An experimental and theoretical study of reaction mechanisms between nitriles and hydroxylamine

Article abstract of DOI:10.1039/c4ob00854e

The industrially relevant reaction between nitriles and hydroxylamine yielding amidoximes was studied in different molecular solvents and in ionic liquids. In industry, this procedure is carried out on the ton scale in alcohol solutions and the above transformation produces a significant amount of unexpected amide by-product, depending on the nature of the nitrile, which can cause further analytical and purification issues. Although there were earlier attempts to propose mechanisms for this transformation, the real reaction pathway is still under discussion. A new detailed reaction mechanistic explanation, based on theoretical and experimental proof, is given to augment the former mechanisms, which allowed us to find a more efficient, side-product free procedure. Interpreting the theoretical results obtained, it was shown that the application of specific imidazolium, phosphonium and quaternary ammonium based ionic liquids could decrease simultaneously the reaction time while eliminating the amide side-product, leading to the targeted product selectively. This robust and economic procedure now affords a fast, selective amide free synthesis of amidoximes.

Full text of DOI:10.1039/c4ob00854e

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
An experimental and theoretical study of reaction
mechanisms between nitriles and hydroxylamine†
Cite this: Org. Biomol. Chem., 2014,
2, 8036
1
a,b
b
a
a
a
Attila Vörös,* Zoltán Mucsi,* Zoltán Baán, Géza Timári, István Hermecz,
b
a
Péter Mizsey and Zoltán Finta
The industrially relevant reaction between nitriles and hydroxylamine yielding amidoximes was studied in
different molecular solvents and in ionic liquids. In industry, this procedure is carried out on the ton scale
in alcohol solutions and the above transformation produces a significant amount of unexpected amide
by-product, depending on the nature of the nitrile, which can cause further analytical and purification
issues. Although there were earlier attempts to propose mechanisms for this transformation, the real
reaction pathway is still under discussion. A new detailed reaction mechanistic explanation, based on
theoretical and experimental proof, is given to augment the former mechanisms, which allowed us to find
a more efficient, side-product free procedure. Interpreting the theoretical results obtained, it was shown
that the application of specific imidazolium, phosphonium and quaternary ammonium based ionic liquids
could decrease simultaneously the reaction time while eliminating the amide side-product, leading to the
targeted product selectively. This robust and economic procedure now affords a fast, selective amide free
synthesis of amidoximes.
Received 25th April 2014,
Accepted 1st August 2014
DOI: 10.1039/c4ob00854e
www.rsc.org/obc
are frequent substructures of drug substances, and some of
the relevant examples are listed in Scheme S2 in the ESI.†
Introduction
1
–4
Amidoximes (I, Scheme 1) are typical key intermediates for the Besides, amidoximes proved to be a good final intermediate
5
synthesis of numerous heterocycles and are used to introduce for the preparation of amidines (Scheme S1 in the ESI†),
selected important functional groups (Scheme S1†). Such among which numerous examples are known as biologically
6
building blocks can be found in many organic and biologically active molecules such as antimicrobial agents, nitric oxide
7
8,9
active molecules. For example, oxadiazole and oxadiazolone inhibitors and tyrosine kinase inhibitors (Scheme S2 in the
five membered heterocycles are generally prepared from ESI†), whose functional group can be prepared via Pinner reac-
amidoximes by ring closure with carbonyl derivatives. These tion of amidoximes. Note that in this paper, general structures
are labeled with roman numerals, while specific structures are
characterized by Arabic numerals.
Since, most of the time, these amidoxime intermediates (II)
are among the last isolated compounds in the synthesis of
active pharmaceutical ingredients, the selectivity of the amid-
oxime formation, as well as the purity of the prepared amid-
oxime, is crucial. The most common synthetic route to
amidoximes (II) is based on the reaction of a nitrile (I) with
Scheme 1 Reaction of arylnitrile (1) with hydroxylamine to yield the
expected amidoxime (II) and unexpected amide byproduct (III) by two
hydroxylamine. Hydroxylamine is mainly used as a hydrate
Method B in Scheme 1) or in situ released from its salt in a
methods.
(
1
0
typical protic solvent (Method A in Scheme 1). However,
whilst a good yield of the expected amidoxime product (II)
exists, typically a significant amount of amide side product
(III) is formed ranging from a few percent up to 20–30%. Due
to the analogous physical/chemical properties, this amide
side-product (III) is usually the main and critical impurity of
the last isolated intermediate, which has the potential to be a
toxic or genotoxic ingredient, giving enormous workload for
the analytical and synthetic departments during the process of
a
Sanofi, Budapest H-1045 Tó u. 1-5, Hungary. E-mail: attila.voros@sanofi.com
b
Budapest University of Technology and Economics, Department of Chemical and
Environmental Process Engineering, Budafoki út 8., H-1111, Budapest, Hungary.
E-mail: zoltanmucsi@gmail.com
†
Electronic supplementary information (ESI) available: Table S1 contains the
experimental conditions of the HPLC methods, applied. Tables S2–S5 show
the row and calculated kinetic data, measured by HPLC. Table S7 contains the
computed energies (E), zero-point energies (EZPE), internal energies (U) and
enthalpies (H) in hartree at various levels of theories for compounds involved.
See DOI: 10.1039/c4ob00854e
8036 | Org. Biomol. Chem., 2014, 12, 8036–8047
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
drug development. From that point of view, an amide free syn-
Comparing the two aprotic solvents Me-THF and NMP
thesis could be a more economical and shorter process, thus applied, a significant increase can be observed in NMP having
improving the efficacy. a larger relative permittivity, and a low conversion in apolar
In this study, we aimed to understand the mechanism of Me-THF (ε = 6.97). The combination of a protic solvent (EtOH,
formation of the amide side product (III) and develop a selec- n-PrOH and H O) and an organic base, such as Et N, provided
2
3
tive amidoxime (II) synthesis from nitriles (I) by means of generally a better conversion than the combination of a protic
hydroxylamine. Finally, the novel reaction mechanism is solvent and Na CO . Due to the enhanced solubility of Na CO
checked and proved by experimental means. According to the in water (ε = 80.1), similarly improved conversion was achieved
2
3
2
3
1
1
3
earlier proposed mechanism, hydroxylamine may acts as an by inorganic bases analogously to Et N in alcohols. The
ambient nucleophile by its N or O sides in the course of the amount of base was found to be a crucial parameter from the
reaction and we aimed to differentiate this dual reactivity by selectivity point of view. In general, a slightly decreased conver-
the quality of the solvent applied. A marked influence was sion and an increased amount of amide by-product were
1
2,13
expected from ionic liquids.
Ionic liquids have melting observed with increasing the base from one to two eq.
points below 100 °C, low vapor pressure and inflammability (Table 1).
14–19
contrary to conventional volatile organic molecular solvents.
An interesting solvent effect can be observed on the conver-
Their physical properties, such as melting point, viscosity, and sion (30 min at 50 °C) and amide formation for different types
density, are limited generally by the organic anion, controlling of alcohols, such as n-PrOH or EtOH (primary; entries 11 and
2
0,21
the chemical reactivity of the solute substantially.
