Mechanistic pathways of aromatic nucleophilic substitution in conventional solvents and ionic liquids

Article abstract of DOI:10.1039/c4nj00130c

Solvation effects on the reaction mechanism of the title reactions have been kinetically evaluated in 21 conventional solvents and 17 ionic liquids. Solvent polarity affects the catalyzed and non-catalyzed S_NAr pathways differently. The ambiphi

Full text of DOI:10.1039/c4nj00130c

NJC
View Article Online
View Journal | View Issue
PAPER
Mechanistic pathways of aromatic nucleophilic
substitution in conventional solvents and ionic
liquids†
Cite this: New J. Chem., 2014,
38, 2611
Marcela Gazitua,*a Ricardo A. Tapia,b Renato Contrerasc and
´
Paola R. Campodonicoa
´
Solvation effects on the reaction mechanism of the title reactions have been kinetically evaluated in
21 conventional solvents and 17 ionic liquids. Solvent polarity affects the catalyzed and non-catalyzed
SNAr pathways differently. The ambiphilic character of water and formamide, which act as a hydrogen
bond donor/acceptor, induces nucleophilic activation at the nitrogen center of the nucleophile. The
ionic liquid EMIMDCN appears to be the best solvent for the SNAr route probably due to the high
polarizability of the dicyanamide anion.
Received (in Porto Alegre, Brazil)
24th January 2014,
Accepted 12th March 2014
DOI: 10.1039/c4nj00130c
www.rsc.org/njc
Solvent effects on SNAr reactions have previously been
studied in conventional organic solvents (COS)4,8–11 and more
Introduction
Nucleophilic substitution is an addition–elimination process recently in room temperature ionic liquids (RTILs).12–16 The
that depending on the nature of the substrate, the attacking main focus in these studies was put on the bulk and specific
nucleophile and the solvent effect may lead to a nucleophilic solute–solvent interactions that determine selectivity, reaction
substitution (SN) product, a nucleophilic aromatic substitution rates and mechanisms in these systems. Nudelman et al.9
(SNAr) product, or both.1 Scheme 1 summarizes the possible studied leaving group abilities in the SNAr reactions of halo-
routes towards any of both reaction products. When the reaction benzenes towards amines in aprotic COS. These authors found
center is a heteroatom (the left branch in Scheme 1) the reaction that the rate-limiting step of the reaction mechanism changes
mechanism is SN and it depends on the stability of the pentavalent when the reaction proceeds in solvents that exhibit hydrogen-
intermediate (PÆ in Scheme 1) formed. If PÆ is not stabilized under bond basicity (HBB) properties. Wang et al.10 reported that HBB
the reaction conditions, the process is concerted.2 Conversely, if PÆ of solvents significantly affects the regiochemistry of the SNAr
is stabilized, a stepwise pathway may be operative.1 Furthermore, if reaction between polyfluoroarene derivatives and amines. Park
the nucleophilic attack occurs at the aromatic ring (the right branch et al.11 investigated the mechanism of SNAr of fluorination
in Scheme 1), the reaction is classified as SNAr.1,3–8 This reaction reactions under the influence of protic solvents and charged
proceeds via a stepwise mechanism1,3–6 that involves: (i) a nucleo- nucleophiles. They found that counterions or protic solvents
philic attack with re-hybridization of the ipso carbon atom on the alone retard the SNAr reactions, while together they promote
aromatic ring from sp2 to sp3 to form a Meisenheimer complex1,7,8 the enhancement of reactivity. For these systems, the protic
(MC1 in Scheme 1) and (ii) the leaving group (LG) departure in solvent may affect the reaction acting over the counterion as a
a second step to regenerate the sp2 center through catalyzed or non- Lewis base and the nucleophile acting as an ion pair. The
catalyzed pathways1,4–8 (k3 or k2, respectively in Scheme 1).
presence of charges, as a separated cation–anion pair or as ion
pairs, strongly suggests that these results may be used to study
solvent effects on chemical reactivity (reaction rates in the
present case) in RTILs. D’Anna et al.14 recently reported on
the reaction of thiophene derivatives with N-nucleophiles in
RTILs. They proposed that the intramolecular interactions at the
transition state structure are strongly affected by the reaction
medium, determining the selectivity and catalysis of the process.
