betaFeedback

WorksPhysics and Astronomy

Piezomagnetism in the Ising ferromagnet URhGe

Abstract

Piezomagnetism, linear response between strain and magnetic field, is relatively unexplored cross-correlation but has promising potential as a novel probe of time-reversal-symmetry breaking in various classes of materials. Interestingly, there has been no report of piezomagnetism in ferromagnets, most archetypal time-reversal-symmetry-broken materials. This half-century absence of piezomagnetic ferromagnets is attributable to complications originating from multiple-domain states, as well as from changes in the magnetic point group by rotation of magnetic moment. Here, we report characteristic V-shaped magnetostriction in the Ising itinerant ferromagnet URhGe, observed by simultaneous multi-axis strain measurement technique utilizing optical fiber Bragg grating sensors. This novel magnetostriction occurs only under fields along the c axis and does not scale with the square of magnetization. Such unconventional feature indicates piezomagnetism as its origin. Our observation, marking the first report of piezomagnetism in ferromagnets, is owing to the mono-domain switching and the Ising magnetization. The obtained piezomagnetic coefficients are fairly large, implying that Ising ferromagnets are promising frontiers when seeking for materials with large piezomagnetic responses.

Cross correlations, coupling between physical quantities with orthogonal symmetry properties in systems lacking the corresponding symmetry, have been attracting much attention [1]. In particular, breakings of the most fundamental symmetries, namely the inversion and time-reversal symmetries, have been studied extensively, and novel cross-correlation phenomena such as magneto-electric effects in multiferroics have been established [2, 3]. Nevertheless, there are many unexplored cross correlations, which are worth extensive investigations both for fundamental and applicational points of view. One of such unexplored cross correlations is the piezomagnetism (PZM). The PZM or piezomagnetic effect refers to the phenomenon that the strain ε\varepsilon of a magnetic material responds linearly to the external magnetic field HH (i.e. ε∝H\varepsilon\propto H), or its inverse effect, namely the magnetization MM induced linearly by external stress σ\sigma (M∝σM \propto\sigma). The former is also called the linear magnetostriction. These effects were first predicted in 1928 [4] and its basic theory was established in 1956 [5]. It is now understood that materials having symmetry groups without independent time-reversal operation or those with time-reversal operation but only in combination with lattice reflections or rotations can exhibit PZM [6]. Experimentally, PZM was first discovered in antiferromagnets CoF2_2 and MnF2_2 in 1959 [7, 8] following theoretical prediction [9]. In these materials, time-reversal symmetry is lost due to the characteristic magnetic structure with up and down magnetic moments sitting respectively on different crystalline sublattices. This is in clear contrast with ordinary antiferromagnets, whose symmetry groups possess time-reversal operation coupled with lattice translations. Recently, non-collinear antiferromagnet UO2_2 is reported to exhibit hard PZM with a coercive field of as large as 18 T [10]. More recently, PZM is attracting renewed attention as a powerful tool to detect non-trivial time-reversal-symmetry breaking (TRSB) in novel states such as magnetic multipole orders [11] and altermagnetism [12, 13]. Nevertheless, PZM has been only reported in several limited materials so far [6, 10, 11, 12], and urges further investigations.

As mentioned above, PZM is allowed in systems that exhibit TRSB [5, 6]. Because of this principle, we can easily predict that ferromagnets should also exhibit piezomagnetism. However, surprisingly, PZM has been observed only in antiferromagnets [10, 11, 12, 13]. Observation of PZM in ferromagnets is perhaps hindered by the complex domain configurations with various magnetization directions. In such multi-domain states, bulk PZM would be cancelled among domains with opposite magnetic moments. In addition, ordinary ferromagnetic (FM) magnetostriction, namely strain due to domain configuration change, dominates [14]. Moreover, if directions of magnetic moments vary due to domain formation and/or applied magnetic fields, the magnetic point group can also change, making detection of PZM even more complicated. Thus, most ferromagnets exhibit ordinary magnetostriction behavior approximated as ε∝(H−Hcoer)2\varepsilon\propto(H-H_{\mathrm{coer}})^2 in magnetic fields near the coercive field HcoerH_{\mathrm{coer}} [15]. Therefore, it is not straightforward to detect the naively expected piezomagnetism in ferromagnets.

In this Letter, we report magnetostriction in the itinerant ferromagnet URhGe measured with the multi-axis simultaneous strain measurement technique based on fiber Bragg grating (FBG). We found unusual “V-shaped” magnetostriction in the FM phase only for specific combinations of field and strain directions. This is attributed to the ferromagnet PZM, which has been overlooked for more than a half century.