18), i-PrOH (secondary; entry 13) and t-BuOH (tertiary; entry
4), in the presence of 1 eq. TEA. In that series, the largest con-
1
versions were detected for primary alcohols (81% in n-PrOH,
76% in EtOH), while the conversion in secondary alcohol
proved to be moderate under the same conditions (44% in i-
PrOH). Tertiary alcohol exhibited only low conversion at
Results
Amidoxime type molecules (II) can be prepared on a large
scale under various conditions, in most of the cases in hetero-
geneous reaction mixture; however, recently, concentrated
aqueous hydroxylamine (50%) was applied successfully under
30 min (22% in t-BuOH), which is close to the value observed
in the aprotic Me-THF (4%). As is known, the HB donor and
acceptor ability of the alcohols is strongly influenced by their
orders, the OH group of primary alcohols is exposed, while the
OH groups of tertiary alcohols are sterically hindered. This
series of values shows that HB ability plays an important role
in the mechanism, due to the fact that the sterically more hin-
dered the OH group in the alcohol, the less the conversion
that occurs.
2
2
homogeneous conditions in a continuous micro reactor.
Here, we aimed to study the heterogeneous reaction of
nitrile derivatives with hydroxylamine hydrochloride in
different molecular and ionic solvents. The reaction mechan-
ism was investigated and compared to proposals of Stephen-
1
1,23
son et al. and Srivastava et al.
2
,2,2-Trifluoroethanol (TFE, entry 20 in Table 1) and
Experimental findings and conversion obtained in molecular
solvents
1
,1,1,3,3,3-hexafluoro-2-propanol (HFIP, entry 15 in Table 1),
two extremely strong hydrogen bonding (HB) solvents, some-
what decreased the conversion; however, the percentage of the
amide by-product remained. This effect can be explained by
the much stronger HB effect between the OH group of the
The preparation of benzamidoxime (BAO, 2) was investigated
from benzonitrile (BN, 1) in the presence of two equivalents of
hydroxylamine hydrochloride in various molecular solvents
(
Scheme 2) such as aprotic N-methyl-pyrrolidone (NMP;
entries 1–4) and 2-methyltetrahydrofuran (Me-THF; entries
–8) as well as different types of alcohols (entries 9–20) and
2
solvent molecule and the nitrogen of NH OH, decreasing its
nucleophilicity. In the light of the results of the different alco-
hols, it can be concluded that mostly the HB acceptor ability
influences the reaction in the beneficial direction, instead of
the HB donor ability. This observation is confirmed by other
HB donor solvents, such as NMP, which gave good conversion,
and the sterically hindered ether type MeTHF providing poor
conversion. Water represents a compromise between the good
HB acceptor and donor abilities, providing moderate conver-
sion. Temperature has also a significant effect on the conver-
sion; at higher temperatures larger conversions were observed
than at lower temperatures for the same reaction time.
5
water (entries 1–4). The reaction temperature was set to 50 °C
and 80 °C. In the course of the reactions, one or two equi-
valents of Na CO or Et N were used as the base (Table 1). The
2 3 3
conversions observed were typically high and reaction rates
were nearly the same in dipolar aprotic NMP (ε = 32.2) and in
protic solvents, such as propanol (ε = 20.33) or ethanol
(
ε = 24.55) in the presence of Et N as the base.
3
Experimental findings and conversion obtained in ionic
liquid solvents
In our study, as shown in Table 2, two different types of ionic
liquids (ILs) were applied, depending on the type of the anion.
Ionic liquids, ranked in the first group (entries 25–26 in
Scheme 2 Reaction of aromatic nitrile derivatives (1a–m; BN) with
hydroxylamine hydrochloride providing aromatic amidoxime (2a–m;
BAO) and aromatic amide (3a–m; BA).
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8036–8047 | 8037

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 1 Reaction composition (%) of unsubstituted benzonitrile (BN; R = H; 1a) to benzamidoxime (BAO, 2a) and benzamide (BN, 3a) by
NH
2
OH·HCl in different organic solutions in the presence of inorganic (Na
2
CO
3
) or organic bases (Et
3
N), in 1 or 2 eq., using 30 min reaction time at
80 °C as well as 50 °C reaction temperature
80 °C
50 °C
1a, BN
2a, BAO
3a, BA
1a, BN
2a, BAO
3a, BA
Entry
Base
Base eq.
a
1
2
3
4
5
6
7
8
9
NMP
Na
Na
2
CO
CO
N
3
1
2
1
2
1
2
1
2
1
2
1
2
1
1
1
1
2
1
2
1
1
2
1
2
80
71
0
16
24
96
96
1
1
8
17
23
36
96
95
—
—
—
75
70
94
90
—
82
82
95
88
4
5
4
4
1
1
2
3
3
4
3
5
—
—
—
3
3
3
87
—
11
—
100
—
95
—
94
—
16
—
54
76
66
70
—
23
—
38
42
—
42
—
12
—
87
—
0
—
4
1
—
2
—
0
—
1
—
1
—
3
—
2
2
1
1
—
1
—
2
2
—
2
a
NMP
2
3
a
NMP
Et
Et
3
a
NMP
3
N
0
b
b
b
b
Me-THF
Me-THF
Me-THF
Me-THF
n-PrOH
n-PrOH
n-PrOH
n-PrOH
i-PrOH
t-BuOH
Na
Na
2
CO
CO
3
98
98
90
80
74
60
1
2
3
Et
Et
3
N
N
3
—
5
Na
Na
2
CO
CO
3
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
0
2
3
—
81
—
44
22
33
29
—
76
—
60
56
—
56
—
1
2
3
4
5
6
7
8
9
0
1
2
3
4
Et
Et
Et
Et
Et
3
N
3
3
3
3
N
N
N
N
0
—
—
—
22
27
3
c
HFIPA
EtOH
EtOH
EtOH
EtOH
Na
Na
2
CO
CO
3
2
3
Et
Et
Et
3
N
N
N
3
3
4
6
—
3
5
2
d
TFE
—
15
13
3
H
2
H
2
H
2
H
2
O
O
O
O
Na
Na
2
CO
CO
3
2
3
Et
Et
3
N
N
3
7
5
—
a
b
c
d
N-Methyl-pyrrolidone. 2-Methyl-tetrahydrofuran. 1,1,1,3,3,3-Hexafluoro-2-propanol. 2,2,2-Trifluoroethanol.
2
4
Table 2), possess anions, where their conjugate acids are involved as the positively charged counterpart of the IL, such
strong (pK < 2); consequently these anions are weak bases, as imidazolium, phosphonium and ammonium. In general,
a
−
−
−
such as trifluoromethanesulfonate (TFMSI), Cl , AlCl
4
, BF
4
,
independently from the cations applied, low conversion was
PF6 , alkylSO4−, OTs , and trifluoroacetate (TFA). The second achieved (0–20%) in ionic liquids belonging to the first group;
−
−
group (entries 27–29 in Table 2) contains ILs, where the conju- only TFA (entry 18 in Table S2 in the ESI†) showed a moderate
a
gate acids are weak (pK > 2); therefore the anion is a stronger transformation (40%). This inactivity remained in the course
−
−
base, such as carboxylates (OAc , OBz , lactate). In these of the variation of various reaction parameters, such as reac-
cases, a real equilibrium exists between the cation–anion pair tion time, temperature, excess and type of base. However, the
and their neutral counterparts. Many different cations were second group of ILs provided high conversion rates (99–100%)
Table 2 Conversion to unsubstituted benzamidoxime (BAO, 2a) in different ionic liquid solutions from benzonitrile (1a) by NH
ence of inorganic (Na CO ) or organic bases (Et N) in 1 or 2 eq. with 30 min reaction time at 80 °C as well as 50 °C reaction temperature. 30 min
reaction time in different ionic liquids with 1 eq. Na
2
OH·HCl in the pres-
2
3
3
2
CO
3
at 80 °C
80 °C
50 °C
1. eq.
