The structural influence in the stability of polymer-bound diazonium salts

Article abstract of DOI:10.1002/chem.200400386

Various new diazonium ions on a polymeric support that have a variety of counterions and complexation with crown ethers have been prepared, and the thermal stability of these resins over a larger temperature interval has been investigated. Nonisothermal k

Full text of DOI:10.1002/chem.200400386

S Bräse, C. Popescu, et al.
Chem. Eur. J. 2004, 10, 5285 – 5296
DOI: 10.1002/chem.200400386
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5285

FULL PAPER
The Structural Influence in the Stability of Polymer-Bound Diazonium Salts
Stefan Bräse,*[a] Stefan Dahmen,[b] Crisan Popescu,*[c] Maarten Schroen,[a] and
Franz-Josef Wortmann[c]
Abstract: Various new diazonium ions on a polymeric support that have a variety
of counterions and complexation with crown ethers have been prepared, and the
thermal stability of these resins over a larger temperature interval has been inves-
tigated. Nonisothermal kinetics applied to DSC data have been used predicting
the lifetime of the resins.
Keywords:
differential scanning calorimetry ·
kinetics · resins · solid supports
azo compounds
·
Introduction
Stability of diazonium ions: Diazonium ions represent one
of the most versatile functional groups in organic chemis-
try.[1,2] The most important reactions of aromatic diazonium
ions are Sandmeyer, Schiemann, Meerwein, Pschorr, and
Gomberg–Bachmann reactions, but also the Heck reaction[3]
and the formation of triazenes[4] and azo compounds[5] rep-
resent important processes. Diazonium ions are also of inter-
est in theoretical chemistry. In more recent theoretical
works they are considered as electron-donor/acceptor com-
plexes between a phenyl cation and a nitrogen molecule.[6,7]
The property that limits the application of diazonium ions
most is their lability. Aromatic diazonium ions undergo frag-
mentation to nitrogen and reactive phenyl intermediates.
The decomposition of diazonium salts was investigated in
1940 by Crossley,[8] who showed that the rate of nucleophilic
substitution of the diazonio group is independent from the
concentration of the nucleophiles and is proportional to the
rate of hydrolysis. This reaction proceeds by an SN1 mecha-
nism, for which the rate-determining step is the irreversible
dissociation of the diazonium ion into an aryl cation and ni-
trogen, followed by a fast reaction of the cation with nucleo-
philes in the reaction mixture (Scheme 1).
Scheme 1. SN1-like decomposition of diazonium ions.
Ionic[9] and radical[10] pathways, initiated by single-electron
transfer reactions, are also known.
In general, an SN1 type mechanism is considered, although
there are indications for an SNAr type mechanism in the
presence of nucleophiles. In 1952 Lewis[11] and Hinds found
that the decomposition of 4-nitrobenzenediazonium ion in
water is proportional to the concentration of the added nu-
cleophile. Although various other examples favor an SN2
mechanism, the effect of substitution as determined by
Crossley and later reconfirmed by Schulte-Frohlinde and
Blume[12] cannot explain this mechanism. Instead of an in-
crease of decay caused by substituents (e.g., 4-NO2), as re-
quired for sucha mechanism, nearly all substituents in meta-
or para-position led to a higher stability. These substitution
effects can be large, a factor of 105 can be found between
the fastest and the slowest decomposition of monosubstitut-
ed diazonium salts.
The activation energy for the dediazotation of the benze-
nediazonium ion and substituted derivative was determined
by Crossley[11] in water (Ea =114 kJmolꢀ1 for C6H5N2+);
later other groups determined it for various other solvents.
The activation enthalpy is quite high (DHꢀ =108 kJmolꢀ1
for C6H5N2+) and is within the same range in for most of
the nucleophilic aliphatic substitution reactions that follow
an SN1-type mechanism. In contrast, the corresponding acti-
[a] Prof. Dr. S. Bräse, Dipl.-Chem. M. Schroen
University of Karlsruhe (TH), Fritz-Haber-Weg 6
76131 Karlsruhe (Germany)
Fax : (+49)721-608-8581
E-mail: braese@ioc.uka.de
[b] Dr. S. Dahmen
vation entropies have
a
high positive value (DSꢀ =
cynora GmbH, 52134 Herzogenrath(Germany)
33 JKꢀ1 molꢀ1), compensating partly the high value of DHꢀ.
The kinetics of dediazotation of 3- and 4-substituted are-
nediazonium ions cannot be explained by a Hammett equa-
tion, but witha DSP (dual substituent parameter) equation
[c] Prof. Dr. C. Popescu, Prof. Dr. F.-J. Wortmann
German Wool ResearchInstitute, Veltmanplatz 8
52062 Aachen (Germany)
Fax : (+49)241-4469100
E-mail: popescu@dwi.rwth-aachen.de
5286
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400386
Chem. Eur. J. 2004, 10, 5285 – 5296

5285 – 5296
as demonstrated by Dickinson and Eaborn[13] in 1959. This
equation takes the mesomeric and inductive effects of the
substituents independently into account and leads to a
nearly perfect correction of calculated and experimentally
verified rate constants[1] (See Figure 1).
plexation with suitable crown ethers. The first diazonium
salt crown ether adduct was isolated in 1975 by Haymore
and identified as a 1:1 complex.[1,17] The ORTEP structure
of this benzenediazonium hexafluorophosphate complex
[18]
with[18]crown-6 is accessible.
The structures of 4-meth-
[19]
oxybenzenediazonium tetrafluoroborate with[21]crown-7
(Figure 2) or dibenzo[24]crown-8, respectively, have also
been published.[20] All these structures show that the crown
ether complexes the diazenyl group.
Figure 1. DSP-Plot of para-substituted benzenediazonium ions in 0.1m
HCl at 258C.[14] The equation log(kx/kH)=sF1F +sRþ1R was used.[1]
The DSP plot explains effects of substitution on the basis
of an SN1 reaction; however, it does not explain the bimo-
lecular reaction type, which is frequently observed for such
systems.
Figure 2. Structure of the complex of 4-methoxybenzenediazonium tetra-
fluoroborate with[21]crown-7; teh counterion tetrafluoroborate was
omitted for clarity. Taken from data of reference [19]
Glaser et al. developed a method that explains the change
of stability of two fragments X and Y, which result from the
[15]
ꢀ
decomposition of X Y. The enthalpy of dediazotation of
a diazonium ion would result as a sum of two components:
the energy difference DE1 between the diazonio group in
the diazonium ion and a free nitrogen molecule (N2), and
the energy difference DE2 between the phenyl cation and
the C6H5 fragment of the diazonium ion. The energy values
DE1 and DE2 are called fragment-transfer energies. The sum
of DE1 and DE2 is the bond dissociation energy DEdiss. Stabi-
lization of the bond results in a positive dissociation energy.
For the benzenediazonium ion, the following values were
obtained: DE1 =ꢀ59.0 kJmolꢀ1, DE2 =164.9 kJmolꢀ1, and
DEdiss =105.9 kJmolꢀ1. This means, that throughout the
decay the fragment N2 is stabilized and the fragment C6H5 is
destabilized. This is in agreement with chemical intuition
and the experiment. The bond stabilization of both frag-
ments is therefore not because of their association; it is a
result of a destabilization of the phenyl cation rather than
the stabilization of the N2. The bond dissociation energy is
close to the experimental activation energy in solution (Ea =
114 kJmolꢀ1 for C6H5N2+[26]). When taking the electron cor-
relation and vibrational zero-point energy into account, the
value one obtains is even closer: Ea =115 kJmolꢀ1.
However, there is a significant difference for the crown
ether sizes. While [18]crown-6 is the smallest crown ether
suitable for complexation of diazonium ions, [21]crown-7 is
the strongest complexing and stabilizing agent for benzene-
diazonium ion.[21] Scheme 2 demonstrates the kinetics of the
Scheme 2. Decay and reaction rates of the dediazotation of substituted
arene diazonium tetrafluoroborates (k1 and k2) and complex equilibrium
constant (K) withcrown ether.
complex diazonium ion decay. The complexation equilibri-
um constant (K) characterizes the stability of these com-
plexes. The thermal decay of the uncomplexed (k1) and the
complexed (k2) ions reflects their respective stabilities.
The Table 1 shows the values for the complex equilibrium
constant and the decay rates for different substituted arene-
diazonium ions. The stability clearly depends on the substi-
tution pattern; this results in the increase in stabilization by
three orders of magnitude by complexation. It has been
shown that a linear Hammett relationship (1=0.65) was
Results and Discussion
Stabilization of diazonium ions by crown ethers: Besides the
change to larger non-nucleophilic counterions,[16] another
possibility for the stabilization of diazonium ions is the com-
Chem. Eur. J. 2004, 10, 5285 – 5296
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5287

