A product analytical study of the thermal and photolytic decomposition of some arenediazonium salts in solution

Article abstract of DOI:10.1246/bcsj.75.789

Products of thermal and photochemical reactions of eleven arenediazonium tetrafluoroborates in various solvents have been analyzed. All compounds in most solvents undergo unimolecular heterolysis to give singlet aryl cations which are captured by solvent. This mechanism is dominant for arenediazonium ions without electron-withdrawing substituents in all solvents, and the only reaction observed in water. Additionally, appreciable yields of fluoroarenes are obtained by fluoride abstraction by the aryl cation from fluorinated solvents and from tetrafluoroborate in fluorinated solvents. Yields from photochemical processes are very similar to those from thermal reactions indicating that the main reactions proceed through common or very similar intermediates. Aryl cations formed from ion-paired diazonium ions may react with the counterion, but fragmentation of dissociated diazonium ions leads only to solvent-derived product. Some arenediazonium ions in some solvents undergo an alternative radical reaction leading principally to hydrodediazoniation. It is proposed that this reaction involves initial rate-limiting electron transfer from ethanol to the arenediazonium ion followed rapidly by homolysis of the resultant aryldiazenyl radical. Within the same solvent cage, the aryl radical then either abstracts an α-hydrogen from the ethanol radical cation generated in the first step to give the reduction product and protonated acetaldehyde, or combines with it at the oxygen to give a protonated aryl ethyl ether.

Full text of DOI:10.1246/bcsj.75.789

c. Jpn.
© 2002 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 75, 789–800 (2002)
789
e Chemical
ciety of Ja-
n
e Chemical
ciety of Ja-
n
A Product Analytical Study of the Thermal and Photolytic
Decomposition of Some Arenediazonium Salts in Solution
Decomposition of Arenediazonium Salts
P. S. J. Canning et al.
Peter S. J. Canning, Howard Maskill,* Katharine McCrudden, and Brian Sexton
02
9
0
Chemistry Department, University of Newcastle, Newcastle upon Tyne, NE1 7RU, U. K.
(Received August 13, 2001)
SJA8
09-2673
270
Products of thermal and photochemical reactions of eleven arenediazonium tetrafluoroborates in various solvents
have been analyzed. All compounds in most solvents undergo unimolecular heterolysis to give singlet aryl cations which
are captured by solvent. This mechanism is dominant for arenediazonium ions without electron-withdrawing substitu-
ents in all solvents, and the only reaction observed in water. Additionally, appreciable yields of fluoroarenes are obtained
by fluoride abstraction by the aryl cation from fluorinated solvents and from tetrafluoroborate in fluorinated solvents.
Yields from photochemical processes are very similar to those from thermal reactions indicating that the main reactions
proceed through common or very similar intermediates. Aryl cations formed from ion-paired diazonium ions may react
with the counterion, but fragmentation of dissociated diazonium ions leads only to solvent-derived product. Some arene-
diazonium ions in some solvents undergo an alternative radical reaction leading principally to hydrodediazoniation. It is
proposed that this reaction involves initial rate-limiting electron transfer from ethanol to the arenediazonium ion fol-
lowed rapidly by homolysis of the resultant aryldiazenyl radical. Within the same solvent cage, the aryl radical then ei-
ther abstracts an α-hydrogen from the ethanol radical cation generated in the first step to give the reduction product and
protonated acetaldehyde, or combines with it at the oxygen to give a protonated aryl ethyl ether.
01
01
.2
Reactions of arenediazonium salts are both synthetically
useful and mechanistically interesting,1 they are exploited in-
dustrially in the manufacture of, for example, pharmaceuticals
and dyestuffs, their mechanisms have been studied by wide-
ranging techniques from the earliest days of organic chemistry,
and they still find new applications in contemporary organic
synthesis.2 The usual simple equations of correlation fail to
yield satisfactory analyses of their kinetics in aqueous and
nonaqueous solution,3,4 and few of the reported rate studies
have been accompanied by reliable product analyses. Conse-
quently, mechanistic understanding of the relationships be-
tween molecular structure of reactant, solvent, and reaction
conditions is less than complete.
The origin of this investigation was a search for a cleaner,
safer, and more efficient procedure for fluorodediazoniation
(replacement of the diazonium group by a fluorine) than meth-
ods currently employed, i. e. a practical method which could be
employed on a manufacturing scale for introducing fluorine
into an aromatic system in a site-specific manner. The Balz–
Schiemann reaction is the synthetic route of fluoroarenes (ArF)
generally favoured on a laboratory scale;5 the isolation of the
arenediazonium tetrafluoroborate is followed by its thermal
decomposition to form the fluoroarene. No solvent is required
for this reaction, although one may be used, or it may be car-
ried out on a slurry using an organic liquid to facilitate con-
trolled heat transfer. This site-specific introduction of the fluo-
rine into the aromatic ring proceeds by heterolytic loss of ni-
trogen and transfer to the aryl cation of a fluoride from the tet-
rafluoroborate counter-ion. In dichloromethane, 4-t-butylben-
zenediazonium tetrafluoroborate gives chloroarene (ArC1), as
well as ArF,6 so the electrophilic intermediate is sufficiently re-
active to abstract chloride from a nonpolar haloalkane. Simi-
larly, when 1,2-dichloroethane is used as the reaction medium
for arenediazonium tetrafluoroborate thermolysis, chloroare-
nes are obtained as well as fluoroarenes.4,7 In alkane systems,
only the most reactive carbenium ions will abstract halide from
a halogenated solvent.8 Variations to the original Balz–
Schiemann procedure have been reported, including the use of
counter-ions other than tetrafluoroborate,9 and photochemical
methods have been reported to give higher yields of fluoroare-
nes.10
The widely accepted mechanism in Scheme 1 is based upon
kinetics results (including kinetic isotope effect measure-
ments)3,4,11 and diverse evidence for two intermediates (an ion-
molecule pair and a fully dissociated aryl cation)12 supported
by theory.13 The proportion of product from each of the se-
quentially formed intermediates and the degree of return of
each to its precursor depend upon the solvent and the external
nitrogen pressure. In water, the first-formed ion-molecule pair
does not live long enough to undergo significant dissociation
Scheme 1. Heterolytic mechanism for dediazoniation of dis-
sociated arenediazonium ions in solution.

790 Bull. Chem. Soc. Jpn., 75, No. 4 (2002)
Decomposition of Arenediazonium Salts
whereas, in less nucleophilic but still ionizing solvents such as
2,2,2-trifluoroethanol,14 it does. Under normal atmospheric
pressure, capture of the dissociated aryl cation by nitrogen in
2,2,2-trifluoroethanol is negligible hence the second step may
normally be regarded as essentially irreversible. Return from
the ion-molecule pair to diazonium ion was detected by mea-
suring rearrangement in unreacted starting material when the
two nitrogens are distinguishable by isotopic labelling.12
Clearly, return without rearrangement will be easier than with,
consequently the extent of rearrangement necessarily underes-
timates the true extent of internal return. The rate-enhancing
effects of some potent nucleophiles are now more convincing-
ly interpreted as reducing the extent of internal return rather
than in terms of a direct bimolecular displacement on the cova-
lent arenediazonium ion.15
Fig. 1. Structures of 1, 2, and 3.
Considerable effort has been applied to the direct observa-
tion of aryl cation intermediates,16 and their generation by
methods other than dediazoniation.17 In 1998, Steenken and
colleagues reported laser flash photolysis studies on para sub-
stituted benzenediazonium ions in the weakly nucleophilic
1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) with and without an
added electron-rich arene such as 1,3,5-trimethylbenzene.18
Absorption bands for triplet cations and radicals were not seen
but, in reactions when the electron-rich arene was added,
bands corresponding to the formation of substituted cyclo-
hexadienyl cations were observed. These cations were stable
in the microsecond to millisecond range in the absence of add-
ed nucleophiles, and represent proof (albeit indirect) that sin-
glet aryl cations were produced even though their lifetimes
were too short to allow direct observation. Most recently, the
parent phenyl cation has been generated at 8 K in an argon ma-
trix (but not from the benzenediazonium ion) and shown to be
a singlet.19
2,2,2-trifluoroethyl ethers (3, R = CF3CH2) (Fig. 1), were
made from the appropriate arenediazonium tetrafluoroborate
(1) and TFE with final product isolation by preparative GLC;
purity was checked by capillary GLC and identification con-
firmed by 1H NMR. In this manner, nine aryl 2,2,2-trifluoro-
ethyl ethers and (using other solvents) one 1,1,1,3,3,3-hexaflu-
oro-2-propyl ether and five ethyl ethers were prepared, isolat-
ed, characterized, and shown to be pure enough to be used for
GLC molar response factor determinations, or construction of
HPLC correlation graphs. The yield of ethyl 3-fluorophenyl
ether was too low by this method, so was prepared by reaction
of ethyl 4-toluenesulfonate with 3-fluorophenol in alkaline so-
lution. Other aryl ethers, arenes, haloarenes, and phenols re-
quired were available either commercially or from our previ-
ous work.
Qualitative GLC Product Analysis. Solutions of arene-
diazonium salts (with and without solutes as required) in the
solvolytic media—TFE,2,2-difluoroethanol(DFE), 2-fluoroeth-
anol (FE), ethanol, aqueous TFE, water, HFIP, and trifluoro-
acetic acid (TFA)—were reacted for at least ten half lives. The
reaction mixtures were then either analyzed by GLC directly,
or quenched and products extracted into ether, pentane, or
dichloromethane for analysis. Although all other major poten-
tial solvolytic reaction products were available for identifica-
tion of peaks in chromatograms of product mixtures, this was
not true of aryl 2-fluoroethyl and aryl 2,2-difluoroethyl ethers
(3, R = FCH2CH2 and F2CHCH2). Consequently, ether ex-
tracts of products from reactions in DFE and FE were first ana-
lyzed by GLC-mass spectrometry to facilitate peak identifica-
tion. 3-Fluoroiodobenzene, a possible product from reactions
of 3-fluorobenzenediazonium tetrafluoroborate (1f) in the pres-
ence of iodoacetic acid (a well known radical trap),29 was com-
mercially available. No major peaks remained unidentified in
chromatograms from any reactions.
Other methods of decomposing arenediazonium ions in po-
tentially useful ways include reactions with transition metal
salts20,21 and nucleophilic anions,22 and reductive methods (hy-
drodediazoniation) at an electrode23 or using, for example, eth-
anol or hexamethylphosphoric triamide as reductant and sol-
vent.24,25 These reactions, especially of substrates with elec-
tron-withdrawing substituents and using nucleophilic solvents
in an oxygen-free environment, unquestionably involve radical
intermediates.20,26 Understanding of the relationships between
the ionic and homolytic mechanisms of fragmentation of are-
nediazonium ions, however, is incomplete.
We reported recently our kinetics studies on thermal sol-
volytic decomposition of arenediazonium ions;27 much of that
report’s introductory account of previous studies relates to this
complementary report on our product analytical investigation.
This report also includes results of photochemical decomposi-
tions of arenediazonium ions in the same solvolytic media but
not considered earlier.
Photolysis reactions were carried out at room temperature
with water cooling as appropriate using an immersion well
photochemical reactor with a 6 watt or 125 watt low-pressure
mercury lamp with maximum emission at 254 nm following
preliminary investigation using a semi-micro scale apparatus.
Product solutions from the photolysis reactions were analyzed
in the same way as those from the corresponding thermal reac-
tions. Since the photochemical reactions were much faster
Methods
Preparations. Preparations of our arenediazonium tet-
rafluoroborates (1) and benzenediazonium chloride (2) fol-
lowed well established methods and have already been de-
scribed.27,28 Other preparations followed literature methods
with only occasional minor modifications. Samples of the ma-
jor solvolytic products in 2,2,2-trifluoroethanol (TFE), i.e. aryl