1
. eq.
2. eq.
Ionic
liquids (IL)
a
Entry
1a, BN
2a, BAO
3a, BA
1a, BN
2a, BAO
3a, BA
1a, BN
2a, BAO
3a, BA
25
26
27
28
29
[BMIM][Cl]
[MOIM][BF
[BMIM][OAc]
[EMIM][OAc]
[Ch][OAc]
81
96
0
0
0
13
2
99
100
100
6
2
1
0
0
—
—
0
0
0
—
—
99
99
98
—
—
1
1
2
—
—
1
3
2
—
—
98
96
97
—
—
1
1
1
4
]
b
a
b
Abbreviations listed in the Experimental section. The reaction took place with 0 eq. of the base (99%, BAO).
8038 | Org. Biomol. Chem., 2014, 12, 8036–8047
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
of the expected major product, within a short reaction time at Table 3 The calculated initial quasi first order reaction rates of the
product BAO formation in various solvents at 50 °C
8
0 °C. More importantly, the formation of the amide by-
product was very limited in all these cases.
2
Solvent
k′(initial) [1/s]
Error [1/s]
R
2 3
Both the organic (TEA) and inorganic (Na CO ) bases
exhibited the same high yields and fast reaction rates. An MeTHF
0.0066
0.2028
0.1126
0.0502
0.1993
0.0026
0.0170
0.0190
0.0123
0.0298
0.756
0.966
0.895
0.796
0.768
n-PrOH
excess of base had also no impact on the results observed. It is
i-PrOH
t-BuOH
important to highlight that the reaction took place in the
2
4
absence of external bases using [BMIM][OAc] as the solvent, [BMIM][OAc]
which can be explained by the presence of anions that behave
as a strong base. Papers of Nyulászi et al. and R. D. Rogers
2
5,26
initial reaction rates of the product BAO [k′(initial)], which
et al. proved that,
when an imidazolium IL has a carboxy-
were estimated from the very beginning of the reaction
late counterion, the reaction mixture is in equilibrium with
the corresponding N-heterocyclic carbene (NHC), which can
(
Table 3). In agreement with the earlier results, the aprotic
MeTHF shows the lowest reaction rate, in contrast to the protic
alcohols and [BMIM][OAc]. The highest and nearly the same
reaction rates are exhibited by n-PrOH and [BMIM][OAc], a
lower rate was calculated for i-PrOH, while almost the quarter
value for the t-BuOH. The earlier conclusion also confirmed
that the HB donor ability of the solvent applied influenced sig-
nificantly the reaction rates. It was determined quantitatively
that the formation of BA side-product is also more preferred in
fewer HB donor solvents. In secondary and tertiary alcohols
act as a base or nucleophile, releasing the NH OH base from
its salt.
2
The influence of temperature was also investigated at two
points (80 and 50 °C) and no significant effect was found
either on conversion or yield. From this result, it can be seen
that the cation has also no influence on the outcome of the
reaction; in all the cases we observed high selectivity toward
the amidoxime and a fast reaction rate. Contrary to molecular
solvents, ILs always showed a higher reaction rate and better
selectivity, independent of the amount of base, reaction temp-
erature and concentration applied. This highly robust pro-
cedure offers a new possibility of amide free amidoxime
synthesis.
(
e.g. t-BuOH) and MeTHF, 10.4% and 21.6% of BA/BAO
product ratios were observed, respectively, in percentages
Fig. 1C, upper line). IL proved to be the best choice, with the
(
fastest reaction rate and the lowest BA concentration.
Reaction kinetics
Substituent effect
Up to this point, only the 30 min conversion was studied in In the next part, the influence of various ortho and para-phenyl
various solvents. However, it is worth examining the kinetics substituents of 1 was studied on the reaction rate and the
of the reaction in different media (Fig. 1). In this experiment amide concentration, in a protic solvent (ethanol) alongside
3
the same reaction conditions were applied (50 °C; 2 eq. of the best IL [BMIM][OAc], with Et N base at 50 °C reaction
NH OH·HCl; 2 eq. of TEA) as those used previously in solvent temperature (Table S2, in the ESI†). The electron withdrawing
2
screening. The relative concentrations of the reaction com- (EW) and electron donating (ED) properties of the phenyl sub-
ponents were measured by HPLC (at 210 nm) using small but stituents were observed to modify the reaction rate and the
precisely added toluene as an internal standard. Due to the product distribution to a large extent. In the case of a para
excess of the hydroxylamine reactant, the quasi first order EW group, such as halogens (4-Cl 1b; 4-F 1c) and nitro
2
assumption can be a good approximation to compare the groups (4-NO 1d), increased yields were observed compared to
Fig. 1 (A) The reaction kinetics for the transformation of BN to the BAO product and BA side-product in five selected solvents at 50 °C in the pres-
ence of 2 eq. of NH OH·HCl; 2 eq. of TEA, given in relative concentration. (B) Determination of the relative reaction rate by the linear fitting on the
2
relating data points. (C) The segment of the reaction in the light of the distribution of components at 30 min reaction time. The upper line shows the
BA/BAO ratio in percentage for the solvents studied.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8036–8047 | 8039

View Article Online
Paper
Organic & Biomolecular Chemistry
unsubstituted benzonitrile, while ED groups slowed down the
reaction, thus decreasing the conversion to the product. The
product rate is also slightly varied by the quality of the para
substituent; in the case of the EW group the amide rate is
lower, while for the ED group it is higher (entries 2–6;
Table S2†). In the case of ED groups, ILs could increase the
yield of the reaction significantly compared to ethanol as the
solvent, whilst keeping the amide concentration low.
Lower conversions were reached by the use of ortho substi-
tuted derivatives (entries 7–12 in Table S2†) in both solvent
systems, presumably due to the steric hindrance. This steric
effect can be verified by 2-pyridinecarbonitrile (entry 14 in
Table S2;† 1l), whereas the reaction behaves similarly to the Scheme 3 Schematic reaction mechanism for (ROUTE A–D) the for-
mation of benzamidoxime (BAO, 2) and benzamide (BA, 3) formation
from benzonitrile (BN, 1) and NH OH.
2
para-substituted derivatives (1b–d), giving fast and selectively
the amidoxime derivative. However, for all these ortho substi-
tuted compounds, ILs were able to provide better conversions
and a lower amide formation rate.
The addition of water was not able to initiate the amide for-
mation, even in trace amounts (entry 5 in Table S5†). This
created a non-reactive, or in other words, apparent stability of
the nitrile proving that the hydrolysis of the BN does not
provide the amide by-product. An interesting experiment was
carried out between BN and NH OH in the presence of one
2
equivalent of NaH. In that case, the OH side of the hydroxyl-
amine deprotonated exclusively, due to its larger acidity and
The most extreme example proved to be 2-nitrobenzamid-
oxime (entry 8 in Table S2;† 2h), which provided very poor con-
version (51%) with the highest amide product rate (17 : 51) in
ethanol, but it could be prepared effectively (conv. = 99%) and
selectively (1 : 99) within an IL. ED groups in the ortho position
decreased further the reaction rate.