Our target material is the Ising-like itinerant ferromagnet URhGe. URhGe has the orthorhombic PnmaPnma crystal structure [16]. This material exhibits ferromagnetism below the Curie temperature TC=9.5T_{\mathrm{C}} = 9.5 K [16, 17] and subsequently superconductivity below Tsc=0.28T_{\mathrm{sc}} = 0.28 K [18]. Due to the strong spin-orbit coupling, the magnetic moment in URhGe shows strong Ising features with the easy axis along the cc direction [19]. Such Ising nature is inherited in the FM state. Recently, Mineev discussed PZM in the FM and superconducting states of URhGe and its related materials UCoGe and UGe2_2 [20]. Because of the Ising magnetic anisotropy, the magnetic point group of the FM state is well-defined to be D2h(C2h)D_{2h}(C_{2h}) irrespective of magnetic-field directions as long as the field is not too strong. For this magnetic point group, PZM is indeed allowed; the piezomagnetic effect obeys εμ=∑kQkμHk\varepsilon_\mu= \sum_k Q_{k\mu}H_k with the non-vanishing piezomagnetic tensor

Qkμ=(0000Q150000Q2400Q31Q32Q33000)(1)Q_{k\mu} = \begin{pmatrix} 0 & 0 & 0 & 0 & Q_{15} & 0 \\ 0 & 0 & 0 & Q_{24} & 0 & 0 \\ Q_{31} & Q_{32} & Q_{33} & 0 & 0 & 0 \end{pmatrix} \tag*{(1)}

where εμ\varepsilon_\mu is the strain expressed using the Voigt notation (ε1\varepsilon_1 corresponding εaa\varepsilon_{aa}, etc), and HkH_k is the kk-component of the magnetic field (k=1k = 1, 2, and 3 corresponding to the aa, bb, and cc directions) [6, 20]. This tensor indicates that cc-axis field will produce piezomagnetic linear magnetostriction in εaa\varepsilon_{aa}, εbb\varepsilon_{bb}, and εcc\varepsilon_{cc}, whereas no PZM in these normal strains for fields along aa or bb axis.

For the present study, we used a high-quality single crystals of URhGe prepared with the Czochralski method. Strain εbb\varepsilon_{bb} and εcc\varepsilon_{cc} of this sample was simultaneously measured by using FBG sensors. The FBG is a periodic grating embedded to the core of an optical fiber and we can sensitively measure the strain from the change in the wavelength of the light reflected from an FBG pasted to the sample [21]. The strain transmission rate between the sample and FBG through glue is calibrated based on thermal-expansion measurements. To introduce light to FBGs and measure the spectra of reflected light, we used a commercial interrogator (KYOWA EFOX-1000B-4). The strain measurements were performed in a commercial cryostat (Quantum Design, PPMS), whereas magnetization MM was measured with a commercial magnetometer (Quantum Design, MPMS). Details of experimental method are explained in Supplemental Material [22].

First, we show in Figure 1 MM of the URhGe sample for HH along the easy-magnetization axis (cc axis) measured at various temperatures. In the FM state below TC∼9.5T_{\mathrm{C}} \sim9.5 K, M(H)M(H) exhibits step-like change around H=0H = 0 with very narrow hysteresis width. The step-like change indicates that all magnetic moments flip simultaneously without forming FM domains. Thus, this compound is free from complicated phenomena originating from multi-domain configurations. Note that the saturated moment is about 0.4 μB0.4\ \mu_{\mathrm{B}}/U, which is much smaller than the value expected for U 5f25f^2 or 5f35f^3 configuration (∼3 μB\sim3\ \mu_{\mathrm{B}}/U), indicating the weak itinerant ferromagnetism.

Image not extracted

Figure 1. (color online) Magnetization MM of URhGe measured with H∥cH \parallel c. The inset is an enlarged view of the 4.0-K data around H=0H = 0. Below TCT_{\mathrm{C}}, M(H)M(H) exhibits step-function like behavior with a narrow hysteresis width, indicating that the multi-domain state is almost negligible.

Next, in Figure 2, we show our main result, namely the magnetostriction measured under H∥cH \parallel c at various temperatures. Above TCT_{\mathrm{C}} (red curves), both εbb\varepsilon_{bb} and εcc\varepsilon_{cc} exhibit quadratic behavior around zero field. This is ordinary behavior for paramagnets; if MM is proportional to HH, the ordinary magnetostriction should obey the relation ε∝M2∝H2\varepsilon\propto M^2 \propto H^2. The difference in the sign of the magnetostriction reflects the anisotropic thermal expansion at zero field (Figure S2): When the magnetic moment grows as temperature decreases, εbb\varepsilon_{bb} tends to expand and εcc\varepsilon_{cc} prefers to shrink. The application of magnetic field, bringing the sample closer to the FM state, causes the same effect. This tendency that a shorter bb axis tends to disfavor the FM order is consistent with the fact that TCT_{\mathrm{C}} decreases under uniaxial pressure along the bb axis [23]. This uniaxial-pressure effect is known to be consistent with the negative jump in the bb-axis thermal-expansion coefficient at TCT_{\mathrm{C}} [24] through the Ehrenfest relation.