S Bräse, C. Popescu, et al.
FULL PAPER
Table 1. Rate of dediazotation of substituted arenediazonium tetrafluor-
oborates (k1 and k2) and complex equilibrium constant (K) with
[18]crown-6 (18C6), [21]crown-7 (21C7) and dicyclohexyl-24-crown-8
(DCH24C8) in 1,2-dichloroethane at 508C.[22]
Table 3. IR spectra of arenediazonium tetrafluoroborates (p-
XC6H4N2BF4) and of their crown ethers in Nujol[26]
n˜N N[cmꢀ1
]
ꢁ
Substituent
Uncomplexed
Complex withComplex with
Substituent
Crown
ether
K
k1
k2
[18]crown-6
[21]crown-7
104 Lmolꢀ1
]
[10ꢀ3 sꢀ1
]
[10ꢀ5 sꢀ1
]
H
2300
2277
2286
2297
2247
2320
2306
2315
2319
2300
2282
2283
2302
2261
p-CH3
18C6
21C7
DCH24C8
18C6
21C7
DCH24C8
18C6
21C7
DCH24C8
18C6
21C7
DCH24C8
18C6
21C7
DCH24C8
18C6
2.47
30.8
1.80
3.36
36.0
1.48
4.69
50.6
2.76
6.03
70.2
3.53
7.72
15.4
2.05
10.4
0.144
0.22
0.032
1.4
6.2
3.0
16
1.8
0.16
6.3
10
0.63
19
0.023
0.004
0.039
1.4
p-tBu
p-CH3
p-Cl
m-CH3
H
5.68
p-CH3O
1.13
of p-CH3O, in within the region of Æ5 cmꢀ1 for the uncom-
plexed ions.
To investigate the influence of a polymeric backbone, di-
azonium salts were immobilized on solid supports.
m-CH3O
p-Cl
8.23
0.00235
2.01
Diazonium ions on solid supports: Diazonium ions have
been synthesized previously on cellulose[27,28] and macropo-
rous supports,[29] and have been used for the immobilization
of enzymes.[28] Benìs et al. showed,[30] that (p-aminophenyl-
sulfonyl)ethyl-modified cellulose could be diazotized to
give, according to the authors, a not very stable diazonium
salt that could be coupled with b-naphthol to give the corre-
sponding azo dye. Lipophilic supports are have been used in
a sequence consisting of diazotation and subsequent reduc-
tion for the synthesis of hydrazines.[31] The diazotation of ali-
phatic amines in the presence of halide ions is a convenient
method and has been used for the synthesis of Merrifield-
m-COCH3
found between log K and s++ values associated withthe
para substituent.[23]
The rate of dediazotation for the complexes with
[21]crown-7 is the slowest and the complex equilibrium con-
stant K is the highest. The absolute value of the decay rate
of the free ions as well as the complex with crown ether is
at minimum with p-Cl-C6H4N2BF4. This means that, in this
evaluation, the diazonium ion p-Cl-C6H4N2BF4 is the most
stable both in the free form and in the complexed state. Ac-
cording to the results in Table 1, p-Cl-C6H4N2BF4 in the un-
complexed form is approximately 3500 times more stable
than m-CH3O-C6H4N2BF4, which indicates that the T2
linker[24,25] is the most unstable diazonium salt. Furthermore,
p-Cl-C6H4N2BF4 is stabilized with[18]crown-6 by a factor of
10 and with[21]crown-7 by a factor of 60. The calculated
half-life time (t1/2) of the decay (for first-order kinetics)
under the given conditions (508C in 1,2-dichloroethane) is
listed in Table 2.
type resins on a Multipin-supportTM [32]
.
We have synthesized diazonium ions with different substi-
tution pattern and used them for the synthesis of tria-
zenes.[33–36] Rademann et al. used the T2-para resin for the
synthesis of triazenes,[37] while Struber et al. developed solid-
phase bound triazenes, based on nitration and further func-
tionalization of polystyrene, and investigated their chemical
and thermophysical properties.[38]
There is, however, less knowledge on the stability of di-
azonium ions on solid supports. In particular, the influence
of the substitution pattern is of interest.
Synthesis of immobilized diazonium ions: Our linker for
secondary amines (T2),[8] withan m-methoxy-type substitu-
tion, was only synthesized as an intermediate and had to be
stored and handled below 08C to avoid decomposition. Our
aim was to develop a temperature-stable diazonium ion on a
solid support, that is, to combine the broad chemical usage
of the diazo chemistry and the simple handling of solid-
phase organic chemistry. Therefore, various immobilized di-
azonium resins were prepared. Particularly, based on the
theoretical considerations by Glaser,[39] as well as on experi-
mental results[14] showing that para-chloro- and methoxy-
substituted diazonium ions are highly stable, the chlorine-
stabilized immobilized diazonium ion T2*[40] and the so-
called T2-para resin were prepared.
Table 2. Estimation of half-life time of the dediazotation (taken from
data in Table 1).[a]
Diazonium ion
t1/2 (508C) [s]
t1/2 (208C) [s]
m-CH3O-C6H4N2BF4
p-Cl-C6H4N2BF4
84
3798
0.3”106
(ꢂ3.5 days)
17”106
13.3”106
(ꢂ154 days)
781”106
p-Cl-C6H4N2BF4·[21]crown-7
(ꢂ200 days)
(ꢂ9040 days)
[a] The half-life time was calculated according to t1/2 =ln2/k. The value
A/RT
for t1/2 (208C) was calculated via the Arrhenius equation (k=AeꢀE
)
from the value for t1/2 (508C). An activation energy of EA =100 kJmolꢀ1
was assumed.
The complexation with crown ethers can be followed by
IR spectroscopic methods. The IR spectra show that the NꢁN-
vibration upon complexation with[18]crown-6 is shifted to
higher wavenumbers. Therefore, the four complexes in
The resins were prepared according standard methods by
using the Merrifield resin (chloromethylated polystryrene,
cross-linked with1–2% divinylbenzene, 100 mes,h ca.
1 mmolgꢀ1) as a solid support. For the T2* linker, the re-
quired aminobenzylalcohol 2 was synthesized in two steps
Table 3 show a shift by of 20–29 cmꢀ1. In contrast, the nN
values of the [21]crown-7 complexes are, with the exception
N
ꢁ
5288
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 5285 – 5296

Polymer-Bound Diazonium Salts
5285 – 5296
from commercially available 2-chloro-5-aminobenzoic acid.
The esterification was carried out by using trimethyl chloro-
silane in methanol, which gave rise to the methyl ester. The
ester was reduced then with lithium aluminum hydride
(LiAlH4) in THF to the corresponding benzyl alcohol 2.
Coupling to Merrifield resin proceeds with O-selectivity if
carried out with sodium hydride (Scheme 3).
or 08C, different resins are obtained withdistinct thermo-
chemical features (see below).
The exchange of the counterion tetrafluoroborate was ac-
complished by using the resin 10 to give the hexafluorophos-
phate resin 12 after treatment withpotassium hexafluoro-
phosphate.
Complexation with crown
ethers: The stability of diazo-
nium ions can be increased by
complexation
withcrown
ethers,[25] with[21]crown-7
showing the highest thermal
stability. Therefore the poly-
mer-bound diazonium ion
4
was shaken with [18]crown-6
or [21]crown-7 to give the new
resins 4·18c6, and 4·21c7, re-
spectively (Figure 3).
Scheme 3. Syntheses of the immobilized diazonium ion 4.
A series of new polymer-bound resins (4–10) were de-
signed taking the influence of substituents into account.
Only resin 5 contained the diazonium group in the ortho po-
sition, which might not only lead to side product such as cy-
clization, but would also have a detrimental effect on the
possible coordination of crown ethers.[23] The resins used for
polymer-bound diazonium ions 5–10 were prepared by cou-
pling either the phenolate, alcoholate, or carboxylate onto
Merrifield resin. The loadings of the resins were determined
by CHN analysis.
The subsequent diazotation reaction of the amino resins
with tert-butyl nitrite (tBuONO) and bortrifluoride etherate
(BF3·Et2O) in THF between 0 and ꢀ108C led to the poly-
mer-bound diazonium ions 4–10. The chloride resin 11 was
obtained by starting from the corresponding ammonium
chloride resin.
It is interesting to note that the resin 8 immediately
turned dark after preparation, an indication that the meso-
meric diazocyclohexadieneoxyl structure might bring an im-
portant contribution for stabilization.[41] In addition, the
phenol formed upon dediazotation might also lead to cross-
linking. If the diazotation is performed at room temperature
Figure 3. Part of IR spectra of polymer-bound diazonium ions; n(NꢁN)
of the diazonium functionality being shown, that is a) uncomplexed,
b) partly complexed as [18]crown-6 complex or c) as [21]crown-7 com-
plex.
Spectroscopic properties of the immobilized diazonium ions:
All resins prepared (4·18c6, 4·21c7) were examined by IR
spectroscopy. The complexa-
tion with[18]crown-6 shifts the
NꢁN stretching vibration clear-
ly to higher wave numbers
(shifts of 20–29 cmꢀ1 are men-
tioned in literature[25]), whereas
the
complexation
with
[21]crown-7 leads to small
shifts only (about +5 cmꢀ1 rel-
ative to the uncomplexed di-
azonium ion). In spite of a
high excess of crown ether the
polymer-bound diazonium salt
4 could not be transformed
into its [18]crown-6 complex
4·18c6 quantitatively, because
the rates of complexation are
lower than in the case of the
Chem. Eur. J. 2004, 10, 5285 – 5296
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5289