One of the lowest conversions could be observed for the
2
,6-disubstituted derivative (entry 15 in Table S2;† 1m), where
was expected to produce the reactive intermediate, BCO.
the accumulated influences of the unbeneficial steric and elec-
tronic effects inhibit the product formation, which also
doubled by using an IL. In all the cases, when conversion was
low in IL, somewhat better conversion and selectivity was
−
However, when the H
2
NO species reacts with BN (1), it
resulted that only BA and not BAO or BCO was detected,
strongly confirming the hypothesis that the origin of BA is
coming from the O attack of the hydroxylamine yielding BCO
and the consequent reaction with an additional hydroxyl-
amine. In the second series (entries 7–10 in Table S5†), the
possible transformation of the BAO product was examined
under several conditions.
3
reached with an increased amount of base (2. eq. Et N;
Table S2†); however, it was not a crucial parameter in the case
of high yields. In two cases, where the conversions were quite
low (entries 10 and 12, Table S2;† 1i and 1j) in both solvents, a
longer reaction time increased the yields somewhat.
In the third part, the simple refluxing of BAO in the
absence or in the presence of Na CO (entries 7 and 8 in
Variation of reaction conditions
2
3
Table S5†) did not result any BA even in trace amounts and the
addition of water did not initiate it (entry 9 in Table S5†).
Finally, applying the standard reaction conditions with BAO
(entry 10 in Table S5†), no transformation of any kind was
observed within hours. These simple experiments undoubtedly
exclude any alternative source of BA; consequently, it can be
concluded that BA can form exclusively from the reaction of
We examined most combinations of the reaction conditions
applied in order to find a way of forming BA. BA can form
directly from BN under basic or aqueous conditions by the
−
attraction of the OH anion. Theoretically, BAO can be hydro-
lyzed analogously by a strong base. Finally, NH OH can react
2
with the BN in two ways due to its ambient nucleophilic char-
acter on the N as well as O atoms, which can lead to the
expected BAO and a new N-amino benzenecarboximidate
2
BN and NH OH.
(
BCO; 4) intermediate. This unstable BCO intermediate was
1
5
N Isotope experiment
Two isotope experiments were planned to clarify the origin of
In the first part, the starting compound (BN) was treated the N atom in BA. In the first experiment, BN was treated with
supposed earlier to be able to react with a new NH OH, fur-
nishing the BA side-product.
2
1
1
1
5
under various conditions. The standard reaction conditions isotope labeled NH
2
OH·HCl under standard reaction con-
(entry 1 in Table S5 in the ESI†) provide the expected product ditions in EtOH and in the presence of TEA (Scheme 4), and
distribution. When NH
2
OH·H
2
O was applied in excess (entry 2 the reaction was followed by HPLC-MS. As expected, the
1
5
+
in Table S5†), besides the fast reaction rate, the amide ratio isotope labeled BAO* formed including one N atom (M + H
was significantly higher than that for the salt form.
1
4
= 137.1 Da); however, the BA formed also contained only
N
+
Secondly, the simple reflux of the BN in ethanol in the (M + H = 121.1 Da). For the reverse experiment, isotope
absence or in the presence of Na
Table S5†) did not yield BA (Scheme 3).
2
CO
3
(entries 3 and 4 in labeled BN* was prepared from a commercially available
isotope labeled BA* by propylphosphonic anhydride (T3P) in
8040 | Org. Biomol. Chem., 2014, 12, 8036–8047
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
2
7–32
acids or bases.
All these components have to be involved
in the calculation in order not only to find the real mechan-
ism, but to access the correct and accurate energy values
taking place in a real medium. Furthermore, an inappropriate
chemical model provides not only inaccurate energy values,
but typically could lead to completely wrong or opposite
2
7
results. Reactions, taking place in solution, usually require
the consideration of a catalyst together with many solvent
2
7,31,32
molecules in an appropriate 3D arrangement.
For the
sake of simplicity and easier understanding in the discussion,
numbers indicate the molecular complexes (reagent, reactant,
solvent, additives) in different arrangements and combi-
nations, while to identify the chemical species we introduced
some letter codes.
1
5
Scheme 4 Experimental N isotope labeling to prove the origin of the
N atom in the reaction of BN with hydroxylamine to BAO and BA. The
1
5
stars show the N isotope labeled compounds.
EtOAc at RT and purified carefully from the remaining starting
Thermodynamical aspects. The benzonitrile (BN; 1) can
material. In the next step, it was reacted with NH OH·HCl react with the NH OH in two ways. When the N-side attacks
2
2
under standard reaction conditions, resulting in the alterna- the nitrile carbon (ROUTE A), the expected benzamidoxime
+
15
tively isotope labeled BAO** (M + H = 137.1 Da) and also
N
(BAO; 2) forms, but the O-side attack (ROUTE B) results in the
+
labeled BA* (M + H = 122.1 Da). According to these two experi- N-amino benzenecarboximidate (BCO; 4) intermediate
ments, one can conclude undoubtedly that the origin of the N (Scheme 6). From a purely thermodynamic aspect, neglecting
atom of the BA is the N atom of the BN and not the NH
2
OH, the explicit solvent effect, both routes show significant exother-
supporting the mechanism proposed previously (Scheme 4).
micity and the formation of BAO (3) is preferred to BCO (4).
1
1,23
According to earlier assumptions
findings, the origin of BA can be attributed to the reaction of
Finally, the mechanism of the amide by-product was checked the formed BAO with an additional NH OH and may result in
experimentally; therefore BN was reacted with the Me
NOH two product states (5 and 6) involving two possible side pro-
reagent, which is able to react only by the O side of the mole- ducts H N–NH–OH (NNO) via ROUTE C or H N–O–NH (NON)
and our experimental
2
Reaction with Me NOH
2
2
2
2
2
cule, forcing the formation exclusively of the dimethyl deriva- via ROUTE D. In both cases, the formation of BA from BN exhi-
tive of BCO (BCOM), as shown in Scheme 5. In that bits large exothermicity (Table 4), which means a strong prefer-
experiment, the formation of BA was observed exclusively, but ence toward this product. The significant difference between
with much lower conversion (9% for BA and 91% for BN after the two end states is attributed exclusively to the different
2
h) and consequently slower reaction rates, compared to the heats of formation of the NNO and NON.
NH OH reagent. This significant difference in reaction rates
Reaction mechanism in a molecular solvent (alcohol). To
2
can be explained by the suggested mechanism (see Theoretical model the real mechanism, the combined explicit–implicit
results, Scheme 11), where the BCOM intermediate is able to solvent model was applied in order to obtain accurate values.
react with only the O side of the Me NOH, yielding BA and However, different solvents require various solvent models.
2
3
Me NONMe . However, as computational results show later, The protic media with an organic base (Et N) was modeled by
2
2
+
the O-side attack of the hydroxylamine exhibits a much higher one EtOH solvent molecule, one Et
N and one Et NH , in the
3
3
activation enthalpy than the N-side attack, supporting the arrangement shown in Scheme 7. The best solvent model for
much slower reaction rate observed.
acetate counterion containing ionic liquids was represented by
an acetate anion (AcO ) and one acetic acid (AcOH) molecule,
−
Theoretical results
shown in Scheme 10 in the next paragraph. In both cases,
the relative permittivity (εrel) was set to 30.0 to consider the
It is a typical situation, in most of the cases the real mechan-
ism of an experimentally simple chemical reaction is very
complex, involving many species in each elementary step, like
reactants, reagents, solvent molecules, catalysts, as well as
Scheme 6 Schematic reaction mechanism for (ROUTES A–D) the for-
mation of benzamidoxime (BAO, 2) and benzamide (BA, 3) formation
from benzonitrile (BN, 1) and NH OH.