P. S. J. Canning et al.
Bull. Chem. Soc. Jpn., 75, No. 4 (2002) 791
than the thermal ones, products from the latter are unlikely to
contaminate the former.
the phenols and, when quantified by UV, HPLC, or GLC,
yields were close to 100%. Fluorobenzene and 1,3-difluo-
robenzene were rigorously sought from the reaction of 3-fluo-
robenzenediazonium tetrafluoroborate (1f) in water, and the
upper limit on their formation is estimated by their non-detec-
tion by GLC to be < 0.1%. Similarly, no trace of 4-fluorotolu-
ene was observed from thermolysis of 4-toluenediazonium tet-
rafluoroborate (1h) in water containing ammonium hexafluo-
rophosphate; virtually 100% yields of phenols were also ob-
served from the hydrolysis of benzenediazonium and 3-meth-
oxybenzenediazonium tetrafluoroborates (1a) and (1c), respec-
tively, and the photolysis of 4-fluorobenzenediazonium tet-
rafluoroborate (1j) in water.35
Molar Response Factors (MRF). Molar response fac-
tors of major reaction products were determined by analyzing
standard solutions of pure authentic samples with either unde-
cane or octane as internal standard in the usual way.30 Results
were reproducible to about 2%, but we observed systematic
differences between gas chromatographs, so always used
MRFs appropriate to the particular instrument. Pure samples
of aryl 2-fluoroethyl ethers and aryl 2,2-difluoroethyl ethers (3,
R = FCH2CH2 and F2CHCH2) were not available in sufficient
quantities to the purity required for molar response factor de-
termination. However, since MRFs of corresponding aryl eth-
yl and aryl 2,2,2-trifluoroethyl ethers (3, R = Et and R =
CF3CH2) had been determined, reliable estimates of the MRFs
of FCH2CH2OAr and F2CHCH2OAr could be made.31 MRFs
of aryl 1,1,1,3,3,3-hexafluoro-2-propyl ethers and aryl trifluo-
roacetates were also estimated,32 and it was established by the
non-detection of phenols from reactions in TFA that the triflu-
oroacetate esters were hydrolytically stable to the work-up and
analytical conditions.33
Quantitative Analysis by GLC. Quantitative GLC analy-
ses were carried out from reactions on accurately known
amounts of arenediazonium salts as described for the corre-
sponding qualitative analyses except that a known amount of
internal standard was added either at the very beginning or to
the product mixture immediately before analysis. Significant
known possible products that were not detected were assigned
maximum possible yields based on the largest peak at the
known retention time that could have escaped detection; this
was always < 1% and generally < 0.1%. Absolute recoveries
were usually in the range 85–105% and have been normalized
(Σ = 100%) to facilitate comparison and for presentation in
the Tables; except when a product is shown to be absent (see
above), estimated maximum uncertainties are 3% or 15% of
the yield quoted, whichever is the larger.
Reactions in 2,2,2-Trifluoroethanol.14 As shown in Ta-
ble 1, yields of ArF are generally in the range 30–45%, the re-
mainders being exclusively the 2,2,2-trifluoroethyl (TFEt)
ether, with no obvious trend according to the nature of the sub-
stituent, and no systematic difference between thermal and
photolytic modes of decomposition. In all cases where the
products of hydrodediazoniation (reduction) were sought, they
were not detected. Also included in Table 1 are the results for
benzenediazonium chloride (2) which was used to estimate the
yield of fluorobenzene which is formed by the phenyl cation
abstracting fluoride from TFE (rather than from tetrafluorobo-
rate);36 chlorobenzene was considered as a possible product,37
but was shown to be absent. The yield of fluorobenzene from
2 is lower than from the tetrafluoroborate (1a) in this solvent,
but still appreciable. Significantly, reaction of 3-fluoroben-
zenediazonium tetrafluoroborate (1f) in TFE saturated with io-
Table 1. Product Analysis from Thermal (∆) and Photolytic
(hν) Reactions of Arenediazonium Tetrafluoroborates (1)
in 2,2,2-Trifluoroethanol
Substituent
∆ or hν
ArF/%
ArOCH2CF3/%
Ha)
∆
∆
hν
∆
hν
∆
hν
∆
hν
∆
hν
∆
hν
∆
hν
∆
28
10
43
42
44
29
31
27
33
35
33
37
33
44
43
37
45
~35
49
72
90
57
58
56
71
69
73
67
65
67
63
67
56
57
63
55
~65
51
Hb)
Analysis by UV. Yields of phenols from some reactions in
water were determined by comparison of the UV absorption at
the λmax of the isolated phenol, or of the reaction mixture,
made up to a known volume in water or methanol with that of a
standard solution of the authentic pure phenol in the same sol-
vent. All such reactions were carried out at least in duplicate
and yields of phenols were generally close to 100%.
Analysis by HPLC. For HPLC analyses, correlation
graphs were constructed from pure samples of phenols and flu-
oroarenes which allowed their absolute yields to be deter-
mined. Yields of aryl 2,2,2-trifluoroethyl ethers by this method
were calculated by difference assuming the total recovery to be
100%.
H
3-Mea)
3-Me
3-MeO
3-MeO
3-CF3
3-CF3
3-Fa),c)
3-Fa)
3-CN
3-CN
3-NO2
3-NO2
4-Me
4-Me
4-OMed)
4-OMe
Some reactions were analyzed quantitatively by more than
one method and agreement between the methods was generally
satisfactory.31,32,34
hν
∆
hν
Results
a) < 0.1% ArH detected. b) Reactant = PhN2Cl rather
than PhN2BF4; PhCl was sought but not detected. c) No 3-
fluoroiodobenzene was detected (< 1%) when the reaction
mixture was saturated with ICH2CO2H. d) Approximate
results; the thermal reaction was too slow to allow a reli-
able analysis.
Reactions in Water. 4-Methyl- and 4-methoxy-benzene-
diazonium tetrafluoroborates (1h,i) were decomposed thermal-
ly and photolytically in water, and water saturated with potas-
sium fluoride (ca. 10 mol dm−3). In all reactions, fluoroarenes
were sought and not detected (< 1%); the only products were