S Bräse, C. Popescu, et al.
FULL PAPER
Table 4. Decomposition temperatures (Tdecomp) for the resins 4–12 as de-
corresponding [21]crown-7 complexes. Therefore, it is possi-
ble that the [18]crown-6 ether is partly washed off during
workup.
termined from their DSC plots (heating rate 10 Kminꢀ1).
Resin
Tdecomp
Resin
Tdecomp
4
114
122
130
147
105
8[a]
8[b]
9
10
11
12
156
147
93
134
87
4·18C6
4·21C7
5
6
7
The stability of immobilized diazonium salts
Elemental analysis: The stabilities of the diazonium ions
were investigated by thermal decomposition. Isothermal de-
composition studies were performed to give an experimental
value for the decay of nitrogen during normal laboratory
use. The studies where carried out at 608C, and the conver-
sion was examined by CHN combustion analysis.[42] Resin
samples were kept isothermally at 608C and after 1, 6, 9, 34,
and 48 hsamples were taken and cooled, and the nitrogen
content examined. The experimentally determined half-life
time of resin 4 is eleven hours at 608C. Analogously the
resins 4·18c6 and 4·21c7 were examined. It is quite puzzling
that, although thermochemically more stable than the un-
complexed resin 4, the resin 4·18c6 is less kinetically stable.
We explain this fact that the crown ether, which is not per-
fectly suitable, might facilitate the reduction to the parent
hydrocarbon. The reduction of diazonium salts using alicy-
clic and cyclic ethers is known.[1]
[c]
–
128
[a] Preparation of the resin at 08C. [b] Preparation of the resin at room
temperature. [c] Rapid decomposition.
This is not likely since the resins have been washed and
dried before usage.
The structural change of the diazonium ions results in
changes of stability. The resin 7 was quite unstable and de-
composes prematurely preventing the thermal analysis. The
change to a para-alkoxy group dramatically increases the
stability (resin 8). However, depending on the diazotation
method (see Experimental Section), apparently two differ-
ent structures can be obtained. A possible, yet to be con-
firmed, hypothesis would be that the two resins inhibit dif-
ferent location of the counterion. Since the mobility of the
counterion is limited because of being on a solid support
without a solvent, it is possible that the two tautomeric
structures (Scheme 5) are differently favored with different
Thermal analysis: The thermal stability of the resins 4–6, 8–
12, 4·18c6, and 4·21c7 has been investigated by the aid of
DSC (differential scanning calorimetry). The resin 6 was
also submitted to TGA (thermogravimetrical analysis)
measurements.[43,44]
The TGA investigations of the resin 6 revealed that the
full decomposition proceeds through a two-step process. The
stoichiometric liberation of dinitrogen, recorded by TGA, is
accompanied by an exothermic effect at around 1408C ob-
served on the DSC curve. Boron trifluoride apparently is
still bound somehow to the resin and is liberated from the
resin above 2008C.
Scheme 5. The alkoxybenzenediazonium–diazocyclohexadienyloxy meso-
merism.
We may consider that for all resins, when heated, the
compound releases first a nitrogen molecule resulting in het-
erolytic cleavage and a subsequent Schiemann-type reac-
tion[45] (Scheme 4).
location of the counterion. However, the IR stretching
ought to be the same. Alternatively, the alkoxybenzenedia-
zonium–diazocyclohexadienyloxy mesomerism plays a major
role for the stabilization of diazoniumphenolate. Further in-
vestigation might explain this finding.
A carboxy group leads to greater stabilization than an al-
koxymethyl group (6 vs 9). Comparing the resins 6/9 with 4/
10, we notice that the substitution of an alkoxymethyl group
against a carboxy group is more efficient when placed in the
meta position.
Scheme 4. Decomposition of diazonium ions 13.
The resin 5 combines the stabilizing effects of the nitro
group and the carboxy moiety, and together with the resin 8
is one of the most stable resins. All the above-mentioned ef-
fects are in good agreement withthe DSP plot and Table 1.
The counterion increases the stability of the resin in the
order chloride !hexafluorophosphateꢂtetrafluoroborate.
The influence of the complexation is quite remarkable, al-
though in the range of the above-mentioned values. While
the complexation with [18]crown-6 shifts the peak tempera-
ture to 1228C, complexation with[21]crown-7 results in an-
other increase, as already noticed in conventional diazonium
salts.
All resins exhibit a first exothermal decomposition be-
tween 90 and 1808C (Table 4), withthe exception of resin 7,
which is much less thermostable.
The DSC data show a reaction enthalpy of DRH=
ꢀ120 kJmolꢀ1 for the fragmentation reaction, which is inde-
pendent of complexation. The decomposition reaction fol-
lows a first-order kinetic, which is expected from both
Scheme 4 and comparision to literature.[13] This reconfirms
an SN1-type mechanism of decomposition. The SNAr-type
mechanism would require the presence of a nucleophile.
5290
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 5285 – 5296

Polymer-Bound Diazonium Salts
5285 – 5296
The T2* resin 4 is, from a practical standpoint, the most
useful, because the ether functionality is by far more stable
than the hydrolytically sensitive ester group. The T2-para
resin 8, although thermally quite stable, is less attractive,
due to its strong coloration and the formation of byproducts.
energy, pre-exponential factor and half life-time, are
summed up in Table 6.
The results in Table 6 show that the values calculated for
the activation energy are all around of 113 kJmolꢀ1 (within
Table 6. The values of the activation energies, pre-exponential factors,
and half life times for the resins 4–12 as calculated by the described
methodology.
Thermochemical evaluation and half-life time predictions:
The values for the activation energy (Ea), the pre-exponen-
tial factor (A) and the function of conversion degree a from
Equation (1) have been calculated by different methods
from the thermodynamical data obtained by DSC measure-
ments (see Table 5). For more details of the mathematical
calculations of these methods, see the Experimental Section.
Resin
EA
A
Half life-time at 208C
[kJmolꢀ1
]
[”1011 sꢀ1
]
Relative value
Days
vs resin 4
4
5
6
105.6
129.9
112.9
126.2
116.9
91.34
115.5
87.53
130.9
25.53
1940
742.8
299
143.2
1.28
92.56
0.0641
15530
1.00
284.17
0.69
403.41
18.48
0.06
16.09
0.24
53.53
21
6100
15
8700
400
1.3
350
5
8[a]
8[b]
9
ꢀ
ꢁ
da
dt
E
RT
¼ fð1ꢀaÞ A exp ꢀ
ð1Þ
10
11
12
1150
Table 5. The values of the kinetic parameters E and A, calculated by var-
ious methods, for resin 4.[46] Investigations on the thermal stability of a
diazonium ion on solid support.
[a] Preparation of the resin at 08C. [b] Preparation of the resin at room
temperature. [c] Rapid decomposition.
Æ13%), very close to 114 kJmolꢀ1 obtained by Crossley
through isothermal measurements.[3] On the other hand, the
pre-exponential factor values range from 0.06 to 15530”
1011 sꢀ1. These results support the idea that increasing the
stability of the diazonium ion can be done at the expense of
decreasing the entropy/decreasing the value of the pre-expo-
nential factor.
The effect of the nature of the bond between the diazoni-
um ion and the resin on the stability can be estimated by
comparing the ester and ether bonds. Therefore we com-
pared life-times at 208C for resins 4 and 10, and 6 and 9, re-
spectively. In bothcases the ratio is about 13 in favor of
ester, suggesting an ester bond increases the stability more
than the ether one does.
Methods
E
lnA
f(1ꢀa)
[kJ molꢀ1
]
[minꢀ1
]
iso-conversional methods
differential
integral
Kissinger
invariant kinetic parameters
“isothermal” parameters [Eq. (15)] 104.4Æ1.2 30.62Æ2.1
30.24Æ1.2
1ꢀa
100.0Æ6
104.4Æ5.1
101.4Æ5.3
–
–
95.7Æ1.7 28.55Æ0.68
1ꢀa
1ꢀa
For resin 4 (example resin) an activation energy (Ea) be-
tween 95.7 and 104.4 kJmolꢀ1, and ln(A) between 28.55 and
30.62 minꢀ1 was determined for a reaction obeying a first-
order decay. The half-life time was analytically determined
by elemental analysis by of isothermal decomposition at
608C to be 11.5 h (see Figure 4). The agreement of the cal-
culated values withtohse determined experimentally is
quite good. We concluded from the example (resin 4) that
the nonisothermal kinetic analysis allows us to predict the
stability of the diazonium ions accurately enough. The re-
sults obtained for all the resins, in terms of activation
By comparing the resins 4 and 9, and 10 with 6, we no-
ticed that the presence of chlorine in the ortho position
leads to about 20 times higher stability of the resin. Also, by
comparing the values of life-times at 208C for resins 10, 11,
and 12 we observed that the increase of the volume of the
counterion linked to diazonium ion leads to an increase of
ꢀ
the stability, according to the rank: Clꢀ <BF4ꢀ ꢂPF6
.
The results of calculations, as well as of the isothermal
test, show that the synthesized diazonium ions have a pretty
good stability over a quite large interval of temperatures,
[2]
and muchabove the values known for similar compounds.
Conclusion
We synthesized and investigated the thermal stability of a
diazonium ion bound to polymeric solid supports over a
large range of temperatures from 0 to 2008C by the help of
DSC. In addition we used nonisothermal kinetic methods to
infer the kinetic parameters. Two iso-conversional calcula-
tion methods (differential and integral) were applied and
bothled to close results. We also used Kissingerꢁs method
and the invariant kinetic parameters method and calculated
the corresponding values of E, which were found to lie close
to the range of values calculated by the iso-conversional
Figure 4. DSC measurements on solid-phase bound diazonium ions 4,
4·18c6, and 4·21c7; the measurements were performed with 8–12 mg
resin and heating rates of 10.08Cminꢀ1
.
Chem. Eur. J. 2004, 10, 5285 – 5296
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5291