2
Scheme 5 Formation of benzamide (BA) from benzonitrile (BN) by
means of N,N-dimethylhydroxylamine through the BCOM intermediate.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8036–8047 | 8041

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 4 Computed thermodynamic parameters for reaction ROUTES
A, B, C and D, shown in Scheme 6
Computed parameters
Process
Reactions
Δ
r
H°
Δ
r
G°
r
Δ S°
1
1
1
1
→ 2 (ROUTE A) BN + NH
→ 4 (ROUTE B) BN + NH
→ 5 (ROUTE C) BN + 2 NH
→ 6 (ROUTE D) BN + 2 NH
2
OH → BAO
OH → BCO
−78.6 −25.8 −177.3
−44.9 6.0 −170.8
2
2
OH → BA + NON −130.5 −79.7 −170.3
OH → BA + NNO −203.3 −150.6 −176.8
2
Scheme 8 Possible interconversion (ROUTE J) of the solvated benz-
amidoxime (BAO; 12) and BCO (15) via the formation of an oxaziridine
ring (14), including their enthalpy values in kJ mol−
1
.
2
and an additional NH OH; however, it was not confirmed by
any computational or experimental study. In a recent paper,
this hypothesis was checked by theoretical methods. For the
modeling of the situation occurring in the protic solvent, an
implicit–explicit solvent model was applied; however, a new
solvent network should be introduced involving BCO, two
Scheme 7 Novel, direct solvation mechanism for the reaction of benzo-
nitrile (BN, 1) to produce the expected benzamidoxime product (BAO; 12)
and the key intermediate (BCO, 13) for the benzamide by-product in
EtOH and one NH
2
OH in a proper arrangement (13). It was a
OH attacks the
NH group of BCO (ROUTE L). Stephenson and colleagues pro-
protic media (ROUTES H and I), including their enthalpy values in debate in the literature which side of the NH
2
−1
kJ mol
.
2
1
1
posed the attack of the nitrogen atom of NH
2 2
OH on the NH
group of BCO with the formation of BA and NNO as side pro-
2
3
continuum media as an implicit solvent model. The necessity ducts (ROUTE K). In contrast, a novel paper assumed an
of the hydrogen bond solvent network was proved experi- opposite attack of the NH OH by its O side, resulting in BA
2
mentally as well; when the aprotic Me-THF was used as and NON; however, they did not study and prove this process
a solvent only marginal conversion of BN (1) was observed by any calculation of the corresponding TSs.
(
ca. 7%) in the presence of NH
2
OH, irrespective of the base
Consequently, both possible ROUTE K and L were con-
applied.
sidered in our theoretical study. In the first case, the attacking
Using the defined solvent model to model the protic media NH OH approaches by its O side (ROUTE K) the NH group of
2
2
(Scheme 7; 11), we could obtain both products (12 and 13) BCO in an appropriate solvent complex (15), resulting directly
smoothly in one elementary step, without any fraction of the in the lower energy complex of BA and NON as the product
potential energy surface, through reaction coordinates. In con- state (16) through one well defined TS [TS(15→16)], exhibiting
trast with the thermodynamic data, the transition state [TS- a low 86.3 kJ mol− activation enthalpy calculated from the
1
(
11→12) and TS(11→13)] toward BAO (12) is larger by 12.2 kJ BCO level. When NH OH attacks by its N side (ROUTE L), an
2
−
1
mol than that for BCO (13) [TS(11→13)], which shows an analogue TS [TS(15→17)] leads to another product state (17)
unquestioned kinetic preference of the side product.
involving NNO and BA in one elementary step; however, the
Earlier, the literature supposed a possible direct equili- enthalpy of the TS was found to be lower (63.9 kJ mol ). In
brium reaction between BAO and BCO, which can be an both cases, the two EtOH molecules, forming a solvent
alternative possibility for the formation of the side product network, mediate the proton transfer in the course of the reac-
−1
BCO. This process involves one low energy cyclic intermediate tion from NH OH to the product. According to these activation
2
(
14) and two TSs [TS(12→14); TS(14→13)], which could be gaps, obtained in protic media, we confirm Stephenson’s
modeled using the same solvent network. However, the first theory, namely the N attack via ROUTE L (Scheme 9).
activation energy was found to be too high to be considered a
Reaction mechanism in an ionic liquid (IL). From the
real transformation of BCO; consequently it was skipped from experimental results, we concluded that irrespective of the
the mechanistic study (Scheme 8).
cation, only the carboxylate-type ionic solvent could provide
Earlier and our recent experimental studies supposed that the amidoxime product (BAO); consequently only this type of
the amide side product BA forms by the reaction between BCO ionic solvent was selected for theoretical study. In order to
8042 | Org. Biomol. Chem., 2014, 12, 8036–8047
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 9 Details of finishing reaction steps for two alternative
mechanisms (ROUTES K and L) of BA from BCO in EtOH solvents,
including their enthalpy values in kJ mol−
1
.
Scheme 11 Three possible mechanisms (ROUTES O–Q) of the trans-
model properly the environment of a carboxylic-type ionic formation of BCO to BA via O and N attack of hydroxylamine, including
liquid, one acetate and one acetic acid were considered, their enthalpy values in kJ mol−
1
.
supposing that the acid–base equilibrium between the cation
or water and the acetate anion can provide a low concentration
of acetic acid. Using the solvent–reagent–reactant structure in
theoretical method. In these cases, no explicit solvent was con-
sidered; only the implicit shell was included. ROUTE O and P
model the O and N attack of hydroxylamine, respectively,
toward the N atom of the BCO.
Analogously to the mechanism in molecular solvents, the
N-side attack of the hydroxylamine is more preferred from
both the kinetic and thermodynamic aspects than the O-side
attack (Scheme 11), due to the lower TS [TS(21→22) vs. TS-
the arrangement (18) as shown in Scheme 10, the two known
reaction routes toward BAO [18→TS(18→19)→19; ROUTE M]
and BCO [18→TS(18→20)→20; ROUTE N] were modeled
successfully.
In the ionic solvent model, also the activation barrier of the
−
1
ROUTE N proved to be lower by 5 kJ mol , preferring the for-
mation of BCO in contrast to BAO, showing analogy to the
result obtained in protic media. As was concluded in the pre-
vious section, the formation of BA starts from the BCO inter-
(
21→23)] and the product state (22 vs. 23). An alternative route
(ROUTE Q) was also considered and computed, where the zwitter-
2
mediate by an additional NH OH attack.
Three possible attacks (Scheme 11) of the hydroxylamine
were considered (ROUTE O–Q) toward BCO and computed by a
ionic form of NH OH attacks by its O atom the BCO inter-
2
mediate. However, here, not the usual BA forms in the course
of the reaction, but a very unusual “dead-end” product 27, by
3
the elimination of NH (ROUTE Q). The activation enthalpy of
the process TS(24→25) is higher than the previous two,
making it more unlikely (Fig. 2).
