Long nanowires of degenerate semiconductor InSb in asbestos matrix (wire diameter is around 50 A, length 0.1 -1 mm) were prepared. Electrical conduction of these nanowires is studied over a temperature range 1.5 -350 K. It is found that a zero-field electrical conduction is a power function of the temperature G ∝ T α with the typical exponent α ≈ 4. Current-voltage characteristics of such nanowires are found to be nonlinear and at sufficiently low temperatures follows the power law I ∝ V β . It is shown that the electrical conduction of these nanowires cannot be accounted for in terms of ordinary single-electron theories and exhibits features expected for impure Luttinger liquid. For a simple approximation of impure LL as a pure one broken into drops by weak links, the estimated weak-link density is around 10 3 − 10 4 per cm. Electron-electron correlation effects being negligible in three-dimensional case play the dominant role in one dimension. As a result, the physical properties of a onedimensional (1D) metal are expected to be dramatically different from properties of usual metals with a Fermi liquid of electrons. One of the main consequences of the electron-electron Coulomb repulsive interaction is the decrease of the energy density of states around the Fermi energy. The resulting electron state depends on details of electron-electron interaction. In the absence of the long-range interaction an 1D electron liquid (so-called "Luttinger liquid" [5] (LL)) is formed, whereas longrange Coulomb interaction leads to 1D Wigner crystal [6]. Transport properties of 1D electron systems are the subject of very high interest. It was shown that in Coulomb blockade systems the tunnelling transparency of a barrier vanishes due to electron-electron interaction resulting to the power-law zero-bias anomaly for the conduction [7] (see also [8]). For the particular case of LL similar result was also obtained by more strong methods [2][3][4]9,10]. A role of impurity scattering in the presence of shortrange [4,11] and long-range [10,12] interactions was also intensively studied.Luttinger-liquid-like behaviour for tunnelling into fractional Hall edge states predicted in [9] was recently observed in a GaAs-Al 0.1 Ga 0.9 As heterostructure in quantum Hall regime with the filling factor 1/3 [13]. However each of these edge states have properties of so-called "chiral Luttinger liquid" but not those of LL. They are similar but not identical. Luttinger-liquid behaviour was predicted also for carbon nanotubes [15,16]. Small (below one order of magnitude) voltage and temperature variations of the conductance of single-walled nanotubes consistent with the expectations for LL were observed experimentally [14].In the present paper we report an electric conduction of InSb nanowires in an asbestos matrix. It has been found that it follows the power law over 5 orders of conduction variation and can be described as conduction of a pure LL broken into drops by weak links, the estimated weak-link density being around 10 3 − 10 4 per cm. Natural as...

A careful analysis of certain observations concerning the polaron nature of current carriers in rutile is presented. The criticism of our paper [l] given in [2] is shown to be unfounded.KpuTasecmx a a n i e r a~~f i , Some arguments confirming the polaron nature of current carriers in rutile were given in [l]. I n the two first sections of his paper Klinger [2] enumerates the well-known results of small polaron theory. I n Section 3 (see items 1' to 5') he criticizes the analysis of experimental data on rutile given in [l].We shall prove in this paper that all arguments given in 1' to 5' of [2] are misleading and do not invalidate our conclusions obtained in [l].1.' I n 1' one may read that "at T = const, a('P)(w) N (T N Ni (1 -K ) as well as a@p)(w) IV 0'' l ) , where d i p ) is the absorption coefficient for polarons localized on impurity ions, d S p ) the absorption coefficient for free polarons, Ni the impurity concentration, K the compensation degree, and (T the dc conductivity. It means that free carrier absorption and absorption due t o nonionized impurities are both proportional to (T, and if one takes G to be a measure of concentration it is impossible to distinguish between them a t T = const. Separation of these both types of absorption can be achieved in particular by taking into account their different T-dependence. On page 484 in [2] Klinger also mentions about this possibility. It is very strange indeed that he did not notice that precisely this simple method was used in [l], where it was shown that G ( O ) increases with increasing T (whereas diP)(w) should decrease with increasing T). The simultaneous increase of cr and a(w) with T for 300 O K < T < 600 OK in samples with deeper donor levels is shown in our recent papers [3].2'. I n 2 we read that "for the SP Seebeck coefficient in" pure "samples, y@P) = const should be observed a t T > T, ( 5 100 OK), while the experiment shows this rather a t T 2 300 OK". Herewith Klinger refers to papers [4] to [7].We reproduce Fig. 6 from [7] which clearly shows that y in rutile is independent of T till T = 70 to 100 OK (cf. Fig. 1). Thus the T-dependence of y is just the same as it should be in Klinger's opinion for small polarons. 1)The original text quoted from [2] is put in inverted commas.

No abstract

We show that the conclusion on the breakdown of the standard small polaron theory made recently by E. V. de Mello and J. Ranninger ͓Phys. Rev. B 55, 14 872 ͑1997͔͒ is a result of an incorrect interpretation of the electronic and vibronic energy levels of the two-site Holstein model. The small polaron theory, when properly applied, agrees well with the numerical results of these authors. Also we show that their attempt to connect the properties of the calculated correlation functions with the features of the intersite electron hopping is unsuccessful. ͓S0163-1829͑99͒02014-7͔