S Bräse, C. Popescu, et al.
FULL PAPER
(5-Amino-2-chlorophenyl)methyl-1-oxymethyl-
polystyrene (3): dry 500 mL, three-necked
round-bottom flask was fitted witha mechanical
stirrer, gas-inlet, and addition funnel. The appara-
tus was purged withargon and charged withdry
methods. With the values of E obtained by the iso-conver-
sional methods we calculated the kinetic function that fits
best the experimental curve, and the resulting value of A.
On the other hand, the values of the invariant kinetic pa-
rameters, Ei and Ai, were used for choosing the most suita-
ble kinetic function out of a list of kinetic functions from
the literature. In both cases the reaction order model, that
is, first-order, was found to be most appropriate to describe
the reaction both from fitting and from a physical point of
view. An isothermal-like method was also used, giving simi-
lar values of the parameters and supporting the choice of
the first-order model as the kinetic function. The values ob-
tained for the kinetic triplet were finally shown to give a
fairly good reproducibility of the initial DSC signal.
A
DMF (100 mL) and sodium hydride (0.64 g,
26.8 mmol, 60% in paraffin). After addition of
Merrifield-resin (1; 5 g, 5 mmol, loading=0.64 mmolgꢀ1), 5-amino-2-
chlorbenzylalcohol (2.1 g, 13.4 mmol) was added portion wise (caution:
H2 evolution). After a reaction time of 4 hat 40–45 8C, it was quenched
with saturated ammonium chloride solution (10 mL), and the resin was
purified by washing on an inert gas frit with solvents (three times with
each, ca. 20 mL for each solvent per 1.00 g resin): THF, Et2O, and
MeOH. Subsequently the resin was dried in vacuo. Elemental analysis
calcd (%) for C85H86NOCl (1172): C 87.03, H 7.34, N 1.19; found: C
86.38, H 8.20, N 0.77; loading: 64%.
(4-Aminophenoxy)methylpolystyrene:
Th
e
The resins 4–12 have good stability at room temperature
and above withthe exception of resin 7, which decomposes
at lower temperature. We can conclude that ester-bound di-
azonium-ions are more stable than ether-bound ones (4/10
and 6/9). Chlorine in ortho position to the linkage on the
resin stabilises the diazonium resin. The size of the counter-
ion plays an important role in the stability; bigger ones give
more stability than smaller ones (10, 11, 12).
The calculated parameters E and A, and kinetic function,
f(1ꢀa), were further used for a half-life time prediction at
various temperatures. The results of calculations were com-
pared and suggestions on how to design a diazonium ion
with high thermal stability were underlined.
procedure described for the preparation of 3
was followed. The reaction starting from NaH
(0.64 g, 26.8 mmol), Merrifield resin (5 g,
5 mmol)
and
4-aminophenol
(1.46 g,
13.4 mmol) in DMF (ca. 100 mL) at 408C for
4 hyielded the resin. Elemental analysis calcd (%) for C 84H85NO (1123):
C 89.76, H 7.57, N 1.25; found: C 89.16, H 7.13, N 0.92; loading: 73%.
(5-Amino-2-chlorophenyl)carboxymethyl-
polystyrene: The procedure described for the
preparation of 3 was followed. The reaction
starting from NaH (0.64 g, 26.8 mmol), Merri-
field resin (5 g, 5 mmol), and 5-amino-2-chlor-
benzoic acid (2.29 g, 13.4 mmol) in DMF (ca.
100 mL) at 408C for 4 hyielded the colorless
resin. Elemental analysis calcd (%) for C85H84O2NCl (1185.5): C 86.04, H
7.09, N 1.18; found: C 88.17, H 7.61, N 0. 84; loading: 71%.
(4-Aminophenyl)methyl-1-oxymethylpolystyrene: The procedure descri-
bed for the preparation of 3 was followed. The reaction starting from
NaH (0.64 g, 26.8 mmol), Merrifield resin (5 g, 5 mmol), and 4-aminoben-
zylalcohol (1.65 g, 13.4 mmol) in DMF (ca.
Experimental Section
100 mL) at 40–458C for 16 hyielded the col-
All resins were characterized by IR spectroscopy, loading and conver-
sions were valued by CHN combustion analysis. Experimentally exam-
ined loading of the resins can be correlated very well with the results
from CHN-combustion analysis. Loadings of the used chloromethylpoly-
styrene were between 0.6 and 1.4 mmolgꢀ1. All new compounds were
characterized completely or compared to known substances.
ꢀ
orless resin. IR (KBr): n˜ =3464 (m, nas(N
ꢀ
H)), 3381 (m, ns(N H)), 3081 (m), 3058 (m),
ꢀ
3023 (m, nAr(C H)), 2912 (s, n(CH2)), 2850
(s, n(CH2)), 1943 (m), 1872 (m), 1803 (m),
ꢀ
ꢀ
ꢀ
1678 (m, d(N H)), 1598 (s, nAr(C=C)), 1444 (m, d(C H)), 1067 (s, n(C
O)), 840 (m), 820 (m), 743 (s), 694 cmꢀ1 (s). elemental analysis calcd (%)
for C85H87NO (1137): C 89.71, H 7.65, N 1.23; found: C 89.99, H 6.62, N
0.81; loading: 66%.
3-Diazoniophenyl-1-oxymethylpolystyrene tetrafluoroborate (7): A dry
500 mL, three-necked round-bottom flask was fitted with a mechanical
stirrer, gas-inlet, and addition funnel. The
(4-Aminophenyl)carboxymethylpolystyrene: The procedure described for
the preparation of 3 was followed. The reaction starting from Cs2CO3
(3.05 g, 9.36 mmol), Merrifield resin (5 g,
5 mmol), and anthranilic acid (2.05 g,
15 mmol) in DMF (ca. 100 mL) at 508C for
apparatus was purged withargon and charg-
ed withdry DMF (250 mL) and sodium hy-
dride (63.0 mmol, 1.51 g, 60% in paraffin;
2.52 g in total). After addition of the Merri-
field-resin (1; 20.0 g, 12.8 mmol, loading=
0.64 mmolgꢀ1),
m-aminophenol
(6.9 g,
1 hyielded the colorless resin. IR (KBr): n˜ =
ꢀ
ꢀ
3485 (m, nas(N H)), 3388 (m, ns(N H)),
63 mmol) was added portionwise (caution:
H2 evolution). After a reaction time of 20 h, the resin was purified by
washing on an inert gas frit with solvents (three times with each; ca.
20 mL for eachsolvent per 1.00 g resin): THF, Et 2O, and MeOH. Subse-
quently the resin was dried in vacuum. IR (KBr): n˜ =3390, 3200, 3080,
3060, 3020, 2910, 2840, 2640, 2600, 2310, 2340, 2380, 2260, 2110, 1940,
1870, 1800, 1710, 1670, 1620, 1590, 740, 690 cmꢀ1; elemental analysis
calcd (%) for C125H125ON (1655.0): C 90.63, H 7.55, N 0.85; found: C
89.90, H 8.03, N 0.89.
ꢀ
3081 (m), 3058 (m), 3024 (m, nAr(C H)),
2925 (s, n(CH2)), 2850 (s, n(CH2)), 1944 (m),
ꢀ
1873 (m), 1804 (m), 1701 (s, n(C=O)), 1618 (m, d(N H)), 1598 (s, nAr(C=
ꢀ1
ꢀ
ꢀ
C)), 1444 (m, d(C H)), 1070 (s, n(C O)), 840 (m), 744 (s), 693 cm (s).
elemental analysis calcd (%) for C85H85NO2 (1151): C 88.62, H 7.38, N
1.22; found: C 87.28, H 7.36, N 0.91; loading: 75%.
(2-Amino-5-nitrophenyl)carboxymethylpolystyrene: The procedure de-
scribed for the preparation of 3 was followed. The reaction starting from
Cs2CO3 (3.05 g, 9.36 mmol), Merrifield resin (5 g, 4.7 mmol) and 5-nitro-
anthranilic acid (2.5 g, 14.1 mmol) in DMF (ca. 100 mL) at 508C for 1.5 h
yielded the colorless resin. IR (KBr): n˜ =
The amine resin (10.0 g, 6.4 mmol) was suspended in dry THF (150 mL)
and cooled by means of a cold bath(EtOH/dry ice) to ꢀ208C. After
20 min, BF3·Et2O (6.9 mL, 7.7 g, 54 mmol) was added, followed after
5 min by tBuONO (5.7 mL, 5.0 g, 49 mmol). After a reaction time of
ꢀ
ꢀ
3473 (m, nas(N H)), 3355 (m, ns(N H)),
ꢀ
3081 (m), 3058 (m), 3023 (m, nAr(C H)),
30 min, the mixture was collected in an inert gas frit, filtered, and washed
ꢀ1
3002 (m), 2921 (s, n(CH2)), 2850 (s, n(CH2)),
1944 (m), 1871 (m), 1803 (m), 1697 (s, n(C=
withchilled THF (4”15 mLg
resin). The resin was washed on an inert
gas frit withsolvents (three times witheach; ca. 20 mL for eachsolvent
per 1.00 g resin): THF, Et2O, and MeOH. Subsequently the resin was
dried in vacuo.
ꢀ
O)), 1597 (s, nAr(C=C)), 1449 (m, d(C H)),
1321 (s, ns(N=O)), 829 (m), 746 (s), 695 cmꢀ1
5292
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 5285 – 5296