This background prompted us to perform a comparative
study of solvation effects in COS and a series of RTILs. However,
some preliminary considerations are worth stressing. First of
all, the physical interpretation of solvent effects in RTILs is an
a
´
´
´
Centro de Quımica Medica, Facultad de Medicina, Clınica Alemana Universidad
´
del Desarrollo, Codigo Postal 7710162, Santiago, Chile. E-mail: migazitu@uc.cl,
pcampodonico@udd.cl; Fax: +56 2 23279639; Tel: +56 2 23279682
b
c
´
´
Facultad de Quımica, Pontificia Universidad Catolica de Chile,
´
Codigo Postal 7820436, Santiago, Chile. E-mail: rtapia@uc.cl;
Tel: +56 2 23544429
´
Departamento de Quımica, Facultad de Ciencias, Universidad de Chile,
Casilla 653, Santiago, Chile. E-mail: rcontrer@uchile.cl; Tel: +56 2 29787272
† Electronic supplementary information (ESI) available: Copies of 1H NMR
spectra, rate constant values and experimental conditions. See DOI: 10.1039/
c4nj00130c
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 2611--2618 | 2611

View Article Online
Paper
NJC
Scheme 1 Possible nucleophilic substitution reactions of DNBSCl with secondary alicyclic amines.
extremely complex problem, because in general, good RTILs a reasonably large number of RTILs (17 in the present case) and
(as reaction media) are those species where the anion and evaluate their performance as reaction media by fixing the
cation are associated to a very low extent.17 This result implies anion and varying the cation and vice versa. The criteria for
that the target solute–solvent interactions will in general be selecting the series of RTILs were based on: (i) the solubility of
masked by the leading solvent–solvent interactions that are substrates and nucleophiles; (ii) to have a reasonable number
coulombic in nature. A usual experimental model to rationalize of anions and cations to assess anion and cation effects and
solvation effects on reaction rates in RTIL media is to consider to ensure that these RTILs do not interfere with the reaction
Scheme 2 Structures and acronyms of the RTILs used in this study.
2612 | New J. Chem., 2014, 38, 2611--2618
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
Table 1 pKa and kN values for the reactions of SA amines with DNBSCl in
watera
under study. Scheme 2 shows structures and acronyms of the
RTILs used in this study.
104 kN/sÀ1 MÀ1
21
Amine
pKa
Piperidine
Piperazine
1-(2-Hydroxyethyl)piperazine
Morpholine
1-Formylpiperazine
11.24
9.94
9.39
8.78
7.68
306.4 Æ 4.8
51.9 Æ 0.78
16.1 Æ 0.22
4.95 Æ 0.14
0.786 Æ 0.04
Results and discussion
In this work we use the reaction of 2,4-dinitrophenylsulfonyl
chloride (DNBSCl) with secondary alicyclic (SA) amines in
aqueous media as a model system to analyze the effect of
solvent polarity on the reaction mechanism. We started with
the synthesis of 1-(2,4-dinitrophenylsulfonyl)piperidine and
1-(2,4-dinitrophenyl)piperidine (compounds 2 and 3 in Scheme 1),
corresponding to the SN and SNAr reaction products, respectively
(synthetic details are given in the Experimental section). The
product analysis in water and in acetonitrile (MeCN) shows that 3
is the unique product at the end of the reaction, thereby suggesting
an SNAr pathway. HPLC analyses were performed to compare the
retention times and UV-visible spectra with those of authentic
samples in water and in MeCN as reaction media at the end of
the reaction (see Fig. S1–S9 in ESI†). However HPLC analyses at
different times show that the SN product is formed in minor
proportion and decomposed in excess of amine to give the SNAr
product in both cases (see Fig. S8 and S9 in ESI†).
The formation of a unique SNAr product discards the possibility
of nucleophilic attack at the unsubstituted position on the ring.18
Under amine excess, the pseudo first order rate constant
(kobs) was found for all the reactions. The values of kobs for all
the reactions are in accordance with eqn (1) where k0 and kN are
the rate coefficients for solvolysis and aminolysis of the substrate,
respectively. These values were obtained as the intercept (k0) and
slope (kN) of linear plots of eqn (1) for the reactions of DNBSCl with
all amines in water, some COS and all RTILs.
a
The value accompanying the kN values correspond to error of the slope
of a plot of kobs vs. free amine concentration required to build up
Brønsted plots.
in water as a reference. For all reaction media, the final product
was confirmed to be compound 3, thereby ratifying an SNAr
pathway.