792 Bull. Chem. Soc. Jpn., 75, No. 4 (2002)
Decomposition of Arenediazonium Salts
doacetic acid gave no detectable 3-iodofluorobenzene which
confirms that radical intermediates are not involved in these re-
actions.29 Table 2 shows the product yields from reactions of
benzenediazonium tetrafluoroborate (1a) in TFE containing
various fluoride-containing salts—NaPF6, NaBF4, and
NaSbF6—as potential fluoride donors that are soluble in TFE.
Only NaBF4 has an effect possibly outside experimental uncer-
tainty. As the results in Table 3 show, there is a modest in-
crease in the yield of ArF from 4-toluenediazonium tetrafluo-
roborate (1h) as the proportion of trifluoromethoxybenzene
(TFMOB) as cosolvent increases. More limited studies with
(1h) using difluoromethyl 2,2,2-trifluoroethyl ether indicated
that this cosolvent has a smaller effect.
ratio of 3.8:1. In the least aqueous medium, 97TFE, the yield
of fluoroarene from 1h is 32%, and the ratio of TFEt ether to
phenol (9:1) is again higher than the molar ratio of TFE:H2O
in the solvent (5.8:1).
Reactions in 1,1,1,3,3,3-Hexafluoro-2-propanol.
HFIP
has been widely used as a highly-ionizing, weakly-nucleo-
philic solvent,39,40 the low nucleophilicity of the hydroxy being
ascribed to the electron-withdrawing effect of the six β-fluo-
rine atoms. So many fluorines will also reduce the ability of
one of them to act as a nucleophilic centre compared with one
in TFE. It was difficult, therefore, to predict which of the two
sites (OH or F) would be most effective in capturing the (high-
ly reactive) electrophile in the dediazoniation. Table 5 shows
the product yields obtained for the thermal and photolytic reac-
tions of various arenediazonium tetrafluoroborates. The yields
of ArF are systematically lower than from reactions in TFE,
there is little variation with the nature of the substituent, and
photolysis generally gives a lower yield of the fluoroarene than
the thermal process.
Reactions in Trifluoroacetic Acid. TFA was identified as
a weak fluoride donor which is also only weakly nucleophilic
at oxygen;41 products of thermal and photolytic reactions are
given in Table 6. They are very similar to those in HFIP (re-
sults from the thermal reaction of the parent compound appear
out of line and may be inaccurate) except that now photolysis
Reactions in Aqueous 2,2,2-Trifluoroethanol.38
Only
the 4-methyl compound (1h) was solvolysed in H2O:TFE =
4:1 (molar radio), Table 4; a low yield of fluorodediazoniation
is observed (11%) and the ratio of phenol to phenyl TFEt ether
(2.7:1) is lower than the ratio of water to TFE in the solvent.
In H2O:TFE = 1:1, the yields of fluoroarenes range from 5%
for the parent (1a) to 24% for the 4-methyl-substrate (1h). For
(1h), addition of ammonium hexafluorophosphate (but not tet-
rabutylammonium tetrafluoroborate) leads to an increased
yield of fluorodediazoniation. Ratios of aryl TFEt ethers to
phenols range from about 1:1 to 2:1 although the 2,6-dimeth-
yl-substrate (1k), the most reactive one, is exceptional with a
Table 2. Product Analysis from Thermal Reactions of Benzenediazonium Tetrafluoroborate
(1a) in 2,2,2-Trifluoroethanol with Added Inorganic Salts
Salt (concn/mol dm−3
)
PhF/%
PhOCH2CF3/%
0
28
27
34
29
72
73
66
71
NaPF6 (1.76)
NaBF4 (0.63)
NaSbF6 (0.60)
Table 3. Product Analysis from Thermal Reactions of 4-Methylbenzenediazonium Tetrafluoro-
borate (1h) in 2,2,2-Trifluoroethanol:Trifluoromethoxybenzene (TFE:TFMOB)
Solvent molar ratio
4-MeC6H4F/%
4-MeC6H4OCH2CF3/%
TFE (100%)
28
34
39
40
72
66
61
60
TFE:TFMOB (10:1)
TFE:TFMOB (5:1)
TFE:TFMOB (2.5:1)
Table 4. Product Analysis from Thermal Reactions of Substituted Benzenediazonium Tetra-
fluoroborates in Aqueous Trifluoroethanol
Substituent
Water:TFEa)
ArF/%
ArOCH2CF3/%
ArOH/%
4-Me
H
4:1
1:1
1:1
1:1
1:1
1:1
1:1
1:1
3:97d)
11
5
24
40
42
44
34
65
61
44
61
65
55
34
35
23
17
32
34
7
4-Me
4-Meb)
4-Mec)
2,6-di-Me
3-F
24
21
43
18
7
3-Cl
4-Me
22
32
a) Molar ratio. b) Containing 0.7 mol dm−3 Bu4NBF4. c) Containing 1.2 mol dm−3
NH4PF6. d) By weight.

P. S. J. Canning et al.
Bull. Chem. Soc. Jpn., 75, No. 4 (2002) 793
Table 5. Product Analysis from Thermal (∆) and Photolytic
(hν) Reactions of 3-Substituted Benzenediazonium Tet-
rafluoroborates in 1,1,1,3,3,3-hexafluoro-2-propanol
Table 7. Product Analysis from Thermal Reactions of 3-Sub-
stituted Benzenediazonium Tetrafluoroborates in Ethanol
3-Substituent
ArH/%
ArF/%
EtOAr/%
Ha)
7
4
3
4
5
2
15
17
88
94
82
79
20 (15)
8
3-Substituent
∆ or hν
ArF/%
ArOCH(CF3)2/%
H
H
Me
∆
hν
∆
hν
∆
hν
∆
hν
26
15
24
15
12
13
33
11
74
85
76
85
88
87
67
89
Hb)
Me
MeO
Me
Fc)
80d) (85)
92e)
< 0.1 (< 0.1)
< 1
MeO
MeO
CF3
CF3
NO2
a) [PhN2BF4]0 = 0.0066 mol dm−3
.
b) [PhN2BF4]0 =
0.013 mol dm−3; previously reported < 1:4:96.36 c) An
unquantified yield of 3-fluoroidobenzene was observed
when the reaction was carried out in the presence of 0.18
mol dm−3 iodoethanoic acid;29 yields in parentheses are
for photolysis. d) 3-Deuteriofluorobenzene obtained in
CH3CD2OH. e) 3-Deuterionitrobenzene obtained in
CH3CD2OH.
Table 6. Product Analysis from Thermal (∆) and Photolytic
(hν) Reactions of 3-Substituted Benzenediazonium Tet-
rafluoroborates in Trifluoroacetic acid
3-Substituent
∆ or hν
ArF/%
CF3CO2Ar/%
H
H
Me
∆
hν
∆
hν
∆
hν
∆
hν
∆
4
18
15
20
11
15
18
19
14
21
~9
21
96
82
85
80
89
85
82
81
86
trobenzene established that the nitrobenzene from the reaction
in 1,1-dideuterioethanol has a single deuterium atom. The
1
2
chemical shifts in the H spectrum, and the details of the H
NMR spectrum, confirmed that the deuterium is meta to the ni-
tro group, i.e. where the diazonium group was originally bond-
ed. 3-Fluorobenzenediazonium tetrafluoroborate (1f) was also
solvolyzed in 1,1-dideuterioethanol but preliminary studies
had shown that we were unable to isolate by preparative GLC
or microdistillation from a small scale solvolysis a sample of
fluorobenzene sufficient in quantity and purity for NMR analy-
sis. Consequently, the products from the quenched reaction
were first extracted into pentane, then back-extracted into
Me
MeO
MeO
CF3
CF3
CN
CN
hν
∆
hν
79
~91
79
a)
NO2
NO2
a) This reaction is very slow and the results may be inaccu-
rate.
1
DMSO-d6. The H NMR spectrum of this sample was then
compared with the spectrum from an identical run using unla-
belled ethanol, and with that of a sample of authentic pure un-
labelled fluorobenzene in the same solvent. It was possible to
generally gives the higher yield of ArF.
Reactions in Ethanol. Results in Table 7 show that sub-
strates with strongly electron-withdrawing substituents give
high yields of ArH and no ArF, the residue being the aryl ethyl
ether (ArOEt). In contrast to what was observed in TFE, when
iodoacetic acid was added to the ethanol for the reaction of 3-
fluorobenzenediazonium tetrafluoroborate (1f), 3-iodofluo-
robenzene was detected amongst the products—a positive and
well established test for the involvement of radical intermedi-
ates.29 Only very low yields of ArH are observed for the parent
compound (1a) and analogues with electron-releasing substitu-
ents.42 In these reactions, the major product is ArOEt, with a
low yield of PhF from the parent compound and higher yields
of ArF from substrates with electron-releasing substituents. In
ethanol, therefore, unlike reactions in less nucleophilic but
more ionizing solvents, there is unambiguous evidence for rad-
ical and heterolytic pathways, and which route is dominant de-
pends upon the substituents in the arene ring and the experi-
mental conditions.43
1
conclude from the integrated H NMR spectra that a single
deuterium was present in the fluorobenzene from the run in
1,1-dideuterioethanol, but overlapping signals from contami-
nants prevented more detailed deductions.
Reactions in 2-Fluoroethanol and 2,2-Difluoroethanol.
In view of the dramatic differences in product analyses be-
tween reactions in ethanol (Table 7) and 2,2,2-trifluoroethanol
(Table 1), solvolyses were carried out for a limited range of
substrates in mono- and difluoro-ethanols, and results are
shown in Tables 8 and 9. For benzenediazonium tetrafluorobo-
rate (1a), the results in both solvents are virtually the same as
for reaction in TFE; for the 3-methyl analogue 1b, the only dif-
ference is an inversion in the proportions of fluorobenzene and
the ether in these two solvents compared with reaction TFE. In
neither solvent for neither substrate, 1a or 1b, was there any
Table 8. Product Analysis from Thermal Reactions of 3-Sub-
stituted Benzenediazonium Tetrafluoroborates in 2-Fluoro-
ethanol
Reactions of 3-Fluoro- and 3-Nitro-benzenediazonium
Tetrafluoroborates (1f, g) in 1,1-2H2-Ethanol. The major
product from the thermal dediazoniation reaction of 3-ni-
trobenzenediazonium tetrafluoroborate (1g) in 1,1-dideuterio-
ethanol was isolated and purified by preparative GLC, then in-
3-Substituent
ArH/%
ArF/%
ArOCH2CH2F/%
H
Me
F
< 0.1
< 0.1
2
36
76
32
64
24
66
1
2
vestigated by H and H NMR. Comparison of the integrated
1H NMR spectrum with that of a sample of unlabelled ni-