Polymer-Bound Diazonium Salts
5285 – 5296
(s); elemental analysis calcd (%) for C85H84N2O4 (1196): C 85.28, H 7.02,
N 2.34; found: C 86.423, H 7.87, N 1.91; loading: 81%.
at 08C for 30 min yielded the light yellow resin. IR (KBr): n˜ =3395 (m,
ꢀ
ꢀ
n(N H)), 3082 (m), 3059 (m), 3024 (m, nAr(C H)), 3001 (m), 2910 (s,
ꢀ
n(CH2)), 2852 (s, n(CH2)), 2290 (w, n( NꢁN)), 1943 (m), 1872 (m), 1803
(5-Diazonio-2-chlorophenyl)methyl-1-oxymethylpolystyrene tetrafluoro-
borate (4): The amine resin (2 g) was suspended in dry THF (50 mL).
BF3·Et2O (0.33 mL) was added, followed
ꢀ
(m), 1709 (s, n(C=O)), 1601 (s, nAr(C=C)), 1450 (m, d(C H)), 840 (m),
748 (s), 696 cmꢀ1 (s); elemental analysis calcd (%) for C85H83N2O2BF4
(1250): C 81.60, H 6.64, N 2.24; found: C 85.72, H 7.95, N 0.048; loading:
2%.
after 5 min by tBuONO (0.35 mL). After a
reaction time of 60 min, the mixture was col-
(2-Diazonio-2-nitrophenyl)carboxymethylpolystyrene tetrafluoroborate
(5): The procedure described for the preparation of 4 was followed. The
reaction starting from amino resin (1.6 g), BF3·OEt2 (0.4 mL), and
tBuONO (0.4 mL) in THF (50 mL) at 08C for
lected in an inert gas frit, filtered, and washed
ꢀ1
withcihlled THF (4”15 mLg
resin). The
resin was washed on an inert gas frit with sol-
vents (three times with each, ca. 20 mL for
1.5 hyielded a light yellow resin. IR (KBr):
eachsolvent per 1.00 g resin): THF, Et O, and
2
ꢀ
n˜ =3082 (m), 3058 (m), 3022 (m, nAr(C H)),
3001 (m), 2909 (s, n(CH2)), 2850 (s, n(CH2)),
MeOH. Subsequently the resin was dried in
vacuo. 19F NMR (C6D6): d=-148.3 ppm (s,
BF4ꢀ). IR (KBr): n˜ =3081 (m), 3058 (m), 3021 (m, nAr(C H)), 3002 (m),
ꢀ
ꢀ
2306 (m, n( NꢁN)), 1945 (m), 1872 (m), 1804
ꢀ
(m), 1702 (s, n(C=O)), 1600 (s, nAr(C=C)),
2928 (s, n(CH2)), 2851 (s, n(CH2)), 2278 (s, n( NꢁN)), 1944 (m), 1872
ꢀ
ꢀ
ꢀ
1450 (m, d(C H)), 1323 (s, ns(N=O)), 841 (m),
(m), 1803 (m), 1600 (s, nAr(C=C)), 1441 (m, d(C H)), 1052 (s, n(C O)),
746 (s), 693 cmꢀ1 (s); elemental analysis calcd
(%) for C85H82N3O4BF4 (1295): C 78.76, H 6.33, N 3.24; found: C 83.23,
H 10.19, N 0.95; loading: 30%.
1013 (s, n(C Cl)), 842 (m), 820 (m), 741 (s), 692 cmꢀ1 (s); elemental anal-
ꢀ
ysis calcd (%) for C85H84N2OClBF4 (1270.5): C 80.28, H 6.61, N 2.20;
found: C 83.59, H 6.53, N 1.10; loading: 50%.
(5-Diazonio-2-chlorophenyl)carboxymethylpolystyrene hexafluorophos-
phate (12): Resin 10 (0.5 g) was added to a saturated solution of KPF6
(1.38 g, 7.5 mmol) in acetonitrile (6.4 mL). Agi-
(4-Diazoniophenoxy)methylpolystyrene tetrafluoroborate (8): a) Reac-
tion at room temperature: The procedure described for the preparation
of 4 was followed. The reaction starting from
tation for 64 h, followed by washing with water
amino resin (2 g), BF3·OEt2 (0.33 mL), and
and organic solvents, as described above yield-
tBuONO (0.35 mL) in THF (50 mL) at room
ed the resin. IR (KBr): n˜ =3081 (m), 3059 (m),
temperature yielded a dark resin. IR (KBr):
ꢀ
3024 (m, nAr(C H)), 2925 (s, n(CH2)), 2851 (s,
n˜ =3102 (m), 3081 (m), 3057 (m), 3022 (m,
ꢀ
ꢀ
n(CH2)), 2283 (s, n( NꢁN)), 1945 (m), 1872
nAr(C H)), 2924 (s, n(CH2)), 2851 (s, n(CH2)),
ꢀ
(m), 1803 (m), 1736 (s, n(C=O)), 1601 (s,
nAr(C=C)), 1450 (m, d(C H)), 1026 (s, n(C
2243 (s, n( NꢁN)), 1944 (m), 1872 (m), 1803 (m), 1600 (s, nAr(C=C)),
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
1441 (m, d(C H)), 1052 (s, n(C O)), 1013 (s, n(C Cl)), 834 (m), 741 (s),
Cl)), 829 (m), 743 (s), 697 cmꢀ1 (s). elemental analysis calcd (%) for
C85H82N2O2ClPF6 (1342.5): C 75.98, H 6.11, N 2.09; found: C 76.503, H
6.92, N 1.10; loading: 53%.
692 cmꢀ1 (s).
b) Reaction at 08C: Similarly, the reaction starting from amino resin
(2 g), BF3·OEt2 (0.33 mL), and tBuONO (0.35 mL) in THF (50 mL) at
08C yielded a dark resin. IR (KBr): n˜ =3081 (m), 3058 (m), 3022 (m,
(5-Diazonio-2-chlorophenyl)carboxymethylpolystyrene chloride (11): Th e
amino resin (1 g) was suspended in dry dichloromethane (ca. 25 mL) and
treated withHCl (5 mL, 1 m). After 1 h, the
resin was washed on an inert gas frit with sol-
vents (three times with each, ca. 20 mL for
ꢀ
nAr(C H)), 2915 (s, n(CH2)), 2849 (s, n(CH2)), 2243 (m, n(NꢁN)), 1943
ꢀ
(m), 1872 (m), 1803 (m), 1747 (w), 1598 (s, nAr(C=C)), 1449 (m, d(C H)),
1067 (s, n(C O)), 825 (m), 745 (s), 694 cm (s); elemental analysis calcd
ꢀ1
ꢀ
(%) for C84H83N2OBF4 (1222): C 82.49, H 6.79, N 2.29; found: C 83.92, H
6.97, N 1.40; loading: 61%.
eachsolvent): THF, Et 2O, and MeOH. Subse-
quently the resin was dried in vacuo. The resin
was the suspended in THF and cooled to
(5-Diazonio-2-chlorophenyl)carboxymethylpolystyrene tetrafluoroborate
(10): The procedure described for the preparation of 4 was followed. The
reaction starting from amino resin (1 g), BF3·OEt2 (0.33 mL), and
tBuONO (0.35 mL) in THF (50 mL) at 08C for
ꢀ208C. Isoamylnitrite (0.35 mL) was added,
stirred for 1 hat ꢀ20 to ꢀ108C. After 1 h, the resin was washed on an
inert gas frit withsolvents (three times witheach, ca. 20 mL for eachsol-
vent): THF, Et2O, and MeOH. Subsequently the resin was dried in vacuo.
30 min yielded the orange resin. IR (KBr): n˜ =
ꢀ
3081 (m), 3059 (m), 3024 (m, nAr(C H)), 2920 (s,
ꢀ
IR (KBr): n˜ =3081 (m), 3058 (m), 3022 (m, nAr(C H)), 2903 (s, n(CH2)),
ꢀ
n(CH2)), 2850 (s, n(CH2)), 2286 (s, n( NꢁN)),
ꢀ
2850 (s, n(CH2)), 2253 (w, n( NꢁN)), 2116 (m), 1944 (m), 1872 (m), 1802
1944 (m), 1872 (m), 1803 (m), 1736 (s, n(C=O)),
ꢀ
(m), 1719 (s, n(C=O)), 1598 (s, nAr(C=C)), 1441 (m, d(C H)), 1022 (s,
ꢀ
1601 (s, nAr(C=C)), 1451 (m, d(C H)), 1052 (s,
ꢀ
n(C Cl)), 816 (m), 739 (s), 694 (s); elemental analysis calcd (%) for
ꢀ
ꢀ
n(C O)), 1013 (s, n(C Cl)), 821 (m), 747 (s),
C85H82N2O2Cl2 (1233): C 82.73, H 6.65, N 2.27; found: C 82.851, H 6.53,
N 1.38; loading: 61%.
697 cmꢀ1 (s); elemental analysis calcd (%) for
C85H82N2O2ClBF4 (1284): C 79.42, H 6.38, N 2.18; found: C 79.538, H
6.43, N 1.55; loading: 71%.
Thermal measurements: The thermal measurements were carried out on
a Perkin–Elmer DSC 7 withpolymer-bound diazonium ions witha load-
ing of 1.0 mmolgꢀ1. The runs were conducted in perforated closed alumi-
num crucibles containing approximately 10 mg of the sample. An empty
crucible was used as the reference. The study was carried out over the
temperature range of ambient to 2008C, at seven heating rates of 1, 3, 5,
(4-Diazoniophenyl)methyl-1-oxymethylpolystyrene tetrafluoroborate (9):
The procedure described for the preparation of 4 was followed. The reac-
tion starting from amino resin (1 g), BF3·OEt2
(0.33 mL), and tBuONO (0.35 mL) in THF
(50 mL) at 08C for 30 min yielded a dark to
brownishyellow resin. IR (KBr): n˜ =3081 (m),
7, 10, 15 and 20 Kminꢀ1
.
The DSC analysis of the used chloromethylpolystyrene resin resulted an
endothermic melting enthalpy of DMH=2.6 Jgꢀ1 at the melting point of
1188C. The melting point is in the region of the exothermal decomposi-
tion of the diazonium ions, but it is about 50 times smaller. Because the
tolerance of the method is at about 10%, the melting enthalpy was not
taken into account for the examination of the reaction enthalpy.
ꢀ
3058 (m), 3022 (m, nAr(C H)), 2907 (s, n(CH2)),
ꢀ
2850 (s, n(CH2)), 2274 (w, n( NꢁN)), 1944 (m),
ꢀ
ꢀ
1872 (m), 1803 (m), 1599 (s, nAr(C=C)), 1442 (m, d(C H)), 1060 (s, n(C
O)), 840 (m), 743 (s), 694 cmꢀ1 (s); elemental analysis calcd (%) for
C85H85N2OBF4 (1236): C 82.52, H 6.88, N 2.27; found: C 86.133, H 7.706,
N 1.065; loading: 47%.
Mathematical methods
(4-Diazoniophenyl)carboxymethylpolys-
tyrene tetrafluoroborate (6): The proce-
dure described for the preparation of 4
was followed. The reaction starting from
amino resin (1 g), BF3·OEt2 (0.33 mL)
and tBuONO (0.35 mL) in THF (50 mL)
Kinetic analysis: The stability of the resins is given by the rate at which
the first decomposition step occurs. Therefore we used DSC runs at
seven different heating rates for investigating the decomposition reaction
occurring around 90–1608C (Figure 5); the kinetics of this reaction were
analyzed with the help of both the iso-conversional differential and inte-
Chem. Eur. J. 2004, 10, 5285 – 5296
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5293

