CC BY
16
НАУЧНЫЕ СТАТЬИ
ХИМИЯ, ФИЗИКА, МАТЕМАТИКА, ТЕХНИКА
УДК 539.2; 538.9-405; 548 Р. Mancad, G. Mula* P. Sirigu* A.B. Kuz'menko**, E.A. Tischenko**, V.G. Tyuterev***
INFRARED REFLECTIVITY IN OXYGEN-CHAIN-EQUALIZED AND ORDER STABILIZED PAIRS
OF 123 COPPER OXIDES
*INFM and Dipartimento di Fisica, Universita di Cagliary, Cittadella Universitaria **P.L. Kapitza Institute for Physical Problems "Tomsk State Pedagogical University
1. Introduction
The recognition by Jorgensen et al. [1] of the crucial role of the O-Cu-O chain formation for achieving a high-temperature superconducting transition (T) in YBa2Cu306+x, and vanishing of superconductivity when the basal oxygen Ob sites become symmetry equivalent (O-T transition), represents an important step in the understanding of the interrelation between the structure and the superconductivity. A second step, reached by Cava et al. [2], was the recognition of the importance of microscopic oxygen arrangement in determining the physical properties and of the role of the CujOx chain as a charge reservoir which controls the carriers density in the Cu02 planes. The main features of microscopic oxygen arrangement, which are often found to depend strongly upon preparation conditions, have been established in a number of diffraction experiment [3-5] as well as modelled theoretically [6,7]. The holes density induced in the Cu02 layers is connected with the oxygen ordering in the Cuj plane, which results in the enhancement of a conductivity and Tc [1]. This connection is manifested most strongly in the increasing of jT during the room-temperature annealing (aging) of samples produced by fast quenching [1]. The aging-effect has provided a clear experimental basis for the model and experimental investigations on the microscopic origin of the room temperature-increasing Tc in fast quenched samples with the same oxygen content [8]. The room-temperature aging-reordering of oxygen in the Cuflx subsystem has been found to produce a number of interrelated structural and electronic effects.
Useful information about the kinetics of the oxygen ordering in annealing experiments can be provided by the investigation of lattice vibration spectra by the spectroscopic methods [9-11]. Apart from the problems of characterization the investigation of phonon peculiarities in ordered oxygen-vacancy su-
perstructures could contribute to the state of knowledge in the high-temperature superconductivity phenomenon [12—14].
The aim of the present work is to study the oxygen ordering effects through the monitoring of lattice vibration spectra by infrared spectroscopy in a comparative way on oxygen-chain-equalized poly-crystalline pair-samples. The organization of paper is as follows: Section 2.1 is devoted to description of the growth and preparation of oxygen-chain-equalized polycrystalline ¿-pairs. The infrared measurement technique and fitting procedure are described in Sec. 2.2 and Sec. 2.3. We present in Sec. 3 the microscopic lattice dynamics calculations of infrared reflectivity and optical conductivity and discuss in Sec. 4 the interpretation of experiment. Sec. 5 is the conclusion.
2. Experiment
2.1. Sample preparation. Oxygen-chain-equalized deintercalated [OCD]k and intercalated [OCI]k pair-samples (hereafter ¿-pairs) have been prepared by equilibrating the oxygen-chain content of the stoi-chiometry end terms x~l and x~0 of the YBa2Cu3Os+x system [15-16]. The samples [OCD]k started from almost fully doped ortho-I state (x « 0.96) while the other ones [OCI]t where initially in tetragonal T-phase with low oxygen content (0.07). This technique, which is described in details in [15,16], is able to produce pairs of samples with the same oxygen content (6+k) but having non-equivalent oxygen ordering in basal Cu1Ob plane as a result of different initial states. Equivalence in oxygen concentrations was achieved by long time (30-40 days) annealing of a set of [OCD]k and [OCI]k specimens together in a closed volume under vacuum at low temperature (450 °C).