794 Bull. Chem. Soc. Jpn., 75, No. 4 (2002)
Decomposition of Arenediazonium Salts
Table 9. Product Analysis from Thermal Reactions of 3-Sub-
stituted Benzenediazonium Tetrafluoroborates in 2,2-Di-
fluoroethanol
whereas simple alkyl carbenium ions generally show selectivi-
ty in favour of reaction with water over TFE.40,46 However, flu-
oroarenes are significant additional products formed from the
TFE (see below), i.e. the aryl cation combines with and ab-
stracts a fluoride from the trifluoromethyl end of a TFE mole-
cule—a reaction previously detected in the gas phase.47 Con-
sequently, the ratio of TFEt ether to phenol under-estimates the
ratio of total reaction of the aryl cation with TFE compared
with water, and there are two issues: (i) why is there selectivity
in favour of TFE? and (ii) why is there so much fluoride ab-
straction in aqueous TFE? The selectivity for TFE cannot be
due to better hydrogen bonding with the nucleofuge by the
more acidic OH of TFE compared with an OH of water, which
accounts for similar selectivities involving oxygen-bonded nu-
cleofuges.46 Nor does the encounter-controlled nature of the
product-forming reactions alone provide a satisfactory expla-
nation.48 Arenediazonium ions are both positively charged and
lipophilic, and consequently will be more effectively solvated
by TFE than by water molecules. Moreover, the interactions
between the arenediazonium ions and solvent molecules must
be predominantly electrostatic ion–dipole interactions, and the
CF3 end will be the more negative end of the TFE molecule di-
pole. Consequently, combination of the extremely reactive
aryl cation with a lone pair on one of three fluorines may well
involve less molecular reorganization than combination with a
lone pair on the hydroxy.
Reactions in 2,2,2-Trifluoroethanol, 1,1,1,3,3,3-Hexaflu-
oro-2-propanol, and Trifluoroacetic acid. Results in Table
1 show that, using chloride as the counter ion in TFE, 10% of
the products is ArF (and no chloroarene was detected) com-
pared with 28% ArF from the tetrafluoroborate; the major reac-
tion for both is capture of the electrophile by the hydroxy of
TFE. The absence of fluoroarene from reactions in water with-
out any TFE rules out reaction between the aryl cation and dis-
sociated tetrafluoroborate. We deduce that, in TFE, the aryl
cation has two routes to fluoroarene—fluoride abstraction from
TFE (as in aqueous TFE, but in higher yield) and fluoride ab-
straction from tetrafluoroborate within an ion pair, these occur
in the ratio of about 1:2, Scheme 2. The phenyl cation, there-
fore, is captured by a TFE molecule, either though the oxygen
(~70%) or a fluorine (~10%) or by an ion-paired tetrafluoro-
borate anion (~20%). The absence of chloroarene from the are-
nediazonium chloride in TFE indicates either a lower degree of
ion pairing with chloride than with tetrafluoroborate, or stron-
3-Substituent
ArH/%
ArF/%
ArOCH2CHF2/%
H
Me
F
< 0.1
< 0.1
0.5
33
77
53
67
23
47
Table 10. Product Analysis from Thermal Reactions of 4-
Toluenediazonium Tetrafluoroborate (1h) in Other Media
Medium
4-MeC6H4F/%
Other products/%
1,2-Dichloroethane
Toluene
74
77
4-chlorotoluene (26)
a)
—
a) Three nonpolar minor products were observed but not
identified.
hydrodediazoniation. Only for 3-fluorobenzenediazonium tet-
rafluoroborate (1f) in these solvents, with low yields of hy-
drodediazoniation in both, is there any tendency towards the
outcome in ethanol—the main reactions are still formation of
the ethers (3) and the fluoroarene. Both FE and DFE give
higher yields of fluoroarenes than TFE.
Reactions in Nonpolar, Nonnucleophilic Media.
Few
reliable complete product analyses have been reported for re-
actions using nonpolar, nonnucleophilic solvents possibly be-
cause reaction reproducibility is difficult to achieve, a difficul-
ty we also encountered. Analysis for 4-fluorotoluene by HPLC
from 4-toluenediazonium tetrafluoroborate (1h) in 1,2-dichlo-
roethane and in toluene led to results in Table 10. In 1,2-
dichloroethane, 4-chlorotoluene was identified as a significant
additional product, but the three minor by-products from reac-
tion in toluene were not identified. Attempts to obtain satisfac-
tory reproducible quantitative analyses from reactions of 3-flu-
orobenzenediazonium tetrafluoroborate (1f) in 1,2-dichloroet-
hane, toluene, 4-fluorotoluene, and chlorobenzene were not
successful.
Discussion
Reaction in Aqueous Solution. Dilute arenediazonium
salts will be fully dissociated in highly aqueous solution,44 so
reactions will be of independently solvated diazonium ions,
Scheme 1. In water with no cosolvent, products of thermal and
photolytic reactions are exclusively phenols; even saturation of
the water with potassium fluoride (ca. 10 mol dm−3) led to no
detectable yield of fluoroarene which, in other solvents, is al-
ways found from the heterolytic mechanism. There is already
evidence that aryl cations are extremely short-lived,45 and will
not survive extensive diffusion through even only weakly nu-
cleophilic media; additionally, fluoride is strongly hydrated in
aqueous solution and consequently a very weak nucleophile.
In water–TFE mixtures, the dilute reactant will also be fully
dissociated,44 consequently the products are overwhelmingly
formed from independently solvated dissociated arenediazoni-
um cations and solvent molecules. In these reactions, the ratio
of TFEt ethers to phenols is generally greater than the ratio of
TFE:H2O in the solvent, Table 4. Similar selectivities in
favour of TFE over water have been reported for vinyl cations
Scheme 2. Formation of fluoroarenes from reaction of aryl
cation with ion-paired tetrafluoroborate and with trifluoro-
ethanol.