Linear plots of kobs vs. [NT] (total amine concentration) for a
series of COS and the whole series of RTILs series are coincident
with the same kinetic responses found in aqueous media
(see Fig. S16–S41 and Tables S16–S42 in ESI†). However, some
COS show upward curvature as a function of increasing amine
concentration. Such curvature is typical for reactions that proceed
through a rate limiting proton transfer (RLPT) mechanism22
(MC2, right side branch in Scheme 1). Details are given in
Fig. S42–S67 and Tables S43–S55 in ESI.† According to Scheme 1,
the reaction proceeds through two intermediates (MC1 and MC2)
corresponding to a zwitterionic adduct and its deprotonated form,
respectively. Table 2 shows the microconstants Kk2 and Kk3 for the
reaction of DNBSCl with piperidine for a series of COS studied that
proceeds through catalyzed pathways. The low contribution of Kk2
(to promote the LG departure) and Kk3 (to promote the stabilization
of MC2) shows that solvents effects are small in these cases. This
result may be traced to the presence of a second molecule of amine
that could establish a more favoring interaction to the MC1
compared to that of the solvent molecules. However, solvent
polarity may become relevant in the non-catalyzed route.
kobs = k0 + kN[N]
(1)
In water, for all amines studied, a linear plot of kobs vs. free
amine concentration ([NF]) is observed. These plots pass
through the origin, thereby suggesting that the contribution
of hydroxide and/or water (i.e., the k0 step in Scheme 1) to the
pseudo first order rate constant (kobs) values is negligible8 and
the reactions occurs via a non-catalyzed route (k2 in Scheme 1).
Kinetic details are shown in the Experimental section and ESI†
(Fig. S10–S15 and Tables S1–S15).
The ‘‘polarity of the solvent’’ is a rather ambiguous concept,
because it encompasses a series of different effects, namely, the
static polarizability (dielectric effects) of the solvent; (electronic)
polarization of the solvent and specific solute–solvent interactions
Table 2 Observed microconstants Kk2 and Kk3 for the reaction of
DNBSCl with piperidine in a series of catalyzed processes in COSa
Previous reports suggest that this kind of reactions proceed
through a rate-limiting formation of MC1.8,19 However, the
statistically corrected Brønsted-type plot for the reference solvent
(water) is linear with bnuc = 0.8 a value close to that observed in the
ester aminolysis, where the LG departure is rate determining.20 It
seems then that the bond formation between the nucleophiles
and the substrate is fully advanced in the rate-limiting transition
state. Table 1 shows the nucleophilic rate constants and pKa
values in water needed to build up the Brønsted type-plot. Note
that piperidine is a basic amine that tends to deprotonate more
easily which is in agreement with kN values found for SA amines.
In order to examine the solvent effects on the reaction
mechanism, we studied the reaction between DNBSCl and
Solvent
103 Kk2/sÀ1 MÀ1
101 Kk3/sÀ1
MeCN
THF
CH2Cl2
CHCl3
C6H6
DMF
1,4-Dioxane
Acetone
Cyclohexane
Diethyl ether
Ethyl acetate
n-Hexane
n-Heptane
7.25 Æ 0.347
0.372 Æ 0.0600
1.31 Æ 0.317
0.500 Æ 0.116
1.86 Æ 0.314
8.75 Æ 0.191
0.403 Æ 0.109
1.53 Æ 0.254
1.94 Æ 0.295
1.76 Æ 0.420
1.14 Æ 0.210
0.0454 Æ 0.431
0.0328 Æ 0.452
0.780 Æ 0.0223
0.211 Æ 0.00350
0.640 Æ 0.0204
0.120 Æ 0.00620
0.500 Æ 0.0170
0.270 Æ 0.0106
0.163 Æ 0.00820
0.452 Æ 0.0190
0.85 Æ 0.0356
0.320 Æ 0.0320
0.300 Æ 0.0136
0.791 Æ 0.0402
1.04 Æ 0.0563
a
The value accompanying the Kk2 and Kk3 values correspond to the
piperidine in a series of 20 COS and 17 RTILs using the reaction error of the slope (for Kk2) and the intercept (for Kk3).
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 2611--2618 | 2613

View Article Online
Paper
NJC
Table 3 Values of kN for the reactions of DNBSCl with piperidine in a
series of non-catalyzed processes in COS and their Kamlet–Taft
parameters
Solvent
104 kN/sÀ1 MÀ1
a23
b23
p* 23
DMSO
Ethanol
FMA
278 Æ 9
0.0
0.76
0.84
0.48
0.18
0.93
0.84
0.84
0.66
1.0
7.2 Æ 0.30
547 Æ 11
0.86
0.71
1.17
0.68
0.76
0.84
0.98
0.48
0.97
1.09
0.41
0.48
0.47
0.60
H2O
156 Æ 2.2
16.4 Æ 0.90
6.4 Æ 0.29
9.2 Æ 0.43
4.8 Æ 0.12
tert-Butanol
2-Propanol
Butanol
Methanol
Fig. 2 General scheme for the possible interaction between H2O and the
MC1 intermediate.
kN values were obtained from the slope of a plot of kobs vs. total amine
concentration.
