Phys. Chem. Chem. Phys., 2009, 11, 341 - 348, DOI: 10.1039/b812597j

Quantum mechanical study and vibrational spectra of indaz olium-3-carboxylateand its decarboxylation product, the N-heterocyclic carbene indazol-3-ylidene

Andreas Schmidt, Bohdan Snovydovych, Juan Casado, Jo sé Joaquín Quirante, Juan Teodomiro López Navarrete and Francisco Javier Ramírez

High quality research in physical chemistry, chemical physics and biophysical chemistry.

Quantum mechanical study and vibrational spectra of indazolium-3-

carboxylate and its decarboxylation product, the N-heterocyclic carbene

indazol-3-ylidene

Andreas Schmidt,aBohdan Snovydovych,

aJuan Casado,

bJose Joaquın Quirante,

b

Juan Teodomiro Lopez Navarreteband Francisco Javier Ramırez*

b

Received 22nd July 2008, Accepted 14th October 2008

First published as an Advance Article on the web 20th November 2008

DOI: 10.1039/b812597j

Indazolium-3-carboxylate is a molecule that can be found as the nucleus of several pseudo-cross-

conjugated mesomeric betaines, such as the alkaloid nigellicine. From a chemical point of view,

one of the more interesting properties of this class of molecules is the possibility of forming an

N-heterocyclic carbene by thermal decarboxylation. In this paper we have studied the carbene

generation by decarboxylation of 1,2-dimethyl indazolium-3-carboxylate, using vibrational

(infrared and Raman) spectroscopy and quantum chemistry calculations. Normal mode analysis

allowed us to analyse the changes in the stretching force constants upon decarboxylation and to

establish spectroscopic-structure relationships. We also investigate the effect of 5-halogen (fluoro,

chloro) substitution on the carbene generation. Decarboxylation energy profiles of the three

derivatives were calculated. Crossing of the energy paths when going from the transition state to

the final product were obtained. The theoretical tendency found for the activation energies agrees

with that observed for the decarboxylation temperatures and for the calculated NICS values of

the benzene moieties.

Introduction

Indazolium-3-carboxylate, Fig. 1a, is a representative case of

the class of alkaloids known as pseudo-cross-conjugated meso-

meric betaines (PCCMB). The mesomeric betaines (MB) are

neutral aromatic molecules that delocalize an even number of

charges within a common p-electron system.1,2 According to the

simple valence bond approach of classification, in a PCCMB

these charges are effectively, but not exclusively, delocalized in

separated parts of the molecule, as the positive charge can be

delocalized into the anionic moiety. Common atoms for the

delocalization of the negative and positive charges exist in

the canonical formulae, although these involve electron sextet

structures with a negligible contribution to the overall electronic

structure. Typically, in PCCMB and CCMB the charged

moieties are joined through an unstarred atom of the anionic

moiety. This atom is a nodal site of the highest occupied

molecular orbital (HOMO) and, consequently, the bond bet-

ween the two charged moieties, commonly named the union

bond, is always a weak single bond which ensures charge

separation within the molecule.

Surprisingly, MB are frequently present in nature.1 In fact

some PCCMB have been isolated from natural sources, such as

shihunine from Dendrobium sp.,3–5 flavocarpine from Pleiocarpa

mutica,6,7 aeroginosine A and B from Pseudomonas aeruginosa8,9

or vincarpine and its dihydro derivative from Vinca major

elegantissima.10 Indazolium-3-carboxylate is a recently syn-

thetised analogue molecule11 of the alkaloid Nigellicine,

Fig. 1b. It is a PCCMB isolated from the herbaceous plant

Nigella sativa Linn.12 The seeds of this plant have been used for

thousands of years as a spice and for the treatment of various

diseases.13

In spite of the fact that the chemistry of conjugated MB has

been widely studied, there is little information about PCCMB.14

Recently a series of halogen derivatives of 1,2-dimethyl

indazolium-3-carboxylate (hereafter MI3C) have been synthe-

sized as new representatives of PCCMB.15 With heating, these

molecules decarboxylate to yield the indazol-3-ylidene ana-

logues (hereafter MI3Y), Scheme 1. In a previous work, these

decarboxylations were studied by electrospray-ionization mass

spectrometry, NMR spectroscopy and differential scanning

calorimetry (DSC).15 The achieved data allowed to analyse

the influence the aggregation state and the halogen substitution

in the decarboxylation temperature.