Discussion
In this section the theoretical and experimental results
obtained for the two types of solvents are analyzed and com-
pared with each other. Fig. 3 shows the simplified mecha-
nisms, where only the allowed routes are summarized. In both
types of solvents, the expected product, BAO (2), forms in one
elementary step from BN (1) and NH OH with one TS, while
2
the mechanism toward the by-product, BA (3), involves one
intermediate BCO between two TSs. In a molecular solvent
(
such as EtOH), the rate determining step for BA is the reaction
of BA with NH OH [TS(11→13)], which is somewhat lower by
2.0 kJ mol− (orange arrow) than that of the competitive route
of BAO [TS(11→12)]. The TS of the step of BCO to BA
TS(15→17)] is significantly lower. These data undoubtedly
2
1
1
Scheme 10 Direct solvation mechanism (ROUTES M and N) for the
reaction of benzonitrile (BN, 1) to produce the expected benzamidoxime
product (BAO; 19) and the key intermediate (BCO, 20) for the benzamide
by-product in carboxylic-type ionic liquid (IL), including their enthalpy
[
show that the formation of BA is preferred a little bit against BAO,
which means parallel formation of both products. This slight
−1
values in kJ mol
.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8036–8047 | 8043

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 2 (A) Direct solvation mechanism enthalpy levels for the various ROUTES G–L along the reaction coordinates in molecular solvent (EtOH) from
−1
benzonitrile (BN, 1) to benzamidoxime (BAO, 2), benzamide (BA, 3) via BCO (4), including their enthalpy values in kJ mol . (B) Enthalpy levels in
−1
kJ mol for the various transformations of benzonitrile (BN, 2) to benzamidoxime (BAO, 2) and benzamide (BA, 3) via BCO (3) along the reaction
coordinates in acetate containing ionic liquid (ROUTES G; M–Q).
Conclusions
For several decades the reaction of hydroxylamine with electro-
philes appeared to be rather confusing due to the O and N
atoms possessing sharply different nucleophilicities. This dual
activity of the reagent was attributed to the significant for-
mation of the unexpected benzamide (BA) side-product in the
course of the reaction of hydroxylamine and benzonitrile (BN),
producing benzamideoxime (BAO) as the expected product. In
this paper the detailed mechanism was explored and proved
by experimental and theoretical means. As a consequence of
the results of the new reaction mechanism, it was shown that
an ionic liquid, composed of a carboxylic type anion, like
Fig. 3 Comparison of the enthalpy levels for the various transform-
[BMIM][OAc], is able to eliminate the BA side-product comple-
ations of benzonitrile (BN, 1) to benzamideoxime (BAO, 2) and benz-
amide (BA, 3) via BCO along the reaction coordinates ROUTES G; M–R
in acetate containing ionic liquid.
tely, in contrast with the frequently used protic-type solvent,
where the concentration of the BA can be up to 40%, depend-
ing on the circumstances (Scheme 12).
Theoretical computations predicted that the two hetero-
atoms in hydroxylamine, exhibiting different characters, can
react with benzonitrile resulting in two different products
benzamidoxime (BAO) and O-(imino(phenyl)methyl)hydroxyl-
amine (BCO). The latter proved to be a reactive intermediate
that was never isolated even though numerous attempts were
made. Due to the reactivity of BCO, it reacted readily with a
preference towards BA, which is in contrast with the experi-
mental results, may be attributed to the small error of the
calculation.
In IL, the situation differs significantly. In this case, the
rate determining step for the BA is the last BCO→BA trans-
formation [TS(21→23)], which controls the rate of formation of
−1
BA, and this value is higher by 16.3 kJ mol compared to the
TS of the transformation of BN→BAO [TS(18→19), blue arrow].
It means that the preferred product in IL is the BAO, and the
formation of BA is hindered, due to the last step. It means that
the formed BCO via TS(18→20) is unable to transform forward
to BA and the process turns back from this intermediate due
to the high TS(21→23) and goes back to BN.
Comparing the two solvents applied, it can be concluded
that the TS of BN→BAO is significantly lower in IL than in
molecular solvent (green arrow), while the TS of BN→BCO
exhibits very similar values (green arrow). In contrast with
that, the TS of BCO→BA is higher in IL than in molecular
solvent (red arrow).
Scheme 12 Summary of the transformation of benzonitrile (BN) to
benzamideoxime (BAO) and its side-product benzamide (BA).
8044 | Org. Biomol. Chem., 2014, 12, 8036–8047
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
further hydroxylamine by its O side, resulting at the first sight
51 µl 1a benzonitrile (51.56 mg, 0.5 mmol) and 1 eq.
the strange benzamide side-product. According to compu- Na CO (0.5 mmol, 0.053 g) were heated in 0.5 ml EtOH and
2
3
tations in protic media, the transformation of BCO exhibits 0.25 ml H
lower TS than the TS of the formation of the BAO product; HPLC (method no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
however, the order of TSs reverses in ionic liquids; conse- 51 µl 1a benzonitrile (51.56 mg, 0.5 mmol) was heated
quently the amide formation is hindered. N isotope labeling in 0.5 ml H O at 50 °C. The reaction was stirred and checked
experiments were designed to prove the validity of the compu- by HPLC (method no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12,
2
O at 50 °C. The reaction was stirred and checked by
1
5
2
ter modeling mechanism.
3a: 8.31).
1 µl 1a benzonitrile (51.56 mg, 0.5 mmol) and 1 eq.
CO (0.5 mmol, 0.053 g) were heated in 0.5 ml H O at
0 °C. The reaction was stirred and checked by HPLC (method
no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
.0695 g (1 mmol) HA·HCl was stirred in 10 ml ACN at RT
and 8 eq. NaH 60% dispersion of mineral oil (4.0 mmol,
.16 g) was added to the mixture. After half an hour 51 µl 1a
5
Na
5
2
3
2
Experimental section
0
All chemicals were purchased from Sigma-Aldrich except for
2
,6-dimethylbenzonitrile, which was bought from Alfa Aesar
0
and 4-chlorobenzonitrile which was bought from Merck.
Ionic liquids were purchased from Fluka except for [EMIM]-
benzonitrile (51.56 mg, 0.5 mmol) was added to the solution.
The reaction was checked by HPLC (method no. 3, 225 nm,
RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
[OAc], [Ch][OAc], and [BMIM][TFA], which were bought from
Sigma-Aldrich and Ammoeng 100 and [BMIM][BF
were bought from Solvent Innovation.
All the reactions were tracked using appropriate HPLC
methods on Agilent 1100 equipment, the column temperature
was 30 °C, and the flow rate was 1.0 ml min in each of the
five methods (see ESI†).
4
], which
68 mg 2a benzamidoxime (0.5 mmol) was heated in 0.5 ml
EtOH at 50 °C. The reaction was stirred and checked by HPLC
(
(
method no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
68 mg 2a benzamidoxime (0.5 mmol) and 1 eq. Na CO
3
0.5 mmol, 0.053 g) was heated in 0.5 ml EtOH at 50 °C. The
−1
2
reaction was stirred and checked by HPLC (method no. 3,
High-resolution MS spectra were measured using an Agilent
225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
6
230 TOF LC/MS spectrometer.
1
13
68 mg 2a benzamidoxime (0.5 mmol) was heated in 0.5 ml
H and C NMR spectra were obtained at room tempera-
EtOH and 0.25 ml H
checked by HPLC (method no. 3, 225 nm, RT: 1a: 10.6, 2a:
.12, 3a: 8.31).