We further performed the kinetic analysis of the reaction in
a 50% w/w ethanol/water mixture and a 90% w/w ethanol/water
mixture under the same experimental conditions (see Table S43
in ESI†), in order to compare the role of HB effects in these
mixtures with water. The rate coefficient H2O vs. 50/50 ethanol/
water mixture ratio is 5 times slower while that of H2O vs. 90/10
ethanol/water mixture is 9 times slower. This result suggests
that there may be an increase in reactivity induced by ‘‘prefer-
ential solvation’’ in favor of the aqueous phase.4 Probably, FMA
can also form effective HBs with the nucleophile and the
leaving group of the MC1 structure. These properties of FMA
and H2O should account for the high kN values observed.
Note that all alcohols have high values of a parameters, but
low kN values. This can be traced to its low polarity compared to
that of DMSO, FMA and H2O.
Fig. 1(B) shows the relationship between log kN and the b
parameter,23 which is a measure of the HBB ability of solvents.
Fig. 1(B) displays the same separation between solvents of
varying polarity. Here again we have two regions independent
of their b values that are worth analyzing. In the first one we
found solvents with HB accepting (DMSO) and HB accepting/
donating (FMA and H2O) abilities that can interact with the
proton of piperidine, thereby increasing the electron density on
the nitrogen atom.25 Note that alcohols appear grouped with
the same pattern as that shown in Fig. 1(A).
(typically, hydrogen bonding effects). Kamlet–Taft’s linear
solvation energy relationship (LSER)23–26 model does the job
in the sense that solvent effects on rate constants may be
described by the Hydrogen Bond Acidity (HBA) parameter a;
the HBB parameter b and the p* parameter21 (see Table 3).
Fig. 1(A) shows the relationships between log kN for the non-
catalyzed reactions and the a parameter23 which is associated
with the HBA ability of solvents. In Fig. 1(A) we have two regions
independent of their a values. In the first one, we found
dimethylsulfoxide (DMSO), formamide (FMA) and H2O grouped
as being the best solvents in terms of rate coefficients. On the
other hand, we found a second group including all alcohols
used in this study. Here there are two points worth mentioning:
one is that DMSO, H2O and FMA are more polar than alcohols
and secondly that they display different hydrogen-bond (HB)
accepting/donating abilities. For instance, DMSO presents only
HB accepting properties, while H2O and FMA can accept and
donate HBs. The high rate value for H2O suggests the presence
of a HB between a hydrogen atom of H2O acting as a bridge
between the nucleophile and the leaving group at the MC1
structure. This bridge facilitates the relay of electron density
from the amine towards the electrophilic centre, thereby
enhancing the nucleophilicity of the amine (see Fig. 2).5
Fig. 1(C) shows the relationships between log kN and p* para-
meter23 which measures the ability of solvents to stabilize a neigh-
boring charge or a dipole by virtue of nonspecific dielectric
interactions. Note that the trend is similar to that found for a and
b values shown in Fig. 1(A) and (B). In this figure we have two regions
differing in solvent polarity and independent of their p* values. Since
H2O, FMA and DMSO have the highest p* values among all the
solvents investigated these reaction media enhance the reaction.
Table 4 shows kN values for the reactions of DNBSCl with
piperidine in RTILs and their Kamlet–Taft parameters.
Fig. 3 displays linear relationships between kobs and [NT] for
the same reaction in the series of 17 RTILs. The comparisons
will be made taking water as a reference. We can identify the
following classification of the ionic solvents: (i) best solvent:
EMIMDCN, (ii) good solvents: BMIMBF4, EMIMSCN, EMIMBF4,
BMIMNTf2, BM2IMNTf2, HMIMNTf2, BMPLNTf2, EMIMNTf2,
BMIMFAP, BMPLFAP, BMIMPF6, BMIMDCN and BMPLDCN,
(iii) similar to water: MOEDEAFAP and EMIMFAP and (iv) very
poor solvent: EAN.
Fig. 1 Plots of log kN and a (A), b (B) and p* (C) for the reactions of DNBSCl
with piperidine in a series of non-catalyzed processes in COS.
