CC BY
69
ХИМИЯ
УДК 541.49
E.L. Shevchenko*, A.V. Fateev*, A. Schulz**, O.Kh. Poleshchuk*
ANALYSIS OF HYPERFINE INTERACTIONS IN GOLD, COPPER AND SILVER COMPOUNDS
'Tomsk State Pedagogical University **Ludwig-Maximilians-Umversität München
Гибридный метод Беке в сочетании с корреляционным функционалом Ли, Янга и Парра использован для расчета констант квадрупольного взаимодействия серии соединений меди, серебра и золота. Полученные значения сравнивались с данными микроволновой спектроскопии в газовой фазе. Проведен анализ качества расчета с использованием псевдопотенциала и расширенного базисного состояния для соединений меди. 11а основании результатов мёссбауэровской спектроскопии и проведенного расчета оценен вклад различных заселенностей орбиталей атома золота в значения мёссбауэровского химического сдвига.
Introduction
The properties of coordination compounds of d10 monovalent ions continue to receive much attention. Antes et al. [1] investigated the stability, structural properties and electron distribution on complex fonnation in the carbonyl complexes of the group 11 chlorides at the ah initio level using relativistic and non relativistic energy adjusted pseudopotentials for the metal atoms. Structural data and vibrational frequencies were predicted with very good experimental agreement. A MO analysis shows that both metal d- and metal p-contribution aie important in metal ligand bonding in contrast to the interpretation given from Môssbauer data [2]. Microwave spectra of the MXL (M =Au, Ag, Cu; X = Cl, Br) complexes and nuclear quadrupole coupling constants (NQCC) for Cu, Au, Br and Cl were presented by Gerry and coworkers [3-7].
In this work we would like to discuss the calculated quadrupole coupling constants and Môssbauer isomeric shills of the MC1L (M = Cu, Ag, Au) complexes depending on the strength of different Lewis bases L. In our recently published studies [8, 9] compounds of non-transition and transition elements containing tin, antimony, titanium and niobium atoms with several organic ligands have been intensively investigated. Parameters such as quadrupole splitting or quadrupole coupling constant (QCC) of the nuclei ,271,35C1,8lBr, l,5Nb, 12,Sb and U9Sn were determined. Moreover, recently some AuCIL complexes were investigated by NQR and Môssbauer effect analyses and their electronic structure has been discussed depending on the base. The influence of the donor ability of the base on the chemical shift and quadrupole splitting [10, 11J has been modeled.
Many chemical applications of Môssbauer spectroscopy use the sensitivity of the Môssbauer
parameters to investigate changes in the electron density at the nucleus [12]. The isomer shift is a function of both nuclear and electronic properties of the molecular systems, which are combined in such a way that independent quantitative information on both these properties, cannot be obtained by Môssbauer spectroscopy alone. Since the electronic properties are usually of interest and because the nuclear parameters are constant, the hyperfine parameters are most frequently used to compare the electronic properties of different molecules. The covalence effects and the shielding of one set of electrons by another also influences the electronic environment of the nucleus and may be reflected in changes in the isomer shift [13].
The quadrupole splitting (QS) involves a nuclear quantity, the quadrupole moment, and an electronic quantity, the electric field gradient (EFG). The EFG contains a number of different contributions. The principal of these arises from the valence electrons of the Môssbauer atom itself and is associated with asymmetry in the electronic structure. This asymmetry results from partly filled electronic shells occupied by the valence electrons. Another contribution stems from the lattice and arises from the asymmetric arrangement of the ligand atoms. Molecular orbitals can also contribute to the EFG. The effects of these contributions at the Môssbauer nucleus are modified by the polarization of the core electrons of the Môssbauer atom, which may reduce or enhance the EFG. It is very important that the usefulness of Môssbauer parameters for yielding significant information is strongly dependent on whether the nuclear parameters are sufficiently favourable to allow differences in the electronic environment to be reflected in significant and interpretable changes in the spectra.
