Magnetic and structural features of RNi2B2C and RNiBC (R=Er, Ho, Dy, Tb, Gd) superconducting compounds

Abstract

Temperature dependent Mössbauer spectroscopy on 57Fe doped (1 at % of Ni) RNi2B2C and RNiBC provided clear evidence of a pair-breaking field at the Ni site for non-superconducting compounds. This field is not present in the superconducting collinear AF DyNi2B2C, however it appears when this compound is diluted with non-magnetic Lu (Dy1-xLu xNi2B2C). Important local information on the spin structure of the R magnetic moments is obtained for both systems. The local symmetries of RNi2B2C and RNiBC compounds, measured through the deltaE Q, were found to be linearly correlated with the c<FONT FACE=Symbol>¢/</FONT>a ratio (c´ is the distance of the R-C layers between which the Ni2-B2 layers are sandwiched). A scaling of Tc with the local symmetry (c<FONT FACE=Symbol>¢/</FONT>a) was found to be the same for all the RNi2B2C and RNiBC (except for Y) compounds.

a11v32n3

Magnetic and Structural Features of RNi₂B₂C and RNiBC (R=Er, Ho, Dy, Tb, Gd) Superconducting Compounds

E. M. Baggio-Saitovitch, D. R. Sánchez, and H. Micklitz^*

Centro Brasileiro de Pesquisas Físicas,

Rua Dr. Xavier Sigaud 150, Urca, CEP 22290-180 Rio de Janeiro, Brazil

Received on 28 February, 2002

Temperature dependent Mössbauer spectroscopy on ⁵⁷Fe doped (1 at % of Ni) RNi₂B₂C and RNiBC provided clear evidence of a pair-breaking field at the Ni site for non-superconducting compounds. This field is not present in the superconducting collinear AF DyNi₂B₂C, however it appears when this compound is diluted with non-magnetic Lu (Dy_1-xLu_xNi₂B₂C). Important local information on the spin structure of the R magnetic moments is obtained for both systems. The local symmetries of RNi₂B₂C and RNiBC compounds, measured through the DE_Q, were found to be linearly correlated with the c¢/a ratio (c¢ is the distance of the R-C layers between which the Ni₂-B₂ layers are sandwiched). A scaling of Tc with the local symmetry (c¢/a) was found to be the same for all the RNi₂B₂C and RNiBC (except for Y) compounds.

I Introduction

The interplay between superconductivity and magnetically ordered structures, which is the main reason for the great interest in the RNi₂B₂C compounds, essentially does not exist for the RNiBC family [1-5], since with the exception of LuNiBC [3] none of these compounds has been found to be superconducting. In order to understand why this is the case, despite the fact that the structures of both classes of compounds are very similar, except for an additional R-C layer (next section), we have studied in detail crystallographic structure, magnetic ordering and transport properties of RNiBC (R= Er, Ho, Dy, Tb, Gd) compounds. Especially ⁵⁷Fe Mössbauer Effect (ME) spectroscopy on 1 at % ⁵⁷Fe-doped RNi₂B₂C and RNiBC has been used to study the local structure at the Fe (Ni) site via the nuclear quadrupole splitting and the magnetic order of the R magnetic moments via the transferred magnetic hyperfine field.

II Structure

The similarities between the RNi₂B₂C and the RNiBC structure are illustrated in Figs. 1(a) and (b): while in RNi₂B₂C [ º (RC)₁(NiB)₂] there are alternating layers of R-C and Ni₂-B₂, an additional R-C layer exists between the Ni₂-B₂ layers in the RNiBC [ º(RC)₂(NiB)₂] series. Thus, in both classes of compounds the Ni₂-B₂ layers, which are responsible for superconductivity, are sandwiched between two R-C layers. These Ni₂-B₂ layers are built from NiB₄ tetrahedra with a B-Ni-B bonding angle j as indicated in figs. 1(a) and (b). In RNiBC series the lattice parameters c and d [see fig 1(a)] turned out to be essentially independent of R with c=7.55Å and d=2.41Å [4,5]. In RNi₂B₂C (see Fig. 1(b)) (c/a) = 2(c¢/a), with c¢ being the distance of the R-C layers between which the Ni₂-B₂ layers are sandwiched and a being the lattice parameter in the basal plane [6]. In the RNiBC compounds c¢ is given by c¢ = c-d (Fig. 1(a)).