2614 | New J. Chem., 2014, 38, 2611--2618
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
Table 4 Values of kN for the reactions of DNBSCl with piperidine in RTILs
and their Kamlet–Taft parameters
RTILs
104 kN/sÀ1 MÀ1
a
b
p*
EAN
8.70 Æ 0.35
207 Æ 10.0
480 Æ 50.0
448 Æ 42.0
489 Æ 28.0
152 Æ 6.30
179 Æ 17.0
186 Æ 5.60
194 Æ 10.0
235 Æ 5.30
299 Æ 17.0
717 Æ 32.0
266 Æ 21.3
360 Æ 19.0
330 Æ 28.0
263 Æ 15.0
291 Æ 20.0
0.8527
0.6528
nd
0.4627
0.2528
nd
1.2427
1.0228
nd
BMIMPF6
EMIMSCN
BMIMBF4
EMIMBF4
MOEDEAFAP
EMIMFAP
BMIMFAP
BMPLFAP
BMIMDCN
BMPLDCN
EMIMDCN
BMIMNTF2
EMIMNTF2
BMPLNTF2
BM2IMNTF2
HMIMNTF2
0.6328
nd
0.3928
nd
1.0428
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
0.5428
nd
0.6028
nd
1.0528
nd
0.5427
0.7228
0.7128
0.5728
0.3828
0.6528
0.6427
0.2428
0.2328
0.2328
0.2628
0.2628
1.0727
0.9028
0.9828
0.8728 Fig. 4 Plots of log kN for the reactions of DNBSCl with piperidine in a
1.0228
series of COS and RTILs against a.
0.9728
nd: no data available in the literature. kN values were obtained from the
slope of a plot of kobs vs. total amine concentration.
latter, a charged probe should be used to account for the
cation–anion interaction and the performance depends on
the solvatochromic dyes used.25 The large variations found in
LSER values given by different dyes used prevent a direct
comparison of results obtained with a different probe in the
same analysis.25
Fig. 4 shows the relationships between log kN and a para-
meter23 for those reactions that proceed in COS via a non-
catalyzed route (shown in Fig. 1) and RTILs. a value of an RTIL
is largely determined by the availability of HBA sites on the
cation.25 Here we have again two regions worth analyzing: the
first corresponds to a group containing COS and RTILs which
present high rate constants independent of their a values.
This result confirms that the polarity of the medium is a
determinant in the stabilization of the intermediate thereby
enhancing the reaction rate. The second group includes alcohols
and EAN. Note that EAN is a unique RTIL that presents a low rate
coefficient comparable with that of alcohols, a result that can
again be traced to their low polarity.
Fig. 3 Plot of kobs against total amine concentration [NT] of piperidine for
the reactions of DNBSCl with piperidine in a series of RTILs with H2O as a
reference (red squares).
Welton et al.30 reported on the reaction of methyl p-nitro-
benzenesulfonate with a series of amines in RTILs and found
EAN is the only protic solvent that decreases the reaction that as the hydrogen bond donor ability (a) of the RTILs is
rate 17 times vs. H2O and 82 times vs. EMIMDCN. EMIMDCN increased the nucleophilicity of the amines is evidently reduced
behaves as the best solvent within the series of 38 reaction but in our case we do not find this behavior for the nucleophilicity
media analyzed in this work, it presents catalytic behavior at of piperidine, except in EAN with a high a value and a low rate
room temperature and therefore it qualifies as the best solvent constant.
for this SNAr reaction. Note that EMIMDCN exerts its catalytic
Fig. 5 shows the relationships between log kN and the b
property for the pathway that does not involve a second parameter23 for COS shown in Fig. 1 and RTILs. The b value of
nucleophilic molecule as a catalyst. The pKa value of dicyanamide an RTIL is controlled primarily by the anion, with basicity
is less than 1.29 As a result, the nucleophilicity on the nitrogen of increasing as the strength of the conjugate acid of the anion
piperidine is enhanced. The MC1 forming step is expected to be decreases; and its antagonistic possibility of HBA on the
fast and the leaving group departure becomes rate determining.
cation.25 Fig. 5 display similar trends to those shown in
If we compare the kN values of Table 3 for COS and Table 4 Fig. 4. Here again we have two regions grouped independent
for RTILs we can see that for this SNAr reactions RTILs perform of their b value.
relatively well in comparison with COS.
The analysis of solvent polarity using Kamlet–Taft’s linear parameter23 for COS shown in Fig. 1 and RTILs.
solvation energy relationship (LSER)23–26 shows that the com- Welton et al.30 found that all the RTILs they used presenting
parison of COS vs. RTILs presents the problem that for the high values of p*, with little variation between them, should
Fig. 6 shows the relationship between log kN and the p*
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 2611--2618 | 2615

View Article Online
Paper
NJC
Fig. 5 Plots of log kN for the reactions of DNBSCl with piperidine in a
series of COS and RTILs against b.