8 mg 2a benzamidoxime (0.5 mmol) and 1 eq. Na
0.5 mmol, 0.053 g) were heated in 0.5 ml EtOH and 0.25 ml
O at 50 °C. The reaction was stirred and checked by HPLC
method no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
8 mg 2a benzamidoxime (0.5 mmol) and 1 eq. Et N
2
O at 50 °C. The reaction was stirred and
ture as solutions in DMSO using tetramethylsilane as an
internal standard. The spectra were recorded using a Bruker
Avance 500 MHz spectrometer.
8
6
2 3
CO
(
General procedure for the preparation of benzamidoxime
derivatives (2a–m)
H
2
(
0
1
2
0
.5 mmol 1a (1 eq.) and 1 or 2 eq. Na CO (0.5 mmol, 0.053 g,
.0 mmol, 0.106 g) or Et N (1.0 mmol, 140 µl; 2.0 mmol,
3
80 µl) were heated in 0.5 ml of solvent at 50 or 80 °C and
.0695 g (1 mmol) HA·HCl was added. The reaction was stirred
2 3
6
3
(
0
1.0 mmol, 140 µl) were heated in 0.5 ml EtOH at 50 °C and
.0695 g (1 mmol) HA·HCl was added. The reaction was stirred
and checked by HPLC (method no. 3, 225 nm, RT: 1a: 10.6, 2a:
.12, 3a: 8.31).
8 mg 2a benzamidoxime (0.5 mmol) was heated in 0.5 ml
and checked by HPLC (method no. 1–4, 225 nm, RT). The
reaction mixture was added to ice water, and the product
was filtered off, washed with water (2 × 0.5 ml) and dried for
analysis. Analytical data of 2a–2m were identical to those
8
6
EtOH and 0.25 ml H O at 50 °C. The reaction was stirred and
2
checked by HPLC (method no. 3, 225 nm, RT: 1a: 10.6, 2a:
2
2,24,32–36
described in the literature.
8.12, 3a: 8.31).
5
1 µl 1a benzonitrile (51.56 mg, 0.5 mmol) and 1 eq. Et N
3
Experiments to reaction mechanism investigations
(
1.0 mmol, 140 µl) were heated in 0.5 ml EtOH at 50 °C and
1 µl 1a benzonitrile (51.56 mg, 0.5 mmol) was heated in 0.0976 g (1 mmol) DMHA·HCl (99%) was added. The reaction
.5 ml EtOH at 80 °C, and 61 µl (0.0661 g; 1.0 mmol) 50% was stirred and checked by HPLC (method no. 3, 225 nm, RT:
5
0
water solution of HA was added dropwise. The reaction was 1a: 10.6, 3a: 8.31) and HPLC-MS.
stirred and checked by HPLC (method no. 3, 225 nm, RT: 1a:
51 µl 1a benzonitrile (51.56 mg, 0.5 mmol) and 1 eq. Et
(1.0 mmol, 140 µl) were heated in 0.5 ml NMP at 50 °C and
3
N
1
0.6, 2a: 8.12, 3a: 8.31).
1 µl 1a benzonitrile (51.56 mg, 0.5 mmol) was heated in 0.0742 g (1 mmol) HA·HCl (95%) was added. The reaction
.5 ml EtOH at 50 °C. The reaction was stirred and checked by was stirred and checked by HPLC-MS.
0.201 g benzamide (BA*) (1.65 mmol) was stirred in 248 µl
1 µl 1a benzonitrile (51.56 mg, 0.5 mmol) and 1 eq. EtOAc at 45 °C and 1.18 ml (1.98 mmol, 1.68 M) T3P was
CO
(0.5 mmol, 0.053 g) were heated in 0.5 ml EtOH at added to the mixture. The reaction was checked by HPLC-MS.
0 °C. The reaction was stirred and checked by HPLC (method
52 µl (BN*) (52.1 mg, 0.5 mmol) and 1 eq. Et N (1.0 mmol,
1
5
5
0
HPLC (method no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
5
Na
5
2
3
3
no. 3, 225 nm, RT: 1a: 10.6, 2a: 8.12, 3a: 8.31).
140 µl) were heated in 0.5 ml EtOH at 50 °C and 0.0695 g
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8036–8047 | 8045

View Article Online
Paper
Organic & Biomolecular Chemistry
(1 mmol) HA·HCl was added. The reaction was stirred and [Ch][OAc]
Choline acetate
checked by HPLC-MS.
[Ch][L]
TBAB][OBz]
2-Hydroxyethyl-trimethylammonium L-(+)-lactate
Tetrabutylammonium benzoate
[
Kinetic measurements
5
10 µl of 1a (516 mg, 5 mmol) and 1 eq. Et
were heated in 55 ml of solvent at 50 °C, and 695 mg
10 mmol) HA·HCl and 100 µl of toluene as the internal stan- This paper is dedicated to the memory of István Hermecz
3
N (5 mmol, 700 µl)
Acknowledgements
(
dard were added. The reaction was stirred and followed by (1944–2011). Special acknowledgement goes to: Edit Alattyáni,
HPLC (method no. 3) at different reaction times for 16 h. Csilla Busa, Péter Hudáky, Attila Németh and Charlotte Wiles.
Kinetic data are listed in Tables S2–S6.†
The author expresses his gratitude to Sanofi for contributing
to his Ph.D. scholarship and is grateful to Pro Progressio and
József Varga Foundations for their financial support.
Computational methods
All computations were carried out with the Gaussian09
3
7
program package (G09), using standard convergence criteria.
Computations were carried out at the B3LYP/6-31G(d,p) level References
3
8,39
40
of theory,
2 2
using the IEF-PCM solvent method; CH Cl
1
D. N. Nicolaides, E. A. Varella and S. Patai, The Chemistry of
Acid Derivatives, Interscience, New York, 1992, vol. 2,
p. 875.
does not require consideration of the solvent effect during the
modelling process. The method and basis sets were chosen for
their reliability in agreement with the studies established
2
7,31,41,42
2 L. B. Clapp, A. R. Katritzky and A. J. Boulton, Advances in
Heterocyclic Chemistry, Academic Press Inc., New York,
earlier.
The vibrational frequencies were computed at
the same levels of theory as those used for geometry optimiz-
ation in order to properly confirm all structures as residing
at minima on their potential energy hypersurfaces (PESs).
Thermodynamic functions U, H, G and S were computed at
1976, vol. 20, p. 65.
3
4
L. B. Clapp and K. T. Potts, Comprehensive Heterocyclic
Chemistry, Pergamon Press, Oxford, 1984, vol. 6, p. 365.
J. C. Jochims, A. R. Katritzky, C. W. Rees and
E. F. V. Scriven, Comprehensive Heterocyclic Chemistry II,
Pergamon Press, Oxford, 1996, vol. 4, p. 179.
2
98.15 K, using the quantum chemical, rather than the con-
ventional, thermodynamic reference state.
