LETTERS
Giant negative thermal expansion in magnetic nanocrystals X. G. ZHENG1,2*, H. KUBOZONO1, H. YAMADA2, K. KATO3, Y. ISHIWATA1 AND C. N. XU2 1
Department of Physics, Faculty of Science and Engineering, Saga University, Saga 840-8502, Japan National Institute of Advanced Industrial Science and Technology, Tosu, Saga 841-0052, Japan 3 Structural Materials Science Laboratory, RIKEN SPring-8 Center, 1-1-1, Kouto, Sayo-cho, Sayo-gun, Hyogo 679-5148, Japan *e-mail:
[email protected] 2
Published online: 19 October 2008; doi:10.1038/nnano.2008.309
Most solids expand when they are heated, but a property known as negative thermal expansion has been observed in a number of materials, including the oxide ZrW2O8 (ref. 1) and the framework material ZnxCd12x(CN)2 (refs 2,3). This unusual behaviour can be understood in terms of low-energy phonons1–6, while the colossal values of both positive and negative thermal expansion recently observed in another framework material, Ag3[Co(CN)6], have been explained in terms of the geometric flexibility of its metal –cyanide–metal linkages7. Thermal expansion can also be stopped in some magnetic transition metal alloys below their magnetic ordering temperature, a phenomenon known as the Invar effect8,9, and the possibility of exploiting materials with tuneable positive or negative thermal expansion in industrial applications has led to intense interest in both the Invar effect and negative thermal expansion. Here we report the results of thermal expansion experiments on three magnetic nanocrystals—CuO, MnF2 and NiO—and find evidence for negative thermal expansion in both CuO and MnF2 below their magnetic ordering temperatures, but not in NiO. Larger particles of CuO and MnF2 also show prominent magnetostriction (that is, they change shape in response to an applied magnetic field), which results in significantly reduced thermal expansion below their magnetic ordering temperatures; this behaviour is not observed in NiO. We propose that the negative thermal expansion effect in CuO (which is four times larger than that observed in ZrW2O8) and MnF2 is a general property of nanoparticles in which there is strong coupling between magnetism and the crystal lattice. CuO is unique among the monoxides of the 3d transition metals in having a monoclinic unit cell and a square planar coordination of copper and oxygen10, rather than the cubic salt structure and the octahedral coordination taken by other monoxides such as NiO, CoO and MnO. CuO is also an antiferromagnetic compound with two successive phase transitions, undergoing incommensurate ordering around TN1 ¼ 230 K with a helix period, followed by a first-order commensurate ordering transition around TN2 ¼ 213 K (ref. 11). In earlier research we have investigated changes in the lattice parameters as the temperature was changed from 100 to 1,000 K and found that the thermal expansion almost vanished below the magnetic ordering temperature12, suggesting a strong magnetostriction effect. For the present experiment, nanoparticles of CuO (5 nm diameter on average) were prepared using energetic ball milling (Fritsch Planetary ball mill) of bulk crystals. The crystalline 724
5 nm
Figure 1 Evidence for the crystalline quality of the nanoparticles. High-resolution electron micrograph of a nanocrystal of CuO. The Debye ring is clearly visible in the electron diffraction image (inset).
quality of the nanoparticles was illustrated by high-resolution electron microscopy and electron diffraction measurements (Fig. 1). The lattice constants of the nanocrystals at various temperatures were investigated by X-ray diffraction (see Methods). Similar experiments were also carried out on nanoparticles of NiO (4 nm) and MnF2 (10 nm), and larger particles of CuO (10 mm) and MnF2 (10 mm). The large negative thermal expansion (NTE) can be directly recognized in the way that the angle of the diffraction peak decreases upon cooling (Fig. 2). The lattice parameters calculated nature nanotechnology | VOL 3 | DECEMBER 2008 | www.nature.com/naturenanotechnology
© 2008 Macmillan Publishers Limited. All rights reserved.
LETTERS
Relative unit cell parameters
Intensity (a.u.)