Fig. 7 Comparison between kobs against total amine concentration [NT]
for the reaction of DNBSCl with piperidine in EMIMNTF2, EMIMSCN,
EMIMDCN, EMIMBF4 and EMIMFAP at 25.0 1C.
Fig. 6 Plots of log kN for the reactions of DNBSCl with piperidine in a
series of COS and RTILs against p*.
Fig. 8 Comparison between kobs against total amine concentration [NT]
for the reaction of DNBSCl with piperidine in BMIMNTF2, EMIMNTF2,
BMPLNTF2, EMIMNTf2, BM2IMNTF2 and HMIMNTF2 at 25.0 1C.
always lead to an increased rate for its and similar reactions.
Apparently, in our case, this applies for all RTILs studied,
except for EAN, that present a high p* value and a low rate
constant. Note that here again the polarity of the medium is the This result may be attributed to the significantly high polarizability
determinant in the stabilization of the intermediate and there- of the dicyanamide anion, which presents a highly rich p electron
fore of the reaction rate.
density. An additional factor seems to be the size of the anion
To close our study, we made an additional analysis per- which shows an inverse relationship with the reaction rate. Even
formed by fixing the cation (EMIM in the present case) and though we do not have enough information to rationalize these
varying the anionic counterpart (i.e. to evaluate the anion results it could be related to steric hindrance effects.
effect). Fig. 7 and 8 summarize the comparison between kobs
and [NT] for the EMIM and NTf2 series, respectively. Other anion [NTF2]À and varying the cation. No significant differences
comparisons are shown in ESI† (Fig. S68–S71). in reaction rates were observed by changing the length of the
We considered a similar analysis but this time fixing the
Following a suggestion made by a reviewer, we performed a alkyl chain of the imidazole from two to six carbons, even if the
complete multiparametric correlation among our log kN and the acidic proton of the imidazole is blocked (see Fig. 8). Apparently,
hydrogen bond acidity (a), the hydrogen bond basicity (b) and the steric hindrance effects going from two to six carbon atoms
solvent polarity (p) Kamlet–Taft (KT) solvent parameters taken from in the alkyl chain in the cation is less important than changes in
the literature. The KT analysis of solvation effects shows similar the size of the anion.
trends to that shown in Fig. 4–6 obtained from a one parameter
plot. These correlations are reported in ESI,† Fig. S72 and S73.
A striking observation is that EMIMDCN and BMIMDCN
only differing in the alkyl chain on the cation display quite
The results show the following order of decreasing quality as different rate coefficients. One possibility is to have both RTILs
reaction media: [DCN]À 4 [SCN]À 4 [BF4]À 4 [NTF2]À 4 [FAP]À. with a significant amount of water. We proceeded then to
2616 | New J. Chem., 2014, 38, 2611--2618
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
repeat the kinetic measurements after drying both solvents for the reaction mixture was diluted with dichloromethane
8 hours in a vaccum drying oven at a pressure of À0.06 MPa. (50 mL), washed with 1 N HCl (15 mL), brine and dried. The
The same kinetic results were obtained. It seems that kinetic solvent was removed under reduced pressure and the residue
data are not sufficient to settle these differences which could be was purified by column chromatography on silica gel using
elucidated with the aid of theoretical studies introducing ethyl acetate–hexane 1 : 5 to give 1.85 g of a white product
electronic structure information. Work along this line is in (59%), mp 128–130 1C (Lit.32 130 1C). 1H NMR d 1.58–1.80
course in our group.
(m, 2H), 1.80–1.84 (m, 4H), 3.54–3.70 (m, 4H), 8.21 (d, J = 8.6 Hz,
1H), 8.45 (d, J = 2.0 Hz, 1H), 8.49 (dd, J = 2.0, 8.6 Hz, 1H).
Conclusions
Kinetic measurements
The kinetics of the reactions were measured using a diode array
spectrophotometer in water at 25.0 1C and an ionic strength of
0.2 M (maintained with KCl), by monitoring (380 nm) the
formation of product 3. The kinetic measurements in COS,
RTILs and 90/10 ethanol/water mixture were made in the
absence of KCl and at 350–400 nm at the same temperature.
The initial substrate concentration was 5 Â 10À5 M. Under
excess amine, pseudo-first-order rate coefficients (kobs) were
found throughout. For the reactions via non-catalyzed the kobs
values were obtained through the kinetic software (for first-order
reactions) of the spectrophotometer. For the reactions occurring
via catalyzed pathways, the validation that the reaction proceeds
as shown in Scheme 1 is done through the observation that the
plots of kobs/[NT] vs. [NT] are linear. Kk2 values are obtained from
the intercept of these graphs and Kk3 is obtained from the slope
of these graphs.