S Bräse, C. Popescu, et al.
FULL PAPER
Parameter E: From the recorded data, we first calculated the value of the
parameter E by the help of two iso-conversional methods, that is, the dif-
ferential method of Friedman,[47] and an integral one, which is a variant
of those proposed by Flynn–Wall and Ozawa.[48,49]
The differential method is based on Equation (1) written in a logarithmic
form [Eq. (6)]
lnðda=dtÞ ¼ fðꢀE=RÞð1=TÞg þ ln½Affð1ꢀaÞgꢃ
ð6Þ
The plot of ln(da/dt) versus 1/T for values measured at the same value of
a on each of the seven runs is a straight line, the slope of which equals
ꢀE/R and the intercept is composed of the value of parameter A and the
unknown kinetic function, f(1-a).
The integral method is based on Equation (5). The method makes use of
the first mean theorem of integrals for evaluating the temperature inte-
gral I(T)[50] [Eq. (7)].
Figure 5. DSC measurements on solid-phase bound diazonium ion 4 at
various heating rates.
IðTÞ ¼ ðTjꢀTiÞ expðꢀE=RTxÞ
ð7Þ
In Equation (7) Tx is a temperature value within the interval (Ti, Tj). For
temperature intervals of up to 108 the approximation given in Equa-
tion (8) is considered good enough[49] and it is used in the present work.
gral methods as well as the invariant kinetic parameters method, Kissing-
er method, and a differential isothermal method. All of them led to simi-
lar values for the kinetic parameters and indicated that the first-order
mechanism fits the data best. The values of the kinetic parameters were
further used for calculating the half-life time of the ion at various temper-
atures. The result of stability prediction at 68C was compared withthose
of an isothermal test monitored analytically, and we noticed a good
agreement of the calculated and measured times. As it appears the de-
composition process is single step, corresponding to Scheme 4. The meth-
odology was exemplified for resin 4, but similar approaches were used
for the other ones.
Tx ¼ ðTi þ TjÞ=2
ð8Þ
With Equation (7), after rearranging and taking the logarithms, Equa-
tion (5) can be written in the following form [Eq. (9)]
ꢀ
ꢁ
ꢀ
ꢁ
b
E 1
R Tz
A
ln
¼ ꢀ
þ ln
ð9Þ
TjꢀTi
gðajÞꢀgðaiÞ
The kinetics of the decomposition are described by the known differen-
tial relationship given in Equation (1):
The plot of ln[b/(TjꢀTi)] versus 1/Tx for values of temperatures measured
at the same value of a on eachof the seven curves leads to a straight
line, the slope of which equals ꢀE/R and the intercept is composed of
the value of parameter A and the integral of the unknown kinetic func-
tion.
ꢀ
ꢁ
da
dt
E
RT
¼ fð1ꢀaÞAexp
ꢀ
ð1Þ
The results, calculated by each of the two methods, are plotted in
Figure 6 against the conversion degree.
in which a, the conversion degree, is calculated from the area ratio of the
DSC peaks [Eq. (2)]
DSCt
DSCTotal
ð2Þ
a ¼
and the temperature time dependency is given by Equation (3) in which
b is the heating rate.
T ¼ T0 þ bt
ð3Þ
Eqn. (1) is used also in its integral form [Eq. (4)]
aj
Tj
ꢀ
ꢁ
Z
Z
da
fð1ꢀaÞ
A
b
E
RT
¼
exp
ꢀ
dT
ð4Þ
ai
Ti
The left-hand side integral is known as “the conversion integral” and de-
noted by g(a) and the right-hand side integral is called “the temperature
integral” and we denote it by I(T). By considering the heating rate (b)
and the parameter A as being temperature independent, Equation (4)
can be finally written as Equation (5)
Figure 6. The dependence of kinetic parameter E, as calculated from in-
tegral (variant of Flynn–Wall–Ozawa) and differential (Friedman) iso-
conversional methods, on the conversion degree (a) of the reaction.
gðaÞ ¼ ðA=bÞ IðTÞ
ð5Þ
The plots look very similar, the results from the integral method giving
slightly higher values than those obtained by the differential one. The
stairs plot of the results of the integral method shows the intervals for
which the constancy of parameter E was used in applying Equation (7)
for the temperature integral. We also note the increase of E withthe con-
version degree; this might be due to the increasing influence played by
the diffusion of produced nitrogen through the newly formed compound.
The standard deviations calculated for both cases, and given in Table 5,
suggest, however, that the variation of E withthe conversion degree is
Equations (1) and (5) were both used for investigating the thermal stabil-
ity of the diazonium ion quantitatively.
Our analysis was conducted, in a first step, to find out the so-called “ki-
netic triplet”; that is, the values of parameters A and E, usually termed
as pre-exponential factor and activation energy, respectively, and the ana-
lytical form of the kinetic function, f(1ꢀa). In the second part we investi-
gated the possibilities of predicting the thermal stability of our com-
pounds based on the obtained kinetic triplet.
5294
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 5285 – 5296