83.0
1.010
(111/200)
1.006
0 1.002
11.6 2θ (deg)
12.0
12.4
12.8
200 400 600 800 1,000 T (K)
a b c β V
1.000
0 11.2
Micrometre CuO
81.0
0.9980
10.8
82.0 81.5
1.004
(110)
10.4
Nano CuO
82.5
1.008
V (Å3)
– (111/002)
20 K 40 K 60 K 80 K 100 K 150 K 200 K 225 K 260 K 280 K
50
100 150 200 Temperature (K)
250
300
4.21
Intensity (a.u.)
Unit cell parameter a (Å)
Nano NiO 4.20
4.19
4.18 Bulk NiO 4.17
4
8
12
16
20
24 28 2θ (deg)
32
36
40
0
44
50
100
150 200 Temperature (K)
Relative unit cell parameters
1
Figure 2 X-ray diffraction data for nanocrystals of CuO. a, Diffraction peaks at 10 different temperatures. Upon cooling the peaks shift towards lower angles, providing evidence for NTE in nanocrystals of CuO. b, An example of the Rietveld refinement for the data taken at 280 K. The raw experimental data is shown by red crosses, the grey line is a simulation of the diffraction peaks, and the black line is a simulated baseline. The small peak at around 8.588 and the broad peak near it are caused by a glass capillary used to carry the sample. The short vertical lines show the peak positions, and the blue line shows the difference between the raw data and the calculated peaks.
300
MnF2
0.999
0.9975
a (Nano ) c (Nano) a (Micrometre) c (Micrometre) V (Nano) V (Micrometre)
0.996
0
from the Rietveld refinement are shown in Fig. 3a, together with a graph showing the variation of the unit-cell volume with temperature for both nanocrystals and micrometre-sized particles of CuO. Although the micrometre-sized particles show an almost constant volume (inset, Fig. 3a) below the magnetic ordering temperature of CuO (210 K, ref. 12), a large NTE effect (b ¼ 21.1 1024 K21) was observed for the nanocrystals at low temperatures. In comparison, the renowned NTE compound ZrW2O8 has b ¼ 22.6 1025 K21 (refs 1 –3). Moreover, the micrometre-sized particles (made from sintered CuO) behaved in a similar manner to 10 mm particles (made by crushing singlecrystal CuO) that we studied in previous experiments (unpublished), so we are confident that the micrometre-sized particles behave like bulk CuO. Because the thermal expansion observed in nanocrystals of CuO at high temperatures was almost equivalent to that observed for bulk CuO, lattice softening can be ruled out as the source of the NTE. A strong positive thermal expansion (a ¼ 1.2 1024 K21) (b ¼3a) has been observed for magnetic nanoparticles of Fe3O4, but no NTE effect has been reported13. In metallic gold nanoparticles a crossover from positive to negative thermal expansion upon heating was reported, which was interpreted to be due to the effects of the valence electron potential on the
250
50
100 150 200 Temperature (K)
250
300
Figure 3 NTE in nanocrystals of CuO and MnF2, but not NiO. a, Temperature dependence of five lattice parameters for nanocrystals of monoclinic CuO relative to the values at 280 K (a ¼ 4.6896(3) A˚, b ¼ 3.4305(3) A˚, c ¼ 5.1448(6) A˚, b ¼ 99.352(4)88 and V ¼ 81.67 A˚3). The values in parentheses represent the error bars for each value. The inset shows how the volume of the unit cell for nanocrystals of CuO increases as it is cooled at low temperatures, whereas it remains approximately constant for micrometre-sized particles of CuO. (The high-temperature data for the micrometre-sized particles are taken from ref. 12). The solid line depicts a negative volume expansion coefficient of 2 1.1 1024 K21. b, The value of the lattice constant a does not change significantly with temperature at low temperature for nanocrystals (4 nm) or micrometre-sized particles of NiO. At higher temperatures, the positive thermal expansion seen in the nanocrystals is similar to that observed for the bulk material in previous experiments18. c, Temperature dependence of three lattice constants for nanocrystals and micrometre-sized particles of MnF2. At room temperature a ¼ 4.881 A˚ and c ¼ 3.299 A˚ for the nanocrystals, and a ¼ 4.887 A˚ and c ¼ 3.302 A˚ for the micrometre-sized particles. NTE can be seen in the nanocrystals below the magnetic ordering temperature of 67 K. There are fewer NTE data points for MnF2 than for CuO because MnF has a much lower magnetic ordering temperature.
