexponential factor and the activation
energy of this nucleophilic substitution from the slope of the plot.
(c) H.J. Wester, M. Schottelius, in: P.A. Schubiger, L. Lehmann, M. Friebe (Eds.), PET
Chemistry – The Driving Force in Molecular Imaging, 2006, 79–111;
(d) L. Chai, S.Y. Lu, V.W. Pike, Eur. J. Org. Chem. 17 (2008) 2853–2873;
(e) P.W. Miller, N.J. Long, R. Vilar, A.D. Gee, Angew. Chem. Int. Ed. 47 (2008) 8998–
9033;
4. Conclusions
(f) S.M. Ametamey, M. Honer, P.A. Schubiger, Chem. Rev. 108 (2008) 1501–1506;
(g) Wa¨ngler C., Schirrmacher R., Bartenstein P., Wa¨ngler B., Curr. Med. Chem. 17
(2010) 1092–1116.
We demonstrated that charged SiFA compounds such as N-(4-
(di-tert-butyl[18F]fluorosilyl)benzyl)-2-hydroxy-N,N-dimethy-
lethylammonium bromide ([18F]SiFAN+Brꢀ, [18F]3) can be obtained
in high RCYs at low temperatures via simple isotope exchange with
SAs between 400 and 550 Ci/mmol. The salt [18F]3 was purified by
simple solid phase cartridge extraction and isolated in high RCY of
27–34% (uncorrected for radioactive decay) in as little as 40 min.
[3] (a) J. Becaud, L. Mu, M. Karramkam, P.A. Schubiger, S.M. Ametamey, K. Graham, T.
Stellfeld, L. Lehmann, S. Borkowski, D. Berndorff, L. Dinkelborg, A. Srinivasan, B.
Koksch, Bioconjugate Chem. 20 (2009) 2254–2261;
(b) U. Roehn, J. Becaud, L. Mu, A. Srinivasan, T. Stellfeld, A. Fitzner, K. Graham, L.
Dinkelborg, A.P. Schubiger, S.M. Ametamey, J. Fluorine Chem. 130 (2009) 902–912.
[4] (a) M.S. Rosenthal, A.L. Bosch, R.J. Nickles, S.J. Gately, Int. J. Appl. Radiat. Isot. 48
(1985) 318–319;
(b) A.R. Studenov, M.J. Adam, J.S. Wilson, T. Ruth, J. Labelled Compd. Radiopharm.
48 (2005) 497–500;
[
18F]3 was proven to be 8-times more hydrophilic (log D = 0.44)
(c) R. Ting, M.J. Adam, T.J. Ruth, D.M. Perrin, J. Am. Chem. Soc. 127 (2005) 13094–
13095;
(d) L.J. Mu, A. Hoehne, P.A. Schubiger, S.M. Ametamey, K. Graham, J.E. Cyr, L.
Dinkelborg, T. Stellfeld, A. Srinivasan, U. Voigtmann, U. Klar, Angew. Chem. Int. Ed.
47 (2008) 4922–4925;
than previously reported SiFA compounds and stable in aqueous
buffers up to pH 7.4. The compound was also found to be stable in
human serum at 37 8C for 2 h. The isotope exchange reaction was
characterized to be very fast even below room temperature
(ꢀ20 8C) and a relatively high Arrhenius preexponential factor of
(e) P. Laverman, W.J. McBride, R.M. Sharkey, A. Eek, L. Joosten, W.J.G. Oyen, D.M.
Goldberg, O. Boerman, J. Nucl. Med. 51 (2010) 454–460.
[5] (a) R. Schirrmacher, G. Brandtmo¨ller, E. Schirrmacher, O. Thews, J. Tillmanns, T.
Siessmeier, H.G. Buchholz, P. Bartenstein, B. Wa¨ngler, C.M. Niemeyer, K. Jurkschat,
Angew. Chem. Int. Ed. 45 (2006) 6047–6050;
7.6 ꢁ 1013
M
ꢀ1 sꢀ1 and low activation energy of 65.6 kJ molꢀ1
(15.7 kcal/mol) were experimentally determined from the corre-
sponding Arrhenius plot. The formation of a carbon–18F bond via
nucleophilic substitution using ethyleneglycol-di-p-tosylate as a
labeling precursor yielded a 1.3 kcal/mol higher activation energy
and a much lower preexponential factor of 2.9 ꢁ 109 Mꢀ1 sꢀ1
making the isotopic exchange reaction kinetically more favorable.
Charged [18F]SiFA compounds have the potential to become
valuable new labeling synthons for the development of SiFA-
based 18F-fluorinated radiopharmaceuticals. Especially peptide
and protein labeling will benefit from the development of
hydrophilic SiFA building blocks. We are currently in the process
of synthesizing additional SiFAN+ derivatives bearing a thio- or
amino group instead of the relatively un-reactive hydroxy moiety.
These compounds are intended for subsequent bioconjugation to
maleimido-derivatized peptides and proteins by click chemistry.
(b) E. Schirrmacher, B.Wa¨ngler, M. Cypryk, G.Brandtmo¨ller, M. Scha¨fer, M. Eisenhut,
K. Jurkschat, R. Schirrmacher, Bioconjugate Chem. 18 (2007) 2085–2089;
(c) 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
Chem. 20 (2009) 317–321;
(d) L. Iovkova, B. Wa¨ngler, E. Schirrmacher, R. Schirrmacher, G. Quandt, G. Boening,
M. Schu¨rmann, K. Jurkschat, Chem. Eur. J. 15 (2009) 2140–2147.
[6] F. Cacace, M. Speranza, A.P. Wolf, R.R. MacGregor, J. Fluorine Chem. 21 (1982)
145–158.
[7] A. Hoehne, L. Mu, M. Honer, P.A. Schubiger, S.M. Ametamey, K. Graham, T.
Stellfeld, S. Borowski, D. Berndorff, U. Klar, U. Voigtmann, J.E. Cyr, M. Friebe, L.
Dinkelborg, A. Srinivasan, Bioconjugate Chem. 19 (2008) 1871–1879.
[8] A. Ianni, S.R. Waldvogel, Synthesis 13 (2006) 2103–2112.
[9] J.-H. Chun, S. Lu, Y.-S. Lee, V.W. Pike, J. Org. Chem. 75 (2010) 3332–3338.
[10] D. Block, H.H. Coenen, G. Sto¨cklin, J. Labelled Compd. Radiopharm. 25 (1987) 201–
216.
[11] G.M. Sheldrick, Acta Crystallogr. A46 (1996) 467.
[12] G.M. Sheldrick, University Goettingen, 1991.
[13] International Tables for Crystallography, 1992, Vol. C, Dordrecht: Kluwer Aca-
demic Publishers.
References
[14] G.M. Sheldrick, 1997, SHELXTL. Release 5.1 Software Reference Manual, Bruker
AXS, Inc., Madison, Wisconsin, USA.
[1] K. Hamacher, H.H. Coenen, G. Sto¨cklin, J. Nucl. Med. 27 (1986) 235–238.