This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Discovery of antiferromagnetic chiral helical ordered state in trigonal GdNi3Ga9

Shota Nakamura1, Takeshi Matsumura2, Kazuma Ohashi1, Hiroto Suzuki1, Mitsuru Tsukagoshi2, Kenshin Kurauchi2, Hironori Nakao3, Shigeo Ohara1 1Department of Physical Science and Engineering, Nagoya Institute of Technology, Nagoya 466-8555, Japan 2Department of Quantum Matter, ADSE, Hiroshima University, Higashihiroshima, Hiroshima 739-8530, Japan 3Photon Factory, Institute of Materials Structure Science, High Energy Accelerator Research Organization, Tsukuba, 305-0801, Japan
Abstract

We have performed magnetic susceptibility, magnetization, and specific heat measurements on a chiral magnet GdNi3Ga9, belonging to the trigonal space group R32R32 (#155). A magnetic phase transition takes place at TNT_{\rm N} = 19.5 K. By applying a magnetic field along the aa axis at 2 K, the magnetization curve exhibits two jumps at 3\sim 3 kOe and = 45 kOe. To determine the magnetic structure, we performed a resonant X-ray diffraction experiment by utilizing a circularly polarized beam. It is shown that a long-period antiferromagnetic (AFM) helical order is realized at zero field. The Gd spins in the honeycomb layer are coupled in an antiferromagnetic manner in the cc plane and rotate with a propagation vector 𝒒{\bf\it q} = (0, 0, 1.485). The period of the helix is 66.7 unit cells (180\sim 180 nm). In magnetic fields above 3 kOe applied perpendicular to the helical cc axis, the AFM helical order changes to an AFM order with 𝒒{\bf\it q} = (0, 0, 1.5).

I Introduction

Asymmetry in space and time often plays an important role in unconventional physical phenomena. Chirality is a property of asymmetry essential in cross-disciplinary subjects and nature at all large scales. In structures with chirality, there are two types of structures, i.e., left- and right-handed mirror copies. The difference in chirality, or the handedness, of crystals often appears in macroscopic physical properties. In magnets, chiral symmetry breaking in crystal structures plays an essential role in stabilizing a macroscopic ordering state with left- or right-handed incommensurate twists of magnetic spins.

Magnetic structures reflecting the crystal chiralities have been identified mainly in dd electron systems and have been actively studied. The skyrmion lattice, which is a magnetic vortex, is observed in MnSi Mühlbauer et al. (2009), and a uniaxial helical magnetic structure with a periodic arrangement of ”twists”, which is called a chiral soliton lattice (CSL), is observed in CrNb3S6 Togawa et al. (2012, 2016).

Refer to caption
Figure 1: (Color Online) (a) The unit cell of GdNi3Ga9. Ga(5), which is located in the Gd honeycomb layer, is colored green. (b) Gd honeycomb layer (z1/6z\sim 1/6). This figure contains 9 unit cells. Gd forms a honeycomb network. Ga(5) forms a small triangle at the center of this network. (c) Spiral in the crystal structure of GdNi3Ga9. The figure shows the cc axis spiral of Ga(5). Hard to see because the changes are so small, atoms in general positions, such as Ni and Ga(6), are also arranged in spirals in addition to Ga(5). In this figure, Ga(1)-Ga(4) was eliminated for simplicity. VESTA was used to draw the figure Momma and Izumi (2011).

Recently, chiral magnets have been explored in 4ff electron system. The examples are uniaxial helical magnetic structures in YbNi3Al9 Ohara et al. (2011); Matsumura et al. (2017), DyNi3Ga9 Ninomiya et al. (2017); Ishii et al. (2018, 2019); Tsukagoshi et al. (2022), and GdPt2Sato et al. (2022), and the skyrmion lattices in EuPtSi Kakihana et al. (2018, 2019); Tabata et al. (2019); Sakakibara et al. (2019). The period of the helical magnetic structure in 4ff chiral magnet, for example in YbNi3Al9 and EuPtSi, is approximately an order of magnitude smaller than that in dd electron systems Togawa et al. (2016). In the case of the helical magentic order of YbNi3Al9, the angle between neighboring spins along the cc axis are more than 90 degrees, indicating that the exchange interaction is mediated predominantly by the conduction electrons (the Ruderman-Kittel-Kasuya-Yosida mechanism), and the weak Dzyaloshinskii-Moriya (DM)-type antisymmetric exchange interaction selects the sense of rotation.

