Competitive demethylation and substitution in N,N,N-trimethylanilinium fluorides

Article abstract of DOI:10.1016/j.jfluchem.2007.03.009

Fluorination of aromatic compounds by nucleophilic displacement of trimethylanilinium salts by fluoride is a commonly used reaction for radiotracer synthesis. Though the liberated trimethylamine is thought to be an excellent leaving group for this type of S_NAr reaction, scattered reports show that amine demethylation (reverse Menschutkin reaction) sometimes dominates over substitution, particularly when relatively electron rich fluoroarenes are the desired targets. Here we provide systematic experimental and theoretical studies of trimethylanilinium demethylation and substitution. Results from these studies highlight the limits of this leaving group in fluoroarene synthesis and have important ramifications for the design of nucleophilic fluorinating agents featuring ammonium cations.

Full text of DOI:10.1016/j.jfluchem.2007.03.009

Journal of Fluorine Chemistry 128 (2007) 806–812
www.elsevier.com/locate/fluor
Competitive demethylation and substitution in N,N,
N-trimethylanilinium fluorides
*
Haoran Sun, Stephen G. DiMagno
Department of Chemistry and Nebraska Center for Materials and Nanoscience, University of Nebraska, Lincoln, NE 68588-0304, USA
Received 1 February 2007; received in revised form 7 March 2007; accepted 8 March 2007
Available online 12 March 2007
Abstract
Fluorination of aromatic compounds by nucleophilic displacement of trimethylanilinium salts by fluoride is a commonly used reaction for
radiotracer synthesis. Though the liberated trimethylamine is thought to be an excellent leaving group for this type of SNAr reaction, scattered
reports show that amine demethylation (reverse Menschutkin reaction) sometimes dominates over substitution, particularly when relatively
electron rich fluoroarenes are the desired targets. Here we provide systematic experimental and theoretical studies of trimethylanilinium
demethylation and substitution. Results from these studies highlight the limits of this leaving group in fluoroarene synthesis and have important
ramifications for the design of nucleophilic fluorinating agents featuring ammonium cations.
# 2007 Elsevier B.V. All rights reserved.
Keywords: Tetrabutylammonium fluoride; Amine deprotection; Demethylation; Nucleophilic substitution; SN2; SNAr; Arene fluorination; PET; Leaving groups;
Reverse Menschutkin
1. Introduction
for electron-withdrawing groups ortho or para to the desired
site of fluorination. In practice, this requirement often
necessitates multistep syntheses in which the activating
substituent is subject to functional group transformations after
the fluorine is introduced into the molecule. This limitation is
among the principal unsolved problems in [18F] radiotracer
synthesis.
Nucleophilic aromatic substitution (SNAr) is a particularly
useful transformation for preparing fluorinated radiotracers for
positron emission tomography (PET) [1–5]. Advantages of
using nucleophilic reactions for the introduction of [18F]
fluoride include the high chemical yields, excellent regios-
electivity, and the high radiochemical purity associated with
radiotracers prepared by this technique. Nucleophilic reactions
are also the only means by which radiofluorination can be
performed under no-carrier-added (n.c.a.) conditions. Impor-
tant experimental parameters for optimizing nucleophilic
aromatic fluorination are reaction time, temperature, solvent,
fluoride counterion, substrate substituents, and leaving group.
Typical requirements for radiofluorination are relatively short
reaction times (<30 min), elevated temperatures, and polar
aprotic solvents (acetonitrile, DMF, DMSO). The fluoride
source commonly employed is Kryptofix [2.2.2] [18F] KF, in
which a cryptand complexes the potassium ion to provide a
relatively nucleophilic fluoride ion. A commonly encountered
problem with nucleophilic aromatic fluorination is the necessity
The nature of the leaving group has a profound influence on
the success of SNAr fluorination reactions. Typical leaving
groups for SNAr reactions are the halides, NO2ꢀ, N(CH3)3, ArI,
and N2. Here we will focus on substitution reactions of
trimethylanilinium triflates. Because these substrates are
readily synthesized and are highly reactive toward substitution,
they have become workhorses for nucleophilic aromatic
radiofluorination.
Early work on the radiofluorination of aryltrimethylammo-
nium salts [6–9] demonstrated that the neutral trimethylamine
leaving group was more active than halides and comparable in
reactivity to the nitro group [10]. For certain substrates, for
example in the radiofluorination of COX-1 and COX-2
inhibitors SC63217 and SC58125, the trimethylammonium
salts seem to be clearly superior to the nitrated derivatives [11].