Polymer-Bound Diazonium Salts
5285 – 5296
small enough to allow considering the mean values as constants over the
whole process.
obtained ln(A)=30.24Æ1.2 minꢀ1, a value close to those obtained from
the invariant kinetic parameter method above (see Table 5).
As the value of parameter E may be considered as quasi-independent of
the conversion degree, we calculated the invariant kinetic parameters of
the reaction Ai and Ei.[51]
The validation of the calculated parameters: Within the framework of the
reaction order model, Equation (1) can be rewritten as Equation 15 in
which n is the reaction order and k(T)=Aexp(ꢀE/RT).
The method is based on the observation that by choosing a different ana-
lytical form of the kinetic function, f(1ꢀa), the values of the two parame-
ters E and A, calculated from the data recorded at a single heating rate,
are linearly related to each other through a so-called “apparent compen-
sation effect” [Eq. (10)] in which c1 and c2 are two appropriate con-
stants:[50]
lnðda=dtÞ ¼ n lnð1ꢀaÞ þ ln½kðTÞꢃ
ð15Þ
Taking the experimental values of da/dt and of the conversion degree (a)
at the same temperature (T) on all the seven DSC curves, a plot of
ln(da/dt) versus ln(1ꢀ a) gives a straight line, the slope of which equals
the reaction order (n) and the intercept is the logarithm of the rate con-
stant. The values of the intercept at several temperatures allows, then,
calculation of the parameters E and A as the slope and the intercept, re-
spectively, of the plot ln[k(T)] versus 1/T.
This procedure, which is similar to an isothermal approach,[52] can be fol-
lowed also for a more general kinetic function, as those given by Equa-
tion (13), by using the nonlinear regression.
lnðAÞ ¼ c1E þ c2
ð10Þ
Similar straight lines are obtained at each heating rate and they were
shown to form a pencil, whose coordinates of the crossing point are the
values of Ei and Ai, which are named “invariant kinetic parameters” as
they no longer depend on the choice of the kinetic function.[50] The
values calculated according to this method are also given in Table 5.
The results obtained from this “isothermal” method are also close to all
the other results and are listed in Table 5.
Finally, withthe peak temperatures ( Tpeak) from DSC curves recorded at
different heating rates, the plot of ln(b/T2peak) versus 1/Tpeak gives a
straight line, the slope of which equals ꢀE/R, from which parameter E is
easily calculated. This is the Kissinger method[52] and the result (the
value of parameter E) is also listed in Table 5.
Comparing the reproduction of the experimental data with the calculated
parameters provides further evidence for the validity of the obtained
values. The result is shown in Figure 7 for one heating rate (5 Kminꢀ1
)
We note, from Table 5, that all the calculated values of parameter E lie
within 10% of each other, no matter which method was used.
The kinetic function, f(1ꢀa): The calculated values of the two invariant
kinetic parameters allow the ready selection of the kinetic function,
f(1ꢀa), out of a group of functions proposed in literature[53] by using
Equation (1) and some statistical criteria, for example, minimizing the
sum of residuals. The best result obtained this way indicated the first-
order reaction model as the most suitable function for fitting the experi-
mental data.
The analytical form of the kinetic function f(1ꢀa) was also determined
by using an average iso-conversional value for parameter E in a method
described previously by Malek.[54] According to this, Equation (1) can be
rewritten as Equation (11).
da
dt
ð11Þ
=expðꢀEi=RTÞ ¼ Aiffð1ꢀaÞg
By substituting Equation (2) into Equation (11) and dividing bothsides
with the maximum values in order to normalize them, one obtains, after
simplifying the constants DSCTotal and Ai, Equation (12).
Figure 7. Reconstruction of the experimental DSC curves with the calcu-
lated parameters (full line is the reconstructed DSC curve, and the dots
are the experimental recorded points).
DSCt=expðꢀEi=RTÞ
fð1ꢀaÞ
GðtÞ ¼
¼
ð12Þ
max½DSCt=expðꢀEi=RTÞꢃ max½fð1ꢀaÞꢃ
and the use of the invariant kinetic parameters; a fairly good reproduci-
bility of the initial DSC signal is noted.
By this operation we obtain a left-hand side function, G(t), ranging from
0 to 1, built out of only experimental data and the value of E. This func-
tion has to be fitted by the right-hand side function, which is the normal-
ized kinetic function only, as A vanishes through division. The kinetic
function f(1ꢀa) should have the general form given in Equation (13),
and a nonlinear estimation program was used for finding the best values
of a and b for fitting the curve G(t). For the seven different heating rates
we obtained a=0Æ0.09 and b=0.9Æ0.13.
Lifetime prediction: The results of the kinetic analysis were further used
for predicting the half-life time of the synthesized product. We used the
first-order kinetic model and the parameters E and A, taken from
Table 5, in Equation (16), obtained by integrating Equation (1) for a
going from 0 to 0.5 under isothermal conditions.
ꢀ
ꢁ
lnð2Þ
E
RT
t1=2
¼
exp
ð16Þ
A
fð1ꢀaÞ ¼ aað1ꢀaÞ
b
ð13Þ
As it appears both from the invariant kinetic method and from this ap-
proach, the reaction order models best the experimental data. While the
0.9 value of the reaction order has no physical meaning, we may expect,
from the dediazonation mechanism proposed, first-order kinetics to
occur. Consequently we have taken the kinetic function as Equation (14)
[1] H. Zollinger, Diazo Chemistry, Vol. 1, VCH, Weinheim 1994.
[2] D. S. Wulfman in The Chemistry of Diazonium and Diazo Groups,
Vol. 1 (Eds: S. Patai), Wiley, New York 1978, pp. 247–339.
[3] S. Bräse, A. de Meijere in Metal-catalyzed Cross-Coupling Reactions
(Eds.: F. Diederich, P. J. Stang), Wiley-VCH, Weinheim, 1998, p. 112.
[4] A. Engel, Methoden Org. Chem. (Houben-Weyl) 4th ed. 1952, Vol. E
16a/2, 1990, pp. 1182–1226.
fð1-aÞ ¼ 1ꢀa
ð14Þ
The parameter A: The parameter A was calculated as a scaling parame-
ter, by dividing the left-hand term of Equation (4) by the kinetic function
obtained above, and taking the mean value. For the seven plots we have
[5] The Chemistry of the Hydrazo, Azo and Azoxy Groups, Vols. 1,2
(Ed.: S. Patai), Wiley, New York 1975.
[6] R. Glaser, C. J. Horan, J. Org. Chem. 1995, 60, 7518–7528.
Chem. Eur. J. 2004, 10, 5285 – 5296
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5295