Two-fold linear coordination is one of the simplest geometries known for metal complexes, and gold (I),
copper (I), and silver (I) are unusual in forming a large number of such complexes. Two different bonding schemes have been proposed for such complexes [2]. The M' is formally «d10 ion, and the simpler explanation supposes donation from the ligands into the empty 6s and 6pz orbitals of the metal atom (the z axis being the chlorine-metal-ligand axis). An alternative scheme, involving the formation of two sdz2 hybrid orbitals (one empty, one occupied), supposes the occupied hybrid to lie in the xy plane, and donation to take place into the empty hybrid orbital directed along the z axis. These two schemes predict di ffcrent trends for the quadrupole splitting of gold: the first predicts charge donation into the p; orbital of a spherical d10 ion, giving an increasing QS with increasing CT-donor power of the ligands. The second model assumes a donation from ligand into the sdrf hybrid orbital and as a result it predicts that the QS will decrease with increasing a-donor power of the ligands. In both models o-acceptance can take place from d^ and dy/ orbitals.
For the benefit of the first model, an increasing quadrupole splitting with increasing a-donor power of the ligand in a number of gold (1) complexes accounts. However, the comparison between quadrupole splitting and proton ability of complexes with ligands such as Cl , CO and PPh, published by Jones et al. [2] displayed an opposite trend. Besides, many results of the calculations of MXCO complexes show that the «p orbital occupations are close to zero.
It is well known that for MOssbauer atoms the magnitude of the isomer shift depends simultaneously on the s and p orbital populations ofthese atoms [ 13]. The multiparameter dependence between the Mossbauer isomer shifts and the s and p populations of antimony, tin and iodine atoms are characteristic for this class of compounds and both one-parameter as well as multiparameter dependencies were investigated in previous works [8,9]. Earlier we have used the multiparameter dependencies for all nuclei considered, which include the direct effect ofthe valencc-shcll s electrons and their shielding of the core electrons, and also the other shielding effects. For various Mossbauer atoms veiy good correlations between isomeric shifts and orbital populations have been found. Thus the main contribution to isomer shift stems from the s orbital population for iodine compounds, but for tin and antimony compounds a considerable contribution arises from the shielding by p orbitals.
Unfortunately interpretation of the NQCC’s and relating them to bonding in the complexes is difficult for transition metal containing species, especially for the heavier metal such as Au, Ag and Cu. The NQCC values for 35C1, 79Br, 63Cu and l,7Au in MXL complexes have been reported and discussed previously [3-7]. Above all we are interested in changes of chlorine and bromine NQCC values upon complex formation. These changes can be expected to be quite large since EFG can be quite sensitive to small changes in electron density around the metal-1 igand centre.
Computational details
All calculations were carried out with the Gaussian 98 program package [14]. For metals we have used the small-core (6s5p3d) Stuttgard-Dresden basis set - relativistic effective core potential combination supplemented by (2fl g) functions for transition metals [15]. The use of all-electron basis set (e.g. 6-31+G(d)) for light atoms is better with respect to accuracy and efficiency. The B3LYP functional with described above basis sets were optimal. We have been used BHandHLYP/6-311-+G(3df,3pd) for copper compounds calculation of NQCC values. The NQCC values were obtained from the principal components of the electric field gradient tensor along the principal axes. The experimental values for the electric quadrupole moments were taken from [16]. Weinhold’s NBO method [17] was used to determine NAO charges of the complexes and for the calculation of the gold orbital populations.