In other cases, particularly when the aromatic ring is relatively
electron rich, the nitro group offers clear advantages.
* Corresponding author. Tel.: +1 402 472 9895; fax: +1 402 472 9402.
E-mail address: sdimagno1@unl.edu (S.G. DiMagno).
0022-1139/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2007.03.009

H. Sun, S.G. DiMagno / Journal of Fluorine Chemistry 128 (2007) 806–812
807
Scheme 1.
2. Results and discussion
An activated aromatic ring (i.e. a ring bearing at least one
electron-withdrawing group) is required to effect substitution of
the trimethylammonium group with fluoride. N,N,N-Trimethyl-
4-nitroanilinium triflates undergo relatively rapid substitution
with Kryptofix [18F] KF. In DMSO, substitution is complete
within 10 min at slightly elevated temperature (55–100 8C)
[10,12]; harsher conditions are required if electron-donating
groups are present on the nitroarene [13]. Anhydrous
acetonitrile has also been used as a solvent for fluorodeamina-
tion of N,N,N-trimethyl-2-chloro-4-nitroanilinium triflate; in
this case the reaction takes place at room temperature [14].
Trimethylanilinium salts bearing cyano substituents are also
rapidly fluorinated, providing a straightforward, high yield
entry into [18F] N-(4-fluorobenzyl)-2-bromoacetamide for
radiolabeling oligonucleotides [15–23]. If electron-withdraw-
ing groups less potent than cyano are present on the
trimethylanilinium ring, radiofluorination yields begin to
decline. Corrected radiofluorination yields for para-trimethy-
lammonium aryl aldehydes [24–27], ketones [28–32], and
esters [33–36] range from excellent (80%) for aryl aldehydes to
fair (50%) for aryl esters. For relatively unactivated rings,
fluorodeamination is difficult; Ermert and coworkers [37]
studied substitution in 4-bromo-N,N,N-trimethylanilinium
triflates and found that only 8–12% yields of the desired 4-
fluorobromobenzenes were obtained under optimized condi-
tions.
Differentiallysubstituted substrateswereselectedtoprovidea
broad range of electron-withdrawing ability. The trimethyl-
anilinium triflates were prepared either by exhaustive alkylation
of the corresponding commercially available anilines with
methyl iodide followed by ion-exchange (for the electron rich
anilines), or by direct alkylation with methyl triflate according to
literature methods [20]. Yields for these alkylation reactions
generally exceeded 80%. TBAF*, synthesized and isolated
according to previously reported procedures [40], was combined
with these substrates in d6-DMSO in temperature controlled,
sealed NMR tubes. The final concentration of the substrate was
0.4 M; all reactions were performed on a 0.2 mmol scale. The
reactions investigated are shown in Scheme 1.
2.1. Ion exchange experiments
Upon mixing of TBAF* (1.4 equiv.) with the trimethyla-
nilinium triflates, two of the salts (X = CF3, CN) formed
precipitates at room temperature, suggesting an instantaneous
ion exchange reaction. (The CF3-substituted derivative proved
so insoluble that further studies on the reactions of this
compound were abandoned.) Changes in the 19F and 1H NMR
spectra observed upon mixing of the TBAF* with the substrates
also gave evidence for ion exchange: the signals for the fluoride
anion shifted upfield and broadened significantly in the 19F
NMR spectra, and the aromatic regions of the 1H NMR spectra
showed significantly larger separations in the chemical shifts of
the ortho and meta proton signals (Fig. 1).
The low yields obtained for SNAr fluorination of relatively
unactivated trimethylanilinium triflates stems, in part, from
formation of methyl fluoride from a competing SN2 (reverse
Menschutkin [38]) reaction. Formation of [18F] methyl fluoride
has been observed in some cases where it is suspected to be the
main deleterious side reaction [11,28]. The reverse Menschut-
kin reaction was also exploited for the preparation of labeled
methyl fluoride; 2-acetyl-N,N,N-trimethylanilinium triflate
readily undergoes microwave-induced [18F] fluorodemethyla-
tion [39]. These literature data suggest that the demethylation/
substitution partition ratios are closely correlated to the nature
of the aryl ring substituents; though no truly quantitative study
examining this partition ratio has yet been published.