nature nanotechnology | VOL 3 | DECEMBER 2008 | www.nature.com/naturenanotechnology
© 2008 Macmillan Publishers Limited. All rights reserved.
725
LETTERS equilibrium lattice separations14. However, nanocrystals of NiO (4 nm), which is another antiferromagnetic oxide with a highsymmetry cubic lattice, do not exhibit NTE (Fig. 3b). Indeed, nanocrystals of NiO show a positive thermal expansion that is nearly equivalent to that observed in bulk NiO over temperatures between 10 and 300 K. Although other explanations might be possible, it seems reasonable to assume that the large NTE effect in nanocrystals of CuO has a magnetic origin, because the crossover temperature at which positive thermal expansion changes to negative thermal expansion is the same as the magnetic ordering temperature TN at which the positive thermal expansion of micrometre-sized particles of CuO reduces to almost zero expansion. Our hypothesis is that the large NTE in nanocrystals of CuO is caused by a magneto-lattice effect. The behaviour of the lattice parameters in micrometre-sized particles of CuO suggests that magnetostriction plays a dominant role, as is clearly demonstrated by the vanishing of thermal expansion below TN (ref. 12). On cooling below the magnetic ordering temperature there must be an expansion caused by magnetic ordering that more than compensates for the contraction of the lattice caused by the fall in temperature. The large NTE effect would therefore be due to the magnetostriction effect being significantly enhanced in the nanocrystals. This view is supported by the behaviour of MnF2, another antiferromagnet with lower-than-cubic symmetry15. For bulk MnF2 the thermal expansion was found to almost vanish below the antiferromagnetic ordering temperature (67 K), whereas nanocrystals of MnF2 displayed an NTE effect like that seen in nanocrystals of CuO (Fig. 3c). In conclusion, we have found experimental evidence for giant NTE in nanocrystals of CuO and, based on further measurements on MnF2 and NiO, we believe that this NTE effect is a general property of nanosized magnetic materials that have strong magneto-lattice coupling, although more data on other magnetic nanoparticles are needed to clarify the detailed underlying mechanism.
Received 2 July 2008; accepted 18 September 2008; published 19 October 2008. References 1. Mary, T. A., Evans, J. S. O., Vogt, T. & Sleight, A. W. Negative thermal expansion from 0.3 K to 1050 K in ZrW2O8. Science 272, 90–92 (1996). 2. Phillips, A. E., Goodwin, A. L., Halder, G. J., Southon, P. D. & Kepert, C. J. Nanoporosity and exceptional negative thermal expansion in single-network cadmium cyanide. Angew. Chem. Int. Ed. 47, 1396 –1399 (2008). 3. Goodwin, A. L. & Kepert, C. J. Negative thermal expansion and low-frequency modes in cyanidebridged framework materials. Phys. Rev. B 71, 140301 (2005). 4. Ramirez, A. P. & Kowach, G. R. Large low temperature specific heat in the negative thermal expansion compound ZrW2O8. Phys. Rev. Lett. 80, 4903–4906 (1998). 5. Ernst, G., Broholm, C., Kowach, G. R. & Ramirez, A. P. Phonon density of states and negative thermal expansion in ZrW2O8. Nature 396, 147– 149 (1998). 6. Hancock, J. N., Turpen, C., Schlesinger, Z., Kowach, G. R. & Ramirez, A. P. Unusual low-energy phonon dynamics in the negative thermal expansion compound ZrW2O8. Phys. Rev. Lett. 93, 225501 (2004). 7. Goodwin, A. L. et al. Colossal positive and negative thermal expansion in the framework material Ag3[Co(CN)6]. Science 319, 794 – 797 (2008). 