Image not extracted

Figure 2. (color online) Magnetostriction of URhGe in (a) εbb\varepsilon_{bb} and (b) εcc\varepsilon_{cc} measured under H∥cH \parallel c. Each curve is vertically offset for clarity. Below TCT_{\mathrm{C}} (blue curves), V-shaped magnetostriction originating from PZM is observed in both strain components, in clear contrast to the quadratic behavior above TCT_{\mathrm{C}} (red curves) or the nonlinear behavior with a cusp due to critical behavior at TCT_{\mathrm{C}} (green curves). Results of fittings using Eq. (2) are shown with black dotted curves.

Image not extracted

Figure S2. (a) Nominal thermal expansion of the URhGe sample measured with FBG (obtained by assuming the strain transmission rate K−1=1K^{-1}=1), compared with the literature data (Aoki 2011: Ref. [24]). The apparent discrepancy is attributable to the actual strain transmission rate K−1K^{-1} smaller than 1. (b) Calibrated thermal expansion data. The transmission rate K−1K^{-1} is determined so that the raw data shown in (a) multiplied by KK match the literature value. From this experiment, we obtained Kc−1=0.433K_c^{-1}=0.433 and Kb−1=0.84K_b^{-1}=0.84.

At TC=9.5T_{\mathrm{C}} = 9.5 K (green curves), εbb\varepsilon_{bb} and εcc\varepsilon_{cc} show sharp kinks at H=0H = 0, followed by non-zero curvatures at finite fields. This at first glance looks anomalous but is mostly attributable to the ordinary magnetostriction ε∝M2\varepsilon\propto M^2 together with the critical behavior in MM at TCT_{\mathrm{C}}. Mean-field theories predict M∝H1/3M \propto H^{1/3} at the transition temperature [25], hence ε\varepsilon is expected to show nonlinear H2/3H^{2/3} behavior. Indeed, such cusp-like magnetostriction at TCT_{\mathrm{C}} has been reported in other ferromagnets such as UIr [26] and TbCo2_2Mnx_x [27]. The behavior at TCT_{\mathrm{C}} will be discussed again later.

Interestingly, in the FM state (blue curves), εbb(H)\varepsilon_{bb}(H) and εcc(H)\varepsilon_{cc}(H) curves are strikingly V-shaped, with robust HH-linear dependences in both H>0H > 0 and H<0H < 0 regions. As a consequence, ε\varepsilon is proportional to ∣H∣|H|. This behavior remains down to 2 K, the lowest temperature of the present study. Compared with the step-like M(H)M(H) curve in the FM phase (Figure 1), it is clear that the ε∝∣H∣\varepsilon\propto|H| dependence cannot be described by the ordinary magnetostriction ε∝M2\varepsilon\propto M^2, as demonstrated in Figs. S3 and S4 [22]. Below, we will discuss other features to conclude that this V-shaped magnetostriction originates from the PZM in URhGe.

We firstly examine the anisotropy of the V-shaped magnetostriction. We show in Figure 3 the magnetostriction measured under H∥bH \parallel b. The strain above TCT_{\mathrm{C}} is weak and quadratic against the magnetic fields. At T=TCT = T_{\mathrm{C}}, the strain showed a cusp at H=0H = 0, which is similar to those observed in H∥cH \parallel c and is attributable to the critical behavior. For T<TCT < T_{\mathrm{C}}, the ε(H)\varepsilon(H) curves become quadratic again, in clear contrast to the V-shaped curves observed in H∥cH \parallel c. This quadratic behavior is consistent with the previous report of εbb\varepsilon_{bb} measured under H∥bH \parallel b [28]. Thus, the V-shaped magnetostriction has strong anisotropy depending on the field direction. This anisotropy is consistent with the PZM tensor shown in Eq. (1): for URhGe, Q21Q_{21}, Q22Q_{22}, and Q23Q_{23} are zero, meaning that no PZM should occur in εaa\varepsilon_{aa}, εbb\varepsilon_{bb}, or εcc\varepsilon_{cc} under H∥bH \parallel b.

Image not extracted

Figure 3. (color online) Magnetostriction of URhGe in (a) εbb\varepsilon_{bb} and (b) εcc\varepsilon_{cc} measured under H∥bH \parallel b. Each curve is vertically offset for clarity. V-shaped magnetostriction was not observed below TCT_{\mathrm{C}} (blue curves), whereas the behavior above TCT_{\mathrm{C}} (red curves) and at TCT_{\mathrm{C}} (green curves) qualitatively resembles those observed under H∥cH \parallel c (Figure 2).

Secondly, we compare ε\varepsilon and MM in more detail. For ordinary magnetostriction, the empirical relation ε∝M2\varepsilon\propto M^2 often holds. We thus plot εbb\varepsilon_{bb} and εcc\varepsilon_{cc} as functions of M2M^2 in Figure 4. One can clearly see that, above TCT_{\mathrm{C}}, the strain is nearly proportional to M2M^2. At TCT_{\mathrm{C}}, where the cusps in ε(H)\varepsilon(H) curves at H=0H = 0 are observed, the strain vs M2M^2 curves acquire non-zero curvature, but maintain smooth relations close to linear. Indeed, when we fit the curves with ε∝M2α\varepsilon\propto M^{2\alpha} using the exponent α\alpha as the fitting parameter, we obtain α=1.095\alpha= 1.095 for T=12T = 12 K and 1.268 for T=9.5T = 9.5 K, both being close to unity. This result manifests that the strain in URhGe above and at TCT_{\mathrm{C}} is attributed mainly to the ordinary magnetostriction. In contrast, the ε(M2)\varepsilon(M^2) curve at 4.0 K does not show power-law relation. Thus, the V-shaped magnetostriction does not have the conventional relation to the magnetization.