P. S. J. Canning et al.
Bull. Chem. Soc. Jpn., 75, No. 4 (2002) 795
ger solvation of chloride within the ion pair which prevents di-
rect access to the chloride by the aryl cation. Although photo-
chemical reactions in TFE are much faster, the product analy-
ses are very similar indicating that the same or very similar in-
termediates are involved. And, although substituents have ap-
preciable effects upon rate constants for solvolysis, product ra-
tios are relatively insensitive to the substituents in the arene
ring, again indicative of a highly reactive short-lived electro-
phile.
In an attempt to increase the yield of fluoroarenes in TFE,
two types of solutes were tried. The marginal increase in ArF
yield in the presence of sodium tetrafluoroborate (Table 2)
probably reflects a higher proportion of reaction through the
arenediazonium tetrafluoroborate ion pair whose dissociation
will have been suppressed. Correspondingly, the increase in
ArF in the presence of NH4PF6 (Table 4) could be due to reac-
tion through the additional ion pair formed from the additional
fluorine-containing additive. Phenyl trifluoromethyl ether has
an oxygen bonded to the fluorine-bearing carbon which, by
resonance, has the potential to assist the delivery of fluoride to
an aryl cation. However, as results in Table 3 show, the en-
hancement of the ArF yield is modest at TFMOB mole fraction
= 0.29, and does not indicate a route to a viable practicable
synthesis of fluoroarenes.
1,1,1,3,3,3-Hexafluoro-2-propanol and trifluoroacetic acid
may be considered together, Tables 5 and 6. No arenediazoni-
um tetrafluoroborate in either solvent shows significantly in-
creased yields of ArF in thermal or photolytic reactions com-
pared with reactions in TFE; indeed, they are generally lower.
There are two credible explanations. The higher ionizing pow-
er of these solvents will cause a greater degree of reaction
through dissociated arenediazonium ions, consequently there
will be less reaction of the aryl cation with tetrafluoroborate
anion. (In agreement, much higher yields of fluoroarene are
obtained in nonpolar solvents such a 1,2-dichloroethane and
toluene (see below) in which ions will be overwhelmingly ion-
paired). Secondly, in both HFIP and TFA, the strongly elec-
tron-attracting rest of the molecule inhibits fluoride transfer
more than in TFE so, by default, the aryl cation is trapped
mainly through oxygen.
Reactions in Ethanol, 2-Fluoroethanol, and 2,2-Difluo-
roethanol. Thermolysis of arenediazonium salts in ethanol is
a well-cited hydrodediazoniation (reduction) method.20,24,29 In
practice, however, yields are variable and depend upon the oth-
er substituents in the arene and the precise experimental condi-
tions. This was corroborated in our studies, Table 7. For the
parent (1a) and derivatives with electron-releasing substitu-
ents, there are only low yields of reduction in ethanol, the ma-
jor products being the ethyl ethers with some formation of ArF
especially for reactants with electron-releasing substituents.
Comparison of rate constants27 and product ratios indicates that
these compounds, e.g. (1b,c), react in ethanol largely by the
same mechanism as in other solvents, Schemes 1 and 2. With
electron-withdrawing substituents, however, reduction yields
are high but still invariably accompanied by some ethyl ether
although, significantly, no fluoroarene. There is an early report
that, using methanol as the reducing solvent, it is one of the
methyl hydrogens which is transferred.49 Our present results
using 3-nitro and 3-fluoro-benzenediazonium tetrafluorobo-
rates in CH3CD2OH establish that, in ethanol, it is a hydrogen
from the methylene group which is transferred. And showing
that the deuterium in the nitrobenzene from 1g is meta to the
nitro group also establishes that, as for reactions of phenyl cat-
ions in solution,50 but not in the gas phase,47,51 hydrodediazoni-
ation in ethanol does not involve rearrangement of the interme-
diate whatever its nature.
The parent and analogues with electron-donating and -with-
drawing substituents were also solvolyzed in mono- and di-flu-
oroethanol to investigate the discontinuity in mechanisms be-
tween TFE and ethanol. We see from Tables 8 and 9 that for the
parent in both solvents the results are virtually the same as in
TFE, i.e. heterolytic nucleophilic substitution. For 3-methyl-
benzenediazoniium tetrafluoroborate, there is also no trace of
hydrodediazoniation although there is an inversion of the ratio
of ArF to ether (for which we have no explanation). Only for
3-fluorobenzenediazonium tetrafluoroborate is any reduction
product detected but even here yields are very low. Of the sol-
vents we have investigated, therefore, only ethanol provides
useful yields of hydrodediazoniation, and this occurs only for
arenediazonium salts with electron-withdrawing substituents.52
Mechanisms. Both the heterolytic nucleophilic substitu-
tion reactions and the hydrodediazoniation reactions reported
here follow first-order kinetics, but their activation parameters
are characteristically different. In general, the former have
large ∆H ‡ values and positive ∆S‡ values, and the previously
reported heterolytic mechanism, Scheme 1, through either dis-
sociated or ion-paired arenediazonium ions (depending upon
the solvent), Scheme 2, satisfactorily accounts for the results.27
Formation of 3-iodofluorobenzene from reaction of 3-fluo-
robenzenediazonium tetrafluoroborate in ethanol containing
iodoacetic acid confirmed the involvement of radical interme-
diates in the hydrodediazoniation reactions,29 but they do not
have the kinetic complexities of some types of radical reac-
tions.53 However, rate constants are much faster than expected
on the basis of results in other solvents, and activation parame-
ters are characteristically different; generally, ∆H ‡ values are
smaller and ∆S‡ values substantially negative. Significantly,
hydrodediazoniations, even when yields are high, are invari-
ably accompanied by the formation of some aryl ethyl ether.
Rate constants for the heterolytic mechanism of ca. 10−6 s−1 at
25 °C can be estimated for the reactions for arenediazonium
tetrafluoroborate with strongly electron-withdrawing substitu-
ents from results in other solvents; actual rate constants for
these compounds in ethanol are between 200 and 900 times
larger.27 It follows, therefore, that the appreciable yields of
aryl ethyl ethers in ethanol cannot have been formed by the
heterolytic route in parallel with the homolytic path as has usu-
ally been assumed in the past. We propose that the ethyl ethers
from arenediazonium ions with electron-withdrawing substitu-
ents are formed by a product-forming path, which competes
with the main hydrodediazoniation path, within a single com-
posite homolytic mechanism.20,26,43,54 Strongly supportive of
this conclusion is the total absence in ethanol from arenediazo-
nium ions with electron-withdrawing substituents of any ArF
which, for all reactants in other solvents (except pure water for
reasons given above), and reactants without electron-with-
drawing substituents in ethanol, characterises the heterolytic
mechanism. Scheme 3 involves a rate-limiting bimolecular

796 Bull. Chem. Soc. Jpn., 75, No. 4 (2002)
Decomposition of Arenediazonium Salts
3-fluorobenzenediazonium tetrafluoroborate only in FE and
DFE, and the heterolytic path predominates in both. For sub-
strates without electron-withdrawing substituents, formation of
fluoroarene and arene indicates that both mechanisms run in
parallel in ethanol, but again the heterolytic path predominates.
In such cases, the lower ∆H‡ values for the homolytic reaction
indicate that hydrodediazoniation will compete more effective-
ly at lower temperatures.
Photolytic Reactions. Product analyses from thermally
and photolytically induced reactions are generally very similar
and are satisfactorily explained in terms of common or very
similar intermediates—either cations or radicals.61 It has al-
ready been shown that arenediazonium ions are excellent trip-
let quenchers,62 and that direct irradiation of an arenediazoni-
um ion can lead to an excited singlet state of the diazonium ion
and then, by intersystem crossing, the triplet state.63 However,
the present results for substrates without electron-withdrawing
groups indicate involvement of only the singlet phenyl cation.
The singlet excited arenediazonium ion, therefore, fragments
faster than it crosses to the triplet state under the experimental
conditions used. Photolysis of the 3-fluoro-substrate in etha-
nol, likewise, gives a product analysis very similar to that ob-
tained by thermolysis which proceeds via electron transfer and
homolysis. Like comparable reactions already reported, this
also appears to be a homolytic reaction.61,64
Scheme 3. Electron transfer-homolysis mechanism for dedi-
azoniation of arenediazonium ions in ethanol.
electron transfer from ethanol to the arenediazonium ion prob-
ably via an encounter complex of the sort identified by
Kochi.55 Then, before the ethanol radical cation can diffuse
away, fragmentation of the very short-lived arenediazenyl radi-
cal gives the aryl radical.56 This then partitions between alter-
native competing product-forming routes: hydrogen atom ab-
straction from the methylene of the proximate ethanol radical
cation with elementary rate constant k1, and combination with
the ethanol radical cation through the oxygen with elementary
rate constant k2. This mechanism for reactions in ethanol ac-
counts for the kinetics parameters,27 the reduction product, the
ether formation, the rate-enhancing effect of electron-with-
drawing substituents in the arene ring, the rate-enhancing ef-
fects of nucleophilic solutes and co-solvents,43 and the isotopic
labelling results,36,49 and is well precedented.20,21,57,58 Solvents
or solutes such as low valence state metal ions and iodide
which are effective electron donors should facilitate this pro-
cess but may, as in the case of iodide, lead to additional prod-
ucts.20,22 An alternative radical mechanism shown in Scheme 4
via intermediate aryl diazo ethers, ArNwNOR,26,59 seems less
likely as such compounds are known to react by acid catalyzed
heterolysis, i.e. the reverse of the first two steps of Scheme 4.60
Formation of both 1,3-fluorobenzene and fluorobenzene in-
dicates that heterolytic and homolytic routes are concurrent for
Experimental
2,2,2-Trifluoroethanol was distilled and stored over 4 Å molecu-
lar sieves. 2-Fluoroethanol (Aldrich), 2,2-difluoroethanol (Lan-
caster), and HPLC grade ethanol (Lancaster) were used as sup-
plied. 2-Fluoroethanol is highly toxic and should be handled
with extreme care. Samples of octane and undecane (Lancaster)
used as GLC internal standards were shown to be pure by GLC.
Ether means diethyl ether and, unless otherwise stated, deuterio-
chloroform was used as solvent for 1H and 13C NMR spectra on a
Bruker WP200 spectrometer, for 19F NMR spectra on a Jeol
2
Lambda 500, and for H NMR spectra on Bruker WM300-WB;
coupling constants (J) are given in Hz. Analysis by GLC was car-
ried out on Pye 104, Hewlett-Packard 5890, Pye Unicam 15503
GCD, Packard 427, and Perkin Elmer Sigma 4b instruments, and
peak integrations were recorded with chromatograms. Preparative
GLC was on a Varian Aerograph 2700. Analysis by reverse phase
HPLC was carried out on a dual pump Gilson system fitted with a
manual Rheodyne injection valve and an Applied Biophysics UV
detector (set at 254 nm) controlled by Gilson software on an Opus
PC IV.
Small scale photolytic work was first carried out using a Photo-
chemical Reactors semi-micro instrument (model RS-55, 4 watt
maximum emission, 254 nm). An immersion well photochemical
reactor was used for larger samples (model RQ 125, 6 watt or 125
watt low-pressure mercury lamp, maximum emission 254 nm).
For most reactions, the solution of reactant was placed in a 150
mL standard reaction flask containing the water-cooled borosili-
cate-glass immersion well. For reactions in TFA, HFIP, and TFE,
a 3 mL quartz UV cell was filled with the reaction solution and
clamped to the outside of the immersion well.
Preparations. Preparations of benzenediazonium chloride
and arenediazonium tetrafluoroborates have already been report-
ed.27,28,65
Scheme 4. Alternative homolytic mechanism for dediazonia-
tion of arenediazonium ions in ethanol.
Ethyl 3-fluorophenyl Ether. A solution of ethyl tosylate