The |DE_Q| values for RNiBC (R= Er, Ho, Dy, Tb, Gd) and for RNi₂B₂C (R=Er, Ho, Dy, Tb, Gd, Nd, Pr) [7,8] are plotted in Fig. 2 as a function of the structural parameter (c¢/a) which is a measure for the B-Ni-B bonding angle j [see Figs. 1(a) and (b)].

Fig. 2 shows that all compounds, RNi₂B₂C as well as RNiBC, have the same linear correlation between (c¢/a) and |DE_Q|. This clearly indicates that both the crystallographic and the electronic structure (which determine DE_Q) at the Fe(Ni) site in RNiBC and RNi₂B₂C compounds are very similar for equal structural parameter (c¢/a). Furthermore, since |DE_Q| is a measure for the deviation from ideal tetrahedral symmetry, we have to conclude from Fig. 2 that the NiB₄ tetrahedra in RNiBC compounds displays stronger deviation from ideal tetrahedral symmetry than the RNi₂B₂C compounds. Considering the band structure calculations by Mattheiss et al [9], which have shown that ideal tetrahedral symmetry of the NiB₄ tetrahedra happens to coincide with a relative high density of states at the Fermi energy, it becomes evident why the structure of the RNiBC compounds is less favorable for superconductivity than that of the RNi₂B₂C compounds (see section IV).

III Magnetism

HoNi₂B₂C orders antiferromagnetically at T_N = 8.5 K and becomes superconductor at 8 K. An incommensurate modulated magnetic structure appears at ~ 6 K, which seems to be responsible for the reentrant behavior, i.e., the suppression of superconductivity between 4.7 K < T < 6 K (Fig. 3(a)).

The loss of superconductivity in this range of temperature is attributed to the appearance of a pair-breaking field at the Ni layer, where ⁵⁷Fe probe is located. The transferred magnetic hyperfine field B_thf at the Ni (⁵⁷Fe) site, resulting from the four nearest neighbor of Ho atoms (two from the Ho layer above and two from the Ho layer below the Ni plane), will not cancel in the interval 4.7 K < T < 6 K, due to the incommensurate antiferromagnetic structure of HoNi₂B₂C. The B_thf as a function of the temperature (Fig. 3(b)), shows clearly that such a magnetic pair breaking field indeed exists in the reentrant region for the HoNi₂B₂C compound. A pair-breaking field (transferred hyperfine field) was observed at the Ni site and also suppresses superconductivity in TbNi₂B₂C [10] and GdNi₂B₂C [11]. Such kind of pair breaking field is not observed in case of DyNi₂B₂C and ErNi₂B₂C, where superconductivity coexists with magnetic order [12].

The superconductivity in the collinear AF superconductor DyNi₂B₂C relies on the delicate canceling of the ordered magnetic moments on the Ni site. As a consequence, no B_thf was observed at any temperature in this compound. On the other hand, the doping of DyNi₂B₂C with other non-magnetic elements (e.g. Lu), Dy_{1 - x}Lu_xNi₂B₂C, disturb this balance and would yield to the appearing of a net B_thf at the Ni site, such that Lu should act like a magnetic impurity. So, a decreasing of T_c with the Lu doping, in agreement with the Abrikosov-Gor'kov theory, was observed [13]. Furthermore, these elements introduce disorder in the 4f -spin system such that simultaneously T_N also decreases with doping concentration. Low temperature Mössbauer experiments on non-superconducting Dy_0.8Lu_0.2Ni₂B₂C shown that in fact a B_thf exist below T_N and act as pair breaking field. Fig. 4 shows the Bt_hf, as a function of the temperature for Dy_0.8Lu_0.2Ni₂B₂C.

