- Email: [email protected]
Spin-lattice relaxation time of Sc metal
Physica B 284}288 (2000) 1708}1709
Spin-lattice relaxation time of Sc metal Haruhiko Suzuki *, Mitsuhiro Nasu , Satoshi Abe , Makoto Hondoh , Dmitrii Tayurskii Department of Physics, Faculty of Science, Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan Department of Physics, Kazan State University, 420008 Kazan, Russia
Abstract The spin-lattice relaxation time ¹ of Sc metal was measured at low temperatures to investigate the nuclear spin ordering of Sc metal. The Korringa constant gives information about the nuclear spin}spin interaction in metals. The reported values of the Korringa constant for Sc by several authors are distributed between 0.09 and 1.6 s K. We measured the temperature dependence, "eld dependence and anisotropy of ¹ of Sc metal. 2000 Elsevier Science B.V. All rights reserved. Keywords: Korringa constant; NMR; Nuclear magnetism
Scandium is an exchange-enhanced Pauli paramagnet which shows a maximum of the susceptibility. The natural abundance of Sc (I" ) is 100%. Since the crystal structure is HCP, the electronic quadrupole interaction between nuclear spins exists. An NQR experiment at very low temperatures by Pollack et al. [1] showed that the ground state is $ in Sc metal with the "rst excited state $ located at 18 lK (390 kHz) above the ground state. Pollack et al. also reported the Korringa constant of 90 ms K which is rather short compared with other reported values at high temperatures in rather high magnetic "eld, 1.6 s K [2,3], 1.1 s K [4]. Usually, the short spin-lattice relaxation time at low temperatures in low magnetic "eld is attributed to magnetic impurities. Small amounts of Fe impurities (19 at ppm) were found to drastically change the properties of scandium metal [5]. The Sc crystal used by Pollack et al. was grown in the same batch at Ames Laboratory as ours. Both specimens, Pollack's and ours, contain 3 ppm Fe. Our specimens contain much smaller Fe impurities than the specimens used in the previous experiments [2}4]. One intrinsic mechanism which gives the short relaxation time is the
* Corresponding author.
spin #uctuation of a correlated electron system [6}8]. However, later work by Pollack [9,10] determined that for their case a novel cross-relaxation mechanism speci"cally related to energy relationships only true at zero "eld was responsible for their fast relaxation times. We measured the temperature dependence, "eld dependence and anisotropy of the Korringa constant of Sc metal. We will compare our results with the previous data. A single crystal used in our experiments was grown in Ames Laboratory. It has a dimension of 3.5; 2.0;27 mm with the long direction paralleled to the a-axis. The relaxation time was measured by using the spin-echo technique. The measured RF frequencies were approximately 5, 7 and 11 MHz. The external magnetic "eld was applied parallel to the a- and c-axis. The temperature range was between 1.5 and 4.2 K. The sample holder with the crystal was immersed in the liquid helium. Nuclear spin energy levels of Sc metal were split due to the nuclear electric quadrupole interaction, so we have to be careful in measuring the spin-echo and analyzing the data. The nuclear spin-lattice relaxation time was measured by recording the recovery of spin-echo intensity at varying delays after saturating comb pulses. In Fig. 1 the values of ¹ ¹ measured for the crystal's c- and a-axis are plotted against temperature. The resonance frequency in this measurement was 11 MHz. Results show the
0921-4526/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 2 6 ( 9 9 ) 0 2 9 7 4 - 9
H. Suzuki et al. / Physica B 284}288 (2000) 1708}1709
Fig. 1. Temperature dependence of the Korringa constants measured parallel to the a-axis (open circle) and the c-axis (closed circle). Dashed lines represent the average values of the data points, respectively.
1709
The spin #uctuation in the correlated electron system gives the magnetic "eld e!ect on the Korringa constant. In the case of the itinerant ferromagnetic compound ZrZn , the magnetic "eld e!ect was observed in the order of several T [6}8]. Our experimental results measured in an applied magnetic "eld between 0.4 and 1.1 T show that the Korringa constant keeps a nearly constant value within experimental error. Any magnetic "eld e!ect is expected to occur at much lower magnetic "elds. We will continue the relaxation measurement in lower magnetic "elds to test for e!ects related to paramagnetic impurities.