The attribution of these spectral changes to ion exchange
1
was supported strongly by H–19F heteronuclear Overhauser
effect spectroscopy (HOESY) experiments performed in our
laboratory. Pregosin and coworkers exploited HOESY experi-
ments to gain insight into the nature of ion pairing [43] for
ammonium salts. Where tight ion pairing between an organic
cation and a fluorinated anion occurs, 1H–19F HOESY data can
also indicate the site(s) of the ion pairing interaction(s). Our
1H–19F HOESY data (Fig. 2) demonstrate the usefulness of this
technique for ammonium fluoride salts. These spectra indicate
that TBAF* is tightly ion paired in DMSO solution, and that the
principal interaction of fluoride is with the a-CH2 group of the
TBA cation. When 4-methyl-N,N,N-trimethylanilinium triflate
is added to this solution, ion exchange occurs and the fluoride is
bound exclusively (within the experimental detection limits,
>100:1) to the anilinium cation. Moreover, the significant cross
peak attributable to the ortho-hydrogen atom on the anilinium
ring suggest an ion pair structure in which the fluoride is
Here we report quantitative experimental and theoretical
studies to investigate the impact of substrate substituents upon
the competing SNAr and SN2 pathways for the reaction of
N,N,N-trimethylanilinium triflates with fluoride ion. We are
aided in these efforts by our recent synthesis of anhydrous
tetrabutylammonium fluoride (TBAF*) [40], a convenient
source of highly nucleophilic fluoride ion [41] that may be used
for radiotracer synthesis [42].

808
H. Sun, S.G. DiMagno / Journal of Fluorine Chemistry 128 (2007) 806–812
Table 1
Hammett spꢀ constants and observed aryl/alkyl fluorination ratios observed in
DMSO at 100 8C
ꢀ
Substituent
sp
Reaction time (min)
Prod. ratio
NO2
CHOb
CN
1.27
1.03
1a
1
>99:1
>99:1
99:1
1.00
1
COPh
CO2Me
Br
0.84
1
99:1
0.64
2
96:4
0.25
10
10
10
30
2:98
H
0.0
<1:99
<1:99
<1:99
Me
ꢀ0.17
OMe
ꢀ0.26
a
Reaction is complete within 10 min at room temperature.
Some decomposition noted at elevated temperature.
b
ꢀ
Table 1 along with the sp values [44] for these substituents.
(Since both the demethylation and substitution reactions
proceed through transition states in which increasing electroꢀn
density is directly conjugated with the aromatic ring, the sp
parameter set most accurately reflects the electronic effects for
the selected substituents.) All reactions were conducted at
100 8C in d6-DMSO in sealed NMR tubes in order to capture
Fig. 1. 1H NMR spectra showing the spectral shifts of 4-methyl-N,N,N-tri-
methylanilinium triflate in d6-DMSO upon addition of TBAF*.
strongly hydrogen bonded to this position. (The integrated cross
peak volumes at 3.8 and 8.2 ppm are almost equal.) The
experiments that confirm fluoride ion transfer from the
relatively weakly coordinating tetrabutylammonium cation to
the relatively strongly ion pairing trimethylanilinium cation are
quite important in the context of SNAr reactions of
trimethylanilinium salts, since they indicate that nothing is
to be gained from the synthesis of cations more weakly
coordinating than TBA for this type of fluorination reaction.
Similarly, the preparation of trimethylanilinium salts featuring
more weakly coordinating anions (PF6ꢀ, tetraarylborates) is
also shown to be unwarranted, since the synthetic studies
reported here are conducted at the reactivity limit for
displacement of trimethylamine in DMSO.