S Bräse, C. Popescu, et al.
FULL PAPER
[7] R. Glaser, C. J. Horan, H. Zollinger, Angew. Chem. 1997, 109, 2324–
2328; Angew. Chem. Int. Ed. Engl. 1997, 36, 2210–2213.
[8] M. L. Crossley, R. H. Kienle, C. H. Benbrook, J. Am. Chem. Soc.
1940, 62, 1400–1403.
[33] S. Bräse, S. Dahmen, M. Pfefferkorn, J. Comb. Chem. 2000, 2, 710–
715.
[34] C. Pilot, S. Dahmen, F. Lauterwasser, S. Bräse, Tetrahedron Lett.
2001, 42, 9179–9181.
[9] H. Eustathopoulos, J. Court, J. M. Bonnier, C. R. Acad. Sci. Ser. II
1984, 298, 687–689.
[35] N. Vignola, S. Dahmen, D. Enders, S. Bräse, Tetrahedron Lett. 2001,
42, 7833–7836.
[10] H. Eustathopoulos, J. Court, C. Lambeaux, J. M. Bonnier, Nouv. J.
Chim. 1985, 9, 737–742.
[36] N. Vignola, S. Dahmen, D. Enders, S. Bräse, J. Comb. Chem. 2003,
5, 138–144.
[11] E. S. Lewis, W. H. Hinds, J. Am. Chem. Soc. 1952, 74, 304–308.
[12] D. Schulte-Frohlinde, H. Blume, Z. Phys. Chem. 1968, 59, 299–315.
[13] a) J. D. Dickinson, C. Eaborn, J. Chem. Soc. 1959, 3036–3040. b) S.
Ehrenson, R. T. C. Brownlee, R. W. Taft, Prog. Phys. Org. Chem.
1973, 10, 1–80.
[37] J. Rademann, J. Smerdka, G. Jung, P. Grosche, D. Schmid, Angew.
Chem. 2001, 113, 390–393; Angew. Chem. Int. Ed. 2001, 40, 381–
385.
[38] B. Erb, J. P. Kucma, S. Mourey, F. Struber, Chem. Eur. J. 2003, 9,
2582–2588.
[14] C. G. Swain, J. E. Sheats, K. G. Harbison, J. Am. Chem. Soc. 1975,
97, 783–790.
[39] R. Glaser, C. J. Horan, M. Lewis, H. Zollinger, J. Org. Chem. 1999,
64, 902–913.
[15] a) R. Glaser, C. J. Horan, E. D. Nelson, M. K. Hall, J. Org. Chem.
1992, 57, 215–228. b) R. Glaser, G. S.-C. Choy, C. J. Horan, J. Org.
Chem. 1992, 57, 995–999.
[16] C. Colas, M. Goeldner, Eur. J. Org. Chem. 1999, 1357–1366.
[17] B. L. Haymore, J. A. Ibers, D. W. Meek, Inorg. Chem. 1975, 14, 541–
516.
[18] R. A. Bartschin The Chemistry of Functional Groups. Supplement C
(Eds.: S. Patai, Z. Rappoport), Wiley, Chichester, 1983, p. 889.
[19] P. Groth, Acta Chem. Scand. Ser. A 1981, 35, 541–544.
[20] G. Xu, N. Zhu, F. Guo, X. Sun, Y. Zhao, Jiegou Huaxue 1986, 5,
131–134.
[21] T. Kuokkanen, J. Phys. Org. Chem. 1997, 10, 67–75.
[22] H. Nakazumi, I. Szele, K. Yoshida, H. Zollinger, Helv. Chim. Acta
1983, 66, 1721–1736.
[40] S. Dahmen, S. Bräse, Angew. Chem. 2000, 112, 3827–3830; Angew.
Chem. Int. Ed. 2000, 39, 3681–3683.
[41] a) B. Quintero, J. J. Morales, M. Quiros, M. I. Martinez-Puentedura,
M. de C. Cabeza, Free Radical Biol. Med. 2000, 29, 464–479;
b) “Quinone Diazides”:V. V. Ershov, G. A. Nikiforov, C. R. H. I.
de Jonge, Studies in Organic Chemistry, Vol. 7, Elsevier, Amsterdam,
1981.
[42] Combustion analysis is an appropiate tool for the determination of
the loading.
[43] For reviews: a) P. G. Laye, in Principles of Thermal Analysis and
Calorimetry (Ed.: P. J. Haines), 2002, pp. 55–93; b) S. L. Simon,
Thermochim. Acta 2001, 374, 55–71. M. P. DiVito, R. B. Cassel, M.
Margulies, S. Goodkowsky, Am. Lab. 1995, 27, pg. 28, 30–32, 34,
36–37.
[23] R. M. Izatt, J. D. Lamb, C. S. Swain, J. J. Christensen, B. L. Haymore,
J. Am. Chem. Soc. 1980, 102, 3032–3034.
[24] S. Bräse, J. Köbberling, D. Enders, M. Wang, R. Lazny, S. Brandtner,
Tetrahedron Lett. 1999, 40, 2105–2108.
[25] S. Bräse, Acc. Chem. Res. 2004, ASAP DOI: 10.1021/ar0200145.
[26] R. A. Bartschin Crown Ethers and Analogs (Eds.: S. Patai, Z. Rap-
poport), Wiley, Chichester 1989, p. 505–517.
[44] For DSC measurement of diazonium salts: a) D. V. Kosynkin, J. M.
Tour, Org. Lett. 2001, 3, 993–995; a) P. D. Storey, Inst. Chem. Eng.
Symp. Ser. 1981, 68, 3/P/1–3/P/9.
[45] a) H. G. O. Becker, G. Israel, J. Prakt. Chem. 1979, 321, 579–586.
b) A. Roe, Org. React. 1949, 5, 193–228.
[46] S. Bräse, S. Dahmen, C. Popescu, M. Schroen, F.-J. Wortmann,
Polym. Degrad. Stab. 2002, 75, 329–335.
[27] M. L. Fontanel, H. Bazin, R. Teoule, Bioconjugate Chem. 1993, 4,
380–385.
[28] P. Docolomansky, A. Breier, P. Gemeiner, A. Ziegelhoffer, Anal.
Lett. 1995, 28, 2585–2594.
[29] G. L. Tao, S. Furusaki, Polym. J. 1995, 27, 111–121.
[30] a) M. J. Benìs, J. Lenfeld, O. Pradova, J. Mol. Recognit. 1996, 9,
701–705. b) J. Lenfeld, M. J. Benìs, A. N. Fraga, React. Polym. 1994,
22, 127–130.
[31] A. N. Semenov, K. Y. Gordeev, Int. J. Pept. Protein Res. 1995, 45,
303–304.
[47] H. L. Friedman, J. Polym. Sci. Part C 1963, 6, 183–195.
[48] J. H. Flynn, L. A. Wall, J. Res. Natl. Bur. Stand. Sect. A 1966, 70,
487–523.
[49] T. Ozawa, Bull. Chem. Soc. Jpn. 1965, 38,1881–1886.
[50] C. Popescu, Thermochim. Acta 1996, 285, 309–323.
[51] A. I. Lesnikovich, S. V. Levchik, J. Therm. Anal. 1983, 27, 89–93.
[52] H. E. Kissinger, Anal. Chem. 1957, 29, 1702–1706.
[53] C. Popescu, E. Segal, Int. J. Chem. Kinet. 1998, 30, 313–327.
[54] J. Malek, Thermochim. Acta 2000, 355, 239–253.
[32] C. T. Bui, N. J. Maeji, F. Rasoul, A. M. Bray, Tetrahedron Lett. 1999,
40, 5383–5386.
Received: April 21, 2004
Published online: September 23, 2004
5296
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 5285 – 5296