Abstract

Various spintronic phenomena originate from the exchange of angular momentum between the spin of electrons and other degrees of freedom in crystalline materials. Many degrees of freedom, such as magnetization1 and mechanical motion2, have already been united into this exchange framework. However, the nuclear spin—a key angular momentum—has yet to be incorporated. Here we observe spin pumping from nuclear magnetic resonance (NMR), in which nuclear spin dynamics emits a spin current, a flow of spin angular momentum of electrons. By using the canted antiferromagnet MnCO3, in which typical nuclear spin-wave formation is established due to the reinforced hyperfine coupling, we find that a spin current is generated from an NMR. Nuclear spins are indispensable for quantum information technology3 and are also frequently used in various sensors, such as in magnetic resonance imaging4. The observed NMR spin pumping allows spin-current generation from nuclei and will enable spintronic detection of nuclear spin states.

Main

The search for new phenomena that create spin current in condensed matter has been a driving force in spintronics research1. In this context, spin pumping5,6,7,8,9,10, which generates spin current by magnetization dynamics, is a powerful technique, which was studied at first in ferromagnetic metals6, and recently found to be applicable to magnetic insulators—for example, Y3Fe5O12 (ref. 10). Spin pumping using various spin waves in Y3Fe5O12 (ref. 10) has received keen attention not only in spintronics but also in electronics for the possible realization of novel magnetic devices.

In quantum information science, on the other hand, a nuclear spin is a promising candidate for a quantum bit3. Coupling between a nuclear spin and an electron spin through the hyperfine interaction has been studied intensively11,12. The hyperfine interaction also leads to indirect interactions among nuclear spins via virtual spin-wave absorption and emission, called the Suhl–Nakamura interaction13,14. The nuclear spin coupling through the Suhl–Nakamura interaction gives rise to a collective excitation mode, called a nuclear spin wave15,16,17,18 (Fig. 1a). Nuclear spin waves belong to the nuclear branch of the nuclear–electron spin coupling modes, which has been established in some antiferromagnets16,17,18 in terms of a shift of the nuclear magnetic resonance (NMR) frequency, called a frequency pulling effect (Fig. 1b).

Fig. 1: Concept of NMR spin pumping and sample properties.
Fig. 1

a, Schematic illustration of spin pumping driven by nuclear magnetic resonance (NMR) and the inverse spin-Hall effect. As a result of the excitation of nuclear spin waves by applying sub-gigahertz-band electromagnetic waves to MnCO3, an electric field, ESHE, arises in the direction of js × σ in Pt via the inverse spin-Hall effect of Pt. Here js and σ are the spatial direction and the spin direction of the injected spin current, respectively. rf, radiofrequency. b, A frequency pulling effect calculated for a nuclear spin wave mode, a nuclear spin dynamics coupled to an electron spin dynamics through the Suhl–Nakamura interaction. c, Rhombohedral MnCO3 crystal structure. d, Temperature (T) and magnetic field (H) dependence of magnetization (M) for a MnCO3 sample. The antiferromagnetic transition temperature TN is estimated to be approximately 35 K. The Mn spins lie in the {111} plane and are canted slightly from the pure antiferromagnetic ordering because of the bulk Dzyaloshinskii–Moriya interaction22,23,24,25. Since the canting angle is about one degree, the MnCO3 crystal exhibits a weak ferromagnetic moment whose magnitude is about 0.01 μB/Mn2+, as shown in the isothermal magnetization curves.

Here, we report the observation of spin-current generation from NMR due to nuclear spin waves in the weakly anisotropic antiferromagnet MnCO3 (Fig. 1c), in which nuclear spin waves are well established18. In MnCO3, a 55Mn nucleus (spin I = 5/2) has a strong hyperfine interaction and a 100% abundance of the magnetic isotope. We found that a nuclear spin wave excited by a radio wave in insulating MnCO3 single crystals pumps a spin current into an attached 5-nm-thick Pt film (surface area: 9 mm2) through electron spin dynamics created from the nuclear spin dynamics in MnCO3 (Fig. 1a). Platinum is a typical metal that exhibits strong inverse spin-Hall effects7,8,9,10—the conversion of a spin current into a voltage. Therefore, when a spin current is injected from MnCO3 into the Pt layer, a voltage should appear in the Pt layer, the sign of which reflects the spin direction of the spin current.