1
any volatiles produced during the reaction. H, 13C, and 19F
NMR spectral data were used to identify and quantify the
principal products of the fluorination reactions. In every case
examined, the reaction ran to completion and only two
fluorinated products were observed by NMR spectroscopy:
fluoroarene and/or methyl fluoride. Inspection of Table 1
indicates that there is a sharp demarcation between reactions
that yield almost exclusively the fluoroarene products of SNAr
reactions (X = NO2, CHO, CN, COPh, CO2Me) and those that
produce methyl fluoride as the principal product (X = Br, H,
CH3, OCH3). Interestingly, even the reaction featuring a
relatively electron rich arene ring (X = OMe) was still
completed relatively rapidly at elevated temperature. The
electron poor systems (X = NO2, CHO, CN, COPh, CO2Me)
are also substituted relatively rapidly at room temperature
under these anhydrous conditions, and the product ratios
obtained are virtually identical to those obtained at 100 8C. The
quantitative yields of the two products indicate that caution
should be exercised in radiochemical preparations employing
relatively electron rich trimethylanilinium salts, as 18F-labeled
methyl fluoride is sure to be produced in good yields in these
reactions. Taken together, the ion pairing and reaction data
suggest that these fluorination reactions proceed through the
collapse of ion pair intermediates, and that the product ratios
obtained in DMSO result only from the difference in activation
barriers for the SNAr and SN2 pathways. To test this hypothesis,
we examined the ion pairing, kinetics, and thermodynamics of
tetraalkylammonium dealkylation using computational meth-
ods.
2.2. Fluorination reactions
The results of fluorination reactions for the series of para-
substituted N,N,N-trimethylanilinium salts are reported in
2.3. Theoretical modeling of tetramethylammonium
fluoride
Fig. 2. 1H–19F HOESY spectra of TBAF* (left) and a mixture of TBAF* and 4-
Though Collie first reported [45,46] the preparation of
methyl fluoride by the decomposition of the ‘‘dry’’ salt of
methyl-N,N,N-trimethylanilinium triflate (right) in d6-DMSO solution.

H. Sun, S.G. DiMagno / Journal of Fluorine Chemistry 128 (2007) 806–812
809
tetramethylammonium fluoride (TMAF) at 180 8C, we are
unaware of any theoretical treatment of this simple reaction.
Two recent theoretical studies of TMAF have focused on the
nature and strength of the ion pairing interaction in the gas
phase and solid state. Harmon argued that IR spectra and ab
initio molecular orbital calculations suggest that TMAF
contains the strongest C–Hꢁ ꢁ ꢁFꢀ hydrogen bond yet observed
[47]. Davies, George and Howard examined the gas phase
structure of the TMAF ion pair and found the most stable ion
pairing arrangement to possess C3v symmetry, with one
hydrogen from each of three methyl groups forming close
contacts to the fluoride ion [48]. In contrast, in a classical SN2
attack trajectory the incoming fluoride ion interacts solely with
the hydrogen atoms of a single methyl group. We investigated
the gas phase and solution phase energetics of ion pairing and
fluorodemethylation at the MP2(6-311G+(d, p)) level. Basis set
superposition error (BSSE) corrections were made for the gas
phase calculations. Energies for reactions modeled in THF and
DMSO solution (PCM solvation treatment) were not corrected
for BSSE. The calculated structures and their free energies at
room temperature (ZPE and thermal corrections employed) are
summarized in Fig. 3.
Inspection of Fig. 3 shows that the separation of the
transition state and the lowest energy ion pair structure
Fig. 3. Results of MP2(6-311G+(d, p)) calculations for the TMAF ion pair and
for the reverse Menschutkin reaction (TMAF ! trimethylamine + CH3F) in the
gas phase, and in solution (THF, DMSO). Free energies are given for the
following calculated structures (from left to right) (1) C3v ion pair, (2)
dissociated TMA and fluoride ion, (3) the reaction side minima (RSM), (4)
transition state (TS), (5) product side minima (PSM), and (6) dissociated
products. The TS structures are labeled; the remaining, unlabeled structures
are true minima derived from gas phase calculations.
(DDG8 = DGꢂTS ꢀ DGiꢂon pair
)
varies only modestly with
solvent, (DDGꢂgas = 25.7 kcal/mol, DDGTꢂHF = 42.1 kcal/mol,
DDGꢂDMSO = 32.6 kcal/mol) with the largest barrier occurring
for THF. This result is easily explained. In THF, the carbon
atom in the transition state is nearly coplanar with its three
attached hydrogen atoms, while the gas phase calculation (early
TS) and DMSO calculation (late TS) show somewhat
pyramidal carbon atoms at the transition state.
salts, as some decomposition of the tetrabutylammonium cation
was observed. Nevertheless, this study indicates that the methyl
groups on TMA cation and related ammonium cations are quite
labile in the presence of anhydrous fluoride.