Results and Discussion
The calculated metal-ligand and metal-chlorine distances for all C1ML species arc reported in Table I together with available experimental data. It is well known that the DFT approach tends to underestimate the strength of atom-atom interaction in describing such kind of system. However, in this case a good agreement between theory (B3LYP) and experiment (accurate gas phase structural data for CuCl, CuBr, CuClCO, CuBrCO, AuCl, AuClCO, AgCI and AgClCO are available [3 7]) was found. The same holds for the MP2 results. For example, MP2 values reported by Antes [ I ] and Fortunelli
[18] for Au-Cl bond distance in AuClCO is 2.266 and 2.265 A, respectively, the Au-C bond length is 1.872 and 1.864 A, respectively. MP2 values ofthe analogous silver complex AgClCO [1] are 2.254 A for the Ag-CI and 1.947 A for the Ag-C bond. The estimated B3I.YP data set of this work slightly overestimates (about 0.03-
0.07 A) the bond lengths in comparison with the experimental values. Moreover, the data of Table 1 show an anomalous trend not only for metal-ligand bonds, that was noticed by Antes [I], but also for metal-chlorine ones. It is known that the longest bond distances are usually measured for the silver compound in group 11
[19]. It can be assumed that this anomalous trend can be partly attributed to the different magnitude of relativistic effects along the series Cu, Ag and Au complexes [20]. Since relativistic effective core potentials have been used, the most important relativistic effect has been introduced.
From Table 1 it can also be seen that in most complexes Au-Cl bond distance increases upon complex formation (e.g. about 0.02 A in AuXCo), whereas Ag-CI and Cu-F bond lengths decrease in the corresponding carbonil complexes (AgXCo: -0.02... 0.03 A; CuFCO: - 0.009 A). For the Cu-Cl and Cu-Br carbonil complexes only a very small increase of the Cu-X bond length was found (Cl: 0.005, Br: 0.009 A). Actually, in weakly bound
Table 1
Bond distances in MC.IL compounds calculated by B3LYP/SDD for Au and Ag compounds, and BHandIILYP/6-311 +G(3df,3pd) level for Cu compounds
Compound Rm.x"P p cnl. KM- X R exp. *VW-L
AuCI 2.199 2.268
AuClCO 2.217 2.259 1.884 1.901
AuCIAr 2.198 2.261 2.469 2.655
AuCIKj 2.210 2.264 2.522 2.657
AuBr 2.318 2.390
AuBrCO 2.337 2.318 1.892 1.910
AuCIPMc, [23] 2.308 2.309 2.234 2.273
AgCl 2.281 2.324
AgClCO 2.253 2.284 2.013 2.009
AgCIAr 2.313 2.610 2.756
AgBr 2.393 2.436
AgBiCO 2.373 2.400 2.027 2.022
AgBrAr 2.427 2.640 2.787
CuF 1.745 1.781
CuFCO 1.736 1.758 1.764 1.833
CuFAr 1.773 2.220 2.325
CuCl 2.051 2.105
CuClCO 2.056 2.096 1.796 1.871
CuClAr 2.101 2.270 2.388
CuBr 2.173 2.232
CuBrCO 2.182 2.225 1.802 1.882
CuBrAr 2.228 2.300 2.410
donor acceptor systems such as the MXCO spccies a M-X bond length increase would be expected in all cases upon complex formation. The above-discussed results can only be explained by an underestimation of the donor-acceptor interactions by DFT methods (changing of metal atom hybridisation and redistribution of the atomic charges upon complex formation).
At the same time a comparison of the geometrical parameters calculated by us with the experimental data of the free molecules and complexes displays that the bond lengths have been overestimated. Analysis leads to the following correlation between the calculated and experimental bond lengths for the compounds studied: Rcal =-0.4 + 1.2 Rexp
(r = 0.989, s = 0.04, n= 18) (1)
for Au and Ag compounds and Rc*' =-0.1 + 1.1 Roxr
(r = 0.994, s = 0.03, n = 12) (2)
for Cu compounds.
It is nccessary to note, that these correlations are valid for all compounds studied, in spite of the different environment of the halogen atoms concerned.