In Fig. 2a, we show NMR spectra for 55Mn nuclei in a Pt(5-nm-thick)/MnCO3{111} sample at 1.52 K—far below the antiferromagnetic transition temperature of MnCO3 (TN = 35 K, as shown in Fig. 1d). Here, the measurement was performed at each magnetic field, H, by applying continuous radio waves in a perpendicular pumping configuration. At each magnetic field, a clear dip structure signalling the NMR absorption appears in the radio-wave reflection spectra, as shown in Fig. 2a. With increasing magnetic fields, the NMR peak-frequency fNMR increases and the peak shape becomes sharper. The peak frequency measured at several temperatures is plotted as a function of external magnetic fields in Fig. 2b. The H dependence of fNMR in Fig. 2b deviates from the values expected from the hyperfine field alone: a sign of the Suhl–Nakamura interaction13,14,15,16,17,18. The strong hybridization of nuclear and electron spins leads to the shift of the NMR frequency (a frequency pulling effect)16,17,18. The solid curves in Fig. 2b show a theoretical calculation of the H dependence of fNMR in which the frequency pulling effect is taken into consideration (see Supplementary Information). The experimentally obtained H dependence of fNMR is in good agreement with the theoretical curves. From a fitting of the theoretical curves shown in Supplementary Information, the effective anisotropic magnetic field Ha originating from the hybridization between nuclear and electron spins was obtained at each temperature, as shown in the inset to Fig. 2b. The Ha values lie in the range 1–10 Oe, and increase monotonically with decreasing temperature, which means that the Suhl–Nakamura interaction becomes important at lower temperatures.

Fig. 2: Observation of spin current induced by NMR.
Fig. 2

a, Nuclear magnetic resonance (NMR) spectra measured at various magnetic fields at 1.52 K. Here, Pref and Pin are the reflected electromagnetic-wave power and the input electromagnetic-wave power, respectively (see also Supplementary Information). The dip magnitude, 1 − Pref/Pin, presents the absorbed electromagnetic-wave power, Pabs, divided by Pin (Pabs/Pin). The rate of the frequency sweep is (750 [MHz]−450 [MHz])/2 [ms] = 1.5×1011 Hz s−1; at faster sweep rates, sharper NMR spectra are observed (see e). b, NMR frequency of 55Mn in MnCO3 as a function of an external magnetic field (H) measured at various temperatures for an incident radio-wave power Pin of 5 mW. The bare NMR frequency without the frequency pulling effect is about 640 MHz. The inset shows the estimated effective magnetic field Ha originating from the hybridization of nuclear and electron spins. The fitted curve is 8.6/T, where T is temperature (see Supplementary Information). c,d, NMR spectra (c) and voltage (V) spectra (d) measured in the Pt layer around the NMR frequency of MnCO3, where the incident radio-wave power Pin is 5 mW. The external magnetic field was fixed at +0.3 T. The rate of the frequency sweep is (750 [MHz]−450 [MHz])/60 [s] = 5×106 Hz s−1. The peaks labelled with an asterisk at around 700 MHz in c are not from the Pt/MnCO3, but from the sample rod. e, Sweep-rate (sweep-time) dependence of NMR spectra at 1.96 K. As the total sweep time increases, the linewidth becomes broader, which can be attributed to a nonlinear effect of NMR. At a sweep time of 60 s, which corresponds to the sweep time of the spin-pumping voltage measurement shown in d, the linewidth is as large as approximately 40 MHz, consistent with that of the inverse spin-Hall voltage. f,g, Temperature (T) dependence of the peak intensity of the NMR absorption power (f) and the voltage signal (g) observed around the NMR frequency in a Pt/MnCO3 sample at +0.3 T; these data are obtained from c and d. The error bars in g correspond to the background noise levels. The theoretical temperature dependence expected from the change in Ha is also shown in g.