P. S. J. Canning et al.
Bull. Chem. Soc. Jpn., 75, No. 4 (2002) 797
(3.64 g, 18 mmol) in acetonitrile (10 mL) was added drop-wise
over 1 h to a magnetically stirred solution of sodium hydroxide
(0.867 g, 22 mmol) and 3-fluorophenol (2.02 g, 18 mmol) in water
(15 mL) at 0 °C. The pale yellow solution was next allowed to
stand at room temperature for 1 h, then was heated with stirring
under reflux at 100 °C for 3 h during which time it became color-
less, and a small amount of white suspension disappeared. The
solution was allowed to cool, then was extracted with ether (2×20
mL), and the extracts were combined, dried (MgSO4), filtered, and
evaporated under reduced pressure until the total volume was ap-
proximately 1 mL. A pure sample of the product was isolated by
preparative GLC (10% Carbowax 20M, 8ft by 1/4 inch column;
1H NMR: δ 1.40 (3H, t, J = 6.9 Hz), 3.99 (2H, q, J = 6.9 Hz),
6.63 (3H, m), 7.17 (1H, m)). Capillary GLC analysis showed that
the same product had been formed in the reaction of 3-fluoroben-
zenediazonium tetrafluoroborate in ethanol.
diluted by a factor of two and analyzed a further six times. This
procedure was carried out two or three times with different
amounts of compounds to allow two or three separate determina-
tions of the MRF from which an average was taken.
Quantitative Product Analysis of Thermal Solvolysis Reac-
tions by GLC: General Aspects. Reactions were carried out at
least twice for not less than ten half-lives, and GLC analyses were
usually carried out six times and average results taken. The fol-
lowing are representative.
(i) 3-Fluorobenzenediazonium Tetrafluoroborate in TFE.
A solution of 3-fluorobenzenediazonium tetrafluoroborate (0.2015
g, 0.9601 mmol) in TFE (3.00 mL) was heated under reflux for 4
h, then extracted between water (8 mL) and ether (8 mL). Unde-
cane (0.1345 g, 0.8605 mmol) was added to the ether phase which
was then analyzed by GLC.
(ii) 3-Fluorobenzenediazonium Tetrafluoroborate in Etha-
nol. A solution of 3-fluorobenzenediazonum tetrafluoroborate
(0.1200 g, 0.5718 mmol) and undecane (0.0640 g, 0.4094 mmol)
in ethanol (6 mL) was heated under reflux for 3 h at 78 °C, then
cooled and analyzed directly by GLC.
Isolation of Reaction Products. The following are represen-
tative.
(i) 3-Fluorophenyl 2,2,2-Trifluoroethyl Ether. A solution
of the arenediazonium salt (4.0 g, 19 mmol) in TFE (45 mL) was
heated under reflux for 4 h then cooled, and extracted between
ether (80 mL) and aqueous sodium hydroxide (2 mol dm−3, 80
mL). The organic phase was separated and combined with two
further ether extracts of the aqueous phase. The dried ether solu-
tion was evaporated under reduced pressure, and a sample of 3-
fluorophenyl 2,2,2-trifluoroethyl ether was isolated from the resi-
due by preparative GLC and shown by capillary GLC to be pure
(1H NMR: δ 4.31 (2H, q, J = 8.1 Hz), 6.72 (3H, m), 7.23 (1H, t, J
= 7.4 Hz).
(ii) Phenyl 2,2,2-Trifluoroethyl Ether. The procedure de-
scribed above was followed using benzenediazonium tetrafluorob-
orate (6.0 g, 31 mmol) in TFE (40 mL); 1H NMR: δ 4.34 (2H, q, J
= 8.2 Hz), 6.95 (2H, d, J = 7.3 Hz), 7.05 (1H, t, J = 7.5 Hz), 7.33
(2H, m).
(iii) Benzenediazonium Tetrafluoroborate in Ethanol.
This reaction was carried out at two different concentrations of
substrate: (a) 0.013 mol dm−3 and (b) 0.0066 mol dm−3
.
(a) The procedure was as described above for 3-fluorobenzene-
diazonium tetrafluoroborate in TFE (except that pentane was used
for the organic extraction) using benzenediazonium tetrafluorobo-
rate (0.0265 g, 0.138 mmol), octane (0.0165 g, 0.1445 mmol), and
ethanol (10.4 mL). The procedure was repeated using the arenedi-
azonium salt (0.0191 g, 0.0995 mmol), octane (0.0337 g, 0.0295
mmol), and ethanol (7.5 mL).
(b) The procedure was as under (a) but using benzenediazoni-
um tetrafluoroborate (0.0240 g, 0.125 mmol), octane (0.0370 g,
0.324 mmol), and ethanol (18.8 mL). The procedure was repeated
twice (i) using the arenediazonium salt (0.0290 g, 0.151 mmol),
octane (0.042 g, 0.368 mmol), and ethanol (22.8 mL), and (ii) us-
ing the arenediazonium salt (0.0166 g, 0.0865 mmol), octane
(0.0346 g, 0.303 mmol), and ethanol (13.1 mL).
(iv) 3-Fluorobenzenediazonium Tetrafluoroborate in 2,2-Di-
fluoroethanol. A solution of 3-fluorobenzenediazonium tet-
rafluoroborate (0.0198 g, 0.0943 mmol), octane (0.0378 g, 0.331
mmol), and 2,2-difluoroethanol (0.5 mL) was heated under reflux
for 2 h at 90 °C then allowed to cool. The reaction mixture was
extracted between ether (2 mL) and aqueous sodium hydroxide (3
mol dm−3, 2 mL), then the organic phase was separated and ana-
lyzed by GLC.
(v) 4-Methoxybenzenediazonium Tetrafluoroborate in Satu-
rated Aqueous Potassium Fluoride. A solution of 4-methoxy-
benzenediazonium tetrafluoroborate (877 mg) in water (200 mL)
containing potassium fluoride (120 g) was heated under reflux for
12 h with constant stirring. The cooled reaction mixture was di-
luted with water (50 mL) and extracted with ether (4×50 mL).
The combined extract was evaporated under reduced pressure and
the residue diluted in methanol to 500 mL. Analysis by GLC
showed 4-methoxyphenol to be the only product, and comparison
by UV spectroscopy (at 292.5 nm) with a standard solution of 4-
methoxyphenol (829 mg) in methanol (500 mL) indicated a yield
of 96%.
(iii) 1,1,1,3,3,3-Hexafluoro-2-propyl 3-Methylphenyl Ether.
A solution of 3-methylbenzenediazonium tetrafluoroborate (3.7 g,
18 mmol) in HFIP (10 mL) was heated at 60 °C for 24 h. The re-
action was worked up in the usual way and a sample of the prod-
uct purified by preparative GLC; 1H NMR: δ 2.3 (3H, s), 4.7 (1H,
sept., J = 5.7 Hz), 6.9 (3H, m), 7.2 (1H, m).
(iv) 3-Deuterionitrobenzene.
3-Nitrobenzenediazonium
tetrafluoroborate (0.25 g, 1.1 mmol) was dissolved in 1,1-2H2-eth-
anol (99%, 2 mL). The reaction mixture was heated under reflux
at 80 °C for 48 h. A sample of the product (1H NMR: δ 7.5 (1H,
m, Ar), 7.7(1H, m, Ar), 8.2(2H, m, Ar); 2H NMR: δ 7.6 (1D, t, J =
1.2 Hz), EIMS m/z 124 (M+, 100%), 94, 78, 52) was isolated and
purified by preparative GLC.
Qualitative Product Analysis by GLC. All reactions were
carried out at least in duplicate. Thermal solvolyses were for 10
half-lives or longer before (in most cases) direct GLC analysis fol-
lowed by GLC and GLC-mass spectrometric analysis of the ether
(or pentane or dichloromethane) extract of an aqueous quench of
the reaction. Photolytic reactions were monitored by UV to en-
sure completion before analysis as for the thermal reactions.
Products were identified and characterized either by the usual
spectroscopic methods following isolation, or by comparison with
authentic samples.
Determination of GLC Molar Response Factors. The fol-
lowing procedure is representative. A solution (25 mL) of unde-
cane (0.1660 g, 1.06 mmol) and 1,3-difluorobenzene (0.1875 g,
1.64 mmol) in ether (in some determinations, ethanol or pentane
were used) was analyzed six times by capillary GLC. It was then
Quantitative Product Analysis of Thermal Reactions by
HPLC: General Aspects. Correlation graphs were constructed
by reverse phase HPLC analysis (aqueous methanol) of standard
solutions of 4-fluorotoluene, p-cresol, and 1,3-difluorobenzene
which were shown to be pure by GLC. Typically, each of two or

