www.elsevier.com/locate/poly

Polyhedron 24 (2005) 1175–1184

Interaction of Na(I), Ni(II) and Cu(II) with 2-cyano-2-(hydroxyimino)acetic acid: Spectroscopic and theoretical studies

Kamilla Malek a, Henryk Kozłowski b, Leonard M. Proniewicz a,c,*

a Faculty of Chemistry, Jagiellonian University, R. Ingardena 3, 30-060 Krakow, Polandb Faculty of Chemistry, University of Wroclaw, F. Joliot-Curie 14, 50-383 Wroclaw, Poland

c Regional Laboratory of Physicochemical Analysis and Structural Research, Jagiellonian University, R. Ingardena 3, 30-060 Krakow, Poland

Received 15 December 2004; accepted 4 April 2005

Available online 23 May 2005

Abstract

Sodium(I) salt as well as nickel(II) and copper(II) complexes of 2-cyano-2-(hydroxyimino)acetic acid (1a) have been prepared and

characterized by infrared, Raman and EPR spectra. Molecular structures of the compounds in the solid state are proposed. The

bidentate 1a ligand chelates the copper and nickel ions through the oxime nitrogen and the carboxyl oxygen atoms to form a trans

bis-complexes. However, two sodium ions are bonded to the deprotonated oximic as well as carboxylic groups. Equilibrium geo-

metries, atomic charges, harmonic vibrational frequencies with potential energy distribution (PED), infrared and Raman intensities

are calculated for all compounds studied here by using the hybrid functional of DFT (B3LYP) with 6-311++G(d,p) and LanL2DZ

basis sets, for sodium and transition metal ions complexes, respectively. The computed properties are compared to the experimental

values.

� 2005 Elsevier Ltd. All rights reserved.

Keywords: Cyanoxime complexes; IR; Raman; DFT; PED; Atomic charges

1. Introduction

The coordination properties of cyanoximes have sig-

nificant impact on their biological activity such as

growth regulatory, antimicrobial or fungicidal proper-ties. For instance, this class of oximes belongs to the

family of Althiomycin antibiotics [1–3]. Also organoanti-

mony(V) cyanoximes have been intensively studied as

potential use in chemotherapy agents with lower toxicity

than Pt(II) and Pd(II) anticancer drugs [4–6].

This group of ligands, with the general formula

HOAN@C(C„N)AR, is characterized by the high acid-

ity that is due to the presence of the cyano group located

0277-5387/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.poly.2005.04.007

* Corresponding author. Tel.: +48 12 663 2253/2064; fax: +48 12 634

0515.

E-mail address: proniewi@chemia.uj.edu.pl (L.M. Proniewicz).

close to the oximic moiety. Deprotonation of the oximic

proton in solution causes formation of stable anions.

This acidity is almost 103–105 greater than the acidity

of oximes with aliphatic or ring substituents at the a-car-bon. Thus, the deprotonated cyanoximes form variouscoordination compounds with different metal ions. The

type of metal ion (s-, p- d-metals) bonded by a

cyanoxime forces firstly a particular complexation struc-

ture. Additionally, H-bonding, conjugation, stabiliza-

tions of certain ligand molecule conformation are

responsible for structures of cyanoximes and their metal

complexes in solution as well as solid state [4–8].

As the extension of our previous studies, we nowpresent a systematic investigation of molecular struc-

tures of sodium(I) salt as well as nickel(II) and

copper(II) complexes of 2-cyano-2-(hydroxyimino)ace-

tic acid (1a, [H2L]). In this ligand, the carboxylic group

1176 K. Malek et al. / Polyhedron 24 (2005) 1175–1184

is adjacent to the oximic a-carbon. Far now, this partic-ular oxime and their complexes with copper(II) and

nickel(II) ions have been studied in solution (potentio-

metric, UV–Vis, EPR studies) only [8]. It is known that

this class of oximes predominantly coordinates to metals

via the nitrogen atom (mainly transition metal ions) orvia the oxygen atom (s-, p-metal ions). Additionally,

deprotonation of the 1a oximic group allows formation

of dimeric species where the nitrogen and oxygen atoms

act as metal donors [8–11].

In the present work, the focus is on elucidating

molecular structures of mentioned-above 1a complexes

in solid state based on molecular spectroscopy methods

(IR, RA and EPR). To give detailed assignment ofvibrational spectra and consequently to solve molecular

structures we applied quantum-mechanical calculations

at the DFT level of theory. Additionally, we performed

potential energy distribution calculations (PED) to show

a detailed description of vibrational modes and atomic

charge distribution to give an explanation of the 1a an-

ionic nature (the APT population analysis).

2. Experimental

2.1. Preparation

2.1.1. Na–1aThe sodium salt of 2-cyano-2-(hydroxyimino)acetic

acid was synthesized by the reported procedure [7]. Ele-mental analysis (C, H, N) was conducted at the Faculty

of Pharmacy of Jagiellonian University according to

standard microanalytical procedures (Anal. Calc. for

Na2C3H0.7N2O3.35: C, 21.9; H, 0.4; N, 17.0. Found: C,

22.2; H, 0.7; N, 16.7%). In the microcrystalline sample,

the sodium-to-1a molar ratio (2:1) was confirmed by a

standard analytical procedure (AAS).