Solvation effects on the reaction mechanism of the title reactions
have been kinetically evaluated for a set of 21 conventional solvents
and 17 RTILs. Solvent polarity affects the catalyzed and non-
catalyzed SNAr pathways differently. The competitive SN product
is not observed at the end of the reaction under pseudo first order
conditions verified by HPLC analyses. The study of solvent polarity
performed on the series of COS plus water and FMA reveals that HB
ability drives the SNAr process in the non-catalyzed route in
Scheme 1. The role of water and FMA is the most significant due
to its ambiphilic character as an HB donor and an HB acceptor that
results in a nucleophilic activation at the nitrogen center of
piperidine. It is relevant to note that RTILs performed relatively
well in comparison with COS. The ionic liquid EMIMDCN appears
to be the best solvent for this SNAr route, a result probably due to
the high polarizability of the dicyanamide anion.
Chromatographic system and conditions
Experimental section
The HPLC system used for the analysis of the samples was a
UV-DAD Elite Lachrom equipped with a quaternary pump
L-2100 with a UV-DAD detector L-2455, an 8 mL injection loop,
an oven column L-2300 and an autosampler L-2200 with a
cooling unit. The column attached was a Chromolith Fast
Gradient RP 18 50–3 mm (Merck). The UV detector was set at
260 nm which was found to be the most suitable wavelength for the
detection of all the substrates, product and internal standard.
The flow-rate of the mobile phase was adjusted to 0.5 mL minÀ1
to keep the column pressure between 47–50 bar. The system was
thermostated at 25 1C to maintain the same reactions conditions.
Chromatograms were recorded in a computer system using
EZChrom Elite software from Agilent.
Materials
Piperidine was purified before use. All the solvent used were
commercially available by Sigma-Aldrich and Merck with purity
Z99%, stored under anhydrous conditions and used as
received. The certificate of analysis given by Merck S.A. of all
RTILs show purity values between 99 and 100%, presence of
halides r0.1% and content of water r1%. To ensure that they
had no water, we put the RTILs into a vaccum drying oven LabTech
Model LVO-2013 for 4 hours at a pressure of À0.06 MPa
before use.
Synthetic protocol of 1-(2,4-dinitrophenyl)piperidine
To a solution of 1-chloro-2,4-dinitrobenzene (200 mg, 0.99 mmol)
in dry DMSO (2.0 mL), containing potassium carbonate (280 mg,
2.03 mmol), was added piperidine (169 mg, 1.98 mmol). The
mixture was stirred for 12 h at room temperature and the
reaction mixture was poured onto ice-water (20 g). The solid
was filtered, washed with water, dried and recrystallized from
ethanol to give 1-(2,4-dinitrophenyl)piperidine (180 mg, 73%),
Acknowledgements
This work was supported by Project ICM-P10-003-F CILIS,
´
granted by Fondo de Innovacionpara la Competitividad del
´
Ministerio de Economıa, Fomento y Turismo, Chile; Fondecyt
mp 92–93 1C (Lit.31 91–92.5 1C). H (400 MHz, CDCl3): d 1.65–
1.80 (m, 6H), 3.20–3.30 1 (m, 4H), 7.08 (d, J = 9.4 Hz, 1H), 8.21
(dd, J = 9.4, 2.7 Hz, 1H), 8.69 (d, J = 2.7 Hz, 1H).
1
grants 1100492 and 1110062. M.G. acknowledges support from
Conicyt under the postdoctoral fellowship 3120060.
Notes and references
Synthetic protocol of 1-(2,4-dinitrophenylsulfonyl)piperidine
To a solution of 2,4-dinitrobenzenesulfonyl chloride (2.67 g,
10 mmol) in dichloromethane (35 mL), piperidine (0.85 g,
10 mmol) at 0 1C. The mixture was stirred for 4 hours and
1 J. F. Bunnett and R. E. Zahler, Chem. Rev., 1951, 49, 273;
J.-H. Choi, B.-C. Lee, H.-W. Lee and I. Lee, J. Org. Chem.,
2002, 67, 1277; I.-H. Um, J.-Y. Hong, J.-J. Kim, O.-M. Chae
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 2611--2618 | 2617

View Article Online
Paper
NJC
and S.-K. Bae, J. Org. Chem., 2003, 68, 5180; I.-H. Um, S.-M. 16 C. C. Weber, A. F. Masters and T. Maschmeyer, Org. Biomol.