As depicted in Scheme 1, the product of this reacction can

be described as the contributions of two relevant resonance

Fig. 1 Chemical structures of the alkaloids indazolium-3-carboxylate

(a) and nigellicine (b).

a Institute of Organic Chemistry, Clausthal University of Technology,Leibnizstrasse 6, D-38678 Clausthal-Zellerfeld, Germany

bDepartamento de Quımica Fısica, Facultad de Ciencias, Universidadde Malaga, 29071 Malaga, Spain. E-mail: ramirez@uma.es;Fax: +34-952-132000; Tel: +34-952-132258

This journal is �c the Owner Societies 2009 Phys. Chem. Chem. Phys., 2009, 11, 341–348 | 341

PAPER www.rsc.org/pccp | Physical Chemistry Chemical Physics

forms. One of these, an all-octet structure, preserves the

separation between the negative and positive charges, which

are now separated by the N2QC3 double bond of the

pyrazolium ring. The second form is anN-heterocyclic carbene

(NHC) in which the carbon atom attached to the carboxylate

group in MI3C possesses an electron sextet structure. In the

solid state, the carbene MI3Y cannot be stabilized after CO2

extrusion.15 On the contrary, in solution of methanol and

deuterated dimethyl sulfoxide the product is able to capture a

proton to give the 1,2 dimethyl indazolium cation. Evidence

has been achieved on the strong relationship between the

phenomenon of pseudo-cross-conjugation and NHC formation

by means of an extrusion reaction.11,16,17

In this paper we present a spectroscopic study of the genera-

tion of the carbene MI3Y by thermal decarboxylation of the

betaine MI3C in the solid state. We used infrared and Raman

spectroscopies, aided by quantum chemistry calculations in the

framework of density functional theory (DFT) calculations, to

get insight into the mechanistic aspects of this process as well as

characterizing the species involved in the reaction and the

dependency with the chemical structure and the 5-halogen

(fluoro, chloro) substitution. Vibrational spectroscopy is a very

advantageous technique for structural characterization. Com-

pared with other methods, infrared and Raman spectra offer

high sensitivity and a great number of experimental data

directly related with the atoms in the molecules and their bonds.

These techniques are able to be applied to any aggregation state

in a non-destructive form and can be used coupled with a wide

catalog of accessories (temperature, pressure, reflection. . .) and

complementary techniques (electrochemistry, microscopy. . .).

These make infrared and Raman spectroscopy greatly superior

to monitor chemical reactions and their reactants and products.

Experimental methods

Samples

The 1,2-dimethyl indazolium-3-carboxylate, or MI3C, was

synthesized by esterification of indazol-3-carboxylic acid

with methanol, Scheme 2. The N,N-dimethyl derivative was

obtained by treating with dimethylsulfate in xylene and using

nitrobenzene as a catalyst. Final saponification with H2SO4

yielded MI3C. The 5-fluoro and 5-chloro derivatives of this

compound were obtained similarly from the 5-fluoro and

5-chloro indazol-3-carboxylic acids, respectively, using KOH

instead of H2SO4 in the saponification reaction. More details

of this synthesis are reported elsewhere.15

Vibrational spectroscopy

Infrared spectra at room temperature were recorded from pure

solid samples, dispersed into a KBr pellet, using a Bruker

VERTEX 70 Fourier-transform (FT) spectrometer purged

with dry nitrogen. The KBr used was dried following the usual

routine for infrared measurements (110 1C, at least 24 h),

thus preventing H2O trapping during the decarboxylation).

Typically, 500 scans at a resolution better than 4 cm�1 were

accumulated to optimize the signal-to-noise ratio. Individual

scans were examined by the recording routine before aver-

aging, being automatically discarded when the mean intensity

deviations were greater than 10% over the full interferogram

length.

Raman spectra were obtained using both the FRA106/S

module coupled to the VERTEX 70 FT system and the Bruker

Senterra dispersive Raman microscope. Excitation wavelengths

used were 1064 nm (from a Nd-YAG laser at 500 mw) for the

FT-Raman and 532 nm (from a high energy laser diodes at

20 mw) for the dispersive micro-Raman spectra. Spectra resolu-

tion were always better than 2 cm�1. Negligible differences were

observed when comparing FT and dispersive Raman spectra.

Spectral measurements were carried out by using the Bruker

OPUSr spectroscopic software.

Decarboxylation reactions were monitorized by infrared

and Raman spectroscopy. A variable-temperature cell Specac

P/N 21525, with interchangeable pairs of KBr windows, was

used for infrared spectra. The variable temperature cell

consists of a surrounding vacuum jacket (0.5 Torr), and

combines a refrigerant Dewar and a heating block as the

sample holder. It is also equipped with a copper constantan

thermocouple for temperature monitoring purposes, so that

any temperature from�170 to 250 1C can be reached. Samples

were inserted into the heating block part or the Dewar/cell

holder assembly in the form of a KBr pellet. Temperature-

controlled Raman spectra were obtained using the heating/

freezing stage THMS600 of the Senterra Microscope. A

temperature controller (Eurotherm) allows us to reach

temperatures from ambient to 600 1C, while temperatures

below ambient were achieved by pumping liquid nitrogen

through the heating/cooling block of the stage.