The magnetic ordering temperatures T_M of RNiBC (R=Er, Ho, Dy, Tb, Gd) have been determined by AC susceptibility measurements [5]. We have plotted these values for the different compounds in Fig. 5 together with those of the corresponding RNi₂B₂C and RRh₄B₄ compounds as a function of the de Gennes factor. While the T_M values of the RNi₂B₂C compounds nicely follow a linear relation with the de Gennes factor (see Fig. 5) this is not the case for the RNiBC compounds where T_M for Dy and Tb are even higher than that for Gd (see Fig.7). This finding is very similar to that found for ferromagnetic (FM) ordered RRh₄B₄ compounds where again the highest T_M does not occur for Gd but for Dy (Fig. 5). Crystal field effects have explained this fact which enhance T_M of FM compounds in such a way that T_M of TbRh₄B₄ and DyRh₄B₄ is higher than that of GdRh₄B₄ [14]. As it is shown below, DyNiBC and TbNiBC have a FM structure just below T_c.

From the analysis of magnetoresistance data for RNiBC (R=Er, Ho, Dy, Tb , Gd) [5] one obtains FM coupling for Er, Dy and Tb, corroborated by recently neutron diffraction data [15]. ⁵⁷Fe ME studies below T_M for RNiBC (R=Er, Ho, Dy, Tb , Gd) show a magnetic hyperfine field at the Fe (Ni) nucleus for Tb and Er, again in agreement with neutron and magnetoresistance data. No magnetic hyperfine field, on the other hand, was observed for Ho, Dy and Gd. While this is in agreement with the other experimental data for Ho and Gd, the absent hf field in DyNiBC is a puzzle.

IV Superconductivity

The T_c of the heavy magnetic rare earth RNi₂B₂C has been found to decrease linearly with increasing de Gennes factor (g_J-1)²J(J+1) [16]. This correlation pointed the magnetic pair-breaking as the dominant effect in determine the depression of T_c in magnetic RNi₂B₂C compounds but does not explain the behavior of non magnetic Lu, Y and La. However, a different approach, in which the structural effects seem to play an important role in determining the variation of T_c in the rare earth-nickel-borocarbides (including the magnetic and non-magnetic RNi₂B₂C, and their counterparts RNiBC), will be shown below.

A deviation of the ideal tetrahedral symmetry of the NiB₄ of LuNi₂B₂C can be obtained experimentally doping the Lu site with the nonmagnetic La. Being the ionic radius of the La different from Lu the doping will lead to a change in the (c¢/a) parameter. Because these samples are nonmagnetic, there is no magnetic influence on superconductivity and the depression of T_c will be determined only by structural effects (change of (c¢/a) parameter). The above arguments are also valid for YNi₂B₂C if doped with La. In view of these facts we decided to prepare the Lu_{1 - x}La_xNi₂B₂C and Y_{1 - x}La_xNi₂B₂C alloys.

In order to see if there is any relation between this structural parameter and the superconducting transition temperature T_c we have plotted in Fig. 6 the T_c value for LuNiBC [3] together with those for the non-magnetic RNi₂B₂C (R=Lu, Sc) compounds, the non-magnetic mixtures (Lu_{1 - x}La_x)Ni₂B₂C [17], (Y_{1 - x}La_x)Ni₂B₂C [18] and the magnetically ordered RNi₂B₂C (R=Tm, Er, Ho, Dy) compounds. The non-magnetic LuNi₂B₂C, ScNi₂B₂C, (Lu_{1 - x}La_x)Ni₂B₂C and LuNiBC show a linear relation between (c¢/a) and T_c (see solid line in Fig. 6). The non-magnetic mixtures (Y_{1 - x}La_x)Ni₂B₂C also show a linear relationship between (c¢/a) and T_c (see dashed line in Fig. 6) with a slope which is identical to that for the other non-magnetic compounds. This slope of T_c vs (c¢/a), thus is a measure for the decrease of T_c with increasing (c¢/a).

The T_c values found for (Y_{1 - x}La_x)Ni₂B₂C are higher (for the same (c¢/a) value) than those of RNi₂B₂C family of borocarbides. As mentioned in section II, this feature is likely related with the fact that the Y does not behave as a rare earth in these series of compounds, which is also observed in the Chevrel phases [19].

The YNiBC becomes superconducting when doped with Cu [20], and their T_c vs c¢/a plot follows quite the same trend observed for the Y_{1 - x}La_xNi₂B₂C series (Fig. 9). Despite the fact that in this case there occurs a change in the density of states at the Fermi level it seems that again a structural change is relevant for the change in T_c when the Cu substitutes the Ni in YNiBC.