The calculations showing ready access to the transition state
for SN2 reactions of TMAF in DMSO prompted us to
investigate decomposition of TMAF in solution. When
TMAPF6 and TBAF were mixed in d6-DMSO and heated to
100 8C for 20 min, methyl fluoride and trimethylamine were
indeed formed in a reaction similar to that observed for the
electron rich trimethylanilinium triflates. Evidently, pyrolysis
of the dry salt is unnecessary for a relatively rapid reverse
Menschutkin reaction. Unfortunately, demethylation of TMAF
was not as clean as the reactions with the trimethylanilinium
2.4. Theoretical modeling of trimethylanilinium fluorides
Table 2 compares the two calculated (B3LYP/6-311G++(d,
p)) lowest energy optimized ion pair structures for trimethy-
1
lanilinium fluoride. In agreement with the H–19F HOESY
experiments, the structure with the fluoride ion located in the
Table 2
DFT (B3LYP/6-311G++(d, p)) calculated ion pair structures and free energies and enthalpies at room temperature
Ion pair structure
a
b
dF-Hph
dF-HMe
m (Debye)
DG8
DH8
dF-HMe
dF-HMe
m (Debye)
DG8
DH8
gas
1.764
1.939
1.979
1.804
2.036
2.125
9.563
13.845
15.071
ꢀ105.3
ꢀ6.4
3.7
ꢀ113.8
ꢀ13.9
ꢀ3.3
1.815
2.021
2.080
1.843
2.061
2.129
10.374
14.405
15.372
ꢀ105.5
ꢀ7.3
2.9
ꢀ113.6
ꢀ14.5
ꢀ4.0
THF
DMSO
a
Hydrogen bond length for the in-plane methyl proton and fluoride.
b
˚
Hydrogen bond length for the out-of-plane protons and fluoride; DG8 and DH8 are in kcal/mol and are referenced to the free ion values; distance in A; the PCM
solvation model is used to for THF and DMSO.

810
H. Sun, S.G. DiMagno / Journal of Fluorine Chemistry 128 (2007) 806–812
plane of the aromatic ring (Table 2, left) is calculated to be of
similar energy to that of the end-on fluoride structure
reminiscent of the C3v structure of TMAF (Table 2, right).
The exceptionally short ortho-aryl C–Hꢁ ꢁ ꢁFꢀ hydrogen bond
indicates that this is a remarkably strong interaction; the
predicted short interaction distance may also account for the
disproportionately large HOESY cross peak observed in the 2D
spectrum (vide supra).
r = ꢀ8.9) than they are for fluorination (gas phase:
r = ꢀ8.4, THF: r = ꢀ3.9, DMSO: r = ꢀ3.1) that the
demethylation product is predicted to be thermodynamically
favored for electron deficient ring systems under all conditions.
Similarly, and also contrary to experiment, the fluorination
product is predicted to be thermodynamically favored for
electron rich rings. The reason for this seeming discrepancy
between theory and experiment is, of course, that the product
formation is under kinetic control. Furthermore, since
development of ring electron density at the transition state is
expected to be larger for formation of the Meisenheimer
intermediate than for demethylation, the rate for SNAr
substitution is expected to be much more strongly affected
by the substituents. This classical kinetic argument is borne out
by experiment.
Previous computational work from our laboratory indicated
that the activation barrier for fluorodenitration reactions was
strongly dependent upon the substituent s-value [41]. This
strong substituent effect arises because the rate-determining
step in nucleophilic aromatic substitution is formation of a
Meisenheimer intermediate. The transition state for the
formation of the Meisenheimer intermediate exhibits extensive
negative charge injection into the aromatic ring. Thus, SNAr
reactions of trimethylanilinium triflates are expected to (and
do) show large substituent effects similar to nitroarenes. Here
we report thermodynamic studies for the competing SNAr and
reverse Menschutkin reactions, which, in conjunction with
experimental studies, indicate that solution phase trimethyla-
nilinium fluoride decomposition is indeed under kinetic
control.
3. Conclusions
Several important conclusions can be drawn from this work.
(1) The tetrabutylammonium cation is significantly more
weakly ion paired to fluoride than trimethylanilinium cations
are. (2) Trimethylanilinium fluorides are tightly ion paired in
DMSO solution, thus these ion pair reactions represent the
reactivity limit for this type of substitution reaction. (3) Under
anhydrous conditions in DMSO, decomposition of these salts is
rapid at 100 8C, regardless of the nature of the substituents on
the ring. (4) Except for the case of extremely electron poor
aromatic substrates ( para-nitro), methyl fluoride is always
generated in nucleophilic aromatic fluorination reactions of
trimethylanilinium salts. Precautions to guard against exposure
to this gaseous byproduct should be implemented if radiotracer
syntheses are conducted using trimethylanilinium precursors.