Image not extracted

Figure 4. (color online) Relation between εcc\varepsilon_{cc} and M2M^2 under H∥cH \parallel c at various temperatures deduced from the data in Figs. 1-3. The strain is almost proportional to M2M^2 above TCT_{\mathrm{C}} (red solid curve) as demonstrated by the linear fit shown with the red dotted line. At TCT_{\mathrm{C}} (green curve), the curvature is non-zero, but the power-law relation still holds between strain and magnetization. In clear contrast, there is no simple power-law relation between them in the FM phase (blue curves).

Now we explain that the observed V-shaped magnetostriction is indeed well attributable to PZM. Generally in magnets showing TRSB, two magnetic structures connected by the time-reversal operation exhibit opposite signs of piezomagnetic coefficients, respectively. Thus, in typical piezomagnetic materials having the piezomagnetic coefficient QQ for one of the magnetic structures, the strain obeys ε=QH\varepsilon= QH as long as the magnetic structure is kept, while the opposite behavior ε=−QH\varepsilon= -QH emerges when the magnetic structure is reversed by strong magnetic field exceeding HcoerH_{\mathrm{coer}}. This results in butterfly-like ε(H)\varepsilon(H) curves [10, 11]. In URhGe under H∥cH \parallel c, HcoerH_{\mathrm{coer}} is nearly zero, as demonstrated in the step-like M(H)M(H) curve (Figure 1). Zero coercive field changes the butterfly-like ε(H)\varepsilon(H) curve to the V-shaped curve, as observed in URhGe.

These analyses and considerations confirm that the V-shaped magnetostriction observed under H∥cH \parallel c in URhGe originates from PZM, marking the first clear example of PZM in ferromagnets. It is interesting that, although ferromagnets are most archetypal examples exhibiting spontaneous TRSB, required feature to realize PZM as theoretically established in 1956 [5], ferromagnetic PZM has been overlooked for more than 60 years. The near absence of multi-domain state near H=0H = 0 and the strong Ising nature fixing the magnetic point group irrespective of the field directions are the keys to avoid various complications characteristic to ferromagnets.

To investigate the temperature evolution of the PZM, we fitted the data with the function

ε(H)=ε0+a2(μ0H)2+aabs∣μ0H∣(2)\varepsilon(H) = \varepsilon_0 + a_2(\mu_0 H)^2 + a_{\mathrm{abs}}|\mu_0 H| \tag*{(2)}

in the field range −0.5-0.5 T ≤μ0H≤+0.5\leq\mu_0 H \leq+0.5 T. As shown with the dotted curves in Figure 2, the fittings are successful for all data sets. The temperature dependence of a2a_2 and aabsa_{\mathrm{abs}} are shown in Figure 5. Above TCT_{\mathrm{C}}, a2a_2 is dominant, whereas aabsa_{\mathrm{abs}} is within the noise level. The peaks at TCT_{\mathrm{C}} are attributable to the critical behavior in ε(H)\varepsilon(H) discussed in previous paragraphs: when we are forced to fit the cusp-like ε(H)\varepsilon(H) curve using Eq. (2), we mathematically need a large aabs∣H∣a_{\mathrm{abs}}|H| term and similarly large a2H2a_2H^2 term with the opposite sign. Much below TCT_{\mathrm{C}}, aabsa_{\mathrm{abs}} becomes dominant and a2a_2 is nearly zero, as expected from the V-shaped magnetostriction. These results quantitatively show that the behavior of strain drastically changes from the ordinary magnetostriction ε∝H2\varepsilon\propto H^2 above TCT_{\mathrm{C}} to the piezomagnetic response ε∝∣H∣\varepsilon\propto|H| in the FM phase. The fitting coefficient aabsa_{\mathrm{abs}} in the FM phase is equivalent to the corresponding piezomagnetic-tensor component. Taking the lowest-temperature values in the FM state, we obtain Q32=8.0×10−6 T−1Q_{32} = 8.0 \times10^{-6}\ \mathrm{T}^{-1} and Q33=−4.1×10−6 T−1Q_{33} = -4.1 \times10^{-6}\ \mathrm{T}^{-1} for URhGe. These values are compared with results of other piezomagnets. As listed in Table I, most of the known piezomagnets exhibit ∣Qkμ∣\lvert Q_{k\mu}\rvert of less than 2×10−6 T−12 \times10^{-6}\ \mathrm{T}^{-1}, whereas recently found piezomagnets such as UO2_2 and Mn3_3Sn exhibit ∣Qkμ∣\lvert Q_{k\mu}\rvert exceeding 10×10−6 T−110 \times10^{-6}\ \mathrm{T}^{-1} [10, 11]. The observed PZM components of URhGe reach around 50% of these values. This comparison implies that ferromagnets with appropriate conditions are candidate materials realizing large piezomagnetic coefficients. We comment that the signs of QQ is well-defined in our case due to the very soft magnetism, whereas other piezomagnets exhibit PZM of both signs depending on magnetic structures.