References anisotropy of the spin-lattice relaxation time between aand c-axis. The anisotropy of the spin-lattice relaxation time in scandium was "rst observed by Ross et al. [4]. Our observed anisotropy of ¹ ¹ was smaller than the value obtained by them. The nuclear spin-lattice relaxation time of Sc was calculated by Asada and Terakura [11]. They found a smaller anisotropy of ¹ ¹ than our result. Our observed average Korringa constant, 1.18 s K, is much longer than the Cornell group result and nearly the same as the previous values measured at high temperatures [2}4]. This is consistent with the observation in the Cornell work [9,10] that their fast relaxation time depended critically on the zero-magnetic-"eld condition, with even 1 mT applied "elds substantially increasing the relaxation times. At elevated "elds, the characteristic "eld is di!erent among the relaxation mechanisms due to magnetic impurities and to the intrinsic spin #uctuation mechanism. In the case of Fe in Pt, the characteristic "eld was about 100 mT [12].
[1] L. Pollack, E.N. Smith, J.M. Parpia, R.C. Richardson, J. Low Temp. Phys. 87 (1992) 753. [2] T. Umemura, Y. Masuda, J. Phys. Soc. Japan 53 (1983) 932. [3] A. Narath, T. Fromhold, Phys. Lett. 25 A (1967) 49. [4] J.W. Ross, F.Y. Fradin, L.L. Isaacs, D.J. Jam, Phys. Rev. 183 (1969) 645. [5] R.J. Stierman, K.A. Gschneider, T.-W.E. Tsang Jr., F.A. Schmidt, P. Klavins, R.N. Shelton, J. Queen, S. Legvold, J. Magn. Magn. Mater. 36 (1983) 249. [6] M. Kontani, Hioki, Y. Masuda, Solid State Commun. 18 (1976) 1251. [7] T. Moriya, K. Ueda, Solid State Commun. 15 (1974) 169. [8] D.A. Tayurskii, private communication. [9] L. Pollack, E.N. Smith, J.M. Parpia, R.C. Richardson, Phys. Rev. Lett. 69 (1992) 3835. [10] L. Pollack, E.N. Smith, R.C. Richardson, Physica B 194}196 (1994) 343. [11] T. Asada, K. Terakura, J. Phys. F 12 (1982) 1387. [12] W. Wendler, T. HerrmannsdoK rfer, S. Rehmann, F. Pobell, J. Low Temp. Phys. 111 (1998) 99.
Spin-lattice relaxation time of Sc metal Haruhiko Suzuki *, Mitsuhiro Nasu , Satoshi Abe , Makoto Hondoh , Dmitrii Tayurskii Department of Physics, Faculty of Science, Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan Department of Physics, Kazan State University, 420008 Kazan, Russia
Abstract The spin-lattice relaxation time ¹ of Sc metal was measured at low temperatures to investigate the nuclear spin ordering of Sc metal. The Korringa constant gives information about the nuclear spin}spin interaction in metals. The reported values of the Korringa constant for Sc by several authors are distributed between 0.09 and 1.6 s K. We measured the temperature dependence, "eld dependence and anisotropy of ¹ of Sc metal. 2000 Elsevier Science B.V. All rights reserved. Keywords: Korringa constant; NMR; Nuclear magnetism
Scandium is an exchange-enhanced Pauli paramagnet which shows a maximum of the susceptibility. The natural abundance of Sc (I" ) is 100%. Since the crystal structure is HCP, the electronic quadrupole interaction between nuclear spins exists. An NQR experiment at very low temperatures by Pollack et al. [1] showed that the ground state is $ in Sc metal with the "rst excited state $ located at 18 lK (390 kHz) above the ground state. Pollack et al. also reported the Korringa constant of 90 ms K which is rather short compared with other reported values at high temperatures in rather high magnetic "eld, 1.6 s K [2,3], 1.1 s K [4]. Usually, the short spin-lattice relaxation time at low temperatures in low magnetic "eld is attributed to magnetic impurities. Small amounts of Fe impurities (19 at ppm) were found to drastically change the properties of scandium metal [5]. The Sc crystal used by Pollack et al. was grown in the same batch at Ames Laboratory as ours. Both specimens, Pollack's and ours, contain 3 ppm Fe. Our specimens contain much smaller Fe impurities than the specimens used in the previous experiments [2}4]. One intrinsic mechanism which gives the short relaxation time is the
* Corresponding author.