At very low temperatures, an unconventional voltage signal clearly appears in the Pt layer at the NMR frequencies. In Fig. 2d, we show the radio-wave-frequency dependence of the d.c. voltage generated between the ends of the Pt layer at 0.3 T. At 3.0 K, no voltage signal is observed. However, as the temperature decreases to 2.4 K, a clear voltage peak appears around the NMR frequency (600 MHz). Broader linewidths of the voltage peaks than those of the NMR spectra shown in Fig. 2a result from slower frequency-sweep rates for the voltage measurements (Fig. 2e). The intensity of the voltage peak becomes greater with decreasing temperature, and reaches approximately 8 nV at 1.52 K. The temperature dependence of the peak height is shown in Fig. 2g. Whereas the absorbed power at NMR changes only slightly with decreasing temperature, as shown in Fig. 2c,f, the voltage peak observed at the NMR frequency grows rapidly below 3.0 K, concomitant with the enhancement of the Suhl–Nakamura interaction at lower temperatures (inset to Fig. 2b). In fact, the increase in Ha with decreasing temperature (Fig. 2b) quantitatively explains the voltage enhancement, as indicated by the green curve in Fig. 2g. We also confirmed that spin Seebeck voltages due to the weak ferromagnetism in MnCO3 are negligibly small (see Supplementary Information). The results rule out spin Seebeck effects19,20,21.

The voltage peak observed in the Pt layer around the NMR frequency exhibits a sign change when the direction of the magnetic field is reversed, consistent with the behaviour of the inverse spin-Hall effect7,8,9,10. In Fig. 3a, voltage spectra measured in external magnetic fields of +0.3 T and −0.3 T are compared at different input power levels, Pin. As Pin increases, the intensity of the voltage peaks becomes greater both at +0.3 T and −0.3 T. The signs of the voltage peaks are opposite for +0.3 T and −0.3 T, although the magnitude is almost the same. The clear sign change of the voltage shows that the voltage signal is due to the inverse spin-Hall effect7,8,9,10 caused by a spin current injected into the Pt layer due to the NMR from the MnCO3. As shown in Fig. 3b, the magnitude of the voltage signal increases almost in proportion to the absorbed radio-wave power Pabs estimated from the NMR spectra. The power dependence is also consistent with previous electron spin-pumping measurements in various ferromagnetic materials5,6,7,8,9,10, whereas, in the present experiment, the spin pumping is driven by nuclear spin dynamics. We note that the sign of the observed inverse spin-Hall voltage is the same as that of the spin-pumping voltages for Pt/ferromagnet bilayers, such as Pt/Y3Fe5O12 (ref. 10).

Fig. 3: Power dependence of spin-pumping voltage.
Fig. 3

a, Frequency dependence of the spin-pumping voltage in a Pt/MnCO3 sample at 1.53 K, where the input radio-wave power is changed from Pin = 0.1 mW to 5 mW. b, Pin dependence of the voltage peak intensity V and the radio-wave absorption power Pabs in a Pt/MnCO3 sample measured at magnetic fields of ±0.3 T. The error bars correspond to the background noise levels.

The static nuclear polarization is small (approximately 1%; see Supplementary Information) even under magnetic fields. Nevertheless, in MnCO3, the electron spin dynamics created from the nuclear spin dynamics enables efficient nuclear spin pumping, as formulated in our theoretical calculation. Below TN in MnCO3, the Mn electron spins are aligned in the {111} plane and canted slightly from the pure antiferromagnetic ordering direction because of the bulk Dzyaloshinskii–Moriya interaction22,23,24,25 (Fig. 1d). The Mn nuclear spins interact with the electron spins through the hyperfine interaction, which favours antiparallel arrangement of electron and nuclear spins (Fig. 1a). At very low temperatures, where the hybridization of nuclear and electron spins is sufficiently strong, the coupled resonant dynamics of nuclear and electron spins is excited under NMR, and spin currents are generated via the electron spin pumping5,6,7,8,9,10. The magnitude of the generated spin current is characterized by an effective magnetic field for electron spins originating from the hybridization with nuclear spins, Ha (Fig. 2b, see Supplementary Information). Since the antiferromagnetic crystalline anisotropy in the {111} plane is very small in MnCO3, the hybridization of nuclear and electron spin dynamics via the strong hyperfine coupling can noticeably affect the electron spin dynamics in MnCO3 at very low temperatures25, giving rise to detectable NMR spin pumping.