Theoretical calculations

The Gaussian’03 package of programs18 was used for DFT

quantum chemical calculations. The Becke’s three parameter

(B3) gradient-corrected exchange functional was used, and the

non-local correlation was provided by the Perdew–Wang’91

Scheme 1

Scheme 2

342 | Phys. Chem. Chem. Phys., 2009, 11, 341–348 This journal is �c the Owner Societies 2009

(PW91) expressions.19,20 All the electronic and vibrational

properties were calculated using the 6-311+G(2d,p) basis

set.21,22 It includes diffuse functions on heavy atoms and

polarization on heavy and hydrogen atoms, a requirement

necessary for calculations of charged-separated states. This is

also required for diradical structures where some atoms bear

very large electron densities. The open-shell methodology for

both singlet and triplet states was always used. Implicit solvent

was incorporated using a polarizable continuum model

(PCM),23,24 in which the solvent is assimilated to a continuum

characterized by its dielectric constant while the solute

molecule is placed into a size-adapted cavity formed from

overlapping atom-centered van der Waals spheres.

The minimum energy structures were achieved by allowing

all the geometrical parameters to vary independently. Harmonic

force constants, in Cartesian coordinates, were evaluated

at the ground state optimized geometry using analytical

second derivatives. Infrared absorption intensities were

analytically evaluated from the atomic polar tensors.25 The

Cartesian force constants were transformed into a set of non-

redundant locally symmetrized internal coordinates, defined

according to the Pulay methodology,26 which allowed for a

more useful description of the vibrational potential energy.

Wavenumbers and normal coordinates were calculated by the

Wilson FG matrix method.27 Obtained transition structures

(TS) were confirmed by frequency calculations. Subsequent

intrinsic reaction coordinate (IRC) calculations28 were

performed to unsure that the computed transition states

correlate with the minima depicted in Scheme 1. Isotropic

nucleus independent chemical shift (NICS) values were

evaluated by the gauge invariant atomic orbital (GIAO)

method.29 NICS scannings were performed over the optimized

structures at the geometric center of the benzene rings of

the decarboxylation product. NICS scanning over vertical

distances above of that point were also calculated.

Results and discussion

Interpreting the vibrational spectra of MI3C

Prior to analyzing the decarboxylation, we have to achieve a

suitable knowledge of the vibrational spectra of MI3C,

especially concerning the region in which vibrations of the

carboxylate group appear. This task has not been reported to

date, so that we have proposed a general assignment of the

vibrational features of this molecule supported by quantum

chemistry and normal mode calculations. The experimental

infrared and Raman spectra of MI3C between 1700 and

700 cm�1 are shown in Fig. 2 together with the theoretical

spectra as predicted by B3PW91/6-311+G(2d,p) quantum

chemistry calculation. The quadratic force constants were

uniformly scaled by a factor of 0.96.30 This methodology does

not modify the calculated infrared intensities, which allowed us

to combine them with the scaled wavenumbers in order to

depict the theoretical spectrum. The only exception to this

general procedure was the correction of the low out-diagonal

force constant of the two C–O stretching vibrations, which does

not describe suitably the strong interaction of these two bonds

in the same carboxylate group.

Table 1 summarizes the more intense infrared and Raman

bands and their assignments on the basis of the calculated

potential energy distribution (PED). As can be seen in Fig. 2,

the theoretical spectra exhibit a reasonably good correlation

with the experimental ones in both wavenumbers and inten-

sities. The PED values, together with the calculated intensities,

allow a reliable assignment of the carboxylate bands. Thus, the

two stretching modes are assigned to the two more intense

infrared bands in this region, measured at 1657 and 1320 cm�1,

while the corresponding bending vibrations are assigned to the

medium intensity bands at 816 and 748 cm�1. The aromatic

stretching vibrations of the benzene and of the pyrazolium rings

and the methyl bending modes are also quite well discriminated.

We would like to emphasize that most of the bands in this

region can be largely assigned, in a good approach, to a single

vibrational local coordinate, which is very useful to monitor the

decarboxylation reaction using vibrational tools.

Once a Cartesian force field is transformed to the space of

locally symmetrized vibrational coordinates, the diagonal force

constants for the stretching vibrations are available. They are

directly interpretable as a measure of the bond strength and can

be also compared with the calculated bond lengths of the

optimized structure. We have depicted in Fig. 3 the stretching

force constants (SFC) and the optimized bond lengths (OBL) of

Fig. 2 Comparison between the theoretical and the experimental

vibrational spectra of MI3C.