5
6
R. Roger and D. G. Nielson, Chem. Rev., 1961, 61, 179.
C. E. Stephens, E. Tanious, S. Kim, D. W. Wilson,
W. A. Schell, J. R. Perfect, S. G. Frazblau and D. W. Boykin,
J. Med. Chem., 2001, 44, 1741.
Ionic liquid abbreviations
[
[
[
[
[
[
BMIM]
TFMSI]
EMIM][Cl]
BMIM][Cl]
THTDP][Cl]
BMIM][AlCl
1-Butyl-3-methylimidazolium
methyl-sulfonyl)imide
bis(trifluoro-
7 J. L. Collins, B. G. Shearer, J. A. Oplinger, S. Lee,
E. P. Garvey, B. G. Salter, J. A. Oplinger, S. Lee, E. P. Garvey,
M. Salter, C. Dufry, T. C. Burnette and E. S. Furtine, J. Med.
Chem., 1998, 41, 2858.
8 H. Nakamura, Y. Sasaki, M. Uno, T. Yoshikawa, T. Ansano,
H. S. Ban, S. Fukazawa and Y. Uehara, Bioorg. Med. Chem.
Lett., 2006, 16, 5127.
9 R. Plapinger and O. Owens, J. Org. Chem., 1956, 21, 1186.
10 F. Eloy and R. Lenaers, Chem. Rev., 1962, 62, 155.
11 L. Stephenson, W. K. Warburton and M. J. Wilson, J. Chem.
Soc. C, 1969, 861.
12 R. D. Rogers and K. R. Seddon, Ionic Liquids: Industrial
Applications for Green Chemistry, American Chemical
Society, Washington, 2002.
1-Ethyl-3-methylimidazolium chloride
1-Butyl-3-methylimidazolium chloride
Trihexyltetradecylphosphonium chloride
1-Butyl-3-methylimidazolium
tetrachloroaluminate
4
]
[
[
[
[
BMIM][BF4]
1-Butyl-3-methylimidazolium
tetrafluoroborate
1-Hexyl-3-methylimidazolium
tetrafluoroborate
1-Methyl-3-octylimidazolium
tetrafluoroborate
Trihexyltetradecylphosphonium
tetrafluoroborate
HMIM][BF
MOIM][BF
THTDP][BF4]
4
]
4
]
[
TBP][BF
4
]
Tetrabutylphosphonium tetrafluoroborate
Trihexyltetradecylphosphonium
hexafluorophosphate
13 P. Wasserscheid and T. Welton, Ionic Liquids in Synthesis,
Wiley-VCH, Weinheim, 2nd edn, 2007.
14 M. J. Earle, J. M. S. S. Esparença, M. A. Gilea, J. N. C. Lopes,
L. P. N. Rebelo, J. W. Magee, K. R. Seddon and
J. A. Widegren, Nature, 2006, 439, 831.
[THTDP][PF6]
[
[
[
EMIM][EtSO
4
]
1-Ethyl-3-methylimidazolium ethylsulfate
EMIM][HeSO4] 1-Ethyl-3-methylimidazolium hexylsulfate
EMIM][BuSO 1-Ethyl-3-methylimidazolium butylsulfate
4
]
15 R. Ludwig and U. Kragl, Angew. Chem., Int. Ed., 2007, 46,
6582.
Ammoeng 100 Tetraalkylammonium methylsulfate
[TIBMP][OTs]
[BMIM][TFA]
[BMIM][OAc]
[EMIM][OAc]
Triisobutylmethylphosphonium tosylate
1-Butyl-3-methylimidazolium trifluoroacetate 17 M. Freemantle, Chem. Eng. News, 1999, 77, 23.
1-Butyl-3-methylimidazolium acetate
1-Ethyl-3-methylimidazolium acetate
16 M. Freemantle, Chem. Eng. News, 1998, 76, 31.
18 D. Bradley, Chem. Ind., 1999, 3, 86.
19 M. Freemantle, Chem. Eng. News, 2000, 78, 37.
8046 | Org. Biomol. Chem., 2014, 12, 8036–8047
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
2
0 N. V. Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008, 34 S. Hirono, S. Shiozawa, H. Chaki, H. Kotsubo, T. Tanaka
3
7, 123.
and Y. Aikawa, Preparation of benzophenone derivatives
as AP-1 inhibitors for treatment of arthritis, EP1445249,
2004.
2
2
1 C. Reichardt, Green Chem., 2005, 7, 339.
2 A. Vörös, Z. Baán, P. Mizsey and Z. Finta, Org. Process Res.
Dev., 2012, 16, 1717.
3 R. M. Srivastava, M. C. Pereira, W. W. Faustino, K. Coutinho,
J. V. dos Anjos and S. J. de Melo, Monatsh. Chem., 2009,
35 D. M. Cottrell, J. Capers, M. M. Salem, K. Deluca-Fradley,
S. L. Croft and K. A. Werbovetz, Bioorg. Med. Chem., 2004,
12, 2815.
2
2
1
319.
36 J. Lessel, Arch. Pharm., 1993, 326, 383.
4 A. Vörös, Z. Baán, G. Timári, I. Hermecz, P. Mizsey 37 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
and Z. Finta, Curr. Org. Synth., 2014, DOI: 10.2174/
5701794113106660082, in press.
5 Zs. Kelemen, O. Hollóczki, J. Nagy and L. Nyulászi, Org.
Biomol. Chem., 2011, 9, 5362.
6 H. Rodríguez, G. Gurau, J. D. Holbrey and R. D. Rogers,
Chem. Commun., 2011, 47, 3222.
7 Z. Mucsi, A. Szabó, I. Hermecz, Á. Kucsman and
I. G. Csizmadia, J. Am. Chem. Soc., 2005, 127, 7615.
8 E. Buncel, R. A. Stairs and H. Wilson, The Role of the
Solvent in Chemical Reactions, Oxford University Press,
Oxford, 2003.
9 F. Ruff and I. G. Csizmadia, Organic Reactions: Equilibria,
Kinetics and Mechanism, Elsevier, Amsterdam, 1994, 8, 232.
0 J. C. Cramer, Essentials of Computational Chemistry, John
Wiley & Sons Ltd., West Sussex, 2001, vol. 7, p. 433.
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski,
J. A. Montgomery, R. E. Stratmann Jr., J. C. Burant,
S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin,
M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi,
R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford,
J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui,
K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-
Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong,
J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle
and J. A. Pople, Gaussian 03 6.0, Gaussian Inc., Pittsburgh,
PA, 2003.
1
2
2
2
2
2
3
3
1 E. Frank, Z. Mucsi, I. Zupkó, B. Réthy, G. Falkay, 38 A. D. Beke, J. Chem. Phys., 1993, 98, 5648.
Gy. Schneider and J. Wölfling, J. Am. Chem. Soc., 2009, 131, 39 M. Head-Gordon, J. A. Pople and M. Frisch, J. Chem. Phys.
3
894.
Lett., 1988, 153, 503.
3
3
2 X. Yang, H. Li, Y. Zhang, C. Li, B. Liu, W. Liang, G. Zhao, 40 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005,
G. Liu, D. Song, R. Wang and Y. Jing, J. Med. Chem., 2010,
3, 1015.
3 D. Holsworth, J. Waaler, O. Machon, S. Krauss and
105, 2999–3093.
41 Z. Mucsi, M. Porcs-Makkay, Gy. Simig, I. G. Csizmadia and
B. Volk, J. Org. Chem., 2012, 77, 7282.
5
L. Golding, Azole derivatives as WTN pathway inhibitors, 42 Z. Mucsi, B. Viskolcz and I. G. Csizmadia, J. Phys. Chem. A,
WO/2010/139966, 2010. 2007, 111, 1123.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8036–8047 | 8047