spin #uctuation of a correlated electron system [6}8]. However, later work by Pollack [9,10] determined that for their case a novel cross-relaxation mechanism speci"cally related to energy relationships only true at zero "eld was responsible for their fast relaxation times. We measured the temperature dependence, "eld dependence and anisotropy of the Korringa constant of Sc metal. We will compare our results with the previous data. A single crystal used in our experiments was grown in Ames Laboratory. It has a dimension of 3.5; 2.0;27 mm with the long direction paralleled to the a-axis. The relaxation time was measured by using the spin-echo technique. The measured RF frequencies were approximately 5, 7 and 11 MHz. The external magnetic "eld was applied parallel to the a- and c-axis. The temperature range was between 1.5 and 4.2 K. The sample holder with the crystal was immersed in the liquid helium. Nuclear spin energy levels of Sc metal were split due to the nuclear electric quadrupole interaction, so we have to be careful in measuring the spin-echo and analyzing the data. The nuclear spin-lattice relaxation time was measured by recording the recovery of spin-echo intensity at varying delays after saturating comb pulses. In Fig. 1 the values of ¹ ¹ measured for the crystal's c- and a-axis are plotted against temperature. The resonance frequency in this measurement was 11 MHz. Results show the
0921-4526/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 2 6 ( 9 9 ) 0 2 9 7 4 - 9
H. Suzuki et al. / Physica B 284}288 (2000) 1708}1709
Fig. 1. Temperature dependence of the Korringa constants measured parallel to the a-axis (open circle) and the c-axis (closed circle). Dashed lines represent the average values of the data points, respectively.
1709
The spin #uctuation in the correlated electron system gives the magnetic "eld e!ect on the Korringa constant. In the case of the itinerant ferromagnetic compound ZrZn , the magnetic "eld e!ect was observed in the order of several T [6}8]. Our experimental results measured in an applied magnetic "eld between 0.4 and 1.1 T show that the Korringa constant keeps a nearly constant value within experimental error. Any magnetic "eld e!ect is expected to occur at much lower magnetic "elds. We will continue the relaxation measurement in lower magnetic "elds to test for e!ects related to paramagnetic impurities.
References anisotropy of the spin-lattice relaxation time between aand c-axis. The anisotropy of the spin-lattice relaxation time in scandium was "rst observed by Ross et al. [4]. Our observed anisotropy of ¹ ¹ was smaller than the value obtained by them. The nuclear spin-lattice relaxation time of Sc was calculated by Asada and Terakura [11]. They found a smaller anisotropy of ¹ ¹ than our result. Our observed average Korringa constant, 1.18 s K, is much longer than the Cornell group result and nearly the same as the previous values measured at high temperatures [2}4]. This is consistent with the observation in the Cornell work [9,10] that their fast relaxation time depended critically on the zero-magnetic-"eld condition, with even 1 mT applied "elds substantially increasing the relaxation times. At elevated "elds, the characteristic "eld is di!erent among the relaxation mechanisms due to magnetic impurities and to the intrinsic spin #uctuation mechanism. In the case of Fe in Pt, the characteristic "eld was about 100 mT [12].
[1] L. Pollack, E.N. Smith, J.M. Parpia, R.C. Richardson, J. Low Temp. Phys. 87 (1992) 753. [2] T. Umemura, Y. Masuda, J. Phys. Soc. Japan 53 (1983) 932. [3] A. Narath, T. Fromhold, Phys. Lett. 25 A (1967) 49. [4] J.W. Ross, F.Y. Fradin, L.L. Isaacs, D.J. Jam, Phys. Rev. 183 (1969) 645. [5] R.J. Stierman, K.A. Gschneider, T.-W.E. Tsang Jr., F.A. Schmidt, P. Klavins, R.N. Shelton, J. Queen, S. Legvold, J. Magn. Magn. Mater. 36 (1983) 249. [6] M. Kontani, Hioki, Y. Masuda, Solid State Commun. 18 (1976) 1251. [7] T. Moriya, K. Ueda, Solid State Commun. 15 (1974) 169. [8] D.A. Tayurskii, private communication. [9] L. Pollack, E.N. Smith, J.M. Parpia, R.C. Richardson, Phys. Rev. Lett. 69 (1992) 3835. [10] L. Pollack, E.N. Smith, R.C. Richardson, Physica B 194}196 (1994) 343. [11] T. Asada, K. Terakura, J. Phys. F 12 (1982) 1387. [12] W. Wendler, T. HerrmannsdoK rfer, S. Rehmann, F. Pobell, J. Low Temp. Phys. 111 (1998) 99.