This journal is �c the Owner Societies 2009 Phys. Chem. Chem. Phys., 2009, 11, 341–348 | 343

MI3C. In our opinion, the most outstanding fact is the low value

obtained for the SFC of the union bond, 2.83 mdyn A�1, which

is appreciably lower than the usual values for a typical C–C

single bond (about 4.5–5.0 mdyn A�1). This result agrees

with the pseudo-cross-conjugation phenomenon and supports

the CO2 extrusion observed at relatively low temperatures. It

also correlates well with the optimized bond length of 1.559 A.

The six skeletal SFC of the benzene ring show a low dispersion,

namely 1.14 mdyn A�1 measured as the difference between the

maximum and the minimum value, being the averaged value

6.64 mdyn A�1. On the contrary, this difference is almost three

times higher, 3.12 mdyn A�1 within the pyrazolium ring. Here,

the weakest skeletal SFC was calculated for the N1–N2 bond,

6.16 mdyn A�1, while the highest one was predicted for the

N1–C5 bond, 9.28 mdyn A�1. Interestingly, the N1–C5 OBL is

0.034 A longer than the N2–C3 one, although its SFC is

0.89 mdyn A�1 greater, thus suggesting a stronger bond. On

the other hand, the N1–C5 stretching vibration is strongly

coupled with the benzene stretching vibrations in the light of

the PED data, and its contribution to the pyrazolium stretching

modes is small.

Monitoring the decarboxylation of MI3C by infrared and

Raman spectroscopy

We depict in Fig. 4 the effect of increasing temperature on the

vibrational spectra of MI3C. It is worth pointing out that, in

the case of the infrared spectra, the carbon dioxide molecules

were trapped in the potassium bromide pellet, as evidenced by

the appearance of the band at 2335 cm�1 (see Fig. 4) which is

due to the antisymmetric stretching vibration of the CQO

bonds. At the same time we observed a noticeable intensity

reduction of the infrared and Raman bands assigned in the

precedent section to the stretching and bending vibrations of

the carboxylate group. The spectral features are also in agree-

ment with large changes in the electronic structure of the

aromatic moiety, especially the pyrazolium ring. This assess-

ment is supported by the fact that no clear correlation can be

Table 1 Experimental and B3PW91/6-311+G(2d,p) wavenumbers and normal mode descriptions for the more intense infrared and Ramanbands of MI3C

Ira Rama Calc.b I,ir I,Ra Potential energy distribution (PED)3

1657 vs 1630 wm 1662 443 27 100 n (CO2�)

1607 s 1638 54 13 72 n (benzene), 23 n (pyraz)1572 s 1587 6 32 81 n (benzene)1517 sh 1523 36 40 54 n (pyraz)

1513 m 1508 m 1511 16 22 29 d(pyraz), 27 n (benzene),23 n (pyraz)

1484 m 1478 2 19 56 d(CH3), 12 n (pyraz)1474 w 1476 sh 1460 44 19 34 d(CH3), 23 d(CH),

22 n (benzene), 20 n (pyraz)1443 m 1449 16 9 88 d(CH3)

1434 w 1436 w 1438 30 51 80 d(CH3)1405 s 1409 m 1391 56 5 74 d(CH3)1379 w 1377 w 1384 7 15 34 n (pyraz), 32 d(CH), 24 n (benzene)1363 wm 1363 s 1373 17 22 90 n (benzene), 17 n (pyraz)1341 s 1342 s 1355 46 16 52 n (pyraz)1320 vs 1317 m 1322 246 54 71 n (CO2

�)1246 m 1247 w 1234 50 2 30 d(CH), 24 n (pyraz)1194 wm 1195 m 1187 11 27 27 d(N–CH3), 27 n (benzene)1163 w 1168 w 1155 7 1 59 d(CH)

1141 wm 1132 0 1 92 r(CH3)1130 w 1124 w 1125 7 5 53 d(CH)

1103 w 1100 3 16 52 r(CH3)1102 w 1084 12 8 52 r(CH3), 24 n (pyraz)991 wm 999 m 1008 14 23 60 n (benzene)962 w 964 w 970 7 7 62 d(benzene), 17 n (benzene)907 wm 908 w 898 14 8 27 n (benzene), 23 d(N–CH3), 22 n (pyraz)860 vw 861 vw 846 2 1 92 g(benzene)816 s 808 w 785 111 5 58 d(CO2

�), 18 n (C–CO2�), 16 d(benzene)

761 s 762 w 744 38 1 100 g(benzene)748 sh 743 m 741 31 2 40 g(benzene),17 g(CO2

�), 13 d(CO2�)

a Wavenumbers in cm�1. Used symbols for relative intensities are: s = strong, m = medium, w = weak, v = very, sh = shoulder. b Potential

energy distribution (values greater than 10%). Symbols used in vibrational coordinates: n (bond stretching), d(aromatic in-plane bending),

g(aromatic out-of-plane bending), r(methyl rocking). Coordinates with the same characters were added to clarify.