Chun, O.-M. Chae, M. Fujio and Y. Tsuno, J. Org. Chem., Chem., 2013, 11, 2334.
2004, 69, 3166; I.-H. Um, J.-Y. Hong and J.-A. Seok, J. Org. 17 C. A. Angell, N. Byrne and J.-P. Belieres, Acc. Chem. Res.,
Chem., 2005, 70, 1438. 2007, 40, 1228.
2 A. Williams, Concerted Organic and Bio-organic Mechanisms, 18 E. Buncel, J. M. Dust and F. Terrier, Chem. Rev., 1995, 95, 2261.
CRC Press, Boca Raton, FL, 2000, ch. 4, p. 43. 19 M. R. Cramptom, Adv. Phys. Org. Chem., 1969, 7, 211.
3 O. Banjoko and I. A. Babatunde, Tetrahedron, 2004, 60, 4645. 20 W. P. Jencks, Chem. Rev., 1985, 85, 511; E. A. Castro, Chem.
´
4 R. Ormazabal-Toledo, J. G. Santos, P. Rıos, E. A. Castro,
Rev., 1999, 99, 3505.
P. R. Campodonico and R. Contreras, J. Phys. Chem. B, 2013, 21 E. A. Castro, M. Aliaga, M. Gazitu´a and J. G. Santos, Tetra-
117, 5908. hedron, 2006, 62, 4863.
5 R. Ormazabal-Toledo, R. Contreras, R. A. Tapia and 22 F. G. Bordwell and D. L. Hughes, J. Am. Chem. Soc., 1986,
´
´
P. R. Campodonico, Org. Biomol. Chem., 2013, 11, 2302.
108, 5991.
6 R. Ormazabal-Toledo, R. Contreras and P. R. Campodonico, 23 M. J. Kamlet, J. L. M. Abboud, M. H. Abraham and
J. Org. Chem., 2013, 78, 1091. R. W. Taft, J. Org. Chem., 1983, 48, 2877.
7 C. F. Bernasconi, MTP Int. Rev. Sci.: Org. Chem., Ser. One, 24 A. Cerda-Monje, A. Aizman, R. A. Tapia, C. Chiappe and
1973, 3, 33. R. Contreras, Phys. Chem. Chem. Phys., 2012, 14, 10041.
8 I.-H. Um, S.-W. Min and J.-M. Dust, J. Org. Chem., 2007, 25 J. P. Hallett and T. Welton, Chem. Rev., 2011, 111, 3508.
´
72, 8797.
26 R. Bini, C. Chiappe, V. L. Mestre, C. S. Pomelli and
T. Welton, Org. Biomol. Chem., 2008, 6, 2522.
27 P. G. Jossop, D. A. Jossop, D. Fu and L. Phan, Green Chem.,
2012, 14, 1245.
28 M. A. Ab-Rani, A. Brant, L. Crowhurst, A. Dolan, M. Lui,
N. H. Hassan, J. P. Hallett, P. A. Hunt, H. Niedermeyer,
J. M. Perez-Arlandis, M. Shrems, T. Welton and R. Wilding,
Phys. Chem. Chem. Phys., 2011, 13, 16831.
´
9 N. S. Nudelman, P. M. E. Mancini, R. D. Martınez and
L. R. Vottero, J. Chem. Soc., Perkin Trans. 2, 1987, 951.
10 X. L. Wang, E. J. Salaski, D. M. Berger and D. Power, Org.
Lett., 2009, 11, 5662.
11 S. Park and S. Lee, Bull. Korean Chem. Soc., 2010, 31, 2571.
12 I. Newington, J. M. Perez-Arlandis and T. Welton, Org. Lett.,
2007, 9, 5247.
13 F. D’Anna, S. Marullo and R. Noto, J. Org. Chem., 2010, 29 A. C. Kenneth, Chemical Kinetics: The study of reaction rates
75, 767. in solution, John Wiley & Sons, New York, USA, 1990, p. 236.
14 F. D’Anna, V. Frenna, R. Noto, V. Pace and D. Spinelli, J. Org. 30 L. Crowhurst, L. Lancaster, J.-M. Perez-Arlandiz and
Chem., 2006, 71, 5144. T. Welton, J. Am. Chem. Soc., 2004, 126, 11549.
15 F. D’Anna, S. Marullo and R. Noto, J. Org. Chem., 2008, 31 J. F. Bunnett and G. T. Davis, J. Org. Chem., 1954, 76, 3011.
´
´
73, 6224.
32 J. D. Loudon and N. Shulman, J. Chem. Soc., 1938, 1926.
2618 | New J. Chem., 2014, 38, 2611--2618
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014