2.1.2. Cu–1a and Ni–1aThe complexes were prepared as described below

using 1:2 molar ratio of metal ions to 1a. An aqueous

solutions of Cu(NO3)2 Æ 2.5H2O and NiCl2 Æ 6H2O

(21.9 and 21.4 mg in 5 ml H2O, respectively) were added

to Na–1a (25 mg) dissolved in water (5 ml). In case of

the Cu–1a complex the clear dark green solution was

stirred (20 min) with heating (�35 �C) and treated drop-wise with 0.5 M HNO3 (to pH 2). The resulting green

solution was still heated and stirred (1 h) then filtered

and concentrated by evaporation at room temperature.

Dark green microcrystalline precipitate began to form

within a week. The procedure for the Ni–1a complex

was analogues. The NiCl2 Æ 6H2O–1a mixture was clear

straw-colored and after treating with 1 M KOH (to

pH 8) the solution became dark red. Finally, after afew days dark red microcrystalline powder was ob-

tained. Despite of co-precipitation of the synthesized

complexes with nitrate and chloride ions, respectively,

elemental analyses indicated clearly that the metal-to-

1a molar ratio is about (1:2).

2.2. Spectroscopy

For FT-Raman measurements, a few milligrams of

the compound was placed in capillary tube and mea-

sured directly; 512 and 3000 scans were collected (with

a resolution of 4 cm�1) for the sodium and the transition

metal ions compounds, respectively. Fourier transform

mid-infrared (FT-MIR, 256 scans) and Fourier trans-

form far-infrared (FT-FIR, 512 scans) spectra were

run in KBr and low molecular weight polyethylene discs,respectively. Resolution was set at 4 cm�1 (MIR) and

2 cm�1 (FIR). Sodium–1a salt FT-Raman spectrum

was recorded on a Bio-Rad step-scan spectrometer

(FTS 6000) combined with a Bio-Rad Raman Accessory

(FTS 40). Excitation at 1064 nm was made by a Spectra-

Physics Topaz T10-106c cw Nd:YAG laser. Raman

spectra of the copper and nickel complexes were col-

lected on a Jobin Yvon spectrometer model T6400equipped with an argon ion laser (excitation at

514.5 nm) and a CCD camera as a detector (Princeton

Instruments). FT-IR spectra were measured on Bruker

(IFS 48) and Bio-Rad (FTS 60V) spectrometers in the

mid and far IR regions, respectively. The accuracy of

the readings was ±1 cm�1.

EPR spectrum of the Cu–1a complex (the microcrys-

talline powder sample) was recorded on a Bruker spec-trometer ELEX SYS X500 [X-band (9.5 GHz)] at

298 K.

3. Computational details

B3LYP [12,13] calculations with 6-311++G(d,p) and

LanL2DZ [14,15] basis sets (for the sodium salt and thetransition metals complexes, respectively) were carried

out using the quantum-mechanical GAUSSIAN-98 set of

programs [16]. Optimum geometries and harmonic fre-

quencies were determined by using SGI 2800 computer

in the Academic Computer Center ‘‘Cyfronet’’ in Kra-

kow and on a Cray SV1ex-1-32 computer in Interdisci-

plinary Center for Mathematical and Computational

Modeling in Warsaw. No symmetry constraints wereimposed during the optimization process. At the

B3LYP/6-311++G(d,p) level a scaling factor of 0.983

was applied to the most harmonic frequencies [except

to m(OH) modes (0.953)] in order to yield the best fit

to the experimental data, as suggested by Michalska

and co-workers [17]. Raman intensities were determined

by RAINT program [18] using values of Raman scatter-

ing activities obtained from the Gaussian outputs. Thepotential energy distributions (PED) were obtained

from the Veda program [19], while atomic charge

K. Malek et al. / Polyhedron 24 (2005) 1175–1184 1177

calculations were performed using APT population

analysis (GAPT, Generalized Atomic Polar Tensor)

[20]. For the latter, the optimized geometries at the

B3LYP/LanL2DZ level of theory were used.

(d)

4. Results and discussion

The 1a complexes can exist as variety of geometrical

isomers, where a metal ion is chelated by the ligand in

the bidentate mode via the carboxylic oxygen and oxi-

mic nitrogen atoms or in the monodentate fashion via

either the carboxylate ion or the oximato fragment

(Scheme 1). Potentiometric and UV–Vis studies havesuggested that the copper complex exists as [CuHL2]

�

species at pH 2, while the nickel complex is present in

the [NiL2]2� form at pH range of 6–10. In these com-

plexes, two ligand molecules are located to each other

in the trans configuration. However, crystallographic

data of Cu(II) complex showed the dimeric structure

[8]. All discussed above molecular structures of 1a com-

plexes are presented in Scheme 1.In order to identify the molecular structures of the 1a

complexes obtained in this work we used theoretical

simulation of their vibrational spectra. It should be

emphasized that this method allows to predict a general

coordination pattern only but it is very valuable in the

case of the lack of crystallographic data. However, this

is more than enough to distinguish between cis and trans

conformation of the metallocomplex with informationwhether coordinated ligands are protonated or not.

Thus, for the purpose of this work, we considered as

models bis-chelate square planar molecules with differ-

ent conformation of the ligand (cis [MHL2]� and trans

M

O

N N

OC

C

C

C

OH O-

OO

NC CN

M

N

O N

OC

C

C

C

OH

ONC

O CN

OH C C

O N

M

O CN

O

M

ON

O

O

NC

O

O

M

NC

NMO

cis bidendate mode

trans bidendate mode

monodendate mode

dimeric mode

Scheme 1.

with the protonated [MH2L2] as well as deprotonated

oximic group [ML2]2�). However, for the sodium com-

plex we assumed that sodium ions can be coordinated

to either oximate or carboxylate moieties [MHL] or to

both of them simultaneously [M2L] (see Scheme 1).

In case of the nickel–1a complex, frequency calcula-tions at equilibrium geometries yield only real values;

hence all models correspond to local minima. The com-

parison of theoretical and experimental IR spectra in the

range of 1000–1800 cm�1 showed (Fig. 1) that [NiL2]2�

model reproduces the best obtained experimental data,

where Ni(II) coordinates to deprotonated 1a ligands in

trans configuration (Fig. 2). The deprotonation of both

ligand molecules is confirmed by the detailed analysis ofthe oximic vibrations as it is discussed below. It is clear

that there has to be a counter cation for the dianionic

form of 2-cyano-2-(hydroxyimino)acetic acid in this

structure. It is known, from our previous works

[22,24], that the type as well as the binding site of a

counterion do not effect significantly experimental and

theoretical vibrational spectra. Hence, in our models

the presence of the counterion is neglected.

(a)

(b)

(c)

Fig. 1. Comparison of theoretical and experimental IR spectra (1000–

1800 cm�1) of the 1a nickel complex: (a) cis [NiHL2]�, (b) trans

[NiH2L2], (c) trans [NiL2]2� and (d) experimental spectrum.

Fig. 2. Model structures of 1a complexes with the atom numbering used in calculations.

1178 K. Malek et al. / Polyhedron 24 (2005) 1175–1184

The same procedure was applied to determinate the

structure of the 1a complex with Cu(II) ions. The di-

meric species can be ruled out based on the EPR spec-

trum typical for a complex with orthorhombicsymmetry (g1 = 2.249, g2 = 2.188 and g3 = 2.051; cf.

Fig. 3). Frequency analysis of trans [CuL2]2� showed

one imaginary value what indicated this model to be a

transition state and excluded it finally from further stud-

ies. Moreover, the differences in experimental IR and

Raman spectra of the nickel and copper complexes evi-

dently indicates that these compounds are not isostruc-

tural. Additionally, the comparison of the calculatedvibrational spectra of the other copper complex models

with experimental IR and Raman clearly indicates that

trans [CuH2L2] form exists in solid state (see Figs. 2

and 4). Also, the presence of the protonated oximic

groups has been established by their positions in the

vibrational spectra (Section 4.3).

Fig. 3. Microcrystalline EPR spectra of the 1a copper complex at the

X band frequency at 298 K.

In case of the sodium–1a, several structures with dif-

ferent binding of this metal ion to the carboxylic and/

or oximic groups were taken into consideration. As men-

tion in Section 2 two sodium atoms are present in thiscomplex of 1a. Furthermore, despite of the elemental

analysis results (water-to-Na–1a molar ratio is 1:3),

two water molecules were added to the model structure

to mimic H-bond presented in experimental IR spectrum

(see Fig. 5). Finally, the comparison of experimental and

simulated vibrational spectra showed that both NOH

and COOH have to be deprotonated to allow coordina-

tion of two sodium ions ([Na2L] Æ 2H2O; cf. Fig. 2).

(a)

(b)

(c)

Fig. 4. Comparison of theoretical and experimental IR spectra (900–

1800 cm�1) of the 1a copper complex: (a) cis [CuHL2]�, (b) trans

[CuH2L2] and (c) experimental spectrum.

Table 1

Selected computed bond lengths [A] for 1a and its complexes with

Na(I), Ni(II) and Cu(II) ions

Bond H2L 6-311

++G(d,p)

[Na2L] Æ 2H2O

6-311++G(d,p)

[NiL2]2�

LanL2DZ

[CuH2L2]

LanL2DZ

C@O 1.199 1.254 1.263 1.236

CAO 1.347 1.298 1.339 1.342

C@N 1.287 1.316 1.357 1.305

NAO 1.359 1.292 1.300 1.393

C„N 1.154 1.157 1.189 1.181

Cox–Ccar 1.506 1.481 1.496 1.559

Cox–Ccya 1.433 1.427 1.423 1.428

MAO 2.256a 1.883b 1.920b

2.197b

2.439b

MAN 2.236 1.935 1.962

Functional groups are denoted as: ox – oximic, car – carboxylic, cya –

cyano groups and M as metal ion. Upper symbols are denoted as:

[H2L] – the fully protonated ligand; [CuH2L2], [Na2L] Æ 2H2O, [NiL2]2�

– applied models of complexes.a The oximic oxygen atom.b The carboxylic oxygens.

K. Malek et al. / Polyhedron 24 (2005) 1175–1184 1179

Geometrical parameters, charge distribution and

vibrational spectra for assigned molecular structures of

1a complexes are discussed below.

4.1. Geometry

Computed bond lengths obtained for structures of

E-isomer [H2L] of 1a and its complexes with metal ions

present here are shown in Table 1. The comparison be-

tween bond length values in all studied complexes with

the uncoordinated protonated 1a molecule shows differ-

ences mainly in ‘‘functional’’ groups coordinating to the

metal ions. The carbonyl bonds in the complexes

lengthen considerably (0.04–0.06 A) after the coordina-tion, whereas the CAO bond involved in metal ion bind-

ing shorten slightly in the transition metal complexes

(�0.01 A) and over 0.05 A in the sodium salt. This re-

sults probably from the bidentate coordination of the

sodium ion by carboxylate group. The observed trend

of changes in bond lengths for this group is typical

for many oxocarboxylic acid oximic complexes

[3,7,21,23,24].The oximic NO and CN bonds in the sodium and

nickel compounds are close to those reported for the

N-coordinated deprotonated oxime group [3–5,23]. This

(a)

(b)

(c)

Fig. 5. Mid-IR spectra (400–4000 cm�1) of (a) the 1a sodium complex,

(b) the 1a copper complex and (c) the 1a nickel complex.

indicates that the CNO� moiety exists in the nitroso-

form (CAN@O). The lengths of the oximic bonds in

the protonated copper complex are longer than those

in the 1a molecule, contrary to the deprotonated com-

plexes, where the C@N bond is elongated and the

NAO bond is shortened upon coordination. The latter

effect was observed for the cis complex of 2-hydroxyimi-

nopropanoic acid [H3CAC(COOH)@NOH, hpa] withCu(II) ion regardless of the ligand protonation state

[22]. Besides, in the 1a copper complex the elongation

of the oximic bonds supports the trans configuration

of the investigated compound like it was found for the

trans nickel hpa complex.

Regarding the cyano group, its bond lengthens upon

the complexation with all metal ions studied here. Inter-

estingly, coordination of the alkali ion causes a slightchange of the C„N length (0.003 A), whereas for tran-

sition metal ions the elongation is rather significant

(�0.03 A). Besides, the deprotonation of the oximic

group causes the decrease of the electron density of

the cyano bond, and consequently, its lengthening.

The B3LYP/LanL2DZ method predicted this bond to

be too long (cf. Table 1), but it is still close enough to

those observed in related cyanoxime complexes (1.13–1.15 A) [3–5,7].

The metal–ligand bond distances were calculated in

the range comparable to other complexes. For sodium

salt of 1a, the distances between sodium and carboxylic

oxygen atoms were calculated at 2.20 and 2.44 A. These

values are consistent with the range of 2.01–2.59 A

found for other oximic complexes with alkali metal ions.

Similarly, the Na–Noximic bond length is 2.24 A, while itstypical value is observed in the 2.22–2.54 A range

[21,25,26]. However, the comparison of calculated

1180 K. Malek et al. / Polyhedron 24 (2005) 1175–1184

metal–ligand bond lengths for proposed structures (trans

[NiL2]2� and [CuH2L2]) indicates the stronger binding of

Ni(II) than Cu(II) ion to 1a. It seems that the transition

metal ion rather than the deprotonation state of the

oximic group is responsible for this effect. As calculated

for trans [NiH2L2] isomer, the metal–nitrogen bondlengths (NiAN = 1.855 A, NiAO = 1.851 A) are still

shorter than respective bonds in [CuH2L2]. All calculated

metal–ligand bond lengths are in the range of the oximic

complexes studied so far [3–5,7,22–26].

4.2. Atomic charges

The atomic charges at all atoms in 1a at the differentprotonation stages and its complexes studied here are gi-

ven in Table 2. The effective charges at the metal ions are

reduced up on binding ligand and are +1.06 [(Cu(II)],

+0.91 and +0.86 [Na(I), for the ion bonded to the car-

boxylate and oximato groups, respectively], +0.72 a.u.

[Ni(II)]. The smallest change of the charge value in the

sodium ion is consistent with the weakest metal–ligand

interaction.As expected, the electron density on the oxygen

atoms in the deprotonated carboxylic group of 1a (de-

noted in Table 2 as [HL]�) is almost equal on both oxy-

gens. The further deprotonation of the oximic group

(see data for [L]2�) causes only a charge increase at

the carboxylic oxygens by � � 0.1 a.u. Upon coordina-

tion of the copper ion, the withdrawal of electronic

charge towards O-atom bonded to the metal ion is ex-pected and observed. The similar distribution of the

electron density on the carboxylic group is found for

the sodium salt. The difference between charges on the

O-atoms for the sodium compound is lower than for

[CuH2L2]. It results from the fact that Na(I) ion coordi-

nates to both oxygen atoms. Interestingly, a reverse

trend is observed for the nickel complex. It appears that

the deprotonation of the oximic group reduces the elec-

Table 2

Computed atomic charges [APT population analysis (GAPT)] of 1a forms a

Atom [H2L] [HL]� [CuH2L2]

Ocarboxylic �0.726 �0.873 �1.010

Ccarboxylic +1.195 +1.199 +1.266

Ocarbonyl �0.678 �0.887 �0.750

Hcarboxylic +0.349

Coximic �0.008 �0.007 +0.210

Noximic +0.251 +0.163 +0.079

Ooximic �0.543 �0.610 �0.521

Hoximic +0.333 +0.254 +0.415

Ccyano �0.003 +0.131 �0.125

Ncyano �0.185 �0.370 �0.094

M +1.059

Upper symbols are denoted as: [HL]� – ligand with the deprotonated carboa Geometry by B3LYP/LanL2DZ for all models.

tron density at the carboxylic O-atom and shifts it to-

wards the Ni-atom.

In the oximic group, the electron density is systemat-

ically shifted toward the oxygen atom. The deprotona-

tion of the carboxylic proton causes an increase of

negative charge at the NOH moiety (see data for[HL]� form of 1a). After coordination of Cu(II), the

negative charge density is accumulated on the N-atom

and partially moved towards the metal ion. Upon the re-

moval of the oximic proton the distribution of electron

density changes significantly. The C and O atoms be-

come strongly negative, whereas positive charge is accu-

mulated on the N atom. This affects the C@N and NAO

bond lengths as discussed above. These changes of thecharge distribution follow the coordination of the so-

dium and nickel ions to the fully deprotonated ligand.

Again, the largest withdrawal of the electronic charge

from the N-atom towards the metal ion is observed in

[NiL2]2�.

It is worth to look at the charge distribution of the cy-

ano group. Upon the deprotonation of the carboxylic

proton, the negative charge of the N-atom increases by0.18 a.u. whereas a charge at the C-atom decreases by

0.13 a.u., due to the withdrawing electron character of

C„N. After coordination of Cu(II), the electron density

flows from the N-atom through the cyano carbon to-

wards the remaining fragment of the ligand. However,

the deprotonation of the oximic proton causes more io-

nic character of the cyano bond than in the [HL]� form.

This group behaves as an electron donating substituentin coordination of sodium and nickel ions. Regardless

the coordination mode and the type of a metal ion,

the same effective charge (�0.08 a.u.) is shifted from

C„N towards the adjacent oximic C-atom. Further-

more, the difference between the effective charges at

the carbon and nitrogen atoms in C„N follows the

trend: [CuH2L2] (0.03 a.u.), [Na2L] Æ 2H2O (0.55 a.u.)

and [NiL2]2� (0.99 a.u.). It worth to note that this differ-

nd its complexes with metal ions [M@Na(I), Ni(II) and Cu(II)]a

[L]2� [Na2L] Æ 2H2O [NiL2]2�

�0.968 �1.229 �0.871

+1.206 +1.433 +1.264

�0.979 �1.052 �0.925

�0.518 �0.413 �0.535

+0.332 +0.376 +0.611

�0.779 �0.792 �0.694

+0.405 +0.169 +0.388

�0.699 �0.381 �0.599

+0.911 +0.724

+0.858

xylic group; [L]2� – the fully deprotonated ligand.

(a)

(b)

(c)

Fig. 6. Far-IR spectra (50–400 cm�1) of (a) the 1a sodium complex,

(b) the 1a copper complex and (c) the 1a nickel complex.

K. Malek et al. / Polyhedron 24 (2005) 1175–1184 1181

ence for models [CuL2]2� and [NiH2L2] are 0.97 and

0.01 a.u., respectively. It clearly indicates that polariza-

tion of the cyano group strongly depends on the depro-

tonation of the oximic group only. That can be traced in

vibrational properties of this group as will be shown

below.

4.3. Vibrational spectra

The calculated and experimental vibrational frequen-

cies and assignments of the characteristic modes for all

investigated complexes are summarized in Table 3 (full

description of the internal coordinates and vibrational

spectra are given in the Supplementary Material, TablesI–VI). Experimental (IR and Raman) spectra of the

complexes are shown in Figs. 5–7. It should be empha-

sized that the simulated spectra refer to the isolated mol-

ecule at 0 K, whereas the assignments are made for the

vibrational spectra of the solid state. Thus, some differ-

ences in the calculated versus experimental frequencies

or intensities are most probably caused by intermolecu-

lar interactions (crystal packing) or by anharmonicity ofvibrations. Vibrational modes which are of great impor-

tance in the determination of discussed molecular struc-

tures are shown below.

4.3.1. m(OH) and d(OH)

The IR spectra of all compounds exhibit broad

absorption bands with maxima at 3456 and 3234 [Na–

1a], 3450 [Cu–1a] and 3430 cm�1 [Ni–1a] (cf. Fig. 5).These are assigned to the stretch vibration of the oximic

OH (in the copper complex) as well as free or weakly H-

bonded hydrated water molecules in the other com-

plexes. Calculations for the applied models predict the

presence of those vibrations at 3720, 3719, 3265 and

2983 cm�1 [mðOHÞH2O] for the sodium salt (cf. Fig. 2)

and at 3496 and 3499 cm�1 [m(OH)ox] for the copper

complex. Additionally, calculations of vibrational spec-tra for Ni and Cu complexes with hydrated waters

showed that mðOHÞH2Oshould appear in the men-

tioned-above experimental range (data not shown).

Moreover, bands at �1620 cm�1 are characterized by

significant values of FWHM resulting from the presence

of bending modes of H2O [d(OH)]. Hence, we can con-

clude that water molecules are not directly coordinated

to the complex structures. Furthermore, the absence ofIR absorptions characteristic for water molecules lo-

cated in the first coordination sphere of a complex con-

firms this assumption (at �900 and �770 cm�1, the

rocking and the wagging vibrations of the OH group,

respectively).

4.3.2. m(C„N)

The stretches modes of cyano bond in the IR and Ra-man spectra of the 1a complexes (cf. Table 3) definitely

indicate the shortening of this bond in the following or-

der: Na–1a, Ni–1a and Cu–1a. The mode frequencies of

the sodium and nickel compounds are typical for theircoordination motifs [3,5]. However, a participation of

the C„N group in the metal binding may be suggested

by the high-shift of the C„N stretch (2242 cm�1, Ra-

man spectrum, see Fig. 7(b)) for the cooper complex

comparing to the latter. Surprisingly, IR and Raman

intensities for this complex are very low, in the contrary

to the other complexes. That is due to the electron den-

sity distribution. As showed above, the smallest polari-zation of the cyano bond is observed for [CuH2L2].

4.3.3. mas(COO) and ms(COO)

A characteristic features of the vibrational spectra of

the 1a complexes is the lowering of frequencies of the

asymmetric carboxylate stretching [mas(COO)] down to

�1610 cm�1. Thus, the change in these mode frequen-

cies can be attributed to the deprotonation of the car-boxylic group as well as the ability to bind the metal

centers. The position of this band in the sodium complex

is typical for the alkali complexes of hpa [21]. That

clearly indicates the absence of a significant influence

of the withdrawing electron cyano group on the coordi-

nation fashion by the carboxylate anion. Furthermore, a

Table

3

Computed[withpotentialenergydistribution(PED,%)]andexperim

entalIR

andRamanfrequencies

(incm

�1)oftheselected

stretchingmodes

forsodium,nickel

andcopper

complexes

of1a

Mode

Na–1a(exp.)

[Na2L]Æ2H

2O

(calc.)

[NiL

2]2�(exp.)

Ni(1a) 2

(calc.)

Cu–1a(exp.)

[CuH

2L2](calc.)

IRRaman

IR/R

aman

IRRaman

IRRaman

IRRaman

IRRaman

m(C„

N)

2209

2208

2275[90]

2226

2229

2210[63]a

2212[62]b

2242

2281[86]a

2281[86]b

m as(COO)

1613

1611

1538[81]

1619

1645

1605[84]a

1614[81]b

1649

1668

1691[64]a

1695[83]b

m s(C

OO)

1383

1381

1359[36]

1152

1160

1126[22]a

1130[36]b

1110

1165

1154[56]a

1143[60]b

1140

1144

1136[30]

m(C@N)

1440

1444

1454[69]

1489

1496

1431[73]a

1423[74]b

1649

1650

1674[48]a

1678[66]b

m(NO)

1214

1211

1260[70]

1252

1263

1269[42]a

1275[45]b

1050

1045

1048[53]a

1038[56]b

m(M–N)

307

294[32]

357

373[46]a

202[12]b

330

344[41]a

156[53]b

m(M–O)

330c

330c

337[46]c

307d

294[23]d

408

356

427[48]a

316[18]b

363

376[39]a

290[41]b

161d

162d

165[19]d

aIn-counter-phase.

bIn-phase.

cTheoxim

icoxygen

atom.

dThecarboxylicoxygens.

(a)

(b)

(c)

Fig. 7. Raman spectra (50–2500 cm�1) of (a) the 1a sodium complex,

(b) the 1a copper complex, (c) the 1a nickel complex.

1182 K. Malek et al. / Polyhedron 24 (2005) 1175–1184

smaller low-frequency shifts of this vibration for 3-d me-tal ions complexes support bidentate coordination of the

2-cyano-2-(hydroxyimino)acetate anion to metal center.

Similar behavior has been found previously for other

transition metals oximates [22]. Additionally, the vibra-

tional spectra of those complexes exhibit the splitting of

the COO stretching vibration. The in-counter-phase

modes are observed in IR whereas the in-phase counter-

parts are present in the Raman spectrum. That is consis-tent with the proposed model where two ligands in trans

configuration coordinate to the metal ion (see Table 3).

Moreover, the deprotonation of the oximic group but

not a metal ion [Cu(II) versus Ni(II)] seems to

be responsible for the lower shift-frequency of the

mas(COO) in the nickel complex. The IR frequency of

the latter is almost the same as for the sodium salt with

the deprotonated C@NO.According to the DFT results, the symmetric stretch-

ing mode [ms(COO)] appears mainly in the range of

1165–1110 cm�1 with its relatively small contribution

in the normal mode (22–60% PED). Merely, the spectra

of Na–1a exhibit additional bands in the typical

frequency range for ms(COO), at �1381 cm�1 with 36%

participation of this vibration.

K. Malek et al. / Polyhedron 24 (2005) 1175–1184 1183

4.3.4. m(C@N) and m(NAO)

The positions of the m(C@N) and m(NO) vibrations in

IR and Raman spectra are dependent to a great extend

on the oxime or nitroso character of this group in a par-

ticular compound and are reflected by the coordination

mode. The vibrational spectra of the sodium and nickelcomplexes exhibit a significant shift of m(C@N) to lower

frequencies, at �1440 and �1490 cm�1, for Na–1a and

Ni–1a, respectively. That clearly indicates that 1a ap-

pears in these complexes as a nitroso-containing anion.

The lower shift of the band of the sodium salt results pri-

marily from a participation in coordination of the oxi-

mic oxygen atom. DFT fully confirms this assignment

(1454 cm�1, 69% PED). The m(C@N) of the nickel com-plex is found in our experiment in the similar range as

for the Ir(III) and Ru(II) complexes of other a-hydrox-yiminocarboxylic acids with the deprotonated oximic

group (1515–1495 cm�1) [23]. On the other hand, the

C@N stretch vibration of the copper complex is assigned

to the band at 1650 cm�1 in the Raman spectrum

[1678 cm�1 (66% PED) calculated]. This frequency indi-

cates that oximic not nitroso group takes part in Cu(II)coordination.

The high-frequency shift of the m(NO) modes in the

vibrational spectra of Na(I) and Ni(II) compounds

confirms additionally that these metal ions are bonded

to the ligand by the nitrogen atom of the nitroso group

(cf. Table 3). B3LYP predicts positions of this mode

very well, with an error less than 4.0%. Additionally,

the simulated spectra of the nickel complex exhibitthe different position of the mode in IR (1269 cm�1)

and Raman (1275 cm�1) spectra, respectively. The

same has been observed in the experimental spectra

and results from the trans ligand conformation. The

difference is equal to 9 cm�1. The most pronounced

feature of vibrational spectra of the copper complex

is in their very low frequencies of the m(NO) vibrations

(1050 and 1045 cm�1 in IR and Raman, respectively)comparing to Na–1a and Ni–1a. Thus, the positions

of these bands are consistent with the protonated trans

structure.

4.3.5. Metal–ligand vibrations

The frequency region below 500 cm�1 provides

information on the metal–ligand stretch vibrations.

The far IR and Raman spectra of the complexes pres-ent here in this particular range are shown in Figs. 6

and 7. The IR absorptions at 307, 330 and 357 cm�1

are assigned to the metal–nitrogen stretching vibration

of Na, Cu(II) and Ni(II) complexes, respectively. The

contribution of this vibration to the normal modes is

less than 50%, as predicted by calculated PEDs (cf.

Table 3). As expected, the coordination by the oximic

nitrogen atom of the s-type metal ions is considerablyweaker. Similar observation has been found for the

metal-carboxylic oxygen stretch modes. Moreover,

the comparison of frequencies of Na–Ooximic with

Na–Ocarboxylic indicates the stronger coordination this

alkali ion by the nitroso moiety than by the carboxyl-

ate anion.

5. Conclusions

The experimental IR and Raman spectra of solid

state supported by the DFT calculations have shown

that the transition metal complexes of 2-cyano-2-

(hydroxyimino)acetic acid have trans bidendate square-

planar geometry with the different protonation state of

the oximic group. This is similar to their structures insolution determined previously by our potentiometric

and spectroscopic studies [8]. However, two sodium ions

are bonded to the carboxylate as well as oximate moie-

ties. Vibrational analysis has clearly indicated the bind-

ing nature of the oximic and carboxylic oxygen atoms

and the oximic nitrogen atom. It also shows the presence

of the nitroso form of the oximic group in Ni–1a and

Na–1a. Furthermore, the latter effects the coordinationstrength in all complexes presented. Namely, this causes

the increase of the stability of the nickel complex com-

paring to the copper one and the binding ability of the

oximic group comparing to the carboxylic one in Na–

1a. The magnitude of all stretch vibrations has been

shown to be well correlated with the atomic charge dis-

tribution. The calculations of atomic charges have sug-

gested that the deprotonation of the carboxylic groupdoes not change significantly the electronic distribution

in the 1a molecule, whereas the removal of the oximic

proton causes the accumulation of the positive charge

on the oximic nitrogen. Moreover, the electron density

is shifted from the metal ion towards the carboxylic oxy-

gens in the copper and sodium complexes in reverse to

Ni–1a. And in the oximic group, the N-atom is an elec-

tron donor for metal ions.

Acknowledgments

The authors thank Academic Computer Centre

CYFRONET of the University of Science and Technol-

ogy in Krakow (grant no. KBN/SPP/UJ/027/1999) and

Interdisciplinary Center for Mathematical and Compu-tational Modeling of Warsaw University in Warsaw

(grant no. G25-8) for the possibility to perform neces-

sary calculations.

Appendix A. Supplementary data

Supplementary data associated with this article can

be found, in the online version at doi:10.1016/

j.poly.2005.04.007.

1184 K. Malek et al. / Polyhedron 24 (2005) 1175–1184

References

[1] H.A. Kirst, E.F. Szymanski, D.E. Doman, J.L. Occolowitz, N.D.

Jones, M.O. Chaney, R.L. Hamill, M.M. Hoehn, J. Antibiot. 28

(1975) 286.

[2] H. Nakamura, Y. Iitaka, H. Sakakibara, H. Umezawa, J.

Antibiot. 27 (1974) 894.

[3] V.V. Ponomareva, N.K. Halley, X. Kou, N.N. Gerasimchuk,

K.V. Domasevich, J. Chem. Soc. Dalton Trans. (1996) 2351.

[4] A.A. Mokhir, K.V. Domasevich, N.K. Dalley, X. Kou, N.N.

Gerasimchuk,O.A.Gerasimchuk, Inorg.Chim.Acta 284 (1999) 85.

[5] K.V. Domasevich, N.N. Gerasimchuk, A.A. Mokhir, Inorg.

Chem. 39 (2000) 1227.

[6] A. Gupta, R.K. Sharma, R. Bohra, V.K. Jain, J.E. Drake, M.B.

Hursthouse, M.E. Light, J. Organomet. Chem. 645 (2002) 118.

[7] T.Y. Sliva, A.M. Duda, T. Glowiak, I.O. Fritsky, V.M. Amir-

khanov, A.A. Mokhir, H. Kozlowski, J. Chem. Soc. Dalton

Trans. (1997) 273.

[8] T.Y. Sliva, A. Dobosz, L. Jerzykiewicz, A. Karaczyn, A.M.

Moreeuw, J. Swia�tek-Kozłowska, T. Głowiak, H. Kozłowski, J.

Chem. Soc. Dalton Trans. (1998) 1863.

[9] Ch.O. Onindo, T.Y. Sliva, T. Kowalik-Jankowska, I.O. Fritsky,

P. Buglyo, L.D. Pettit, H. Kozlowski, T. Kiss, J. Chem. Soc.

Dalton Trans. (1995) 3911.

[10] A.A. Mokhir, E. Gumienna-Kontecka, J. Swia�tek-Kozłowska,

E.G. Petkova, I.O. Fritsky, L. Jerzykiewicz, A.A. Kapshuk,

T.Y. Sliva, Inorg. Chim. Acta 329 (2002) 113.

[11] T.Y. Sliva, T. Kowalik-Jankowska, V.M. Amirkhanov, T.

Głowiak, Ch.O. Onindo, I.O. Fritsky, H. Kozlowski, J. Inorg.

Biochem. 65 (1997) 287.

[12] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.

[13] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.

[14] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299.

[15] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 270.

[16] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A.

Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery,

R.E. Stratmann, 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. Raghava-

chari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, 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, C. Gonzalez, M. Challacombe, P.M.W. Gill, B.

Johnson, W. Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M.

Head-Gordon, E.S. Replogle, J.A. Pople, Gaussian 98 (Revision

A.6), Gaussian Inc., Pittsburgh, PA, 1998.

[17] D. Michalska, W. Zierkiewicz, D.C. Bienko, W. Wojciechowski,

Th. Zeegers-Huyskens, J. Phys. Chem. A105 (2001) 8734.

[18] D. Michalska, R. Wysokinski, Chem. Phys. Lett. 403 (2005) 211.

[19] M.H. Jamroz, Vibrational Energy Distribution Analysis: Veda 3.0

Program, Drug Institute, Warsaw, Poland, 2000.

[20] J. Cioslowski, J. Am. Chem. Soc. 111 (1989) 8333.

[21] A.W. Apblett, G.D. Georgieva, J.T. Mague, Inorg. Chem. 36

(1997) 2656.

[22] K. Malek, PhD dissertation, Jagiellonian University, Krakow,

2003.

[23] R. Bergs, R. Lampeka, C. Robl, W. Beck, Z. Naturforsch. b 49

(1994) 483.

[24] K. Malek, M. Vala, J. Swiatek-Kozlowska, L.M. Proniewicz,

New. J. Chem. 28 (2004) 477.

[25] T. Adatia, J. Chakrabarti, O. Carugo, C.B. Castellani, Polyhe-

dron 15 (1996) 1331.

[26] G.P. Voutsas, K.G. Keramidas, M. Lalia-Kantouri, Polyhedron

15 (1996) 147.