Fig. 3 Optimized bond lengths (A) and stretching force constants

(mdyn A�1) for MI3C.

344 | Phys. Chem. Chem. Phys., 2009, 11, 341–348 This journal is �c the Owner Societies 2009

found for most of the pyrazolium vibrational bands, while

those assigned to the benzene ring are easily correlated with

bands appearing at similar wavenumbers. In consequence,

information about the behaviour of the pyrazolium stretching

vibrations will help us to characterize the properties of the

final product of the decarboxylation.

Characterizing of the decarboxylation product

Quantum chemical modelling of the decarboxylation mecha-

nism has been carried out by assuming the reaction pathway in

Scheme 1. Subsequent elimination of CO2 generates a charge

separated state, MI3Y, consisting of the carbon atom pre-

viously connected to the carboxylate group charged negatively

and the adjacent nitrogen with positive charge. This structure

has a neutral or charge compensated resonance form, also

illustrated in Scheme 1, in which a carbene species is formed

on the relevant carbon atom while the charge over the nitrogen

is originated exclusively from its electronegativity and those of

the neighbour atoms.

Electronic and vibrational properties of the initial and final

products of this decarboxylation reaction have been compared.

OBL, SFC and Mulliken charges of MI3Y are shown in Fig. 5.

OBL and SFC exhibit similar values to those obtained

for MI3C (Fig. 3) for the benzene moitey. On the contrary,

significant deviations were found for the pyrazolium ring, which

suggests the weakening of the skeletal bonds. The greater

changes are localized in the N1–N2 bond, whose OBL and

SFC change by 3.6 and 31.2%, respectively. We must emphasize

that all the skeletal SFC of the pyrazolium ring decrease;

however, their associated OBL do not always enlarge. The

N2–C3 OBL varies from 1.331 to 1.320 A while its SFC

decreased by 14.0%. In addition, four C–C bonds of the benzene

moiety have shorter OBL and lower SCF values than the C3–C4

bond of the pyrazolium ring. This unusual result not only

evidences a significant change in the characteristics of this bond,

and more specially in the common atom C3, but the nature of

the bonds of the whole pyrazolium ring are modified. We have

summarized in Table 2 the calculated wavenumbers of both

MI3C and MI3Y in which the contribution of intrinsic stretch-

ing vibrations of the pyrazolium ring is greater than 10%. These

selected values in both molecules are well supported by experi-

mental wavenumbers. At a first glance, we observe two different

correlations of wavenumbers and assignments. However, we can

appreciate that the wavenumbers for MI3Y are lower, as a

whole, than those calculated for MI3C. Lower wavenumbers are

indicative of weaker bonds, which supports the contribution of

the carbene resonant form of MI3Y since it only presents single

bonds, while the charge-separated form displays pseudo-

aromaticity (see Scheme 2). A non-vanishing contribution of

the carbene form in the structure of MI3C is also supported by

the Mulliken charge distribution (Fig. 5). The charge over the

N2 atom is +0.44 a.u., or intermediate between +1 (charge-

separated form) and 0 (carbene form) and even lower than the

charge of the C4 atom, +0.71 a.u. In addition, the charge over

the C3 (the negative center in the first form) is only �0.09 a.u.

Fig. 4 Evolution of the infrared and Raman spectra of MI3C when

going from 20 to 170 1C. The arrows indicate the main spectral

features. The band of the insert belongs to the carbon dioxide

molecules trapped into the potassium bromide pellet.

Fig. 5 Optimized bond lengths (A), stretching force constants (mdyn A�1) and atomic charges (atomic unit) of MI3Y.

This journal is �c the Owner Societies 2009 Phys. Chem. Chem. Phys., 2009, 11, 341–348 | 345

Theoretical study of MI3Y

Fig. 6 shows the calculated dipole moments (DM) and the

electronic density surfaces (EDS) of the two molecules. As

observed, decarboxylation gives rise to a dramatic reduction of

the DM, which goes from 11.85 D in MI3C to 4.50 D in

MI3Y, and to an appreciable change of its direction. The

normalized EDS explain these changes, as the highest electron

density moves from the oxygen to the nitrogen atoms and the

peripheral hydrogen atoms withdraw electron density from the

inner regions of the molecule. It is worthy to comment that, in

spite of the fact the C3 position does not have any attached

hydrogen atom, an considerable electronic density toward the

outer region of the molecule was obtained. This result is

compatible with the presence of two non-bonded electrons

on this position.

We calculated the infrared and Raman spectra of MI3Y,

which are depicted in Fig. 7 and 8 together with the observed

ones of the decarboxylation product. In this case, the good fitting

when comparing the wavenumbers is not always accompanied

by the intensities. The largest deviations are observed in the

methyl bending region (1300–1450 cm�1) of both the infrared

and the Raman spectra. Methyl group polarization, and a

consequent enhancement of its bending infrared bands, can

not be excluded, especially for the N1-attached CH3 group.

For the remaining regions, the profiles of the calculated and

experimental spectra present a better similarity.

In the gas phase, where the previous discussed calculations

were made, the charge separated form could contribute less than

the carbene form. This fact might be solvent dependent, so that

the whole theoretical treatment of MI3Y was also performed

using a water-like PCM model. By comparing with the results

for the isolated molecule, minimal differences were obtained

for OBL and SFC values. The most noticeable deviation was

found for the charges over the bonded atoms C3 and N2, which

are now�0.34 and+0.32 a.u. This balances the charge separation

around this bond and, simultaneously, increases the contribu-

tion of the first resonant form in MI3Y. The calculated PCM

infrared and Raman spectra, included in Fig. 7 and 8, introduce

subtle improvements with respect to those obtained in gas phase,

which are largely due to minor changes in the calculated

Fig. 6 Electrostatic potential surfaces and dipole moments of MI3C

and MI3Y.

Fig. 7 Comparison between the theoretical B3PW91/6-311+G(2d,p)

and the experimental infrared spectra of MI3Y.

Table 2 B3PW91/6-311+G(2d,p) wavenumbers of MI3C and MI3Yin which the pyrazolium stretching modes have contributions greaterthan 10%

MI3Ca n (py)b Main assign.c MI3Ya n (py)b Main assign.c

1638 (1607) 23 n (benz) 1628 (1624) 16 n (benz)1523 (1517) 54 n (pyraz) 1494 (1516) 19 d(CH)1511 (1513) 23 n (pyraz) 1483 (1482) 24 d(CH3)1478 (1484) 12 d(CH3) 1436 (1446) 28 d(CH3)1460 (1474) 20 d(CH3) 1373 (1382) 32 n (pyraz)1384 (1377) 34 n (pyraz) 1337 (1331) 24 n (benz)1373 (1363) 17 n (benz) 1231 (1250) 67 n (pyraz)1355 (1341) 52 n (pyraz) 1203 (1196) 30 n (pyraz)1234 (1247) 24 d(CH) 1032 (1014) 33 n (pyraz)1084 (1102) 24 r(CH3) 889 (907) 25 d(benz)898 (907) 22 n (benz) 834 (835) 46 n (pyraz)

711 (712) 38 n (pyraz)a Wavenumbers in cm�1. The values in parentheses are the observed

wavenumber, either infrared or Raman bands, having the highest

intensity. b Contribution, in terms of PED, of the stretching modes

of the pyrazolium bonds C4–C3, C3–N2, N2–N1 and N1–C5.c Vibrational coordinates which have the greater contribution to the

PED. See footnotes of Table 1 for symbols.

Fig. 8 Comparison between the theoretical B3PW91/6-311+G(2d,p)

and the experimental Raman spectra of MI3Y.

346 | Phys. Chem. Chem. Phys., 2009, 11, 341–348 This journal is �c the Owner Societies 2009

intensities. The wavenumbers listed in Table 2 for MI3Y are

similar for PCM and gas phase. A noticeable improvement was

only obtained for the observed band at 1250 cm�1, which was

now calculated at 1254 cm�1. For the rest of experimental

wavenumbers deviations are within �5 cm�1.

Taking into account the presence of two non-bonding

electrons in MI3Y, we have investigated the first triplet state

of this molecule. We calculated a S0–T1 gap resulting to be

47.3 kcal mol�1, which is similar to the reported value using the

B3LYP functional and the 6-31G(p) basis.11 This relatively high

exchange energy is justified taking into account the stabilization

energy that aryl groups add to the electronic states of carbenes,

which is greater by several kcal mol�1 for singlet than for triplet

states.31 On the other hand, calculated open-shell spin densities

of MI3Y indicate that the contribution of resonant forms in

which the non-bonding electrons are unpaired, Scheme 3, are

negligible. We are therefore confident that the ground state of

this molecule is a singlet. The T1 state will have, therefore, a

very low population even at the decarboxylation temperatures

higher than 400 K. In addition, no appreciable contamination

by states of higher spin multiplicity was predicted for the S0(open shell) state, while a very low spin contamination of 2.029

was calculated for the T1 state. No significant distortions are

therefore expected for the predicted potential energy surfaces

of this molecule, thus supporting the reliability of further

calculations of reaction energy profiles.32

Influence of 5-halogen substitution on the decarboxylation of

MI3C

The above experimental–theoretical comparison supports the

pathway proposed for the decarboxylation reaction and permits

further insight on the dependence of the reaction energetics with

the substitution of the phenyl group. Fig. 9 shows the energy

profiles obtained for the three compounds from the IRC

calculations. Activation energies around 11–12 kcal mol�1 are

rather low in line with the low temperatures, around 100 1C,

measured in the TGA experiments. This finding is also in

agreement with the low values for the SFC of the union bond

and that breaks down in the reaction under study (i.e., lower

than 3 mdyn A�1). Furthermore the theoretical tendency found

for the activation energies (i.e., 5H 4 5F 4 5Cl) correlates

well with the starting temperatures for the CO2 extrusion (i.e.,

116 1C 4 100 1C 4 94 1C), such as previously reported,15 and

with the SFC provided in this study (i.e., 2.83 mdyn A�1 42.82 mdyn A�1 4 2.80 mdyn A�1). Fig. 9 also shows that the

energy ordering relative to the substituted atom is reversed

when going from activation energies to reaction energies. In

other words, the same atom (Cl) induces the more stable TS and

the less stable final product.

In order to give an explanation of these energies we

calculated NICS over the benzene moiety of the three molecules.

A negative NICS value is a measure of local aromaticity in

polycyclic systems,29 and they are being widely used to evaluate

this property.33–35 Initially, NICS are refered at the geometric

center of the cycle. It has been demonstrated that in this point

the NICS values can be contaminated by contributions from

electrons other than p (s, lone pairs, core. . .).29 In consequence,

we have scanned over a perpendicular height above the center of

the benzene ring to obtain a better evaluation of the p effect.

Typically, the values calculated near 1.0 A are recommended as

suitable indicators of aromaticity.36 The results obtained for

MI3Y and its 5-fluoro and 5-chloro derivatives are depicted in

Fig. 10. The NICS(1.25) values support the relative energies

shown in Fig. 9. The order 5H o 5F o 5Cl is reached for the

NICS(1.00), although the 5H-5F difference is too small when

comparing with its energy difference. NICS(1.00) was also the

more negative value for MI3Y-5H, in accord with previous

studies on five and six membered rings of condensed aromatic

systems.34 The presence of the halogen changes this behavior,

as the 5F and 5Cl derivatives have their minima at 0.5 A above

the center. This result indicates that the lone-electron pairs of the

Scheme 3

Fig. 9 Energy profile diagram of the decarboxylation of 5-H (blue),

5-F (green) and 5-Cl (red) derivatives of MI3Y.

Fig. 10 NICS values of the benzene ring in 5-H (blue), 5-F (green)

and 5-Cl (red) derivatives of MI3Y calculated as a function of the

vertical height over the geometrical center of the cycle.

This journal is �c the Owner Societies 2009 Phys. Chem. Chem. Phys., 2009, 11, 341–348 | 347

halogen, that do not contribute to the p system, have a non-

vanishing effect at this height. In consequence, the NICS values

undergo a noticeable increase when increasing the distance, an

effect not observed for the 5H derivative.

Typically two main effects have to be considered associated

to the halogen substitution in this molecule: the inductive

(electronegativity) and the resonant (lone pairs) or mesomeric

effects. The first one is the predominant; however it largely

involves the s bond of the attached carbon atom. The Mulliken

charges calculated for the three molecules confirm that the

charges on the C5 and C4 atoms (the common atoms between

benzene and pyrazolium rings) are invariant with respect to the

substitution. On the contrary, the mesomeric effect acts on

the p system. In consequence it has the possibility of reaching

the C3 position through the conjugated p bonds, thus account-

ing for the obtained parameters since there is a large distance

between the breaking bond and the substituted position.

Conclusions

The vibrational spectra of indazolium-3-carboxylate and its

decarboxylated derivative have been interpreted using normal

coordinate analysis based on quantum chemistry calculations.

The CO2 extrusion was monitored by infrared and Raman

spectroscopy. A wide set of structural properties (stretching

force constants, optimized bond lengths, atomic charges,

electrostatic potential surfaces, dipole moments) was calculated

for the decarboxylation product. The influence of the medium

was also investigated. In summary, our results indicate that

both the charge-separated and the N-heterocycle carbene forms

noticeably contribute to the chemical properties of this product.

The effect of Cl and F substitution in the 5-position of the

benzene moiety was also investigated. The decarboxylation

temperatures correlated well with the stretching force

constants of the breaking bond and with the calculated

activation energies for the three derivatives. The analysis of

the NICS also agrees with the experimental results and

allowed us to interpret the observed trend in terms of the

contribution of the halogen lone-pairs to the aromaticity.