Refer to caption
Figure 2: (Color Online) Scattering configuration of the RXD experiments with a phase retarder system inserted in the incident x-ray.

The target substance of this study is GdNi3Ga9 Topertser et al. (2019), with the same crystal structure as YbNi3Al9, which has a ErNi3Al9-type crystal structure with a space group RR32 (#155) Nakamura et al. (2020). Figure 1 shows the crystal structure of GdNi3Ga9. We call this structure as right-handed and the mirror image of this structure as left-handed. The handedness can be judged by the Flack parameter of single crystal x-ray experiments. The lattice constants are aa == 0.7264 and cc == 2.7497 nm. The crystal structure of GdNi3Ga9 is characterized by the honeycomb layers of rare earth elements stacked along the helical cc axis, as shown in Figs. 1(a) and 1(b). Note that the Sohncke type space group RR32 is not chiral, but the crystal structure has a chirality. Ga and Ni atoms form spirals, as shown in Fig. 1(c).

Gd compounds are suitable for investigating chiral magnetism because of the zero orbital angular momentum of Gd3+ (4f74f^{7}, S7/28{}^{8}{\rm S}_{7/2}, μeff=7.94μB/Gd\mu_{\rm eff}=7.94\mu_{\rm B}/{\rm Gd}, μJ=gJμBJ=7μB\mu_{J}=g_{J}\mu_{\rm B}J=7\mu_{\rm B}, and gJ=2g_{J}=2). This reduces the effects of crystal-field anisotropy and allows the magnetic moments of Gd to direct to any favorable directions determined by the exchange interactions.

Although neutron diffraction experiments are usually used to investigate the magnetic structure Ishida et al. (1985); Yamasaki et al. (2007), detailed data collection for Gd compounds is difficult because Gd is a strong neutron absorber. To observe the helical magnetic structure and the helicity of GdNi3Ga9, we performed resonant x-ray diffraction (RXD) using circularly polarized x-rays Sutter et al. (1997); Fabrizi et al. (2009); Sagayama et al. (2010).

II Experimental Procedure

We synthesized a single crystalline GdNi3Ga9 by the flux method using gallium as the solvent. The starting materials Gd ingot (purity of 99.9 %), Ni ingot (99.999 %), and Ga ingot (99.9999 %) were prepared in a molar ratio of Gd: Ni: Ga = 1: 3: 30. These materials were placed in an alumina crucible and then sealed into a quartz tube under high vacuum 1×103\sim 1\times 10^{-3} Pa. The ampule was heated to 900C and maintained at this temperature for 5 hours. Then, it was slowly cooled at a rate of 5C/h. The excess gallium was removed by centrifugation at 500 C.

In this work, the magnetization and specific heat measurements have been performed using a commercial magnetic properties measurement system (MPMS, Quantum Design) and physical properties measurement system (PPMS, Quantum Design). The specific heat was measured between 2 and 300 K at zero magnetic field, and the magnetization was measured between 2 and 300 K in magnetic fields up to 5.5 T.

The RXD experiments were performed at BL-3A, Photon Factory at KEK, Japan. The sample with a mirror polished cc plane surface was inserted into an 8 T cryo-magnet and cooled down to 3 K. In this paper, the cc plane denotes the (001) plane. The cc axis was perpendicular to the magnetic field. The scattering geometry is shown in Fig. 2. A diamond phase retarder system was used to tune the horizontally polarized incident beam to a circularly polarized state Matsumura et al. (2017). A phase difference of the incident x-ray between the σ\sigma and π\pi polarization is generated by rotating the angle of the diamond phase plate θPR\theta_{\rm PR} around the 111 Bragg angle θB\theta_{\rm B}, which is 22.3 at 7.932 keV, where the scattering plane is tilted by 45. We can tune the incident linear polarization to right-handed circular polarization (RCP) and left-handed circular polarization (LCP) by changing ΔθPR\Delta\theta_{\rm PR}, where ΔθPR=θPRθB\Delta\theta_{\rm PR}~{}=~{}\theta_{\rm PR}-\theta_{\rm B}.

We used an x-ray beam with a wavelength of 0.1711  and 0.1563 nm (7.246 and 7.932 keV), which are near the L3L_{3} and L2L_{2} edges of Gd, respectively. In GdNi3Ga9, the (1, 0, 3nn+1) reflections are allowed and the (1, 0, 3nn) and (1, 0, 3nn-1) reflections are forbidden, because of the selection rule derived from the crystal symmetry.

III Results and Analysis

Refer to caption
Figure 3: (Color Online) (a) Temperature dependences of magnetic susceptibility obtained in the magnetic fields of 0.05, 1, 10, and 50 kOe along the aa axis, together with that obtained in HcH\parallel c at 1 kOe. (b) Magnetization curves were measured at 2, 4, 6, 10, and 15 K with increasing the magnetic field. (c) Enlarged view of low magnetic field region of Fig. 3(b).
Refer to caption
Figure 4: (Color Online) HTH-T phase diagram of GdNi3Ga9 in HaH~{}\parallel~{}a. The dots in this figure are obtained by magnetization measurements. The α\alpha phase is the AFM helical phase.
Refer to caption
Figure 5: (Color Online) Temperature dependence of specific heat divided by temperature Cmag/TC_{\rm mag}/T, together with the entropy SmagS_{\rm mag}.

Figure 3(a) shows the temperature dependence of magnetic susceptibility (χ\chi-TT) of GdNi3Ga9, which was measured at 0.05, 1, 10, and 50 kOe applied along the aa axis (HaH\parallel a). In this figure, χ\chi-TT for HcH\parallel c at 1 kOe is also shown. This result suggests that the Gd spins in a honeycomb layer of Fig. 1(b) are antiparallel to each other below TNT_{\rm N}\sim 19.5 K. As shown afterwards, this material exhibits antiferromagnetic (AFM) helical order where magnetic propagation vector 𝒒c{\bf\it q}~{}\parallel~{}c^{*}. At low magnetic fields of 0.05 and 1 kOe, a characteristic cusp is observed just below TNT_{\rm N}, and χ\chi decreases monotonically with decreasing temperature. This cusp is similar to the one theoretically predicted to appear when the chiral helix is realized Kishine et al. (2005); Kousaka et al. (2007). This anomaly is not observed in HcH\parallel c. As the magnetic field is increased, the behavior of χ\chi-TT curve changes below TNT_{\rm N}. At 10 kOe, χ\chi decreases below TNT_{\rm N} and then increases slightly at low temperatures. At 50 kOe, bending of χ\chi due to AFM ordering becomes gradual. Warming from TNT_{\rm N} to room temperature, χ\chi-TT curve of this material is in good agreement with the Curie-Weiss law, and the Curie-Weiss temperature is estimated to be -8.7 and -11.2 K for the aa and cc axes, respectively. Although not shown in this paper, this material is isotropic in the cc plane, and similar magnetization curves are obtained when a magnetic field is applied in the aa^{*} direction.

Figure 3(b) shows magnetization curves measured at 2, 4, 6, 10, and 15 K. All magnetization curves jump twice at Hα3H_{\alpha}\sim 3 kOe and Hβ45H_{\beta}\sim 45 kOe. Except for these two jumps, magnetization increases linearly. With increasing temperature, HαH_{\alpha} and HβH_{\beta} shift toward higher and lower magnetic fields, respectively. Figure 3(c) shows an enlarged view of the low magnetic field region in Fig. 3(b). In this figure, we can see the shift of HαH_{\alpha} to the high field side with increasing temperature. The metamagnetic behavior at critical magnetic fields during the linearly increasing magnetization process is a feature that appears in the magnetization process of the AFM helix when a magnetic field is applied perpendicular to the helix axis. Kishine et al. (2005); Kousaka et al. (2007)

Figure 4 shows the HTH-T phase diagram of GdNi3Ga9 for HaH\parallel a obtained by the magnetization measurements. The magnetically ordered phases are named α\alpha, β\beta, and γ\gamma phases in the order from low to high fields. HαH_{\alpha} and HβH_{\beta} are located between the α\alpha and β\beta phases, and the β\beta and γ\gamma phases, respectively. γ\gamma phase changes to the forced-ferromagnetic state at HFFH_{\rm FF}\sim 200 kOe, when we assume that the magnetization curve of Fig. 3(b) behaves linearly up to 7μB\mu_{\rm B}, the saturation magnetization of Gd3+. As will be discussed later, the α\alpha phase is an AFM helical phase, and the β\beta and γ\gamma phases are an AFM phase.

Figure 5 shows the temperature dependence of specific heat divided by temperature (Cmag/TC_{\rm mag}/T) at zero field, together with the entropy (SmagS_{\rm mag}). Cmag/TC_{\rm mag}/T is obtained by subtracting the specific heat of nonmagnetic LuNi3Ga9. SmagS_{\rm mag} is obtained from Cmag/TC_{\rm mag}/T assuming that Cmag/TC_{\rm mag}/T increases linearly from (TT, Cmag/TC_{\rm mag}/T) = (0, 0). A peak of Cmag/TC_{\rm mag}/T appears just below TN19.5T_{\rm N}\sim 19.5 K. The entropy is estimated to be \sim Rln4R{\rm ln4} at TNT_{\rm N}. Below TNT_{\rm N}, a shoulder peak is observed around 7 K. Such a shoulder has often been observed at around TNT_{\rm N}/4 in Gd3+ compounds with JJ = 7/2 Hidaka et al. (2020); Bouvier et al. (1991), and it is attributed to a splitting of the JJ = 7/2 multiplet due to an internal field Blanco et al. (1991).

Refer to caption
Figure 6: (Color Online) X-ray energy EE dependence of the (1, 0, 20.48) reflection around E=7.246E=7.246 keV, together with the background (black curve) obtained with θ\theta shifted slightly from (1, 0, 20.48).
Refer to caption
Figure 7: (Color Online) Temperature dependence of the (-1, 0, LL) peak profile at zero magnetic field in the AFM helical α\alpha phase.
Refer to caption
Figure 8: (Color Online) Temperature variation of the (-1, 0, 17+qq) peak intencity shown in Fig. 7.

To observe the helical magnetic structure and the magnetic helicity of GdNi3Ga9, a RXD study has been performed. The measurement was conducted for the right-handed crystal, whose structure is shown in Fig. 1. The crystal chirality is reflected in helicity of the magnetic structure by a one-by-one relationship in YbNi3Al9 Matsumura et al. (2017), and left-handed crystals are expected to describe the inverted helix of the magnetic structure of this paper.

We found the resonant Bragg reflection from magnetic order at (1, 0, 3n+1±q3n+1\pm q) and (-1, 0, 3n1±q3n-1\pm q) with q=1.485q~{}=~{}1.485 at 3 K. The X-ray energy dependence for the (1, 0, 20.48) peak is shown in Fig. 6. The intensity enhancement due to the E1E1 resonance (2p5d2p~{}\leftrightarrow~{}5d) is observed around the L3L_{3}-edge of Gd, indicating that the resonant signal reflects the magnetic ordering of the Gd 4f4f-spins. The peak position of the intensity is 7.246 keV. It is noted that no reflection was found along the (0, 0, LL) line at (0, 0, 3n±q3n\pm q), indicating that the magnetic moments of the two Gd spins on the same honeycomb layer are antiparallel. The detail will be explained later.

Figure 7 shows the temperature dependence of the (-1, 0, LL) peak profile at zero magnetic field in the AFM helical phase. The two peaks at (-1, 0, 18.485) and (-1, 0, 18.515) at 3 K correspond to (-1, 0, 17+qq) and (-1, 0, 20q-q) with q=1.485q=1.485. The peak intensity decreases with increasing TT, and disappears at TNT_{\rm N} as shown in Fig. 8. As explained later, the magnetic moments lie in the cc plane and exhibit a helical rotation along the cc axis. The qq-value of 1.485 means that the spins rotate by 178.2 between the adjacent Gd honeycomb layers, and 360 with 0.673 unit cells (approximately two Gd honeycomb layers). The deviation of the angle from 180 gives rise to the twist angle of the helix of 1.8 corresponding to qq = 0.015, which is the difference between 1.485 and 1.5. The period of the AFM helix with the twist angle of 1.8 is 66.7 unit cells (\sim 180 nm, 200 layers), which is required to come back to the same direction as the original layer.

This long period of the helix is comparable with those of chiral magnets of dd electron systems. The magnetic propagation vector decreases slightly to 𝒒={\bf\it q}~{}= (0, 0, 1.48) by increasing temperature to TNT_{\rm N}. When 𝒒={\bf\it q}~{}= (0, 0, 1.48), the period of the AFM helix becomes 3/4 times shorter than the one at 3 K, indicating that the AFM helix is more twisted at higher temperatures.

To show that the helical order has a unique sense of rotation, which should be determined by the antisymmetric interaction, we investigated the intensity variation by manipulating the circular polarization of the incident beam. The RCP and LCP are obtained by changing ΔθPR\Delta\theta_{\rm PR} = θPRθB\theta_{\rm PR}-\theta_{\rm B}, where θPR\theta_{\rm PR} and θB\theta_{\rm B} denote the angle of the diamond phase plate and the 111 Bragg angle, respectively. It is convenient to express the polarization state of the X-ray photon by the Stokes vector 𝑷{\bf\it P} = (P1P_{1}, P2P_{2}, P3P_{3}), where P1P_{1}, P2P_{2}, and P3P_{3} denote the degrees of ±\pm45 (P1P_{1} = ±\pm1) linear polarization, the RCP (P2P_{2} = 1) or LCP (P2P_{2} = -1), and σ\sigma (P3P_{3} = 1) or π\pi (P3P_{3} = -1) linear polarization state, respectively. In our geometry shown in Fig. 2, the three parameters are expressed as P1P_{1} = 0, P2P_{2} = sin(A/ΔθPR\Delta\theta_{\rm PR}) and P3P_{3} = -cos(A/ΔθPR\Delta\theta_{\rm PR}), where A is a constant determined experimentally. The LCP and RCP are obtained when ΔθPR\Delta\theta_{\rm PR} = -0.0358 and 0.0358 deg., respectively, whose positions are indicated by vertical dotted lines in Fig. 9.

Refer to caption
Figure 9: (Color Online) ΔθPR\Delta\theta_{\rm PR} scans, which show the intensity relations for the RCP and LCP incident x-rays at (a) (-1, 0, 17+qq) and (-1, 0, 20q-q), (b) (1, 0, 19+qq) and (1, 0, 22q-q) reflection peaks.
Refer to caption
Figure 10: (Color Online) Schematic view of the AFM helical and commensurate AFM structures in the α\alpha and β\beta phases of GdNi3Ga9, respectively. In this figure, the twist angle is emphasized about 10 times. VESTA was used to draw the figure Momma and Izumi (2011).

Figure 9 shows the ΔθPR\Delta\theta_{\rm PR} scans, which demonstrate that this material has the helical magnetic structure in the α\alpha phase. The ΔθPR\Delta\theta_{\rm PR} dependences for the (-1, 0, 17+qq) and the (-1, 0, 20q-q) peaks are opposite to each other, indicating that the magnetic structure in the α\alpha phase has a chirality. The intensity of the (-1, 0, 17+qq) peak is larger (smaller) than that of (-1, 0, 20q-q) for LCP (RCP). The same result is obtained for (1, 0, 19+qq) and (1, 0, 22q-q) as well. The solid lines in Fig. 9 are the calculated curves of the ΔθPR\Delta\theta_{\rm PR} scans by assuming an AFM helical magnetic structure as described below. This calculation agrees well with the ΔθPR\Delta\theta_{\rm PR} scan data.

Let us analyze the magnetic structure in the α\alpha phase. The magnetic moments of Gd, which are antiferromagnetically coupled in the honeycomb layer in the cc plane, are helically twisted along the cc axis with a very long pitch. There are two Gd atoms for the 6c6c site of the unit cell of GdNi3Ga9 with a space group of R32R32: Gd-1 at 𝒅1{\bf\it d}_{1} = (0, 0, zz), and Gd-2 at 𝒅2{\bf\it d}_{2} = (0, 0, z¯\bar{z}), where zz~{}\sim 1/6. In the present case, the magnetic moments are expected to be antiparallel to each other in the cc plane. There are six Gd positions in the unit cell, as indicated by the Wyckoff letter 6c6c of Gd in GdNi3Ga9. The set of Gd-1 and Gd-2 are chosen to be closest to each other, and the one-to-three assignments are determined, where positions are connected by the stacking vector (2/3, 1/3, 1/3) each other. The unit vector 𝝁1,j{\bf\it\mu}_{1,j} and 𝝁2,j{\bf\it\mu}_{2,j} parallel to magnetic moments of Gd-1 and Gd-2, respectively, on the jj-th lattice point at 𝒓j{\bf\it r}_{j} = (n1n_{1}, n2n_{2}, n3n_{3}), (n1n_{1}+2/3, n2n_{2}+1/3, n3n_{3}+1/3), and (n1n_{1}+1/3, n2n_{2}+2/3, n3n_{3}+2/3), where n1n_{1}, n2n_{2}, and n3n_{3} are integers, are expressed as

𝝁1,j=𝒙^cos2π𝒒𝒓j+𝒚^cos(2π𝒒𝒓j+φ),\displaystyle{\bf\it\mu}_{1,j}=\hat{{\bf\it x}}~{}{\rm cos}2\pi{\bf\it q}\cdot{\bf\it r}_{j}+\hat{{\bf\it y}}~{}{\rm cos}(2\pi{\bf\it q}\cdot{\bf\it r}_{j}+\varphi), (1)
𝝁2,j=𝒙^cos(2π𝒒𝒓j+δ)+𝒚^cos(2π𝒒𝒓j+φ+δ).\displaystyle{\bf\it\mu}_{2,j}=\hat{{\bf\it x}}~{}{\rm cos}(2\pi{\bf\it q}\cdot{\bf\it r}_{j}+\delta)+\hat{{\bf\it y}}~{}{\rm cos}(2\pi{\bf\it q}\cdot{\bf\it r}_{j}+\varphi+\delta). (2)

The 𝒙^\hat{{\bf\it x}} and 𝒚^\hat{{\bf\it y}} denote the unit vectors along the xx- and yy-axis, respectively, which are perpendicular to the cc (zz) axis. φ\varphi takes values of ±π\pm\pi/2, and describes a right- or left-handed helical structure. When δ\delta = 178.2, the two spins on the same layer are perfectly antiferromagnetic for qq = 1.485. Note that 𝝁1,j{\bf\it\mu}_{1,j} and 𝝁2,j{\bf\it\mu}_{2,j} are different layers, and they have a relation of 𝝁1,j=𝝁2,j+1{\bf\it\mu}_{1,j}=-{\bf\it\mu}_{2,j+1} on the same layer. A schematic view of the magnetic structure in the α\alpha phase is shown in Fig. 10(a). The twist angle of the helix is emphasized by 10 times.

The calculated ΔθPR\Delta\theta_{\rm PR} dependence of the intensity for the above AFM-helical model well explains the data in Fig. 9, which were obtained at the Gd L2L_{2} edge. The E1E1 resonant scattering amplitude from the magnetic dipole order is proportional to i(𝜺×𝜺)𝒁dip(1)i({\bf\it\varepsilon}^{\prime}\times{\bf\it\varepsilon})\cdot{\bf\it Z}_{\rm dip}^{(1)} Hannon et al. (1988); Lovesey et al. (2005), where

𝒁dip(1)=d,j𝝁d,jei2π𝑸(𝒓j+𝒅)\displaystyle{\bf\it Z}_{\rm dip}^{(1)}~{}=~{}\sum_{d,j}{\bf\it\mu}_{d,j}e^{-i2\pi{\bf\it Q}\cdot({\bf\it r}_{j}+{\bf\it d})} (3)

represents the magnetic dipolar structure factor at the scattering vector 𝑸=𝒌𝒌{\bf\it Q}={\bf\it k}^{\prime}-{\bf\it k}. Here, 𝜺{\bf\it\varepsilon} and 𝜺{\bf\it\varepsilon}^{\prime} denote the polarization vector of the incident and diffracted beam, respectively. In the horizontal scattering plane configuration in this experiment, the incident linear polarization is π\pi when |ΔθPR|\absolutevalue{\Delta\theta_{\rm PR}} is large. At (-1, 0, 17+qq), 𝒁dip(1){\bf\it Z}_{\rm dip}^{(1)} is calculated to be (1, i\mp i, 0) for φ=±π\varphi=\pm\pi/2. The sign is reversed at (-1, 0, 20q-q), where 𝒁dip(1){\bf\it Z}_{\rm dip}^{(1)} = (1, ±i\pm i, 0) for φ=±π\varphi=\pm\pi/2. For (0, 0, 3n+-q), 𝒁dip(1){\bf\it Z}_{\rm dip}^{(1)} vanishes when the two Gd moments on the same layer are antiferromagnetic by setting δ\delta = 178.2.

We use the scattering-amplitude-operator method to analyze the polarization properties of the experimental results Lov . We consider a 2×22\crossproduct 2 matrix F^\hat{F}, consisting of four elements of the scattering amplitudes for the polarization arrangements σσ\sigma-\sigma^{\prime}, πσ\pi-\sigma^{\prime}, σπ\sigma-\pi^{\prime}, and ππ\pi-\pi^{\prime}:

As shown in Appendix B in Ref. Matsumura et al. (2017), the scattering cross section (dσ\sigma/dΩ\Omega) is expressed as

(dσdΩ)=12(|Fσσ|2+|Fσπ|2+|Fπσ|2+|Fππ|2)+P1Re{FπσFσσ+FππFσπ}+P2Im{FπσFσσ+FππFσπ}+12P3(|Fσσ|2+|Fσπ|2|Fπσ|2|Fππ|2)=C0+C1P1+C2P2+C3P3,\begin{split}(\frac{d\sigma}{d\Omega})&=\frac{1}{2}(|F_{\sigma\sigma^{\prime}}|^{2}+|F_{\sigma\pi^{\prime}}|^{2}+|F_{\pi\sigma^{\prime}}|^{2}+|F_{\pi\pi^{\prime}}|^{2})\\ &+P_{1}~{}{\rm Re}\{F_{\pi\sigma^{\prime}}^{*}F_{\sigma\sigma^{\prime}}+F_{\pi\pi^{\prime}}^{*}F_{\sigma\pi^{\prime}}\}\\ &+P_{2}~{}{\rm Im}\{F_{\pi\sigma^{\prime}}^{*}F_{\sigma\sigma^{\prime}}+F_{\pi\pi^{\prime}}^{*}F_{\sigma\pi^{\prime}}\}\\ &+\frac{1}{2}P_{3}(|F_{\sigma\sigma^{\prime}}|^{2}+|F_{\sigma\pi^{\prime}}|^{2}-|F_{\pi\sigma^{\prime}}|^{2}-|F_{\pi\pi^{\prime}}|^{2})\\ &=C_{0}+C_{1}P_{1}+C_{2}P_{2}+C_{3}P_{3},\end{split} (4)

where P1P_{1} = 0 in the present geometry of the phase retarder. By using the scattering amplitude matrix F^\hat{F} = i(𝜺×𝜺)𝒁dip(1)i({\bf\it\varepsilon}^{\prime}\times{\bf\it\varepsilon})\cdot{\bf\it Z}_{\rm dip}^{(1)}, (C2C_{2}/C0C_{0}, C3C_{3}/C0C_{0}) is calcurated to be (-0.574, -0.562) and (0.574, -0.562) for (-1, 0, 17 + qq) and (-1, 0, 20 - qq), respectively, where φ=π/2\varphi=\pi/2. The solid lines in Fig. 9 are the calculated curves using the above-mentioned magnetic structure model and agree with the experimental data.