798 Bull. Chem. Soc. Jpn., 75, No. 4 (2002)
Decomposition of Arenediazonium Salts
three stock solutions would be diluted to give three or four differ-
ent solutions for analysis.
8
E. H. White, H. P. Tiwari, and M. J. Todd, J. Am. Chem.
Soc., 90, 4734 (1968).
Quantitative Photolytic Reactions. The following are repre-
sentative.
9
K. G. Rutherford, W. Redmond, and J. Rigamonti, J. Org.
Chem., 26, 5149 (1961); C. Sellers and H. Suschitsky, J. Chem.
Soc. C, 1968, 2317; D. J. Milner, Synth. Comm., 22, 73 (1992);
M. Tamura, M. Shibakami, S. Kurosawa, T. Arimura, and A.
Sekiya, J. Fluorine Chem., 78, 95 (1996); K. Sasaki, M. Oishi,
and N. Imaki, J. Fluorine Chem., 76, 59 (1996); N. Yoneda and T.
Fukuhara, Tetrahedron, 52, 23 (1996); M. Tamura, M. Shibakami,
and A. Sekiya, Eur. J. Org. Chem., 1998, 725.
(i) 3-Fluorobenzenediazonium Tetrafluoroborate in TFE.
A solution of the arenediazonium salt (0.0185 g, 0.0882 mmol)
and undecane (0.022 g, 0.142 mmol) in TFE (3.0 mL) in a UV cell
was clamped to the outside of the photochemical reactor and irra-
diated. The reaction was monitored by UV, then, upon comple-
tion, was extracted between water and ether, and the ether phase
was analyzed.
(ii) 4-Methoxybenzenediazonium Tetrafluoroborate in Sat-
urated Aqueous Potassium Fluoride. A stirred solution of 4-
methoxybenzenediazonium tetrafluoroborate (730 mg) in saturat-
ed aqueous potassium fluoride solution (10.2 mol dm−3, 200 mL)
was photolyzed in the usual way and monitored by UV spectros-
copy. Upon completion (ca. 2 h), the products were extracted into
ether (4×50 mL); the combined extract was dried (MgSO4), fil-
tered, and evaporated under reduced pressure. GLC analysis
showed 4-methoxyphenol to be the only product.
10 K. Shinhama, S. Aki, T. Furuta, and J. Minamikawa, Synth.
Comm., 23, 1577 (1993).
11 E. A. Moelwyn-Hughes and P. Johnson, Trans. Faraday
Soc., 36, 948 (1940); L. L. Brown and J. S. Drury, J. Chem. Phys,
43, 1688 (1965); C. G. Swain, J. E. Sheats, and K. G. Harbison, J.
Am. Chem. Soc., 97, 783 (1975); C. G. Swain, J. E. Sheats, D. G.
Gorenstein, and K. B. Harbison, J. Am. Chem. Soc, 97, 791
(1975); C. G. Swain, J. E. Sheats, and K. B. Harbison, J. Am.
Chem. Soc, 97, 796 (1975); I. Szele and H. Zollinger, Helv. Chim.
Acta, 64, 2728 (1981).
12 J. M. Insole and E. S. Lewis, J. Am. Chem. Soc., 85, 122
(1963); E. S. Lewis and J. M. Insole, J. Am. Chem. Soc., 86, 32
(1964); E. S. Lewis and R. E. Holliday, J. Am. Chem. Soc., 88,
5043 (1966); E. S. Lewis and R. E. Holliday, J. Am. Chem. Soc.,
91, 426 (1969); E. S. Lewis and P. G. Kotcher, Tetrahedron, 25,
4873 (1969); R. G. Bergstrom, G. H. Wahl, and H. Zollinger, Tet-
rahedron Lett., 1974, 2975; P. Burri, G. H. Wahl, and H. Zollinger,
Helv. Chim. Acta, 57, 2099 (1974); R. G. Bergstrom, R. G. M.
Landells, G. H. Wahl, and H. Zollinger, J. Am. Chem. Soc, 98,
3301 (1976); I. Szele and H. Zollinger, J. Am. Chem. Soc., 100,
2811 (1978); Y. Hashida, R. G. M. Landells, G. E. Lewis, I. Szele,
and H. Zollinger, J. Am. Chem. Soc., 100, 2816 (1978); M. D.
Ravenscroft and H. Zollinger, Helv. Chim. Acta, 71, 507 (1988);
M. D. Ravenscroft, P. Skrabal, B. Weiss, and H. Zollinger, Helv.
Chim. Acta, 71, 515 (1988).
We thank the EPSRC for studentships to KMcC, BS, and
PSJC, and Zeneca (formerly ICI) for financial support. We
also thank Drs J.H. Atherton and D. J. Moody for helpful dis-
cussions, and Mss H. Dowling and N. Martin for supportive
experimental work. HM thanks members of the Institute for
Fundamental Research in Organic Chemistry (IFOC) at
Kyushu University for hospitality during the writing of this
manuscript, and Professors S. Kobayashi and T. Sonoda for
many helpful discussions.
References
1
H. Zollinger, Acc. Chem. Res., 6, 335 (1973); H. Zollinger,
“Diazo Chemistry ꢀ: Aromatic and Heteroaromatic Compounds,”
VCH, New York (1994); “The Chemistry of Diazonium and Diazo
Groups,” ed by S. Patai, Interscience, New York (1978); “Chemis-
try of Triple Bonded Functional Groups,” ed by S. Patai, Wiley,
New York (1983).
13 W. Maurer, I. Szele, and H. Zollinger, Helv. Chim. Acta,
62, 1079 (1979); M. A. Vincent and L. Radom, J. Am. Chem. Soc.,
100, 3306 (1978).
14 V. J. Shiner, W. Dowd, R. D. Fisher, S. R. Hartshorn, M. A.
Kessick, L. Milakofsky, and M. W. Rapp, J. Am. Chem. Soc., 91,
4838 (1969); M. D. Bentley and J. A. Lacadie, Tetrahedron Lett.,
1971, 741; J. M. Harris, D. J. Raber, W. C. Neil, and M. D. Dukes,
Tetrahedron Lett., 1974, 2331.
15 E. S. Lewis and W. H. Hinds, J. Am. Chem. Soc., 74, 304
(1952); E. S. Lewis and J. M. Insole, J. Am. Chem. Soc., 86, 34
(1964); E. S. Lewis, L. D. Hartung, and B. M. Mckay, J. Am.
Chem. Soc, 91, 419 (1969); P. Burri and H. Zollinger, Helv. Chim.
Acta, 56, 2204 (1973).
16 A. Cox, T. J. Kemp, D. R. Payne, M. C. R. Symons, D. M.
Allen, and P. P. De Moira, J. Chem. Soc., Chem. Commun., 1976,
693; A. Cox, T. J. Kemp, D. R. Payne, M. C. R. Symons, and P. P.
de Moira, J. Am. Chem. Soc., 100, 4779 (1978); H. B. Ambroz and
T. J. Kemp, J. Chem. Soc., Perkin 2, 1979, 1420; H. B. Ambroz
and T. J. Kemp, Chem. Soc. Rev., 8, 353 (1979); H. B. Ambroz
and T. J. Kemp, J. Chem. Soc., Faraday 1, 78, 1269 (1982); H. B.
Ambroz and T. J. Kemp, J. Chem. Soc., Chem. Commun., 1982,
172; E. Uggerud, D. Arad, Y. Apeloig, and H. Schwarz, J. Chem.
Soc., Chem. Commun., 1989, 1015.
2
See, for example, D. F. Oliveira, E. A. Severino, and C. R.
D. Correia, Tetrahedron Lett., 40, 2083 (1999), and S. Sengupta
and S. Bhattacharyya, J. Org. Chem., 62, 3405 (1997), for recent
synthetic applications.
3
M. L. Crossley, R. H. Kienle, and C. H. Benbrook, J. Am.
Chem. Soc., 62, 1400 (1940); L. P. Hammett, “Physical Organic
Chemistry,” McGraw Hill, New York (1940), p. 295; E. S. Lewis
and E. B. Miller, J. Am. Chem. Soc., 75, 429 (1953); D. F. DeTar
and A. R. Ballentine, J. Am. Chem. Soc., 78, 3916 (1956); R. W.
Taft, J. Am. Chem. Soc., 83, 3350 (1961).
4
T. Kuokkanen and P. O. I. Virtanen, Acta Chem. Scand.,
B33, 725 (1979); H. Nakazumi, I. Szele, K. Yoshida, and H.
Zollinger, Helv. Chim. Acta, 66, 1721 (1983); H. Nakazumi, T.
Kitao, and H. Zollinger, J. Org. Chem. 52, 2825 (1987); T.
Kuokkanen, T. Slotte, and V. Virtanen, Acta Chem. Scand., 45,
674 (1991).
5
G. Balz and G. Schiemann, Ber. Dtsch. Chem. Ges., 60,
1186 (1927); A. Roe, Org. Reac., 5, 193 (1949).
6
(1975).
7
C. G. Swain and R. J. Rogers, J. Am. Chem. Soc., 97, 799
17 H. W. Leung and A. G. Harrison, J. Am. Chem. Soc., 101,
3168 (1979); M. Hanack and W. Holweger, J. Chem. Soc., Chem.
Commun., 1981, 713; M. Hanack, Pure Appl. Chem., 56, 1819
(1984); H. Holweger and M. Hanack, Chem. Ber., 117, 3004
R. A. Bartsch, H. Chen, N. F. Haddock, and P. N. Juri, J.
Am. Chem. Soc., 98, 6753 (1976).