Quite surprisingly the magnetically ordered RNi₂B₂C (R=Dy, Ho, Er) compounds with T_N close to T_c are on the same line as the non-magnetic compounds. Only T_c of TmNi₂B₂C with T_N << T_c is slightly below the line (see Fig. 6), indicating that there is a clear paramagnetic pair-breaking effect due to exchange scattering in the framework of Abrikosov-Gorkov theory for this compound. Thus, the pair breaking effect of the R magnetic moment seems to be very small for Dy, Ho, Er and it is D T_c » -2 K for Tm. This finding is quite in contrast to the generally accepted opinion that T_c in all magnetically ordered RNi₂B₂C compounds scales with the de Gennes factor [6,16]. However, it is in perfect agreement with very recent theoretical calculations [21] which have shown that the Ni(3d), electrons which are mainly responsible for superconductivity, do not feel the R magnetic moments in the case that the R magnetic moments are antiferromagnetically ordered, i.e. for Dy and Ho.

In the R_1-xR'_xNi₂B₂C pseudoquaternary alloys, where one of them is magnetic (e.g. R = Lu or Y and R' = Gd [11,22]), T_c will decrease according to Abrikosov-Gorkov theory. In this case, the pair breaking effect is more effective in reduction of T_c than the structural effect (change of c¢/a) and cannot be considered in our linear relation of T_c with (c¢/a).

Taking into account what was discussed above, the (c¢/a) parameter (a measure of the distortion of the NiB₄ tetrahedra) seems to be a relevant parameter, which determines T_c in all non-magnetic and antiferro-magnetically ordered RNi₂B₂C and RNiBC compounds. On the other hand, the T_c scale with (c¢/a) ratio and the de Gennes factor in a similar way for the heavy rare earth RNi₂B₂C. However, in the Abrikosov-Gorkov framework is not possible to explain the differences in T_c of the nonmagnetic LuNi₂B₂C, ScNi₂B₂C, YNi₂B₂C, LaNi₂B₂C, and the R-C double layer LuNiBC compound.

The superconducting properties of the RNi₂B₂C phases are attributed to an electron-phonon mechanism [9]. Although the Fermi electrons in these materials have predominant Ni(3d) character, the superconductivity occurs only when a special s-p band is optimally aligned relative to the Fermi level (i.e., when the NiB₄ tetrahedra angles are nearly ''ideal'') and which exhibits strong electron-phonon coupling [9].

Since we expect that the structure by itself can not determine T_c, more work is necessary to elucidate the role of the structural effects on the electron-phonon coupling, which seems to be essential to describe the nature of the depression of T_c in the rare earth-nickel-borocarbides series of compounds.

V Conclusion

⁵⁷Fe Mössbauer spectroscopy on ⁵⁷Fe doped RNi₂B₂C show a transferred hyperfine field at the Fe (Ni) nucleus for the compounds where superconductivity is not observed (TbNi₂B₂C and GdNi₂B₂C) and where the superconductivity disappears in certain region of temperature (HoNi₂B₂C in the reentrant region) and doping (Dy_0.8Lu_0.2Ni₂B₂C, below T_N). This transferred hyperfine field is interpreted as a field originating from the magnetic moments of the neighboring R atoms, which are in a spin configuration different that commensurate antiferromagnetic, and acts as a pair-breaking field at the Ni site. ⁵⁷Fe Mössbauer results in RNiBC indicate a ferromagnetic spin structure for TbNiBC and ErNiBC and antiferromagnetic one for HoNiBC. Electrical field gradient measurements at Fe site show that the NiB₄ tetrahedra in RNiBC displays stronger deviation from ''ideal'' tetrahedral symmetry (as that LuNi₂B₂C) than the RNi₂B₂C compounds. The same linear relation between T_c and (c¢/a) was found for the magnetic as well as the non-magnetic RNi₂B₂C and the RNiBC series, indicating that structural effects may be essential for the establishment of the superconducting state in the rare-earth-nickel-borocarbides.

Acknowledgements

We thank the financial support from FAPERJ/RJ, VW Foundation, Capes/Brazil, CNPq/Brazil and DAAD/Germany D. R. S. thanks Latin American Center for Physics (CLAF) for his Post-doc fellowship.