Fig. 4 summarizes the results of B3LYP/6-311G++(d, p)
calculations for the reaction of fluoride ion with substituted
trimethylanilinium salts in the gas phase and in DMSO and
THF solutions. The linear free energy plots show that the
thermodynamically most stable product is not obtained in
DMSO. The slopes of the plots for methyl transfer are so much
steeper (gas phase: r = ꢀ12.7, THF: r = ꢀ9.2, DMSO:
ꢀ
(5) The aryl/alkyl substitution ratio is correlated to the sp
value of the substituent. (6) Theoretical studies are consistent
with the notion that fluorodeamination is under kinetic control;
the stability of the Meisenheimer intermediate, not the overall
driving force for the reaction, determines the rate and extent of
arene fluorination. Taken together, these studies imply that
fluorodeamination of trimethylanilinium triflates may be
intrinsically limited only by competitive demethylation, and
suggest that relatively electron rich anilinium salts featuring
more robust substituent groups could be fluorinated more
effectively than trimethylanilinium triflates.
4. Experimental
4.1. General experimental procedures
All reagents were handled under N2. Hexafluorobenzene
(C6F6) (99%, SynQuest) was passed through a column of
activated (130 8C for 5 h) silica gel and distilled from CaH2.
Acetonitrile (HPLC grade, Aldrich) was distilled from P2O5
and redistilled under reduced pressure from CaH2. THF
(anhydrous, Aldrich) was vacuum transferred from LiAlH4.
DMSO was distilled under N2 from CaH2. Purified solvents
were stored under N2 in Schlenk-style, Teflon-capped storage
flasks. Tetra-n-butylammonium cyanide (TBACN) (97%) was
Fig. 4. Calculated free energy changes for both demethylation and fluorination
of trimethylanilinium salts in the gas phase and in solution (THF, DMSO).
(B3LYP/6-311G++(d, p) with PCM solvation model).

H. Sun, S.G. DiMagno / Journal of Fluorine Chemistry 128 (2007) 806–812
[2] Y.-S. Ding, J.S. Fowler, Drug Dev. Res. 59 (2003) 227–239.
811
obtained from Fluka Chemical Co. and dried under vacuum at
40 8C overnight prior to use. TBAF* was prepared from
TBACN as was described previously [40]. Tetramethylammo-
nium hexafluorophosphate (TMAPF6) was obtained from Fluka
and dried under vacuum. All other reagents were of analytical
grade, from Aldrich Chemical Company. 1H, 13C and 19F NMR
spectra were collected and analyzed in the Instrumentation
Center at the University of Nebraska-Lincoln. 400 MHz (QNP
probe) and 600 MHz (HF probe) NMR spectrometers were
used in this study.
[3] J.S. Fowler, N.D. Volkow, G.-J. Wang, Y.-S. Ding, S.L. Dewey, J. Nucl.
Med. 40 (1999) 1154–1163.
[4] M.-C. Lasne, C. Perrio, J. Rouden, L. Barre, D. Roeda, F. Dolle, C.
Crouzel, Top. Curr. Chem. 222 (2002) 201–258.
[5] R. Weissleder, Science 312 (2006) 1168–1171.
[6] T. Irie, K. Fukushi, T. Ido, Int. J. Appl. Rad. Isotopes 33 (1982) 445–448.
[7] T. Irie, K. Fukushi, O. Inoue, T. Yamasaki, T. Ido, T. Nozaki, Int. J. Appl.
Rad. Isotopes 33 (1982) 633–636.
[8] G. Angelini, M. Speranza, A.P. Wolf, C.Y. Shiue, J. Fluorine Chem. 27
(1985) 177–191.
[9] C.Y. Shiue, J.S. Fowler, A.P. Wolf, D.W. McPherson, C.D. Arnett, L.
Zecca, J. Nucl. Med. 27 (1986) 226–234.
[10] M.S. Haka, M.R. Kilbourn, G.L. Watkins, S.A. Toorongian, J. Labell.
Compd. Radiopharm. 27 (1989) 823–833.
4.2. Computational studies
[11] T.J. McCarthy, A.U. Sheriff, M.J. Graneto, J.J. Talley, M.J. Welch, J. Nucl.
Med. 43 (2002) 117–124.
All calculations were performed on a 128 node computer
(Prairie Fire) at the University of Nebraska-Lincoln computing
facility. The Gaussian-view 03W interface and Gaussian-03
programs were used for the calculations.
[12] S. Oya, S.R. Choi, H. Coenen, H.F. Kung, J. Med. Chem. 45 (2002) 4716–
4723.
[13] A.A. Wilson, R.F. Dannals, H.T. Ravert, M.S. Sonders, E. Weber, H.N.
Wagner Jr., J. Med. Chem. 34 (1991) 1867–1870.
[14] Y. Seimbille, M.E. Phelps, J. Czernin, D.H.S. Silverman, J. Labell.
Compd. Radiopharm. 48 (2005) 829–843.
4.3. Syntheses of trimethylanilinium triflate salts [20]
[15] G. Getvoldsen, A. Fredriksson, N. Elander, S. Stone-Elander, J. Labell.
Compd. Radiopharm. 47 (2004) 139–145.
To
a
solution of para-substituted dimethylaniline
[16] B. Kuhnast, S. Klussmann, F. Hinnen, R. Boisgard, B. Rousseau, J.P.
Fuerste, B. Tavitian, F. Dolle, J. Labell. Compd. Radiopharm. 46 (2003)
1205–1219.
(4.0 mmol) dissolved in dry benzene (10 mL) was added
methyl triflate (0.63 mL) at room temperature. The reaction
mixture was stirred at room temperature to 40 8C for 30 min to
8 h. At the conclusion of the reaction the trimethylanilinium
triflate salt precipitated from solution. The precipitate was
isolated by filtration, washed with dry benzene (5ꢃ 5 mL), and
residual solvent was removed in vacuo. Excellent yields (>80%
isolated) were obtained for all of the methylation reactions.
These previously reported compounds were characterized by
1H, 19F, and 13C NMR spectroscopy.
[17] B. Kuhnast, F. Hinnen, R. Boisgard, B. Tavitian, F. Dolle, J. Labell.
Compd. Radiopharm. 46 (2003) 1093–1103.
[18] E.F.J. de Vries, J. Vroegh, P.H. Elsinga, W. Vaalburg, Appl. Radiat.
Isotopes 58 (2003) 469–476.
[19] B. Kuhnast, F. Dolle, B. Tavitian, J. Labell. Compd. Radiopharm. 45
(2002) 1–11.
[20] B. Kuehnast, F. Dolle, S. Terrazzino, B. Rousseau, C. Loc’h, F. Vaufrey, F.
Hinnen, I. Doignon, F. Pillon, C. David, C. Crouzel, B. Tavitian, Bioconj.
Chem. 11 (2000) 627–636.
[21] B. Kuhnast, F. Dolle, F. Vaufrey, F. Hinnen, C. Crouzel, B. Tavitian, J.
Labell. Compd. Radiopharm. 43 (2000) 837–848.
4.4. Reaction of TBAF* with trimethylanilinium triflate
salts
[22] T. Haradahira, Y. Hasegawa, K. Furuta, M. Suzuki, Y. Watanabe, K.
Suzuki, Appl. Radiat. Isotopes 49 (1998) 1551–1556.
[23] F. Dolle, F. Hinnen, F. Vaufrey, B. Tavitian, C. Crouzel, J. Labell. Compd.
Radiopharm. 39 (1997) 319–329.
To a solution of the appropriate trimethylanilinium triflate
salt (0.2 mmol in 0.3 mL of d6-DMSO) was added a solution of
TBAF* (0.26 mmol) in 0.2 mL d6-DMSO. The resulting
solution was mixed well and transferred into an NMR tube
equipped with a Teflon screw cap closure. The NMR tube was
placed into a preheated (100 8C) oil bath for the time indicated
in Table 1. The mixture was cooled by immersing the tube in a
cold water bath before the solution was assayed by 1H, 19F, and
13C NMR spectroscopy. The identities of all products were
confirmed by comparison of the spectra to those of authentic
samples.
[24] T. Toyokuni, J.C. Walsh, A. Dominguez, M.E. Phelps, J.R. Barrio, S.S.
Gambhir, N. Satyamurthy, Bioconjug. Chem. 14 (2003) 1253–1259.
[25] T. Poethko, M. Schottelius, G. Thumshirn, U. Hersel, M. Herz, G.
Henriksen, H. Kessler, M. Schwaiger, H.-J. Wester, J. Nucl. Med. 45
(2004) 892–902.
[26] H.T. Ravert, I. Madar, R.F. Dannals, J. Labell. Compd. Radiopharm. 47
(2004) 469–476.
[27] R.R. Kavali, B.C. Lee, B.S. Moon, S.D. Yang, K.S. Chun, C.W. Choi, C.-
H. Lee, D.Y. Chi, J. Labell. Compd. Radiopharm. 48 (2005) 749–758.
[28] E.D. Hostetler, S.D. Jonson, M.J. Welch, J.A. Katzenellenbogen, J. Org.
Chem. 64 (1999) 178–185.
[29] J. Ermert, K. Hamacher, H.H. Coenen, J. Labell. Compd. Radiopharm. 43
(2000) 1345–1363.
[30] Z. Li, Y.-S. Ding, A. Gifford, J.S. Fowler, J.S. Gatley, Bioconjug. Chem.
14 (2003) 287–294.
Acknowledgments
[31] W.R. Banks, R.D. Borchert, D.R. Hwang, Appl. Radiat. Isotopes 45 (1994)
75–81.
We thank the Nebraska Research Initiative for financial
support. This research was performed in facilities renovated with
funds from the National Institutes of Health (RR16544-01).
[32] W.R. Banks, D.R. Hwang, Appl. Radiat. Isotopes 45 (1994) 599–608.
[33] L. Lang, E. Jagoda, B. Schmall, B.-K. Vuong, H.R. Adams, D.L. Nelson,
R.E. Carson, W.C. Eckelman, J. Med. Chem. 42 (1999) 1576–1586.
[34] A.R. Jalilian, S.A. Tabatabai, A. Shafiee, H. Afarideh, R. Najafi, M.
Bineshmarvasti, J. Labell. Compd. Radiopharm. 43 (2000) 545–555.
[35] K.J. Guenther, S. Yoganathan, R. Garofalo, T. Kawabata, T. Strack, R.
Labiris, M. Dolovich, R. Chirakal, J.F. Valliant, J. Med. Chem. 49 (2006)
1466–1474.
References
[1] M.R. Kilbourn, Fluorine-18 Labeling of Radiopharmaceuticals, National
Academy Press, 1990.

812
H. Sun, S.G. DiMagno / Journal of Fluorine Chemistry 128 (2007) 806–812
[36] Y. Seimbille, J. Czernin, M.E. Phelps, D.H.S. Silverman, J. Labell.
Compd. Radiopharm. 48 (2005) 819–827.
[41] H. Sun, S.G. DiMagno, Angew. Chem. Int. 45 (2006) 2720–
2725.
[37] J. Ermert, C. Hocke, T. Ludwig, R. Gail, H.H. Coenen, J. Labell. Compd.
Radiopharm. 47 (2004) 429–441.
[42] H. Sun, S.G. DiMagno, Chem. Commun. (2007) 528–529.
[43] I. Fernandez, E. Martinez-Viviente, P.S. Pregosin, Inorg. Chem. 44 (2005)
5509–5513.
[38] J.-L.M. Abboud, R. Notario, J. Bertran, M. Sola, Prog. Phys. Org. Chem.
10 (1993) 1–183.
[44] C. Hansch, A. Leo, R.W. Taft 91 (1991) 165–195.
[45] N. Collie, J. Chem. Soc. 55 (1889) 110–113.
[46] A.T. Lawson, N. Collie, J. Chem. Soc. 53 (1888) 624–636.
[47] K.M. Harmon, S.L. Madeira, J. Mol. Struct. 560 (2001) 179–188.
[48] A.S. Davies, W.O. George, S.T. Howard, Phys. Chem. Chem. Phys. 5
(2003) 4533–4540.
[39] W.R. Banks, M.R. Satter, D.R. Hwang, Appl. Radiat. Isotopes 45 (1994)
69–74.
[40] H. Sun, S.G. DiMagno, J. Am. Chem. Soc. 127 (2005) 2050–2051;
The asterisk is used here to denote the anhydrous reagent and to
distinguish it from the commercially available hydrated salt.