Image not extracted

Figure 5. (color online) Temperature dependence of the fitting coefficients a2a_2 and aabsa_{\mathrm{abs}} (Eq. (2)) for the magnetostrictions (a) εbb\varepsilon_{bb} and (b) εcc\varepsilon_{cc} measured under H∥cH \parallel c. The vertical dotted line indicate TCT_{\mathrm{C}}. The dominant magnetic field dependence of the strain changes from ε∼H2\varepsilon\sim H^2 (ordinary magnetostriction) above TCT_{\mathrm{C}} to ε∼∣H∣\varepsilon\sim\lvert H\rvert (PZM) below TCT_{\mathrm{C}}. The singular behaviors at TCT_{\mathrm{C}} is attributable to the critical non-linear behavior in ε(H)\varepsilon(H) as discussed in the text.

One open question is the microscopic mechanism of the observed large PZM in URhGe. Naively, sufficient magnetolattice coupling is necessary to induce PZM. Such magnetolattice coupling should originate from spin magnetic moments with strong spin-orbit coupling (SOC) and/or orbital magnetic moments. Uranium compounds, exhibiting strong SOC and thus angular-momentum coupling between spin and orbitals, should satisfy both criteria. The Ising magnetic nature in URhGe indeed originates from SOC [38]. More recently, it is revealed that the non-symmorphic crystalline structure, resulting from the zig-zag uranium chain, leads to anisotropic pseudo-spin texture pointing perpendicularly to the Brillouin-zone boundaries under SOC [39]. Such pseudo-spin texture may explain microscopic origins of the Ising ferromagnetism in URhGe, and may further provide bases toward clarifying mechanism of PZM.

To summarize, we revealed the piezomagnetism (PZM) in the itinerant Ising ferromagnet URhGe. To our knowledge, this is the first report of ferromagnetic PZM, which has been overlooked for many years. The observed piezomagnetic coefficients range around 50% of the largest ones ever known. This work demonstrates that Ising ferromagnets without multi-domain states can be good candidates when seeking for materials with large piezomagnetic responses, which can be utilized for novel sensors or actuators. This new finding would stimulate further experimental and theoretical studies toward understanding of microscopic mechanisms of PZM, which can be a novel probe of TRSB phenomena occurring in various materials.

Acknowledgments

We acknowledge V. P. Mineev, Y. Yanase, G. Mattoni, S. Kitagawa, D. Aterberg, K. Ogushi and Y. Maeno for fruitful discussion. The work at Kyoto Univ. was supported by Grant-in-Aids for Scientific Research on Innovative Areas “Quantum Liquid Crystals” (KAKENHI Grant Nos. JP20H05158, JP22H04473) from the Japan Society for the Promotion of Science (JSPS), Academic Transformation Area Research (A) “1000 Tesla Science” (KAKENHI Grant No. 23H04861), Grant-in-Aids for Scientific Research (KAKENHI Grant Nos. JP17H06136, JP20H00130, JP22H01168) from JSPS, by Core-to-Core Program (No. JPJSCCA20170002) from JSPS, by Bilateral Joint Research Projects (No. JPJSBP120223205) from JSPS, by research support fundings from The Kyoto University Foundation, by ISHIZUE 2020 and 2023 of Kyoto University Research Development Program, and by Murata Science and Education Foundation. The work at Tohoku Univ. was supported by Grant-in-Aids for Scientific Research (KAKENHI JP22H04933, JP20KK0061, and JP20K20889). S. Yonezawa acknowledges support for the construction of experimental setups from Research Equipment Development Support Room of the Graduate School of Science, Kyoto University; and support for liquid helium supply from Low Temperature and Materials Sciences Division, Agency for Health, Safety and Environment, Kyoto University.

MaterialsQkμQ_{k\mu} (10−6 T−110^{-6}\ \mathrm{T}^{-1})TT (K)MagnetismRefs.
CoF2_2Q14=±21∗Q_{14}=\pm21^{*}20AFM[8]
Q36=±9.8Q_{36}=\pm9.84[29]
MnF2_2Q14=±0.2∗Q_{14}=\pm0.2^{*}20AFM[8]
Q14=±0.07∗Q_{14}=\pm0.07^{*}60[30]
α\alpha-Fe2_2O3_3Q22=±1.9Q_{22}=\pm1.978AFM[31]
Q22=±3.2∗Q_{22}=\pm3.2^{*}77[32]
Q22=±1.3Q_{22}=\pm1.3100[33]
Q23=±2.5∗Q_{23}=\pm2.5^{*}292canted AFM[31]
YFeO3_3Q15=±1.7Q_{15}=\pm1.76AFM[34]
YCrO3_3Q15≃±1Q_{15}\simeq\pm16AFM[35]
DyFeO3_3Q36=±6.0Q_{36}=\pm6.06AFM[36]
UO2_2Q14=±10.5Q_{14}=\pm10.52.5AFM[10]
MnTeQave=0.0050∗Q_{\mathrm{ave}}=0.0050^{*}250AFM / ALM[12]
Mn3_3SnQ11=±4.4∗Q_{11}=\pm4.4^{*}300AFM / MOO[37]
Q11=±14.6Q_{11}=\pm14.6[11]
URhGeQ32=+8.0Q_{32}=+8.02FMThis work
Q33=−4.1Q_{33}=-4.1