• [01] T. Siegrist, H. W. Zandbergen, R. J. Cava, J. J. Krajewski, and W. F. Peck Jr. Nature (London) 367, 254 (1994).
• [02] L. J. Chang, C. V. Tomy, D. McK Paul, N. H. Andersen, and M. Yethiraj, J. Phys.: Condens. Matter 8, 2119 (1996).
• [03] L. Gao, X. D. Qui, Y. R. L. Cao, Meng, Y. Y. Sun, Y. Y. Xue, and C. W. Chu, Phys. Rev. B 50, 9445-9448 (1994).
• [04] Q. Huang, J. W. Lynn, A. Santoro, B. C. Chakoumakos, R. J. Cava, J. J. Krajewski, and W. F. Peck Jr, Physica C 271, 311 (1996).
• [05] M. B. Fontes, J. C. Trochez, M. A. Continentino, B. Giordanengo, S. L. Bud'ko, D. R. Sánchez, and E. M. Baggio-Saitovitch, Phys. Rev. B 60, 6781 (1999).
• [06] J. W. Lynn, S. Skanthakumar, Q. Huang, S. K. Sinha, Z. Hossain, L. C. Gupta, R. Nagarajan, and C. Godart, Phys. Rev. B 55, 6584 (1997).
• [07] Z. Zeng, D. R. Sánchez, Diana Guenzburger, D. E. Ellis, E. M. Baggio-Saitovitch, and H. Micklitz, Phys. Rev. B 55, 3087 (1997).
• [08] D. R. Sánchez, H. Micklitz, and E. M. Baggio-Saitovitch, J. Phys. : Condens. Matter 12, L177 (2000).
• [09] L. F. Mattheiss, T. Siegrist, and R. J. Cava, Solid State Commun. 91, 587 (1994).
• [10] D. R. Sánchez, S. L. Bud'ko, E. M. Baggio-Saitovitch, M. Hillberg, W. Wagener, H.-H. Klauss, G. H. Walf, and F. J. Litterst. Phys. Rev. B, 57 10268 (1998).
• [11] D. R. Sánchez, H. Micklitz, and E. M. Baggio-Saitovitch, J. Phys.: Condens. Matter 9, L299 (1997).
• [12] D. R. Sánchez, H. Micklitz, M. B. Fontes, S. L. Bud'ko, and E. M. Baggio-Saitovitch, Phys. Rev. Lett. 76, 507 (1996).
• [13] B. K. Cho, P. C. Canfield, and D. C. Johnston, Phys. Rev. Lett. 77, 163 (1996).
• [14] B. D. Dunlap, L. N. Hall, F. Behroozi, G. W. Crabtree, and D. G. Niarchos, Phys. Rev. B 29, 6244 (1984).
• [15] F. Bourdarot, J. W. Lynn, E. Baggio-Saitovitch, Q. Huang, D. R. Sánchez, and S. Skanthakumar. Phys. Rev. B 64, 104410 (2001)
• [16] P. C. Canfield, B. K. Cho, and K. W. Dennis, Physica B 215, 337 (1995).
• [17] D. R. Sánchez and E. M. Baggio-Saitovitch, to be published.
• [18] D. R. Sánchez, S. L. Bud'ko, and E. M. Baggio-Saitovitch, J. Phys.: Condens. Matter 12, 9941 (2000)
• [19] Ř. Fischer, Appl. Phys. 16, 1 (1978).
• [20] A. K. Gangopadhyay and J. S. Scilling, Phys. Rev. B 54, 10107 (1996) and Cava, R.J. in this proceeding.
• [21] Hyeonjin Doh, Manfred Sigrist, B. K. Cho, and Sung-Ik Lee, Phys. Rev. Lett. 83, 5350 (1999).
• [22] M. El Massalami, S. L. Bud'ko, B. Giordanengo, M. B. Fontes, J. C. Mondragon, E. M. Baggio-Saitovitch, and A. Sulpice, Phys. Stat. Sol. (b) 189, 489(1995).

^*

Permanent address: II. Physikalisches Institut, Universität zu Köln, Zülpicherstr. 77, 50937 Köln, Germany

Publication Dates

• Publication in this collection
  04 Dec 2002
• Date of issue
  Sept 2002

History

• Received
  28 Feb 2002