In Fig. 4, the voltage signals measured at different external magnetic fields are compared with a result of our theoretical calculation based on an antiferromagnetic macro-spin model, in which the bulk Dzyaloshinskii–Moriya interaction and the hyperfine coupling with nuclear spins are taken into consideration (see Supplementary Information). For all the H values, a clear sign reversal of the voltage peaks is observed between +H and –H in the frequency scan, as shown in Fig. 3. The linewidth of the voltage peak decreases monotonically with H at higher fields (Fig. 4c), consistent with that of the NMR spectra shown in Fig. 2a. In the low-H range where the frequency pulling is clearly observed (Fig. 2b), the hybridization of nuclear and electron spins enhances the linewidth of the inverse spin-Hall voltage, as calculated in Fig. 4d. The voltage peak height normalized by the resonance absorption intensity |V|/(Pabs/Pin) is plotted as a function of magnetic fields in Fig. 4a. As the magnetic field increases, |V|/(Pabs/Pin) shows a broad maximum around 0.3 T, and then decreases gently. The magnetic-field dependence of the spin pumping voltage is reproduced well by a theoretical calculation based on the above model shown in Fig. 4b. The inverse spin-Hall voltage generated at the NMR frequency is greater at lower magnetic fields because of the stronger hybridization of nuclear and electron spins, but it tends to be suppressed towards H = 0. A very high spin polarization (approximately 34%) of nuclear spins was reported by means of dynamic nuclear polarization26, and the NMR spin pumping may exhibit a high conversion efficiency by using dynamic nuclear polarization in the future27.

Fig. 4: Magnetic field dependence of spin-pumping voltage.
Fig. 4

a, Magnetic field (H) dependence of the voltage peak height |V| normalized by the resonance absorption intensity Pabs/Pin (see Methods and Supplementary Information for raw data). The error bars correspond to the background noise levels. b, Calculation results of the inverse spin-Hall voltage induced by the NMR spin pumping (see text and Supplementary Information). The voltage is nonzero at zero magnetic field because canted magnetization is assumed in the calculation. c, Magnetic field (H) dependence of the linewidths of the voltage peak and the NMR absorption power. The linewidths are defined as the widths of the frequency ranges where the NMR absorption/voltage signals are observed (see also Supplementary Information). The error bars for voltage data correspond to the difference in the linewidths between positive and negative magnetic fields. d, Calculation results of the linewidth (half-width at half-maximum) of the inverse spin-Hall voltage induced by the NMR spin pumping (see text and Supplementary Information). a.u., arbitrary units.

The observed NMR spin pumping will be essential in nuclear spintronics28 and spintronic detection of information stored in nuclear spins will be a key ingredient in information science combined with spintronics. The NMR spin pumping is largely controlled by an external magnetic field, and it can be used to make a switchable electromagnetic-wave detector. If a nucleus retains spin angular momentum for a much longer time than electrons, NMR spin pumping and its reciprocal process could be used to make a capacitor for electron spin currents.

Methods

Measurements of spin pumping

We used commercially available MnCO3 single crystals with a size of 3 × 3 × 0.5 mm3 (Surface Net GmbH). The largest plane is {111} in the rhombohedral representation (Fig. 1c). The magnetic properties of the MnCO3 crystals were measured using the RSO option of a magnetic property measurement system (MPMS, Quantum Design, Inc.). On top of the {111} plane of the MnCO3, a 5-nm-thick Pt film was sputtered at room temperature for the spin-pumping experiments.

The spin-pumping experiments were performed with an in-house-made sample rod in a superconducting magnet (SM4000, Oxford Instruments). Sub-gigahertz electromagnetic waves were applied through a coplanar waveguide, for which the width of the signal line was designed to be about 5 μm, to the Pt/MnCO3 samples using a network analyser (N5230C, Keysight Technologies). In a static magnetic field at low temperatures, voltages that show up between the edges of the Pt film were measured using a nanovoltmeter (K2182A, Tektronix, Inc.) while the frequency of incident continuous electromagnetic waves was swept. To improve the signal-to-noise ratio, the voltage measurements were repeated 60 times, and averaged data sets are plotted in the figures.