8. Wasserman, E. F. Invar: Moment –volume instabilities in transition metals and alloys, in Handbook of Magnetic Materials: A Handbook on the Properties of Magnetically Ordered Substances, Vol. 5, ch. 3 (eds Bushow, K. H. J. & Wohlfarth, E. P.) (Elsevier Science, 1990). 9. Khmelevskyi, S., Turek, I. & Mohn, P. Large negative magnetic contribution to the thermal expansion in iron – platinum alloys: quantitative theory of the Invar effect. Phys. Rev. Lett. 91, 037201 (2003). 10. Asbrink, S. & Norrby, L. J. A refinement of the crystal structure of copper(II) oxide with a discussion of some exceptional e.s.d.’s. Acta Crystallogr. B 26, 8 –15 (1970). 11. Forsyth, J. B., Brown, P. J. & Wanklyn, B. M. Magnetism in cupric oxide. J. Phys. C 21, 2917 –2929 (1988). 12. Yamada, H., Zheng, X. G., Soejima, Y. & Kawaminami, M. Lattice distortion and magnetolattice coupling in CuO. Phys. Rev. B 69, 104104 (2004). 13. Nikolaev, V. I. & Shipilin, A. M. On the thermal expansion of nanoparticles. Phys. Solid State 42, 112 –113 (2000). 14. Li, W. H., Wu, S. Y., Yang, C. C., Lai, S. K. & Lee, K. C. Thermal contraction of Au nanoparticles. Phys. Rev. Lett. 89, 135504 (2002). 15. Bragg, E. E. & Seehra, M. S. Magnetic susceptibility of MnF2 near TN and Fisher’s relation. Phys. Rev. B 7, 4197–4202 (1973). 16. Nishibori, E. et al. The large Debye –Scherrer camera installed at SPring-8 BL02B2 for charge density studies. Nucl. Instrum. Methods A 467 –468, 1045– 1048 (2001). 17. Izumi, F. & Ikeda, T. A Rietveld-analysis program RIETAN-98 and its applications to zeolites. Mater. Sci. Forum 321 – 324, 198 –204 (2000). 18. Bartel, L. C. & Morosin, B. Exchange striction in NiO. Phys. Rev. B 3, 1039– 1043 (1971).
Acknowledgements The authors would like to thank W. J. Moon and E. Tanaka at the High Voltage Electron Microscopy Laboratory, Kyushu University, for taking the electron micrographs of the nanoparticles.
METHODS The lattice constants of the nanocrystals at various temperatures were measured by X-ray diffraction at the BL02B2 beamline at the SPring-8 synchrotron using a Debye– Scherrer camera with an imaging plate16. The incident X-ray beam was monochromatized by a double-crystal monochromator tuned to a wavelength of approximately 0.5 A˚ for all experiments. The temperature of the sample was controlled by a helium gas flow cryostat with a temperature deviation of ,1 K.
726
Powder diffraction patterns at temperatures ranging from 300 to 10 K were collected over the 2u range of 0– 688 in a step angle of 0.018. The unit cell parameters were first roughly calculated from the observed d-value and hkl peak data of the powder-diffraction data using a least-squares method. Then the unit cell parameters were refined by the Rietveld method, using the computer-assisted program RIETAN-200017. It was confirmed that both methods gave similar results, although the latter reduced the error bars for the unit cell parameters.
Author contributions X.G.Z. conceived and designed the experiments. H.K., H.Y., X.G.Z. and K.K. performed the experiments. H.K. and X.G.Z. analysed the data. H.Y., C.N.X. and Y.I. contributed NiO and analysis tools. X.G.Z. wrote the paper. All authors discussed the results and commented on the manuscript.
Author information Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/. Correspondence and requests for materials should be addressed to X.G.Z.
nature nanotechnology | VOL 3 | DECEMBER 2008 | www.nature.com/naturenanotechnology
© 2008 Macmillan Publishers Limited. All rights reserved.