Table I. Properties of known piezomagnetic materials [6, 10, 11, 12]. The values with * were obtained by using strain-induced magnetization measurements, while the others by linear-magnetostriction measurements. In the column of magnetization, AFM refers to an antiferromagnet, MOO a magnetic-octupole ordered material, ALM an altermagnet, and FM a ferromagnet.

References

Supplemental Material

Details of experimental methods

In this study, we used a high-quality single crystal of URhGe grown with the Czochralski method in a tetra-arc furnace. A single crystal ingot was oriented by taking Laue photographs, and then cut to the dimensions 1.49×1.46×1.48 mm31.49 \times1.46 \times1.48\ \mathrm{mm}^{3} using a spark cutter. The crystal was subsequently annealed under ultra high vacuum at 900∘C900^{\circ}\mathrm{C} for 10 days. Before strain measurements, we investigated field and temperature dependence of the magnetization using a commercial magnetometer (Quantum Design, MPMS) in order to characterize magnetic properties of the sample.

We used FBG as an optical-fiber-based strain sensor. FBG is a periodic grating embedded to the core an optical fiber. Small modulation of the refractive index nn is formed with the period dd. When light is introduced to this core with the grating, only light with the particular Bragg wavelength (λB=2nd\lambda_{\mathrm{B}} = 2nd) is reflected while others transmit through the fiber. To measure the strain of a sample, FBG is pasted to the sample using a glue. The Bragg wavelength shifts due to changes in dd, which follows the expansion or shrinkage of the sample. Thus we can detect the strain of the sample from the shift of the reflected Bragg wavelength. Including additional contribution due to small changes in nn caused by strain and temperature, we obtain the relation between strain and wavelength as

ΔLL=K0.78(ΔλBλB−ΔλB,freeλB,free),(3)\frac{\Delta L}{L} = \frac{K}{0.78}\left(\frac{\Delta\lambda_{\mathrm{B}}}{\lambda_{\mathrm{B}}} - \frac{\Delta\lambda_{\mathrm{B,free}}}{\lambda_{\mathrm{B,free}}}\right), \tag*{(3)}

where λB,free\lambda_{\mathrm{B,free}} is the Bragg wavelength of the FBG not pasted to the sample, K−1K^{-1} is the strain transmission rate that will be explained later, and the factor 0.78 originates from strain-optic tensor [21].

In this study, we used FBGs made of a bending insensitive fiber. One fiber contains four FBGs whose Bragg wavelengths were 1530 nm, 1540 nm, 1550 nm, and 1560 nm. Two of the FBGs were pasted along the bb and cc axes of the sample with cyanoacrylate glue (Konishi #31701) as illustrated in Figure S1. The other FBGs were placed in free space close to the sample and the light wavelength reflected from them, λB,free\lambda_{\mathrm{B,free}}, were used for the background calibration. Only a corner of the sample was fixed to the sample stage by varnish (GE7301) in order to minimize extrinsic strain from the sample stage. To measure λB\lambda_{\mathrm{B}}, we used a commercial interrogator (KYOWA EFOX-1000B-4), which is equipped with the variable-wavelength laser and an optical spectrometer. The scanning frequency of the interrogator was set to be 100 Hz. The sample was cooled down with a commercial cryostat (Quantum Design, PPMS) equipped with a horizontal superconducting split-coil magnet. To measure the accurate temperature and magnetic field in the sample space, we placed a thermometer (Lake Shore, Cernox CX-1050) and a Hall sensor (Toshiba, THS118) in the vicinity of the sample. These two sensors were calibrated before measurements.

Image not extracted

Figure S1. Schematic to explain how to paste FBG sensors to the sample. We pasted two FBGs parallel to the bb and cc axes of the sample, respectively, in order to perform multi-axis simultaneous strain measurements. We performed two sets of experiments with the applied magnetic field along either the bb or cc axis.

Before the measurement of magnetostriction, it is necessary to calibrate the strain transmission rate K−1K^{-1}. For that purpose, we measured zero-field thermal expansion using our setup and compare it with results of a previous research [24]. Figure S2 compares the raw observed strain assuming K=1K=1 and the literature data. The apparent difference in the two data sets originates from imperfect transmission of the strain through the glue. We determined K−1K^{-1} so that our data multiplied by KK matches the literature value. As shown in Figure S2, the corrected data with K−1=0.840K^{-1}=0.840 for the bb axis and K−1=0.433K^{-1}=0.433 for the cc axis both match the literature data in particular below TCT_{\mathrm{C}}. Strain data shown in the rest of this Letter are corrected by using these KK values.