Data availability

The data that support the plots within this paper and other findings of this study are available from the corresponding authors upon reasonable request.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Maekawa, S., Valenzuela, S. O., Saitoh, E., Kimura, T. (eds) Spin Current (Oxford Univ. Press, Oxford, 2012).

  2. 2.

    Matsuo, M., Saitoh, E. & Maekawa, S. Spin-mechatronics. J. Phys. Soc. Jpn. 86, 011011 (2017).

  3. 3.

    Nielsen, M. A. & Chuang, I. L. Quantum Computation and Quantum Information: 10th Anniversary Edition (Cambridge Univ. Press, Cambridge, 2010).

  4. 4.

    McRobbie, D. W., Moore, E. A., Graves, M. J. & Prince, M. R. MRI from Picture to Proton (Cambridge Univ. Press, Cambridge, 2007).

  5. 5.

    Tserkovnyak, Y., Brataas, A. & Bauer, G. E. W. Enhanced Gilbert damping in thin ferromagnetic films. Phys. Rev. Lett. 88, 117601 (2002).

  6. 6.

    Mizukami, S., Ando, Y. & Miyazaki, T. Effect of spin diffusion on Gilbert damping for a very thin Permalloy layer in Cu/Permalloy/Cu/Pt films. Phys. Rev. B 66, 104413 (2002).

  7. 7.

    Saitoh, E., Ueda, M., Miyajima, H. & Tatara, G. Conversion of spin current into charge current at room temperature: inverse spin-Hall effect. Appl. Phys. Lett. 88, 182509 (2006).

  8. 8.

    Azevedo, A., Vilela Leão, L. H., Rodríguez-Suárez, R. L., Oliveira, A. B. & Rezende, S. M. dc effect in ferromagnetic resonance: Evidence of the spin-pumping effect? J. Appl. Phys. 97, 10C715 (2005).

  9. 9.

    Costache, M. V., Sladkov, M., Watts, S. M., van der Wal, C. H. & van Wees, B. J. Electrical detection of spin pumping due to the precessing magnetization of a single ferromagnet. Phys. Rev. Lett. 97, 216603 (2006).

  10. 10.

    Kajiwara, Y. et al. Transmission of electrical signals by spin-wave interconversion in a magnetic insulator. Nature 464, 262–266 (2010).

  11. 11.

    Smet, J. H. et al. Gate-voltage control of spin interactions between electrons and nuclei in a semiconductor. Nature 415, 281–286 (2002).

  12. 12.

    Kane, B. E. A silicon-based nuclear spin quantum computer. Nature 393, 133–137 (1998).

  13. 13.

    Suhl, H. Effective nuclear spin interactions in ferromagnets. Phys. Rev. 109, 606 (1958).

  14. 14.

    Nakamura, T. Indirect coupling of nuclear spins in antiferromagnet with particular reference to MnF2 at very low temperatures. Prog. Theor. Phys. 20, 542–552 (1958).

  15. 15.

    de Gennes, P. G., Pincus, P. A., Harmann-Boutron, F. & Winter, J. M. Nuclear magnetic resonance modes in magnetic material. I. Theory. Phys. Rev. 129, 1105–1115 (1963).

  16. 16.

    Tulin, V. A. Nuclear spin waves in magnetically ordered materials. Sov. J. Low. Temp. Phys. 5, 455–469 (1979).

  17. 17.

    Borovik-Romanov, A. S. et al. The spin echo in systems with a coupled electron–nuclear precession. Sov. Phys. Usp. 27, 235–255 (1984).

  18. 18.

    Andrienko, A. V., Ozhogin, V. I., Safonov, V. L. & Yakubovskiǐ, A. Yu Nuclear spin wave research. Sov. Phys. Usp. 34, 843–861 (1991).

  19. 19.

    Uchida, K. et al. Observation of the spin Seebeck effect. Nature 455, 778–781 (2008).

  20. 20.

    Uchida, K. et al. Spin Seebeck insulator. Nat. Mater. 9, 894–897 (2010).

  21. 21.

    Jaworski, C. M. et al. Observation of the spin-Seebeck effect in a ferromagnetic semiconductor. Nat. Mater. 9, 898–903 (2010).

  22. 22.

    Andrienko, A. V., Ozhogin, V. I., Safonov, V. L. & Yakubovskiǐ, A. Yu Influence of electronic-magnon relaxation rate on the damping of nuclear spin waves in antiferromagnets. Sov. Phys. JETP 62, 794–799 (1985).

  23. 23.

    Borovik-Romanov, A. S. & Orlova, M. P. Magnetic properties of cobalt and manganese carbonates. Sov. Phys. JETP 4, 531–534 (1957).

  24. 24.

    Borovik-Romanov, A. S. Investigation of weak ferromagnetism in the MnCO3 single crystal. Sov. Phys. JETP 9, 539–549 (1959).

  25. 25.

    Shaltiel, D. Nuclear magnetic resonance of MnCO3 in the canted spin state. Phys. Rev. 142, 300–306 (1966).

  26. 26.

    Tateishi, K. et al. Room temperature hyperpolarization of nuclear spins in bulk. Proc. Natl. Acad. Sci. USA 111, 7527–7530 (2014).

  27. 27.

    Gregg, J. F. et al. Microscopic explanation of microwave spin pumping in spintronics. Preprint at https://arxiv.org/abs/1711.06048 (2017).

  28. 28.

    Vagner, I. D. Recent Trends in Theory of Physical Phenomena in High Magnetic Fields Ch. 23 (Springer, Dordrecht, 2003).

Download references

Acknowledgements

We thank H. Yasuoka, S. Maekawa, M. Matsuo, H. Chudo, K. Harii and M. Imai for fruitful discussions. This research was supported by JST ERATO ‘Spin Quantum Rectification Project’ (JPMJER1402), JSPS KAKENHI (no. 17H04806, no. JP18H04215, no. 18H04311, no. JP16J03699 and no. 17H02927) and MEXT (Innovative Area ‘Nano Spin Conversion Science’ (no. 26103005)).

Author information

Author notes

    • Yuki Shiomi

    Present address: Department of Applied Physics and Quantum-Phase Electronics Center (QPEC), University of Tokyo, Tokyo, Japan

  1. These authors contributed equally: Yuki Shiomi and Jana Lustikova.

Affiliations

  1. Institute for Materials Research, Tohoku University, Sendai, Japan

    • Yuki Shiomi
    • , Jana Lustikova
    • , Shingo Watanabe
    • , Daichi Hirobe
    • , Saburo Takahashi
    •  & Eiji Saitoh
  2. Center for Spintronics Research Network, Tohoku University, Sendai, Japan

    • Saburo Takahashi
    •  & Eiji Saitoh
  3. Advanced Science Research Center, Japan Atomic Energy Agency, Tokai, Japan

    • Eiji Saitoh
  4. Advanced Institute for Materials Research, Tohoku University, Sendai, Japan

    • Eiji Saitoh
  5. RIKEN Center for Emergent Matter Science (CEMS), Wako, Japan

    • Yuki Shiomi

Authors

  1. Search for Yuki Shiomi in:

  2. Search for Jana Lustikova in:

  3. Search for Shingo Watanabe in:

  4. Search for Daichi Hirobe in:

  5. Search for Saburo Takahashi in:

  6. Search for Eiji Saitoh in:

Contributions

S.W. conceived the experiments in discussions with Y.S. and D.H. Y.S., J.L. and S.W. constructed the experimental set-up, performed the experiments, and analysed the experimental data. S.T. conducted the theoretical calculations. Y.S., J.L., S.W. and E.S. wrote the manuscript. E.S. supervised the project. All authors discussed the results and reviewed the manuscript.

Competing interests

The authors declare no competing interests.

Corresponding authors

Correspondence to Yuki Shiomi or Eiji Saitoh.

Supplementary information

  1. Supplementary Information

    Theoretical calculations; Supplementary Figures 1–8; Supplementary References 1–8

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41567-018-0310-x

Further reading