Electron diffraction measurements performed in [16] discovered the evidences of equivalent and non-
equivalent oxygen-ordering configurations and established a range of existence of Oil and OIII superstructures of oxygen-vacancy ordering in /¿-pairs. In the range of jc(0.55-0.72) two principal orthorhom-bic phases are observed: ortho-II(OII) and ortho-Ill (OIII) [16]. The first phase Oil, which is characterized by the «full-empty-full-empty» oxygen structure in the CufOb chains along the ¿-axis [3], is found to be most stable at 0.4<x<0.63; its phase boundary is located around 125-150 °C. The phase OIII corresponds to the «full-full-empty- full-full-empty» oxygen structure [4, 5]. It exists at 0.65<x<0.8 and up to T~75°C.
Resistance measurements, performed with /c-pairs in [15], has shown that the superconducting transition temperature Tc for [OCD]k and [OCI]k differs and it is a function of doping. It was established that [OCI]k become superconducting at systematically higher temperatures than [OCD]k. This effect of Tc splitting [15] has been attributed to a higher degree of chain oxygen ordering in [OCI]k samples [15] and was explained by the non-symmetry of the oxygena-tion-deoxygenation kinetics.
The existence of samples with the same oxygen content and possible difference of a structural ordering provides a good opportunity to investigate the influence of oxygen ordering onto the phonon subsystem through the monitoring of lattice vibration spectra by infrared spectroscopy. It is presented in the next subsection.
Sixteen ceramic samples (8 fc-pairs) with oxygen content x ranging from 0.55 to 0.72 and different preparation history (Table 1) were chosen to investigate their infrared properties. All samples had rectangular shape with dimensions 2x3x12 mm.
Table 1
Oxygen-chain-equalized k-pairs
Oxygen content in equilibrium Deintercalatcd Intercalated
0.55 Л, a2
0.57 B, b2
0.60 C, c2
0.63 D, D2
0.65 Ei f2
0.67 F, e2
0.70 G; g2
0.72 HI H2
The scanning electron microscope (SEM) images [16] show that the objects under investigation are polycrystalline granular samples with 5-30 (im as typical dimensions of randomly shaped irregularities.
2.2. Infrared reflectivity in ceramic samples. Mid-and far-infrared reflectance spectra in the range 30-4000 cm1 at room temperature were obtained
0CDK
Fig. 1. Infrared reflectivity in [OCD], ceramic samples. For a clarity all curves beginning from G, to A1 are successively shifted up to 0.2 from their measured positions
using two grating-type spectrometers with adjacent spectral intervals: Hitachi FIS-3 (30-400 cm1) and Shimadzu IR-460 (400-4000 cm1)- A non-polarized infrared radiation was reflected from a sample surface with an average incidence angle ~15°.The surface was not treated mechanically or chemically prior to measurements. The initial spectrum normalization was performed using a gold mirror as a reference (as is shown below, an additional normalizing procedure is needed in order to account for a diffuse scattering). Spectra measured on two spectrometers were merged, with the high-frequency reflectivity being slightly scaled in order to obtain a smooth united spectrum.
Due to the surface irregularities a certain part of reflected energy is scattered diffusely thus reducing the reflectivity. This effect is especially pronounced for the wavelengths equal to or smaller than the typical size of irregularities. As dimensions of crystallites are distributed from 5 to 30 ¡tim, one could expect a sufficient reduction of the reflectivity due to the diffuse scattering for wave numbers greater then 100 cm*1.
In order to separate the specular reflection we used the rough mirror methodics. It consists of coating a sample with a thin metal layer with a thickness larger than the light penetration depth, but small
I P. Mancd,, G. Mula, P. Sirigu, A.B, Kuz'menko, E.A. Tischenko, KG. Tyuterev. Infrared reflectivity.
2.4 2.2 2.0 1.8 1.6 § 1.4
1.0
0.8 0.6 0.4 0.2
0 200 400 600
Wavenumber (cm4)
Fig. 2. Infrared reflectivity in [OCI],ceramic samples
enough to exactly repeat the surface profile. The ratio between the reflectivity of a non-coated sample and that of coated one yields approximately the specular component of the reflectivity.
It appeared to be very time-consuming to coat a surface and to measure an additional spectrum for all 16 samples so we have coated only one of the samples (Dj) and used it as a «rough mirror» for the normalizing reflectivities of all other specimens. For an additional control we also polished one of the samples mechanically with the diamond 0.03 ¡am abrasive and have measured the reflectivity after. As two methods do not give the identical results one can estimate the systematic error as 20 %. In the range of phonon resonances (700 cm-1) this deviation is even smaller (5 %) because of larger wavelengths. At Fig. 1 the reflectance spectra of [OCD]k (a) and [OCI]k (b) at T=300 K are shown after the normalization by the «rough mirrorc» Dr
A preliminary qualitative interpretation of the reflectivity spectra Fig. 1,2 in ceramics can be made on the base of general peculiarities of vibration modes in the crystal samples [17]. The layered compound YBa2Cu}06+x is a highly anisotropic one. The polarized measurements of infrared spectra on a single crystal showed [17, 18] that the optical characteristics in the aft-plane and along the c-axis are drastically different. The in-plane optical conductivity a ,(©) is largely governed by the Drude component
and the so-called «mid-infrared» interband transition resonances. The IR-active phonon modes, polarized in the «¿-plane, are effectively hidden (screened) in the infrared spectra by free carriers. On the contrary, the out-of-plane optical characteristics are more die-lectric-like with the clearly expressed phonon modes because of the Drude component along the c-axis is 1-2 orders smaller than that of Cu02 plane.
Hence phonons in ceramics samples can be analyzed using the polarized c-axis spectra of the single crystal sample [17]. The following modes are presented: a barium mode Ba-Cu, (147 cm1), an yttrium mode (190 cm1), a bending mode Ob (286 cm-1), a bending mode 02 (324 cm1), and the modes associated with the motion of the apical oxygen 0\ mode 550 cm4, corresponding to 3-fold and 4-fold coordination of CUj and the mode 630 cm4 corresponding to 2-fold (dumbell) Cu} orientation.
It is natural to suggest that a change of concentration of Ob in the basal plane would most strongly affect those modes, which Ob atoms are most actively involved in. The first such vibration is the low-frequency barium mode. From single-crystal data it follows that while x increases from 0.5 to 0.9 its resonance frequency changes from 145 to 153 cm-1. We also observe this effect. In the range of Oil phase x<0.63 this mode splits into two peaks 147 cm-1 and 150 cm"1. It vanishes at x>0.65 (in the range of the OIII phase existence). The apical Oa oxygen mode 550 cm1, which corresponds to 3- and 4-fold coordination of Cuj becomes more intensive and shifts to higher frequencies with x growing. According to the literature [17] its frequency shifts to 556 cm1 in the limit jc—»1. Another apical mode (630 cm1), corresponding to a motion of 2-fold coordinated Cup decreases in its intensity and slightly softens. According to the same single-crystal data, it vanishes at all when x achieves 0.8. At the structural transition to the tetragonal phase this mode increases and in the limit x—>0 the frequency becomes equal to 643 cm1.
The phonon modes mostly sensitive to the oxygen Ob content and to the ordering are the barium mode 147 cm"1 and two apical oxygen modes, associated with 3-4-fold and 2-fold coordination of Cur We have also tracked the evolution of 550 and 630 cm-1 modes as a function of x. With growth of the oxygen doping the intensity increases for the phonon mode corresponding to 3- and 4-fold Cu} coordination and which is important, a sample conductivity simultaneously increases. It confirms the suggestion that 4-fold coordinated Cu1 yields the additional free carrier (holes) to the Cu02 plane. At the same time an intensity decreases for the mode associated with 2-fold Cu] coordination. The frequency of the former mode sufficiently increases; the frequency of the latter - slightly reduces.
Despite the evidences of oxygen superstructure formation are definitely present in the infrared re-
OCD,
flectivity of ¿-pairs, they are apparently smoothed by the polycrystalline disorder. The next subsection offers a methodics to separate the polycrystalline effects and to derive single crystal parameters from a ceramics reflectivity data.
2.3. Extracting of single-crystal parameters: effective media approximation. For the polycrystalline medium composed of randomly oriented grains, some effective averaged conductivity cr should be considered. According to the generalized effective-medium approximation (EMA) [19, 20] the effective conductivity of a set of randomly oriented ellipsoidal crystallites of the same shape is isotropic one and can be obtained by solving the following equation:
g, —a
ab e
+-
1
a -ci
Ъ{\+Ьс)ае+{\~1с)^ 32[(l-I>e+4<xJ
=0, (1)
CD
pDab
CD
pM
ffl +ыууш> a
M
-CD
-mM
,(2)
<OpDc +YjhiPJ (®)
(O2 +i(ùyDc+YJh]Pj{co)
,(3)
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
(1) Exper. polycryst.
(2) Fitted polycryst. -
(3) Model RBb
(4) Model R
where Lc is the depolarisation factor having the values L =0, L =1/3, L= 1 for the cases of needle-like
C C 3 C
grains (along the c-axis), spherical and plate-like ones (parallel to the «¿-plane) correspondingly. Although this equation was derived in [20] for static (zero frequency) current it is assumed to be valid for a dynamical complex conductivity provided that wavelength X is larger than typical grain size d. For the short wavelengths X«d the reflectivity R itself should be averaged over crystallites instead of dynamical conductivity hence one could expect deviations from formula (1) for frequencies higher than 300-400 cm4. Despite its limitations the EMA equation was found to be semiquan-titatively appropriate to describe the electrodynamics properties of anisotropic ceramics [21,22]. So we also used equation (1) for our spectra modelling.
We modelled the in-plane single crystal dielectric function £ab(o)) by composition of the free-carrier Drude term and the mid-infrared Lorentz oscillator, corresponding to interband transitions [23]:
c-axis reflectivity spectrum is determined by Drude component and c-axis phonon modes' contribution. As it is known [17, 24], two high-frequency apical c-polarized phonons have rather non-Lorentzian shape probably due to the strong coupling to the electronic system. To quantitatively describe this effect we used the following dispersion formulas first introduced in [25] for analysis of coupled system of infrared-active phonons and electronic excitations (instead of simple Drude-Lorentz expression):
0 250 500 750 1000 1250 1500 1750 2000
Fig. 3. Experimental reflectivity fitting in the effective media approximation. Curves (1) and (2) present the c-axis and ab-plane reflectivities according to the fitted parameters of Table 2.
Curve (3) is the model polycrystalline reflectivity. Curve (4) is the experimental polycrystalline reflectivity (sample 62)
where Pj(co)=a>iij2/(o}j2-(oj2-iorYJ) is ordinary phonon Lorentzian, h. is a coupling constant for j-th phonon, o /is the oscillator strength.
The determination of single-crystal parameters and depolarisation factor Lc in (2) and (3) was performed by a least-square fitting of the measured polycrystalline reflectivity spectrum in order to minimize a difference between the model and experimental spectra. The parameter Lc strongly affects the shape of the model polycrystalline spectrum. Phonon peaks weaken with Lc decreasing. It means that c-axis phonons like afe-plane phonons are screened by aft-plane free carriers for the case of sphere-like and especially needle-like grains. The model spectrum corresponding to Lc=1, which is the case of plate-like grains parallel to the «¿-plane, is mostly close to the experiment data.
Even for the L= 1 the model reflectivity for co <1000 cm-1 is 10-15 higher than the experimental one. The possible reason is a conductivity reduction in ceramics compared to the single crystals due to the weak links between the crystallites. At high frequencies a correspondence between the absolute values of a reflectivity is better. However, it is likely that this coincidence is an accidental one because, as it was mentioned, the formula (1) is hardly applicable for the case of wavelength X«d ~5-30 jum.
The fitted values of single-crystal dispersion parameters are as follows. In-plane parameters are: = 4.0, vn =631 cm-1, co =32192 cm-1,
' Dab ' pM '
=9032 cm-1. Out-of-plane parame-
JpDc=1328 Cm"'' ?Dc-sults of fitting for phonons modes for the sample G2
that should be considered as a representative one are
shown in Fig. 3 and in Table 2.
Note a wide band at -400 cm"1, which was introduced in order to fit the experimental spectrum properly.
©Bo=3188 cm"1, yDa= 631 cm"1, «„=32192 cm"1 ®m=1537 cm-1, yM=
:1328 cm"1, у„ =1997 cm"1. Re-
ters are: e "=7.0, <o
IP. Manca], G. Mula, P. Sirigu, AB. Kuz'menko, E.A. Tischenko, V.G. Tyuterev. Infrared reflectivity..
Table 2
Effective media approximation parameters of a dielectric dispersion for the sample G, (x=0.7)
j ojj(cms) mp,{ cm"1) yj (cm"1)
1 43.4 348.6 41
2 144.5 295.8 10
3 153.4 305.9 11
4 192.8 198.6 26
5 287.0 239.3 13
6 315.0 546.2 28
7 402.0 634.7 128
8 511.8 306.3 64
9 544.1 323.6 40
10 617.7 247.0 26
s«(rj
<jX
—. (4)
'^(rj-ffl2-/^
Here saael(m) is the electronic background contribution to the dielectric tensor, T is the number of an
5 cr
irreducible representation in the centre of reciprocal space, X is the number of a branch, £\(To) are the vibrations' frequencies for the corresponding representations calculated with the contribution of long-range forces excluded [29], yXa is damping and Sfirj, which should be associated with a in (3), are the oscillator strengths [30]:
(r„ )2 = 4zFcr' \M: (rCT )|\ (5)
where a-component of cell's electric dipole moment is written as
K (rff )= E Kß (ô| \ea)Qß («I )ma
-1/2
(6)
aß
Here m is the ions mass, F is a unit cell volume,
~ a c
) is the effective dynamical charge matrix [29], Qp{a\qir)is the b-component of (/J'j-th phonon
eigenvector at the limit q-»0 of its wave vector, q = q / q and ea are the orths of Cartesian's coordinate system. The D2h! space group symmetry dictates that only non-vanishing components of dipole moment are M;{I\), A/;(P6), M;{1\).
The effective dynamical charge matrix of a-th ion Zaai5 (q) in the shell model can be written as
0(q) =
Zm4¡^Ja,syy^pY(s')
\l/2
Zw
s
V [IV
¿IV
,(7) (8)
UV
3. Theory: dielectric dispersion in YBa2Cu306^x in the shell model
The effect of oxygen ordering to the infrared active vibrations in the high-order phases of YBa2Cu306+x was investigated in [26]. The calculations in the shell model [27] with transferable parameters predict 13 and 19 IR-active phonon modes for a light polarization along c-axis for ortho-II and ortho-Ill phases correspondingly (irreducible representation Tg of a space group Da] in notation of [28]). These results predict noticeable changes in a form of atomic displacements for a polar lattice vibration eigenmodes. The shell model provides the possibility to obtain the ions' dynamical effective charges and so extract the oscillator strengths from the lattice dynamics calculations. The low-frequency dielectric tensor safo)) in an anisotropic crystal in its main axes can be written as
Here Z(a) and Y(s) are bare ion's and shell's charges correspondingly and Eafl(a,s), Weiss') are atom-shell and shell-shell force constant matrices which correspond to the short-range interaction, including the relevant contribution from the Coulomb interaction [29].
The transferable shell model used in [26, 27] actually includes a microscopic metal-like screening of Coulomb forces with a Lindhard-type wave-vector-dependent dielectric function sL{q). It is assumed to be isotropic so formally eaf (q)—>0. Certainly this approach is only a first-order approximation to the real situation with a pronounced anisotropy of conductivity in YBCO samples. However as the model of [27] gives a reasonable values of non-polar phonon frequencies ft/QH^ ) in YBa2Cu301 one can conclude that it adequately reproduces the short-range contribution into the inter-atomic force field. We suppose that Fafa,s) and Tafa,s) in (7) can be taken as a short-range contribution just as they are calculated in the original model [26, 27]. The anisotropy of dielectric properties than can be taken into account phenomenologically in the macroscopic level assuming a finite value of £/'(<») in (7) and (8) while sxxe'(ca) and e e'(co) still are infinite. Henceforth we assumed
yy
the simple Drude's form for c-axis component of the high-frequency dielectric tensor
si{co)=sl-Q.2pl{co2+icorp). (9)
Here Qp (Tn) and yp are the carriers' plasma frequency and it's damping respectively.
The nature of the phonon dissipation processes in YBa2Cu306+x is not studied enough yet so we are not in position to calculate neither yp nor y'-a. So henceforth we assumed s "-1.2, Q =1600 car1, y=1500 cm4
zz 3 p ' *p
for electronic contributions and somewhat arbitrary but reasonable values of yx<T for the lattice contributions taken from the general considerations. The calculated values of transversal frequencies £\(r8) and oscillator strengths Sf for YBa2Cu306+x in the three ortho phases are listed in Tables 3-5.
The comparison of the theoretical OI reflectance with the known experimental data of [17] shows that the: shell model [27] underestimates the value of the
Table 3
Parameters of a dielectric dispersion for YBa2Cu}0? 0Ortho I)
Theory Expt. [17]
n ©л(Г8) SZU\) у*-" ©A ЮрЛ
1 125 118 15 - -
2 168 362 10 153 378
3 174 107 10 - -
4 199 178 15 194 123
5 314 804 70 283 283
6 363 321 35 314 486
7 590 333 20 567 400
Table 4
Parameters of a dielectric dispersion for YBa2Cu}0(5 0Ortho II)
n Theory
ylcr
1 92 18 10
2 118 118 15
3 127 51 10
4 149 356 10
5 172 209 10
6 176 10 10
7 195 102 15
8 250 65 10
9 320 668 70
10 356 480 70
11 563 128 35
12 594 205 20
13 645 188 10
oxygen apical mode OI(7) (hereafter the top index labels the successive number of a frequency mode in the corresponding table). Our frequencies OI(5), OI(6) are too high and keeping in mind the form of atomic displacements [26,27] seem to have a reverse order comparatively to other theoretical calculations. The oscillator strength of OI(6) is highly overestimated. Note in this connection that the details of inter-atomic interaction in Oil and OIII phases of YBa^Cu,0^ are not known and we have used in
2 3 6+x
[26] the shell model of [27] because it gives a set of transferable force field parameters in a general agreement with all phonon database for the great family of cuprates. Certainly it is less good in reproducing the experimental data for any chosen compound, comparing with other versions of shell model, which is the natural price of transferability. However as one can see from Table 3, the model
[27] displays an overall qualitative agreement both for the values of frequencies in OI phase of
Table 5
Parameters of a dielectric dispersion for YBa2Cu}0 (<Ortho III)
Theory Assignment to expt. (Table 2)
ntenr <Ил(А) sz(rs) ^exper
1 79 30 10 -
2 85 12 10 -
3 118 35 15 -
4 125 100 10 -
5 142 153 10 2
6 154 250 10 3
7 170 269 10 3
8 179 36 10 -
9 195 16 10 -
10 199 137 15 4
11 246 6 10 -
12 267 70 10 -
13 315 592 70 6
14 325 419 70 5
15 359 414 30 5
16 399 66 10 7
17 563 116 15 8
18 590 252 15 9
19 648 153 10 10
YBa2Cu307 and oscillator strengths. Hence it seems to be reasonable to use the merit of transferability at least for the qualitative theoretical prediction of infrared properties in Oil and OIII phases.
4. Discussion
As it was shown in [26] the additional long-wave phonon modes of ortho-II and ortho-Ill are originated from the folded back short-wave X-phonons (ortho-II) and A-phonons (ortho-Ill). The perturbation, produced by the relaxation of atomic positions due to the creation of a single oxygen vacancy, changes both the forms of short wave and the intrinsic long wave ortho-I phonon displacements. However the most of X- or A-like phonons are still close in their form of displacements to the short-wave phonons so the oscillator strength are very small (modes Oil0,3'6■8), OIII03-8' 9>n'12' 16>).
Fig. 5 display behaviour of theoretical reflectivity
(Vfi -1) /(4s +1)| for three orthorhombic modifications of YBa,Cu,0^ calculated with the data of Ta-
2 3 6+x
bles 3-5 in comparison with the single-crystal data extracted from a fitting to experimental reflectivity in a ceramic sample (G,) sf. Table 2. The splitting of the apical oxygen mode OI(7) in OH and OIII phases, which is forecasted in [26], clearly manifests itself in the reflectivity. The YBa2Cu3Oflike modes OII<13), OIII(19) appears which correspond to z-displacement
1.0
0.5 -v
0.0
Expt.
0 200 400 600 Wavenumber (cm-1)
Fig. 4. A comparison of calculated reflectivity of YBa2Cu3Os+xin OI, OII and OIII phases for the light polarized in ab-plane with experimental data for G2 sample Fig. 3. Reflectivity curves are shifted up to 0.2, 0.4, 0.6 respectively. The vibration modes in the notation of [26] and Table 5, which are specific for OIII phase, are labelled by circles
of apical oxygen in T-like sub-cell (hereafter a labelling of sub-cells and atoms in a unit cell corresponds to [26]). The modes OII(12 13) and OIII(17' 18) corresponds to z-displacement of apical oxygen only in OI-like subcells and are split-up both in OII and OIII due to the admixture of short-wave modes with out of phase x-displacement of the oxygen in the position O2 with lessening of an oscillator strength mostly noticeable in OII.
In our calculations the smooth peak in the mid-frequency region of OI results from the contributions of modes OI(5) and OI(6). As it was mentioned above, they most probably have a wrong sequence in the accepted model. Modes OII(9) and OII(10) are essentially of the same form and a small admixture of OI(X)-like modes only redistribute slightly the oscillator strengths. In phase OIII mode OI(5) is splitted-up into OIII(13) and OIII(14) with a considerable admixture of the OI(^)-pomt originated anti-phase x-displacements of apical oxygen in OI-like sub-sells and have a noticeable fall of oscillator strengths. The mode OIII(15) still conserves the form and the oscillator strength intrinsic for OI(10). The weak peculiarities OII(8) and OIII(12) correspond to the anti-phase z-displacements of O3 in OI- and T-like sub-cells while OIII(16) is formed by anti-phase z-displacements of O2.
The low-frequency mode OI(2) manifests itself as OII(4) practically without a change in its form while in ortho-III it splits into OIII(6), OIII(7) differing by weights of the same displacements in T- and OI-like subcells. Mode OI(4) manifests itself as OII(7) and OIII(10) without changes of its form. The same is true for OI(1) and OII(2) and OIII(4).
Mode OIII(5) corresponds to (O2+Y)z displacements in the central sub-cell followed by an antiphase Yz in both neighbouring sub-cells. The mode OII(5) corresponds to (O^+Cu^C^^ displacements in the OI-like sub-cell.
As one can see from the comparison of theoretical curves with the experimental single-crystal reflectivity Fig. 3 the sample G2 definitely displays the characteristic features of a high-order phase OIII. The mode ~620 cm-1 should correspond to T-like modes of apical oxygen OII(13), OIII(19). The second high-frequency peak ~546 cm-1 corresponds to the OI-like modes of apical oxygen OII(12), OIII(18) while modes OII(11) and OIII(17) provide its asymmetry form. Modes OII(1-3), OIII(1-6), having small oscillator strengths, probably are not resolved in the experiment and show itself in the asymmetry of a low frequency experimental peak ~150 cm-1. Despite of a considerable difference of calculated and observed oscillator forces the wide peak around ~400 cm-1 can be associated only with the calculated OIII(16).
5. Conclusion
In conclusion, the infrared reflectivity measurements at the oxygen-chain-equalized and order-stabilized polycrystalline k-pair-samples of YBa^C^3O6+x are performed. The generalized mean field approximation is developed to extract the single-crystal data from the infrared reflectivity data of the ceramics samples. The quantitative agreement of experimental results with microscopic lattice dynamics calculations can be considered as an evidence of creation of oxygen-vacancy ordering superstructures in the ceramic [OCD]k and [OCI]k samples. This conclusion strongly supports the interpretation of electro-physical and electron diffraction experiments [15, 16] in the [OCD]k and [OCI]k -pairs from the point of view of asymmetry in the oxygenation-deoxygenation kinetics and the difference of structural ordering.
Acknowledgements
This work was partially supported by the National Research Council (CNR) of Italy under project: «Superconductive and Cryogenic Technologies» and by INTAS Foundation under project n. 93-1707 ext.
References
1. Jorgensen J.D., Veal B.W., Paulikas A.P. et al. // Phys. Rev. B 41. 1990. 1863,
2. Cava R.J., Hewat A.W., Hewat E.A. et al. // Physica C 165. 1990. 419.
— 9 —
3. Plakhty V., Stratilatov A., Chernenkov Yu. et al. // Sol. St. Comm. 84. 1992. 639.
4. Burlet P., Plakthy V.P., Marin C., Henry J.Y. // Phys. Lett. A167. 1992. 401.
5. Grybos J., Hohlwein D., Zeisko T.H. et al. // Physica C 220. 1994. 134.
6. Ceder G., Asta M., de Fontaine D. // Physica C 177. 1991. 108.
7. Semenovskaya S., Khachaturyan A.G. // Phys. Rev. B 54. 1996. 7545.
8. Shaked H., Jorgensen J.D., Hunter B.H. et al. // Phys. Rev. B 51. 1995. 547.
9. Hadjiev V.G., Thomsen C., Kircher Y., Cardona M. // Phys. Rev. B 47. 1993. 9148.
10. Iliev M., Thomsen C., Hajiev V., Cardona M. // Phys. Rev. B 47. 1993. 2341.
11. Fantini S. et al. // Sol. St. Comm. 93. 1995. 519.
12. Zhao G.L., Callaway J. // Phys. Rew. B 50. 1994. 9511.
13. Evarestov R.A., Kitaev Yu.E., Limonov M.F., Panfilov A.G. // Phys. Stat. Sol.(b) 179. 1993. 249.
14. Andersen O.K., Liechtenstein A.I., Rodriguez O. et al. // Physica C 185-189. 1991. 147.
15. Manca P., Sirigu P., Calestani G., Migliori A. // Il Nuovo Cim. 19. 1997. 1009.
16. Calestani G., Migliori A., Manca P., Sirigu P. // Il Nuovo Cim. 19. 1997. 1075.
17. Schutzmann J., Tajima S., Miyamoto S. et al. // Phys. Rev. B 52. 1995. 13665.
18. Orenstein J., Thomas G.A., Millis A.J. et al. // Phys. Rev. B 42. 1990. 6342.
19. Bruggeman D. // Ann. Phys. 24. 1935. 636.
20. Stroud D. // Phys. Rev. B 12. 1975. 3368.
21. Sulewski P.E., Noh T.W., Whirter J.T., Sievers A.J. // Phys Rev. B 36. 1987. 5735.
22. Noh T.W. et al. // Phys. Rev. B 36. 1987. 886.
23. Timusk T., Tanner D.B. Infrared properties of high T superconductors // Physical properties of high temperature superconductors / Ed. D.M. Ginsebrg.1989. World Scientific.
24. Homes C.C., Timusk T., Bonn D.A. et al. // Physica C 254. 1995. 265.
25. Burlakov V.M., Renk K.F., Gaymann A. et al. // Physica C 221. 1994. 269.
26. Tyuterev V.G. et al. // Physica C 297. 1998. 32.
27. Chaplot S.L., Reichart W., Pintchovius L., Pyka N. // Phys. Rev. B 52. 1995. 7230.
28. Kovalev O.V. Irreducible and Induced Representations and Co-representations of Fedorovs Groups. Nauka, 1986.
29. Maradudin A.A. // Dynamical properties of Solids / Ed. G.K. Horton and A.A. Maradudin. 1974. P. 1-82.
30. Skachkov S.I., Tyuterev V.G. / Dep. VINITI. Reg. № 2682-85. 1985.
yflK 541.6
E.L. Kalinina, Yu.A. Shanina, O.Kh. Poleshchuk
STUDY OF THE COORDINATION EFFECTS FOR SnCl4, SbCl5 AND TiCl4 COMPLEXES
4 5 4
ON THE BASE OF AB INITIO CALCULATIONS
Tomsk State Pedagogical University
Introduction
The analysis of the coordination electronic effects, carried out on the base of X-ray fluorescence spectra [1, 2] and PM3 calculations [3] indicated to following. The redistribution of the electron density on donor and acceptor atoms and the donor-acceptor ability were different in the complexes consisting transition or non-transition elements. In these papers we had investigated the complexes of the transition and non-transition metal halogenides with the organic donors.
It seems interesting to calculate the structural and energetic characteristics by ab initio method. On the other hand the redistribution of electron density in these complexes has characteristic differences of transition from non-transition element complexes [1].
The aim of this paper was to obtain the results of ab initio HONDO calculations of SnCl4, SbCl5 and
TiCl4 with H2S complexes. The calculated characteristics of the molecular systems were compared with the experimental data by photoelectron, X-ray fluorescence and X-ray spectra.
Computation details
The ab initio molecular calculations were carried out by the program package HONDO [4] at the Har-tree-Fock (HF) 6-311G** level of theory. The values of the principal components of the diagonalized electric field gradient (EFG) tensors were calculated using IBM RS/6000 workstation running HONDO.
Results and discussion
Table 1 presents the experimental and the optimized bond length values of the acceptors, donor and valent HSH angles of H2S molecule. As follows
— 1 0