Additional experimental data

Image not extracted

Figure S3. Comparison of εbb(H)\varepsilon_{bb}(H) (red, green, blue curves; left axis) and M2(H)M^2(H) (purple curves; right axis) of URhGe measured under H∥cH \parallel c at various temperatures.

Image not extracted

Figure S4. Comparison of εcc(H)\varepsilon_{cc}(H) (red, green, blue curves; left axis) and −M2(H)-M^2(H) (purple curves; right axis) of URhGe measured under H∥cH \parallel c at various temperatures.

References

  1. [1]N. A. Spaldin and M. Fiebig, “The Renaissance of Magnetoelectric Multiferroics” Science 309, 391 (2005).pubmed.ncbi.nlm.nih.gov/16020720
  2. [2]Y. Tokura and S. Seki, “Multiferroics with Spiral Spin Orders” Adv. Mater. 22, 1554 (2010).DOI
  3. [3]N. A. Spaldin and R. Ramesh, “Advances in magnetoelectric multiferroics” Nature Mater. 18, 203 (2019).DOI
  4. [4]V. Woldemar, Lehrbuch der kristallphysik (mit ausschluss der kristalloptik), nachdruck der ersten aufl. ed., B.G. Teubners, Bd., Vol. 34 (B.G. Teubner, [J.W. Edwards], Leipzig, Berlin, 1928).
  5. [5]B. A. Tavger and V. M. Zaitsev, “Magnetic Symmetry of Crystals” Sov. Phys. JETP 30, 564 (1956).
  6. [6]A. S. Borovik-Romanov, “Piezomagnetism, linear magnetostriction and magnetooptic effect” Ferroelectrics 162, 153 (1994).DOI
  7. [7]A. S. Borovik-Romanov, “Piezomagnetism in the Antiferromagnetic Fluorides of Cobalt and Manganese” Sov. Phys. JETP 9, 1360 (1959).
  8. [8]A. S. Borovik-Romanov, “Piezomagnetism in the Antiferromagnetic Fluorides of Cobalt and Manganese” Sov. Phys. JETP 11, 786 (1960).
  9. [9]I. E. Dzyaloshinskii, “The problem of piezomagnetism” Sov. Phys. JETP 6, 621 (1958).
  10. [10]M. Jaime, A. Saul, M. Salamon, V. Zapf, N. Harrison, T. Durakiewicz, J. Lashley, D. Andersson, C. Stanek, J. Smith, et al., “Piezomagnetism and magnetoelastic memory in uranium dioxide” Nat. Commun. 8, 99 (2017).DOI
  11. [11]Q. Meng, P. Nie, L. Xu, J. Wang, S. Jiang, H. Zuo, X. Li, Z. Zhu, L. Balents, and K. Behnia, “Magnetostriction, piezomagnetism and domain nucleation in Mn₃Sn” (2023), arXiv:2301.06401 [cond-mat.str-el].
  12. [12]T. Aoyama and K. Ogushi, “Piezomagnetic Properties in Al-termagnetic MnTe” (2023), arXiv:2305.14786 [cond-mat.mtrl-sci].
  13. [13]P. A. McClarty and J. G. Rau, “Landau Theory of Altermagnetism” (2023), arXiv:2308.04484 [cond-mat.mtrl-sci].DOI
  14. [14]B. D. Cullity, “Fundamentals of Magnetostriction” J. Met. 23, 35 (1971).DOI
  15. [15]Y. L. Bras and J.-M. Greneche, “Resonance” (IntechOpen Limited, London, 2017) Chap. Magneto-Elastic Resonance: Principles, Modeling and Applications.
  16. [16]V. Tran, R. Troć, and G. Andre, “Magnetic ordering in URhSi and URhGe” J. Mag. Mag. Mater. 186, 81 (1998).DOI
  17. [17]F. De Boer, E. Brück, V. Sechovsky, L. Havela, and K. Buschow, “UTX compounds in high magnetic fields” Physica B 163, 175 (1990).DOI
  18. [18]D. Aoki, A. Huxley, E. Ressouche, D. Braithwaite, J. Flouquet, J.-P. Brison, E. Lhotel, and C. Paulsen, “Coexistence of superconductivity and ferromagnetism in URhGe” Nature 413, 613 (2001).pubmed.ncbi.nlm.nih.gov/11595943
  19. [19]D. Aoki, K. Ishida, and J. Flouquet, “Review of U-based Ferromagnetic Superconductors: Comparison between UGe₂, URhGe, and UCoGe” J. Phys. Soc. Jpn. 88, 022001 (2019).DOI
  20. [20]V. P. Mineev, “Piezomagnetism in Ferromagnetic Superconductors” J. Exp. Theor. Phys. 132, 556 (2021).DOI
  21. [21]M. Jaime, C. C. Moya, F. Weickert, V. Zapf, F. Balakirev, M. Wartenbe, P. Rosa, J. Betts, G. Rodriguez, S. Crooker, and R. Daou, “Fiber Bragg Grating Dilatometry in Extreme Magnetic Field and Cryogenic Conditions” Sensors 17, 2572 (2017).