P. S. J. Canning et al.
Bull. Chem. Soc. Jpn., 75, No. 4 (2002) 799
(1984); M. Hanack and R. Rieth, J. Chem. Soc., Chem. Commun.,
1985, 1487; Y. Himeshima, H. Kobayashi, and T. Sonada, J. Am.
Chem. Soc., 107, 5286 (1985).
18 S. Steenken, M. Ashokkumar, P. Maruthamuthu, and R. A.
McClelland, J. Am. Chem. Soc., 120, 11925 (1998).
19 M. Winkler and W. Sander, Angew. Chem., Int. Ed., 39,
2014 (2000).
E. Nordlander and W. G. Deadman, J. Am. Chem. Soc., 90, 1590
(1968); J. E. Nordlander and W. J. Kelly, J. Am. Chem. Soc., 91,
996 (1969); I. L. Reich, A. Diaz, and S. Winstein, J. Am. Chem.
Soc., 91, 5635 (1969); A. Streitwieser and G. A. Dafforn, Tetrahe-
dron Lett., 1969, 1263; G. A. Dafforn and A. Streitwieser, Tetra-
hedron Lett., 1970, 3159.
42 In our early analyses,27 we failed to detect benzene in the
reaction of (1a) in ethanol. We have now shown that it is formed,
and that its percentage yield is higher at lower initial substrate
concentrations.
20 C. Galli, Chem Rev., 88, 765 (1988).
21 W. A. Waters, J. Chem. Soc., 1942, 266.
22 H. J. Lucas and E. R. Kennedy, Org. Synth., Coll. Vol. 2,
351 (1943); R. Kumar and P. R. Singh, Tetrahedron Lett., 1972,
613; P. R. Singh and R. Kumar, Austr. J. Chem., 25, 2133 (1972);
M. Tilset and V. D. Parker, Acta Chem. Scand., Ser. B, 36, 281
(1982); A. N. Abeywickrema and A. L. Beckwith, J. Org. Chem.,
52, 2568 (1987); L. Friedman and J. F. Chlebowski, J. Org.
Chem., 33, 1636 (1968).
23 L. Eberson, Adv. Phys. Org. Chem., 18, 79 (1982); R. M.
Elofson and F. F. Gadallah, J. Org. Chem., 34, 854 (1969); R. M.
Elofson and F. F. Gadallah, J. Org. Chem., 36, 1769 (1971).
24 N. Kornblum, Org. React., 2, 262 (1944); A. Roe and J. R.
Graham, J. Am. Chem. Soc., 74, 6297 (1952).
43 B. Gloor, B. L. Kaul, and H. Zollinger, Helv. Chim. Acta,
55, 1596 (1972); P. Burri, H. Loewenschuss, H. Zollinger, and G.
K. Zwolinski, Helv. Chim. Acta, 57, 395 (1974); H. Hirose, G. H.
Wahl, and H. Zollinger, Helv. Chim. Acta, 59, 1427 (1976); H.
Loewenschuss, G. H. Wahl, and H. Zollinger, Helv. Chim. Acta,
59, 1438 (1976); I. Szele and H. Zollinger, Helv. Chim. Acta, 68,
1721 (1978).
44 T. R. Griffiths and R. H. Wijayanayake, Trans. Faraday
Soc., 66, 1563 (1970).
45 J. C. Scaiano and N. Kim-Thuan, J. Photochem., 23, 269
(1983).
25 F. Trönlin and C. Rüchardt, Chem. Ber., 110, 2494 (1977).
26 D. F. DeTar and M. N. Turetzky, J. Am. Chem. Soc., 77,
1745 (1955); D. F. DeTar and M. N. Turetzky, J. Am. Chem. Soc.,
78, 3925 (1956), 3928; J. F. Bunnett and H. Takayama, J. Am.
Chem. Soc, 90, 5173 (1968); W. J. Boyle, T. J. Broxton, and J. F.
Bunnett, J. Chem. Soc., Chem. Commun., 1971, 1469.
27 P. S. J. Canning, K. McCrudden, H. Maskill, and B.
Sexton, J. Chem. Soc., Perkin Trans. 2, 1999, 2735.
28 H. Maskill and K. McCrudden, Croat. Chem. Acta, 65, 567
(1992).
46 J. Kaspi and Z. Rappoport, Tetrahedron Lett., 1977, 2035;
J. Kaspi and Z. Rappoport, J. Am. Chem. Soc., 102, 3829 (1980).
47 Y. Keheyan and M. Speranza, Helv. Chim. Acta, 68, 2381
(1985).
48 J. P. Lorand, Tetrahedron Lett., 30, 7337 (1989).
49 L. Melander, Ark. Kemi, 3, 525 (1951), cited by T. J.
Broxton, J. F. Bunnett, and C. H. Paik, J. Org. Chem., 42, 643
(1977).
50 P. Haberfield, C. Cardona, and M. Bell, J. Org. Chem., 49,
78 (1984).
29 F. W. Wassmundt and W. F. Kiesman, J. Org. Chem., 62,
8304 (1997).
30 H. Maskill and M. C. Whiting, J. Chem. Soc., Perkin Trans.
2, 1976, 1462; R. M. Banks and H. Maskill, J. Chem. Soc., Perkin
Trans. 2, 1976, 1506.
31 P. S. J. Canning, Ph.D. Thesis, University of Newcastle,
2000.
32 B. Sexton, Ph.D. Thesis, University of Newcastle, 1998.
33 B. Sexton, unpublished work.
51 M. Speranza, Tetrahedron Lett., 21, 1983 (1980); G.
Angelini, S. Fornarini, and M. Speranza, J. Am. Chem. Soc., 104,
4773 (1982); M. Speranza, Y. Keheyan, and G. Angelini, J. Am.
Chem. Soc., 105, 6377 (1983); A. Filippi, G. Lilla, G. Occhiucci,
C. Sparapani, O. Ursini, and M. Speranza, J. Org. Chem., 60, 1250
(1995).
52 Preliminary rate and product studies for 1a,b,f in 2-pro-
panol indicated reactions very similar to ones in ethanol; N.
Martin, unpublished results.
34 K. McCrudden, Ph.D. Thesis, University of Newcastle,
1994.
53 F. Wilkinson, “Chemical Kinetics and Reaction Mecha-
nisms,” Van Nostrand-Reinhold, Wokingham (1980); J. E. Packer,
R. K. Richardson, P. J. Soole, and D. R. Webster, J. Chem. Soc.,
Perkin Trans. 2, 1974, 1472; J. E. Packer and R. K. Richardson, J.
Chem. Soc., Perkin Trans. 2, 1975, 751; J. E. Packer, C. J. Heigh-
way, H. M. Miller, and B. C. Dobson, Aust. J. Chem., 33, 965
(1980).
54 D. F. DeTar and T. Kosuge, J. Am. Chem. Soc, 80, 6072
(1958).
55 T. M. Bockman, D. Kosynkin, and J. K. Kochi, J. Org.
Chem., 62, 5811 (1997).
35 In preliminary experiments, 2,5-difluorobenzenediazonium
tetrafluoroborate was decomposed in water and the only product
detected by GLC was 1,4-difluorobenzene. Similarly, reaction of
2-fluorobenzenediazonium tetrafluoroborate in TFE:H2O (1:1)
gave only fluorobenzene. These hydrodediazoniations in aqueous
solution remain to be investigated further.
36 P. S. J. Canning, K. McCrudden, H. Maskill, and B.
Sexton, J. Chem. Soc., Chem. Commun., 1998, 1971.
37 E. S. Lewis, J. Am. Chem. Soc., 80, 1371 (1958); E. S.
Lewis and J. E. Cooper, J. Am. Chem. Soc., 84, 3847 (1962).
38 D. J. Raber, M. D. Dukes, and J. Gregory, Tetrahedron
Lett., 1974, 667; A. Cooney and K. W. Morcom. J. Chem. Ther-
modyn., 20, 735 (1988).
56 N. A. Porter, G. R. Dubay, and J. G. Green, J. Am. Chem.
Soc, 100, 920 (1978).
57 W. A. Waters, J. Chem. Soc., 1937, 113, 2007, 2014.
58 E. S. Lewis and D. J. Chalmers, J. Am. Chem. Soc., 93,
3267 (1971); J. F. Bunnett and C. Yijima, J. Org. Chem., 42, 639
(1977); J. F. Bunnett, Acc. Chem. Res., 25, 2 (1992).
59 M. P. Doyle, J. F. Dellaria, B. Siegfried, and S. W. Bishop,
J. Org. Chem., 42, 3494 (1977).
39 D. E. Sunko and I. Szele, Tetrahedron Lett., 1972, 3617; F.
L. Schadt, P. von R. Schleyer, and T. W. Bentley, Tetrahedron
Lett., 1974, 2335.
40 F. L. Schadt, T. W. Bentley, and P. von R. Schleyer, J. Am
Chem. Soc., 98, 7667 (1976).
60 T. J. Broxton and A. C. Stray, J. Org. Chem., 45, 2409
(1980).
41
P. E. Peterson, R. J. Bopp, D. M. Chevli, E. L. Curran, D.
E. Dillard, and R. J. Kamat, J. Am. Chem. Soc., 89, 5902 (1967); J.
61 H. G. O. Becker, G. Hoffmann, and G. Israel, J. Prakt.

800 Bull. Chem. Soc. Jpn., 75, No. 4 (2002)
Decomposition of Arenediazonium Salts
Chem. Soc., 117, 5206 (1995).
Chem., 319, 1021 (1977).
62 J. C. Scaiano and N. Kim-Thuan, Can. J. Chem., 60, 2286
(1982).
64 W. E. Lee, J. G. Calvert, and E. W. Malmberg, J. Am.
Chem. Soc., 83, 1928 (1961).
63 S. M. Gasper, C. Devadoss, and G. B. Schuster, J. Am.
65 E. B. Starkey, Org. Synth., Coll. Vol. 2, 225 (1943).