It is well known that rotational spectroscopy is a powerful and precise method of determining molecular properties in the gase phase. Traditionally, it has been the source of geometries, force field, electric dipole moments and electric field gradients near certain nuclei. The spectroscopic, and as a consequence molecular, properties obtained ti'om rotational spectra when observed in this way refer to the molecule in isolation and are therefore more appropriate for comparison with the results of ab initio calculations.
The rotational spectra of chemical compounds allow obtaining the hyperfine structure arising from interactions between the electric quadrupole moment ofthe nucleus and liFG [21 ]. Alongside with rotational spectra of NQCC it is possible to obtain at use NQR and MSssbauer spectroscopy data. If any significant variation in electronic structure, as for example in the formation of a new chemical bond, then changes will occur in the EFG’s at the quadrupole nuclei, and these will be reflected in their NQCC’s. At the analysis of changes of NQCC by complex formation we have used in the basic data of rotational specira in a gas phase.
Antes et al. [1] pointed out that chlorine EFG in metal-carbonyl complexes are most sensitive to changes in the molecular environment and even predict correctly the trend in the metal-ligand bond stability. Though the agreement with experiment was excellent for the AuClCO and CuClCO complexes, it was rather poorer for the monomers, so the agreement for the complexes may be somewhat fortuitous. Tables 2 and 3 contain the NQCCs of 55CI and 75Br for a number of Au(l),
Table 2
Available experimental (NQR and FTMW) and calculated NQCC data for Au and Ag compounds
Compound NQCC 35CI, 7,Br exp. NQCC 35CI, 7,Br cal. j
AuCI -61.99 -62.36
AuClCO -36.39 -36.23
AuCI2 ’ -35.0 -34.61
AuCIAr 54.05 57.19
AuClKr 52.01 54.09
AuBr 492.3 483.0
AuBrCO 285.1 284.4
AuClPMe, 28.7' 33.5
AuCIPOMe, 29.4* 33.3
AuClSMej 34.0* 40.0
AuClHy 35.4’ «1-3 . I
AgClCO -28.15 -28.9
AgCl -36.45 -40.0
AgCI2‘ -16.7 -24.3
AgClAr 34.48 38.84
AgBr 297.1 324.8
AgBrCO 223.9 230.7
AgBrAr 278.9 308.8
*NQR data in the assumption of zero asymmetry parameter.
Table 3
Available experimental (NQR and FTMW) and calculated NQCC data for Cu compounds
Compound NQCC 35CI, 7,Br exp. NQCC 35C1, 7,Br cal. NQCC “Cu exp. NQCC 63Cu cal.
CuCICO -21.5 -19.1 70.8 41.4
CuCl -32.1 -27.8 16.2 10.1
CuCIf -19.3 -16.6 61.4 34.9
CuClAr -28.0 -24.9 33.2 17.0
CuBr 261.2 258 12.8 7.8
CuBrf 152.8 147 57.7 33.3
CuBrCO 171.6 179 67.5 39.6
CuBrAr 225.6 232 29.9 15.2
CuF 22.0 17.4
CuFCO 75.4 45.9
CuFAr 38.1 21.9
Ag(l) and Cu(l) species. The NQCCs of some halogen nuclei of related systems are also included in Table 2, which were obtained using NQR. There are many notable points of comparison: theoretically obtained NQCCs are in better agreement with experimental ones in cases of the carbonyl complexes than in the free acceptor as well as those presented in Antes’ study [1]. Probably it is connected to more significant deviation in the metal-halogen bond lengths in acceptors, than in carbonyl complexes.
For complexes of all three metals the magnitudes of the halogen coupling constants decrease very significantly upon complex formation. This is consistent with electron donation from the donor ligand e.g. CO to the MX acceptor fragment. For carbonyl complexes the change is once again considerably larger than the changes on formation of noble gas-MX complexes, although somewhat less than those found for the MX2-ions when X- representing the donor ligand resulting in the formation of an anionic complex. These trends nicely agree with the experimental observations. Also a comparison of theoretically obtained and experimentally observed NQR-35CI frequencies for AuClPMe3, AuCIPOMe3, AuCIPy and AuClSMe., complexes can be found in Table 2. In this case the agreement is rather poor, unless FTMW data that connected with intermolecular interactions in solid state for NQR spectra are used. However the overall trend in the NQR frequency values is the same.