  22. [22]See Supplemental Material at [URL will be inserted by publisher] for additional information on this experiment.
  23. [23]D. Braithwaite, D. Aoki, J.-P. Brison, J. Flouquet, G. Knebel, A. Nakamura, and A. Pourret, “Dimensionality Driven Enhancement of Ferromagnetic Superconductivity in URhGe” Phys. Rev. Lett. 120, 037001 (2018).DOI
  24. [24]D. Aoki, F. Hardy, A. Miyake, V. Taufour, T. D. Matsuda, and J. Flouquet, “Properties of ferromagnetic superconductors” C. R. Phys. 12, 573 (2011).DOI
  25. [25]S. Blundell, Magnetism in Condensed Matter (Oxford University Press, Oxford, 2001).
  26. [26]W. Knafo, C. Meingast, S. Sakarya, N. H. van Dijk, Y. Huang, H. Rakoto, J.-M. Broto, and H. v. Löhneysen, “Critical Scaling of the Magnetization and Magnetostriction in the Weak Itinerant Ferromagnet URhGe” J. Phys. Soc. Jpn. 78, 043707 (2009).DOI
  27. [27]E. Gerasimov, A. Inishev, P. Terentev, V. Kazantsev, and N. Mushnikov, “Magnetostriction and thermal expansion of nonstoichiometric TbCo₂Mn compounds” J. Mag. Mag. Mater. 523, 167628 (2021).DOI
  28. [28]S. Nakamura, T. Sakakibara, Y. Shimizu, S. Kittaka, Y. Kono, Y. Haga, J. Pospišil, and E. Yamamoto, “Magnetization study on the ising ferromagnet URhGe with high-precision angle-resolved magnetic field near the hard axis” Prog. Nucl. Sci. Technol. 5, 123 (2018).DOI
  29. [29]A. S. Prokhorov and E. G. Rudashevskii, “Magnetostriction of Antiferromagnetic Cobalt Fluoride” JETP Lett. 10, 175 (1969).
  30. [30]J. Barruchel, A. Draperi, M. El Kadiri, G. Fillion, M. Maeder, P. Molho, and J. L. Porteseil, “Piezomagnetism and Domains In MnF₂” J. Phys. Colloques 49, C8 (1988), ISSN 0449-1947.DOI
  31. [31]V. P. Andratskii and A. S. Borovik-Romanov, “Piezomagnetic Effect In α-Fe₂O₃” Sov. Phys. JETP 24, 687 (1966).
  32. [32]J. C. Anderson, R. R. Birss, and R. A. M. Scott, “Linear magnetostriction in hematite” in Proc. Int. Conf. on Magnetism (Nottingham, London: Institute of Physics and the Physical Society, 1964) p. 597.
  33. [33]R. Z. Levitin and V. A. Shchurov, “Physics and Chemistry of Ferrites” (Russian, Moscow: Izd. Mosk. Gos. Univ, 1964).
  34. [34]A. M. Kadomtseva, A. P. Agafonov, V. N. Milov, A. S. Moskvin, and V. A. Semenov, “Direct observation of a symmetry change induced in orthoferrite crystals by an external magnetic field” JTEP Lett. 33, 400 (1981).
  35. [35]A. Kadomtseva, A. Agafonov, M. Lukina, V. Milov, A. Moskvin, V. Semenov, and E. Sinitsyn, “Nature of magnetic anisotropy and magnetostriction of orthoferrites and orthochromites” Sov. Phys. JETP 54, 374 (1981).
  36. [36]A. K. Zvezdin, I. A. Zorin, A. M. Kadomtseva, I. B. Krynetskii, A. S. Moskvin, and A. A. Mukhin, “Linear magnetostriction and antiferromagnetic domain structure in dysprosium orthoferrite” Sov. Phys. JETP 61, 645 (1985).
  37. [37]M. Ikhlas, S. Dasgupta, F. Theuss, T. Higo, S. Kittaka, B. J. Ramshaw, O. Tchernyshyov, C. W. Hicks, and S. Nakatsuji, “Piezomagnetic switching of the anomalous Hall effect in an antiferromagnet at room temperature” Nat. Phys. 18, 1086 (2022).DOI
  38. [38]F. Wilhelm, J. P. Sanchez, J.-P. Brison, D. Aoki, A. B. Shick, and A. Rogalev, “Orbital and spin moments in the ferromagnetic superconductor URhGe by x-ray magnetic circular dichroism” Phys. Rev. B 95, 235147 (2017).DOI
  39. [39]H. G. Suh, Y. Yu, T. Shishidou, M. Weinert, P. M. R. Brydon, and D. F. Agterberg, “Superconductivity of anomalous pseudospin in nonsymmorphic materials” Phys. Rev. Res. 5, 033204DOI

Paper details

vlnc:W_65e4wdmx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents