Fluorination of aromatic compounds from 1-aryl-3,3-dimethyltriazenes and fluoride anions in acidic medium: 1. A model for 18F labelling

Article abstract of DOI:10.1016/S0022-1139(00)00376-6

Decomposition of 1-aryl-3,3-dialkyltriazenes by strong and non-nucleophilic acids (CF₃SO₃H, H₂SO₄) can be achieved at 90°C in carbon tetrachloride, in the presence of sub-stoichiometric amounts of fluoride anions. Yields of fluoroarenes up to 39% (versus F^-) can be reached under these conditions, within 15-30min and without contamination of the fluorinated compounds by inseparable by-products. Such conditions and results enable an adaptation to radiofluorination with ¹⁸F^-.

Full text of DOI:10.1016/S0022-1139(00)00376-6

Journal of Fluorine Chemistry 107 (2001) 321±327
Fluorination of aromatic compounds from 1-aryl-3,3-dimethyltriazenes
and ¯uoride anions in acidic medium
18
1. A model for F labelling
Thierry Pages, Bernard R. Langlois*
Â
Laboratoire SERCOF (UMR 5622), Universite Claude Bernard-Lyon I, 43 Bd du 11 Novembre 1918, F-69622 Villeurbanne, France
Received 15 May 2000; accepted 3 August 2000
Abstract
Decomposition of 1-aryl-3,3-dialkyltriazenes by strong and non-nucleophilic acids (CF3SO3H, H2SO4) can be achieved at 908C in carbon
tetrachloride, in the presence of sub-stoichiometric amounts of ¯uoride anions. Yields of ¯uoroarenes up to 39% (versus F ) can be reached
under these conditions, within 15±30 min and without contamination of the ¯uorinated compounds by inseparable by-products. Such
conditions and results enable an adaptation to radio¯uorination with 18F . # 2001 Elsevier Science B.V. All rights reserved.
Keywords: Fluorination; Triazene; Fluoride; Fluorine 18
1. Introduction
introducing ¯uorine on aromatic nuclei through the acidic
decomposition of 1-aryl-3,3-dialkyltriazenes in the presence
of ¯uoride anions. The advantage of this technique would
rest on the properties of the triazeno moiety, a stable
equivalent of a diazo group [7], which could be introduced
in the early stages of the synthetic chain. Then, the triaze-
nated product could be isolated, puri®ed and functionalized,
provided that the following reactions are performed under
non-acidic conditions [8]. So, the ¯uorodetriazenation step
could be carried on sophisticated triazenes, in the latest
stages of the synthesis. To verify the key-step of our strategy,
we examine, in this paper, the transformation of simple
aryltriazenes into ¯uoroaromatics with ¯uorides.
Positron emission tomography (PET) is now routinely
used for the medical imaging of the human body because of
the short half-life and huge speci®c radioactivity of positron
emitters which allow repetitive administration of very low
‡
radiation doses [1]. 18F is one of the most popular b -
labelling isotope because its rather long half-time
(t1=2 ˆ 110 min) is compatible with the preparation of
sophisticated [18F]-labelled drugs and the study of slow
biological processes [2,3]. For example, (S)-6-[18F]¯uoro-
DOPA is routinely used, as well as its a-methyl derivative
[5], to study neurotransmission disorders such as epilepsy
[1], Alzheimer's disease [1] or Parkinson's disease [4]. It
also constitutes a potent tool for pharmacological studies in
vivo [6]. Other aromatic aminoacids (or derivatives) and
neuroleptics have been also described [2].
2. Results and discussion
‡
Usually, the preparation of b -radiolabelled tracers must
1-Aryl-3,3-dialkyltriazenes are readily obtained from
arenediazonium chlorides and secondary amines [9±11].
They can be easily isolated and puri®ed. Their transforma-
tion into ¯uoroarenes by a large excess of hydrogen ¯uoride
has been recently reviewed [10]. This technique has been
adapted to 18F incorporation by using limited amounts of
H18F in protic solvents [12,13]. But, as H18F is not easily
available, it has been replaced by Cs18F (in de®ciency) and
an auxiliary acid (i.e. CH3SO3H used in excess), the reaction
being carried out in an aprotic solvent [14]. Nevertheless, the
results are dramatically dependent on the nature of the
auxiliary acid and that of the solvent [15,16]. Especially,
be arranged in such a way that the short-lived radioisotope is
introduced in the later stages of the synthetic chain in order
to get the best speci®c radioactivity. Unfortunately, until
now, the synthesis of [18F]¯uoroaromatic aminoacids does
not respect this criterion since ¯uorine is usually introduced
in the early stages by a nucleophilic substitution with 18F
¯uoride anion, which is the most available source of 18F. In
our opinion, this problem could be circumvented by lately
* Corresponding author. Tel.: ‡33-4-72448163; fax: ‡33-4-72431323.
E-mail address: langlois@univ-lyon.fr (B.R. Langlois).
0022-1139/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 0 0 ) 0 0 3 7 6 - 6

322
T. Pages, B.R. Langlois / Journal of Fluorine Chemistry 107 (2001) 321±327
Scheme 1. Competitive processes during fluorodediazoniation.
in hydrogen-containing solvents, large amounts of ArH are
produced through radical hydrodetriazenation along with
Ar18F (formed by ionic ¯uorodetriazenation). This consti-
tutes a severe limitation since these two compounds are very
dif®cult to separate, especially on the nanomole scale where
radio¯uorination must be carried out. Thus, we reinvesti-
gated the acidic decomposition of aryltriazenes in the pre-
sence of non-radioactive ¯uoride anions (19F ), in order to
rationalise the medium effects and ®nd conditions which
could be realistically transposed to radio¯uorination.
During ¯uorodediazoniation in hydrogen- or chloro-con-
taining solvents, several reactions can be considered
(Scheme 1).
To build a good model for radio¯uorination, the para-
meters must be consequently chosen in order that the radical
processes, which produce ArH, can be suppressed and the
ionic processes maximised, even if heavier products (Ar±Cl,
Ar±A, Ar±Solv) are formed, since they can be easily sepa-
rated from Ar±F. For example, solvents such as DMSO,
HMPA, methanol or pyridine, which are known to reduce
arenediazonium cations [17±23], must be avoided. We also
preferred caesium ¯uoride, which can be ef®ciently dried, to
more soluble but hygroscopic tetraalkylammonium ¯uor-
ides, to avoid phenol formation. Finally, tri¯ic and sulfuric
acids were chosen as auxiliary acids because their conju-
gated bases have a suf®ciently low nucleophilicity to mini-
mise the formation of Ar±A.
The in¯uence of the different parameters have been
studied with simple 1-aryl-3,3-dialkyltriazenes (1±5)
(Scheme 2) prepared according to Wallach [9], either in
the presence of a base (3±5) or not (1±2). They were reacted,
in a closed Te¯on PFA1 vessel, with dry CsF in the presence
of a solvent and an acid. Several experiments were done with
an excess of CsF (to be compared to already reported results)
but, to be consistent with radio¯uorination, others were
carried out with a de®ciency of CsF (in this case, ¯uoroaro-
matic yields are given versus CsF).
The ®rst experiments, carried out with
1
(C ˆ
0:11 mol 1), showed that the ratio between the liquid and
the gas volumes in the closed vessel affected the ratio
between ionic and radical processes. When the liquid
volume increased, that means when the partial pressure of
nitrogen in the gas phase increased, the ionic processes
decreased. This observation is consistent with a postulated
Scheme 2. Starting aryltriazenes (yields of isolated compounds).

T. Pages, B.R. Langlois / Journal of Fluorine Chemistry 107 (2001) 321±327
323
ꢀ Procedure C: (1) CsF ‡ triazene ‡ solvent; (2) acid (AH)
at 108C.
‡
‡
Scheme 3. Equilibrium between ArN2 , Ar and N2.
Procedure A led to the formation of HF in situ before
addition of the triazene. When CsF was more abundant than
‡
AH, the triazene was protonated by a mixture of HF and Cs
HF2 . When CsF was less abundant than AH, the triazene
‡
‡
equilibrium between ArN2 , Ar and N2 (Scheme 3)
‡
‡
[20,24], the better reducibility of ArN2 (versus Ar )
[17] and the fact that most ionic reactions arise from
was protonated by HF along with AH. Procedure B led ®rst
‡
to the formation of ArN2
A which, then, was mainly
‡
Ar . Consequently, all the further reactions were performed
with a constant ratio between liquid and gas volumes
(Vl=Vg ˆ 0:3).
decomposed by F (A was presumed to be less nucleo-
philic) and, eventually, the solvent. Procedure C constituted
an intermediate case in which AH could competitively
protonate F and the triazene. When an excess of CsF
(versus triazene) was used, the comparative results in carbon
tetrachloride and tri¯uoroethanol are summarised in Table 2.
In carbon tetrachloride (entries 1±3), ¯uoroarene forma-
tion was only slightly affected by the order of introduction of
the reagents, though a little higher yield was observed with
procedure A. In such a non-polar and viscous medium, all
the reactive species, except triazene, are almost insoluble
and interfacial processes are important. Reactions in solvent
Then, reaction temperature and duration were examined.
As for arenediazonium ¯uorides [10], the temperature
required to signi®cantly transform aryltriazenes is mainly
dependent on the nature of the substrate. A 0.11 M solution
of the less reactive starting material 1 must be heated up to
908C for 1 h to be completely converted by tri¯ic acid and
caesium ¯uoride. Consequently, this temperature was
adopted for the further experiments all the more so since
we observed that the ratio between ionic processes (Ar±
F ‡ Ar±A ‡ Ar±Solv) and radical processes (Ar±H)
increased when raising the temperature (Table 1). Table 1
also shows that ArF/ArH increased with the reaction time. In
other words, radical processes are both thermodynamically
and kinetically favoured. It can be noticed (especially from
3) that the maximum yield of ArF can be reached within 15±
30 min, a duration that is compatible with radio¯uorination.
As the introduction order of the reaction components,
before heating, could be also important, three procedures
were compared.
cages are also favoured. This phenomenon could explain
‡
why procedure A, in which ArN2 F is ®rst formed in a
solvent cage from triazene and HF/HF2 , delivered a better
‡
yield of ArF. In contrast, procedure B, in which ArN2 TfO
was ®rst formed in a solvent cage, favoured the formation of
ArOTf, despite the low nucleophilicity of TfO .
It must be also noticed that procedure B was the only one
which prevented radical process to occur (entry 2) whereas
ArCl and ArH were formed from procedures A and C
(entries 1 and 3). The occurrence of the latter, in a medium
which did not contain any evident reducer or hydrogen atom
donor, was surprising. To explain rather similar experiments,
Satyamurphy et al. [16] postulated that an intramolecular
single electron transfer could occur, from the anions to
ꢀ Procedure A: (1) CsF ‡ solvent; (2) acid (AH) at 108C;
(3) triazene at 108C.
ꢀ Procedure B: (1) triazene ‡ solvent; (2) acid (AH) at
108C; (3) CsF at 108C.
Table 1
Influence of the reaction temperature and time
Substrate; CsF (x eq); TfOH (y eq)
1; CsF (5 eq); TfOH (4 eq)
y8C
Time (min)
Ar±Fa(%)
Ar±OTfa,b(%)
Ar±Ha(%)
ArH/ArF
45
80
90
60
60
60
13
24
26
7
11
12
4
0.31
e
Trace
0
0
3; CsF (0.2 eq); TfOH: (2.1 eq)
90
90
90
90
60
30
15
9
26
25
24
16
12
7
0
1
1
2
0
0.04
0.04
0.13
7
7
a The major product ArOCH2CF3 (from incorporation of the solvent) was not determined.

324
T. Pages, B.R. Langlois / Journal of Fluorine Chemistry 107 (2001) 321±327
Table 2
Influence of the introduction of the reaction components (CsF in excess)
Entry
Solvent
CCl4
Procedure (see text)
ArF (%)
ArOTf (%)
ArH (%)
ArCl (%)
1
2
3
A
B
C
23
20
20
8
48
40
4
0
5
3
0
4
4
5
CF3CH2OHa
A
C
24
26
13
12
4
0
±
±
a The major product ArOCH2CF3 (from incorporation of the solvent) was not determined.
‡
ArN2 , in arenediazonium tri¯uoroacetates and methane-
sulfonates. This internal redox process can be ruled out for
arenediazonium tri¯ates ®rst because of the very low oxid-
ability of the tri¯ate anion and secondly because no ArH or
When a de®ciency of CsF (0.2 eq) and a slight excess of
‡
auxiliary acid (required stoichiometry: 2.0 eq of H ) was
used in CCl4, procedure A delivered again the best yield of
ArF. However, only traces of ArH were detected whatever
the procedure (Table 3). This is probably due to a greater
acidity of the medium, which was buffered by a smaller
amount of ¯uoride anions (HF only is formed, instead of
HF ‡ HF2 ), and, consequently, there is a better protonation
of the triazene.
‡
ArCl appeared in procedure B where ArN2 TfO was
initially formed. So, we suspected that the non-protonated
triazene could be the hydrogen donor.
This proposition was corroborated by the in¯uence of the
triazene concentration: when the experiment reported in
entry 5 (Table 2) was repeated with a higher concentration
of triazene (0.43 mol l instead of 0.11 mol l 1), the for-
mation of ArF and ArOTf decreased (respective yields: 10
and 8%) whereas ArH was extensively produced
(yield ˆ 12%). Such a hypothesis was also consistent with
the in¯uence of the introduction mode of the reactants
(Table 2, entries 1±3). During procedure B, the triazene
was ®rst completely protonated by strong tri¯ic acid; during
procedure A, the triazene was probably incompletely pro-
tonated by the weaker acid HF/HF2 , in procedure C, tri¯ic
acid and HF competed for the protonation of the triazene.
The same trends were observed in tri¯uoroethanol (Table 2,
entries 4±5) where all reactive species were soluble: radical
processes mainly occurred with procedure A. Thus, reduc-
tion of the diazonium moiety by the non-protonated triazene
can be proposed and rationalised as follows (Scheme 4).
It must be also noticed, from Table 3, that oleum was the
best auxiliary acid. As 1.5 equivalent of sulfuric acid was
suf®cient, it acted as a diacid. Larger amounts decreased the
yield of ¯uoroarene, probably because HF was partly com-
plexed by sulfur trioxide to form ¯uorosulfonic acid and,
consequently, was not available for ¯uorination. If tri¯ic
acid delivered lower yields of ArF than sulfuric acid, it could
be due to its higher acidity and its larger ability to protonate
1
‡
HF and form H2F which is not an ef®cient ¯uorinating
agent.
All the above-reported experiments were carried out in
carbon tetrachloride or tri¯uoroethanol. Other solvents were
also evaluated but none of them gave satisfactory results:
ꢀ ArH was the major product in THF (through a radical
chain process) [25] but also in diglyme or ethyl acetate,
Scheme 4. Reduction of a diazonium group by a triazene.

T. Pages, B.R. Langlois / Journal of Fluorine Chemistry 107 (2001) 321±327
325
Table 3
Influence of the introduction of the reaction components (CsF in deficiency)
Acid (x eq)
Procedure
ArF (%)
ArH (%)
ArF/ArCl
CF3SO3H (2.1 eq)
A
C
26
24
Trace
Trace
1.3
1.4
Oleum (6 wt.% SO3) (2.1 eq) (mol/mol)
Oleum (6 wt.% SO3) (1.5 eq) (mol/mol)
a N.D.: not determined.
A
C
36
25
Trace
Trace
5.9
N.D.a
A
C
39
33
Trace
Trace
4.0
2.7
which are good hydrogen atom donors, as well as in
formic acid, known for its reducing properties. Hydrogen
abstraction was less pronounced in cyclohexane but,
nevertheless, also extensive compared to fluoride incor-
poration.
in the presence of sub-stoichiometric amounts of ¯uoride
anions. Yields of ¯uoroarenes up to 39% (versus F ) can be
reached under these conditions, within 15±30 min and with-
out contamination of the ¯uorinated compounds by insepar-
able by-products. Such conditions and results enable an
adaptation to radio¯uorination with 18F .
ꢀ Nucleophilic solvents such as methanol, ethanol (even
trichloroethanol), formamide, acetonitrile, nitromethane
and DMSO [18] competed too strongly with fluoride to be
suitable for fluorodetriazenation.
3. Experimental section
Caesium ¯uoride (from Acros, purity: 99%) was dried at
2708C under atmospheric pressure, cooled under vacuum,
weighted in the reaction vessel (made of Te¯on PFA1) and
stored there at 1108C overnight. Solvents and acids (except
sulfuric acid) were freshly distilled and stored over mole-
cular sieves.
At the ®rst glance, tri¯uoroethanol seemed more suitable
than carbon tetrachloride because it dissolved all the reac-
tants, often provided better yields of ¯uoroaromatics and
could avoid radical by-products. However, too good yields
(>100%) were sometimes obtained in it (Table 4, entry 2)
and, ®nally, we found that this solvent could act as a ¯uoride
source, even in the absence of CsF (Table 4, entry 7). To a
lesser extent, ethyl tri¯uoroacetate (Table 4, entry 4) and
tri¯uoroacetic acid (Table 4, entry 9) behaved in the same
way, but tri¯ic acid did not (Table 4, entry 10).
Thus, tri¯uoroethanol, ethyl tri¯uoroacetate and tri¯uor-
oacetic acid cannot be used as solvents for radio¯uorination
with 18F. Indeed, when the experiment described in Table 4
(entry 6) was repeated with Cs18F in [19F]tri¯uoroethanol,
no [18F]¯uorobenzene but rather [19F]¯uorobenzene was
formed and the radioactivity was recovered in the solvent
front during HPLC. The ability of tri¯uoroethanol to act as a
¯uoride donor (Scheme 5) is due to the stabilisation of the
resulting carbocation by two other ¯uorine atoms. The same
is true for ethyl tri¯uoroacetate, though to a lesser extent
because of the electron-withdrawing effect of the carboxyl
group, but cannot be invoked for tri¯ic acid.
3.1. Procedure A
The solvent CCl4 (4 ml) was added, through a septum, on
CsF (0.18 mmol) conditioned as above and kept under
nitrogen. The suspension was cooled at 108C then tri¯ic
acid (1.8 mmol) was added through the septum. The result-
ing mixture was stirred at 108C for 5 min then the triazene
(0.9 mmol), dissolved in CCl4 (4 ml), was dropped in. After
this ®nal addition, the vessel was tightly closed, kept 5 min
more at 108C then immersed in an oil bath preheated at
908C. The reaction medium was held at this temperature
under magnetic stirring.
3.2. Procedure B
As above, the triazene (0.9 mmol) and CCl4 (8 ml) were
introduced through a septum into the empty vessel kept
In conclusion, decomposition of 1-aryl-3,3-dialkyltria-
zenes by strong and non-nucleophilic acids (CF3SO3H,
H2SO4) can be achieved at 908C in carbon tetrachloride,
under nitrogen. After cooling at
108C, tri¯ic acid

326
T. Pages, B.R. Langlois / Journal of Fluorine Chemistry 107 (2001) 321±327
Table 4
Influence of the solvent
Entry
Substrate
Solventa
Procedure
CsF (eq)
Time (h)
ArF (%)b
ArOTf (%)
ArH (%)
Ar±Solvent (%)
c
1
2
4
4
4
4
6
6
6
6
6
6
CCl4
TFE
C
C
A
A
C
C
C
A
A
C
0.2
0.2
0.2
0
0.5
0.5
26
110
7.2
7.1
31
30
30
8
74
Trace
1
±
N.D.
39
N.D.
3
ETFA
ETFA
CCl4
TFE
0.25
0.25
0.5
Trace
Trace
0
±
±
4
N.D.
93
c
5
0.2
0.2
0
±
6
1
11
0
60
7
TFE
TFA
0.25
0.25
24
18
0
56
8
0.2
0
N.D.
N.D.
25d
N.D.
N.D.
0
N.D.
N.D.
25d
9
TFA
4
10
TfOH
0
0.5
0
a TFE: trifluoroethanol; ETFA: ethyl trifluoroacetate; TFA: trifluoroacetic acid.
b Versus CsF.
c ArCl not determined.
d Isolated yield.
(1.8 mmol) was dropped in and the resulting mixture was
kept under stirring at this temperature for 5 min. Then the
septum was removed as brie¯y as possible to introduce,
under a stream of nitrogen, CsF (0.18 mmol) which was
previously conditioned as indicated. After this ®nal addition,
the vessel was tightly closed, kept 5 min more at 108C then
immersed for the required time in an oil bath preheated at
908C.
3.3. Procedure C
The triazene (0.9 mmol), dissolved in CCl4 (8 ml), was
added, through a septum, onto CsF (0.18 mmol) conditioned
as above and kept under nitrogen. The suspension was
cooled at 108C then tri¯ic acid (1.8 mmol) was dropped
through the septum. The resulting mixture was stirred at
108C for 5 min then the vessel was tightly closed, kept
Scheme 5. Trifluoroethanol as fluoride source.

T. Pages, B.R. Langlois / Journal of Fluorine Chemistry 107 (2001) 321±327
327
[4] P. Morrish, J. Rakshi, D.J. Brooks, in: D. Comar (Ed.), PET for Drug
Development and Evaluation, Kluwer Academic Publishers, Dor-
drecht, 1995, pp. 179±188.
5 min more at 108C and immersed for the required time in
an oil bath preheated at 908C.
[5] P. Damhaut, Contribution to the synthesis of
radiopharmaceuticals for the in vivo study of the dopaminergic
Á
system with PET, Ph.D. Thesis, University of Liege, Belgium, 1994.
[
18F]fluorinated
3.4. Common work up
[6] K. Jorga, K.L. Leenders, I. GuÈnther, M. Psylla, in: D. Comar (Ed.),
PET for Drug Development and Evaluation, Kluwer Academic
Publishers, Dordrecht, 1995, pp. 189±195.
After cooling, 0.3 M aqueous KOH (5 ml) was added and
the two phases were separated. The aqueous phase was
extracted with ethyl ether (3 Â 5 ml) and the collected
organic phases were washed with water (3 Â 5 ml), dried
over Na2SO4 and ®ltered.
[7] M.L. Gross, D.H. Blank, W.M. Welch, J. Org. Chem. 58 (1993) 2104.
[8] M.D. Theadhill, M.F.G. Stevens, Synthesis (1983) 289.
[9] O. Wallach, Justus Liebigs Ann. Chem. (1886) 233 and 255.
[10] B. Langlois, in: B. Baasner, H. Hagemann, J.C. Tatlow (Eds.),
Houben±Weyl Methods of Organic Chemistry, Vol. E10a, Thieme,
Stuttgart, 1999, Chapter 26, pp. 686±740.
3.5. Analyses
[11] P.J. Das, S. Khound, Tetrahedron 50 (1994) 9107.
[12] J.S. Ng, J.A. Katzenellenbogen, M.R. Kilbourn, J. Org. Chem. 46
(1981) 2520.
Most of reaction media were analysed by GC with an
internal standard (4-chloro¯uorobenzene) against which
most of the simplest products have been calibrated. These
products were either commercially available or prepared
according to the literature (4-methyltri¯ate, 4-(tri¯uor-
oethoxy)acetophenone). Some reaction media were also
analysed by 19F NMR with an internal standard (PhOCF3).
[13] M.N. Rosenfeld, D.A. Widdowson, J. Chem. Soc., Chem. Commun.
(1979) 914.
[14] T.J. Tewson, M.J. Welch, J. Chem. Soc., Chem. Commun. (1979) 1149.
[15] J.R. Barrio, N. Satyamurphy, H. Ku, M.E. Phelps, J. Chem. Soc.,
Chem. Commun. (1983) 443.
[16] N. Satyamurphy, J.R. Barrio, D.G. Schmidt, C. Kammerer, G.T. Bida,
M.E. Phelps, J. Org. Chem. 55 (1990) 4560.
[17] C. Galli, Chem. Rev. 88 (1988) 765.
[18] Y. Hirose, G.H. Wahl, H. Zollinger, Helv. Chim. Acta 59 (1976) 1427.
References
È
È
[19] F. Trondlin, C. Ruchardt, Chem. Ber. 110 (1977) 2494.
[20] H. Zollinger, Angew. Chem. Int. Ed. Eng. 17 (1978) 141.
[21] H. Loewenschuss, G.H. Wahl Jr., H. Zollinger, Helv. Chim. Acta 59
(1976) 1438.
[1] D. Comar, PET for Drug Development and Evaluation, Kluwer
Academic Publishers, Dordrecht, 1995.
[2] J.S. Fowler, in: R. Filler, Y. Kobayashi, L.M. Yagupolskii (Eds.),
Organofluorine Compounds in Medicinal Chemistry and Biomedical
Applications, Elsevier, Amsterdam, 1993, pp. 309±338.
[22] P. Burri, H. Loewenschuss, H. Zollinger, Helv. Chim. Acta 57 (1974)
395.
[23] I. Szele, H. Zollinger, Helv. Chim. Acta 61 (1978) 1721.
[24] Y. Hashida, R.G.M. Landells, G.E. Lewis, I. Szele, H. Zollinger, J.
Am. Chem. Soc. 100 (1978) 2816.
[3] Y.S. Ding, J.S. Fowler, Biomedical frontiers of fluorine chemistry, in:
I. Ojima, J.R. McCarthy, J.T. Welch (Eds.), Proceedings of the ACS
Symposium Series 639, American Chemical Society, Washington,
DC, 1996, Chapter 23, pp. 328±343.
[25] C. RuÈchardt, E. Merz, B. Freudenberg, H.J. Opgenorth, C.C. Tan, R.
Werner, Chem. Soc. Spectrosc. Pub. 24 (1970) 51.