Acknowledgements

This work was supported by the Spanish Ministry of Science

and Technology, grant CTQ2006-14987-C02-01. We would

also like to thank to P.A.I., grant P06-FQM-01678, for

financial support. J.C. is grateful to the MEC of Spain for

an I3 position of Chemistry at the University of Malaga.

References

1 A. Schmidt, Adv. Heterocycl. Chem., 2003, 85, 67.2 W. D. Ollis, S. P. Stanforth and C. A. Ramsden, Tetrahedron,1985, 41, 2239.

3 Y. Inobishi, Y. Tsuda, T. Konita and S. Matsumoto, Chem.Pharm. Bull., 1964, 12, 749.

4 M. Elander, L. Gawell and K. Lander, Acta Chem. Scand., 1971,25, 721.

5 E. Breuer and S. Zbaida, Tetrahedron, 1975, 31, 499.6 G. W. Gribble and E. A. Johnson, Tetrahedron Lett., 1987, 28,5259.

7 G. Buchi, R. E. Manning and F. A. Hochstein, J. Am. Chem. Soc.,1962, 84, 3393.

8 F. G. Holliman, J. Chem. Soc. C, 1969, 2514.9 R. B. Herbert and F. G. Holliman, South African Ind. Chem., 1961,15, 233.

10 E. Ali, V. S. Giri and S. C. Pakrashi, Tetrahedron Lett., 1976, 17,4887.

11 A. Schmidt, L. Merkel and W. Eisfeld, Eur. J. Org. Chem., 2005,36, 2124.

12 Atta-ur-Rahman, S. Malik, H. J. Cun-heng and Clardy, Tetrahe-dron Lett., 1985, 26, 2759.

13 cf. Matthaeus, The Holy Bible, New Testament, Math. 23:23.14 A. Schmidt, Curr. Org. Chem., 2004, 8, 653.15 A. Schmidt, B. Snovydovych, T. Habeck, P. Drottboom,M. Gjikaj

and A. Adam, Eur. J. Org. Chem., 2007, 38, 4909.16 A. Schmidt, A. Beutler, T. Habeck, T. Mordhorst and

B. Snovydovych, Synthesis, 2006, 1882.17 A. Schmidt, T. Habeck, L. Merkel, M. Makinen and P. Vainiotalo,

Rapid Commun. Mass Spectrom., 2005, 19, 2211.18 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,

M. A. Robb, J. R. Cheeseman, J. J. A. Montgomery, T. Vreven,K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi,V. Barone, B. Mennucci, M. Cossi,G. Scalmani, N. Rega,G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich,A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz,Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov,G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, 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, C. Gonzalez and J. A. Pople, GAUSSIAN03, Gaussian, Inc., Wallingford, CT, 2004.

19 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.20 J. P. Perdew and Y. Wang, Phys. Rev. B, 1992, 45, 13244.21 P. C. Hriharan and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213.22 T. Clark, J. Chandrasekhar, G. W. Spitznagel and P. V.

R. Schleyer, J. Comput. Chem., 1983, 4, 294.23 S. Miertus, E. Scrocco and J. Tomasi, Chem. Phys., 1981, 55, 117.24 S. Miertus and J. Tomasi, Chem. Phys., 1982, 65, 239.25 S. J. Hickling and R. G. Wooleey, Chem. Phys. Lett., 1990, 166, 43.26 P. Pulay, G. Fogarasi, F. Pang and J. E. Boggs, J. Am. Chem. Soc.,

1979, 101, 2550.27 E. B. J. Wilson, Chem. Phys., 1939, 7, 1047.28 K. Fukui, Acc. Chem. Res., 1981, 14, 363.29 P. V. R. Schleyer, C. Maerker, A. Dransfeld, H. Jiao and

V. E. Hommes, J. Am. Chem. Soc., 1996, 118, 6317.30 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16502.31 H. L. Woodcock, D. Moran, B. R. Brooks, P. V. R. Schleyer and

H. F. Schaefer III, J. Am. Chem. Soc., 2007, 129, 3763.32 C. Sosa and H. B. Schlegel, Int. J. Quantum Chem., 1986, 29, 1001.33 D. Moran, F. Stahl, H. F. Bettinger, H. F. Schaefer III and P. V.

R. Schleyer, J. Am. Chem. Soc., 2003, 125, 6746.34 S. Suzuki, Y. Morita, K. Fukui, K. Sato, D. Shiomi, T. Takui and

K. Nakasuji, J. Am. Chem. Soc., 2006, 128, 2530.35 J. Poater, J. M. Bofill, P. Alemany and M. Sola, J. Org. Chem.,

2006, 71, 1700.36 A. Stanger, J. Org. Chem., 2006, 71, 883.

348 | Phys. Chem. Chem. Phys., 2009, 11, 341–348 This journal is �c the Owner Societies 2009