On the basis of these results the following correlation between the experimental (from FTMW [3-7] and NQR [10]) and calculated NQCC values for all halogen atoms was derived:
NQCC(cal.) = - 0.3 + 1.03 NQCC(exp.)
(r = 0.998, s= 10, n= 18) (3)
for Au and Ag compounds and
NQCCc"(cal.) - 0.3 + 1.7 NQCCCu(exp.)
(r = 0.990, s = 3, n = 11) (4)
NQCCx(cal.) = 3.6 + 0.97 NQCCx(exp.)
(r = 0.999, s = 5, n = 8) (5)
for Cu compounds.
It is necessary to note, that such correlation is valid for all metal compounds studied by the same method of the calculations.
The calculated NQCC values for Au atoms obtained from our calculations were close to zero and are therefore not shown. This error in the NQCC values of the central atom is apparently due to the polarization of the internal filled orbitals (the Stemheimer effect) [22] that is shown by its influence on the electric field gradients of the central atoms of the complexes when pseudo potentials (such as SDD) are used. In this case the revolving electrostatic potential of the nucleus electric quadrupole moment breaks the spherical symmetry of the closed environment and contributes to the total quadrupole moment. The interaction between valence electrons and this induced quadrupole moment results in a change of the NQCC. Moreover, the effect of the valence electrons contributes to the electric field gradient on the nucleus. The influence of the Stemheimer factor is especially important for ionic crystals, i.e. for compounds with a large degree of bond ionicity. For the investigated complexes we observed a significant ionicity in all donor-acceptor bonds of up to 80 % as estimated by NBO calculations. The relativistic pseudo potential probably does not reproduce the effect of the internal occupied orbitals very well. From our data it is possible to conclude that the NQCC is highly dependent on the quality of the used basis set and the level of the core treatment. At the same time, the use of extended all electron basis set for copper species gives acceptable NQCC values.
Table 4 presents the experimental values of the Mdssbauer isomer shifts [2, 11] and the calculated populations of the 6s-, 6p- and 5d-orbitals of gold atom from NBO approach. It is necessary to note, that 6p orbital of gold atom population is close to zero. Earlier
for various so-called Môssbauer atoms very good correlations between isomeric shifts and orbital populations have been found [8]. For iodine compounds the main contribution to isomeric shift comes from the 5s-orbital population, but for tin and antimony compounds a considerable contribution comes from the shielding by 5p-orbitals. On the basis of our theoretical results the following correlation can be deduced:
8 = UN - 2Nd + 10 (r = 0.962, s - 0.3, n = 11) (6)
that include the direct effect of the valencc-shell s electrons and their shielding of the corc electrons. According to equation (6) it is possible to confirm the conclusion about the greater contribution of the 6s-orbital than 5d-orbital of gold atom to the isomeric shift. Thus chemical bonding in gold compounds is determined basically by s- and to a lower extent by d-orbitals of the central atom.
References
1. Antes J., Dapprich S., Frenking G., Schwerdtfeger P. II Inorg. Chem. 1996. Vol. 35. P. 2089.
2. Jones P.G., Maddock A.G., Mays M J et al. II J Chem. Soc. Dalton Trans. 1977. P. 1434.
3. Walker N.R., Gerry M.C.L. II Inorg. Chem. 2001. Vol. 40. P. 6158
4. Walker N.R., Gerry M.C.L. II Inorg. Chem. 2002. Vol. 41. P. 1236.
5. Evans C.J. et al. II J. Am. Chem. Soc. 2000. Vol. 122. P. 6100.
6. Evans C.J. et al II Inorg. Chem 2001. Vol. 40. P. 6123.
7. Evans C.J., Gerry M.C L. II J. Mol. Spectrosc. 2000. Vol. 203. P. 105.
8 Poleshchuk O.Kh. et al. II Phys. Chem. Chem. Phys. 2000 Vol. 2. P 1877
9. Poleshchuk O.Kh. et al. II2. Naturforsch. 2000. Vol. 55a. P. 271.
10. Jones P.G., Williams A. II J. Chem. Soc. Dalton Trans. 1977. P. 1430.
11. Mössbauer Spectroscopy / Ed. Dickson D.P.E., Berry F.J. Cambridge University Press, 1986.
12. Parish R.V. II Prog. Inorg. Chem 1972. Vol. 15. P. 101
13. Parish R.V. II Coord. Chem. Rev 1982 Vol. 42. P. 1.
14. Gaussian 98 (Revision A.6 and Revision A.7), Frisch M.J., Trucks G W, Schlegel H.B., Scuseria G.E., Robb M.A., Cheesemari J.R.,
Zakrzewski V.G., Montgomery J.A., Stratmann R.E., Burant J.C., Dapprich S., Millam J.M, Daniels A D., Kudin K.N., Strain M.C., Farkas 0„ Tomasi J„ Barone V., Cossi M., Cammi R., Mennucci B., Pomelli C., Adamo C., Clifford S., Ochterski J., Petersson G.A., Ayala P.Y., Cui Q., Morokuma K., Malick D.K., Rabuck A.D., Raghavachari K., Foresman J.B., Cioslowski J., Ortiz J.V., Stefanov B.B., Liu G., Liashenko A., Piskorz P., Komaromi I, Baboul A.G., Gomperts R., Martin R.L., Fox D.J., Keith T., Al-Laham M.A., Peng C.Y., Nanayakkara A., Gonzalez C„
Challacombe M., Gill P.M.W., Johnson B.G., Chen W„ Wong M.W., Andres J.L., Head-Gordon M., Replogle E.S., Pople J.A Gaussian, Inc.,
Pittsburgh PA, 1998.
15. Parr R.G., Yang W. Density-Functional Theory of Atoms and Molecules. Oxford, 1989.
16. Seliger J. Encyclopedia of Spectroscopy and Spectrometry I Ed. J.L. Holmes. L., 2000. P. 1672
17. Glendening E.D., Reed A.E., Carpenter J.E., Weinhold F. NBO Version 3.1.
18. Fortunelli A., Germano G. II J. Phys Chem A 2000. Vol. 104. P. 10834.
19. Huber K.P., Herzberg G. Molecular Spectra and Molecular Structure Constants of Diatomic Molecules. N. Y., 1979.
20. Ab initio Methods in Main Group Chemistry, Klapotke T.M., Schulz A., with an invited Chapter of Harcourt R.D. about VB Theory. N. Y. 1998
21. Legon A.C. II Angew. Chem. Int Ed. 1999. Vol. 38. P. 2686.
22. Sternheimer R.M II Phys. Rev. 1954 Vol. 95. P 736
23. Angermaier K. et al. II J. Organomet. Chem. 1994. Vol. 472. P 371.
Table 4
Population analysis of the An atom from NBO approach
Compound N. N„ Nd 8, mm/s, relative to 1,7Au in Pt
AuCI 0.62 0.02 9.85 -1.42
ЛиСІ2- 0.81 0.00 9.77 0.54
AuClCO 0.90 0.03 9.55 1.88
AuClCNPh 0.93 0.03 9.58 2.92
АиСІІЧОМс), 0.97 0.01 9.72 2.63
AuCISMe2 0.88 0.00 9.74 1.26
AuCIFy 0.86 0.01 9.67 1.70
АиСІРСІз 0.86 0.01 9.69 1.80
AuClPMC] 0.92 0.02 9.73 2.61
AuBr 0.65 0.01 9.88 -1.07
Au(CN)2' 1.02 0.01 9.63 3.25
\