Refer to caption
Figure 11: (Color Online) Magnetic field dependence of the (-1, 0, LL) peak profile at the lowest temperature 3 K.
Refer to caption
Figure 12: (Color Online) ΔθPR\Delta\theta_{\rm PR} scan at (-1, 0., 18.5), where 𝒒={\bf\it q}~{}= (0, 0, 1.5), in the β\beta phase.

The AFM-helical order soon changes to an AFM order by applying a magnetic field in the cc plane, where the AFM moments are aligned perpendicular to the field. Figure 11 shows the magnetic field dependence of the (-1, 0, LL) peak profile at the lowest temperature of 3 K. With increasing magnetic field, the AFM chiral helix peak soon moves to the commensurate position at 𝒒={\bf\it q}~{}= (0, 0, 1.5) in a weak magnetic field of \sim 3 kOe above which the β\beta phase is realized. The propagation vector 𝒒={\bf\it q}~{}= (0, 0, 1.5) corresponds to the antiferromagnetic alignment of adjacent Gd honeycomb layers. The qq-value sharply changes from 1.485 to 1.5 with increasing field, and it is hard to observe the soliton lattice Zheludev et al. (1997). When a magnetic field is further increased, the transition from the β\beta to γ\gamma phases occurs at 45 kOe. Although the magnetic Bragg peak was searched along the (-1, 0, LL) line and several other conditions in the γ\gamma phase, the peak could not be found. This result suggests that the qq-vector may have changed from (0, 0, 1.5) to (0, 0, 0). The details of the γ\gamma phase are still unknown. From the results of the magnetization measurements, however, the M(H)M(H) curve suggests an AFM component remains. ErNi3Ga9 and DyNi3Ga9 having the same crystal structure as this material also exhibit AFM order Tsukagoshi et al. (2022); Ninomiya et al. (2018).

Figure 12 shows the ΔθPR\Delta\theta_{\rm PR} scan at (-1, 0, 18.5), where 𝒒={\bf\it q}~{}= (0, 0, 1.5), in the β\beta phase. The shape is different from the one in the α\alpha phase with qq = 1.485, and symmetrical around ΔθPR\Delta\theta_{\rm PR} = 0, because of the loss of the chirality. The solid line is a calculated one as in Fig. 9 for a magnetic structure as shown in Fig. 10(b), where the antiferromagnetically-aligned magnetic moments of Gd orient perpendicular to the magnetic field direction HaH\parallel a in the Gd honeycomb plane. The experimental data are in good agreement with the calculation.

IV Conclusion

We have performed the magnetization and specific heat measurements, and resonant x-ray diffraction experiments to investigate the helical magnetic structure of GdNi3Ga9 with the Sohncke space group R32R32, where a chiral magnetic structure is allowed because of lacks both of inversion and mirror symmetries. This material exhibits an AFM helical order at 19.5K, where a cusp in the magnetic susceptibility, which is a character of a chiral helix, is observed. In this state, the magnetic moments of Gd twist 178.2 per a Gd honeycomb layer along the helical cc axis with 𝒒={\bf\it q}~{}= (0, 0, 1.485), where the period of the helix is the length of 66.7 unit cells, and antiferromagnetically align in the cc plane. We confirmed that the magnetic Bragg peaks with 𝒒={\bf\it q}~{}= (0, 0, 1.485) exhibit opposite response to the RCP and LCP incident x-rays, indicating that the AFM helix consists of a single sense of rotation. With increasing magnetic field, the AFM helical state changes to an antiferromagnetic phase with 𝒒={\bf\it q}~{}= (0, 0, 1.5) in HaH\parallel a of \sim 3 kOe. This 𝒒{\bf\it q} corresponds to a magnetic structure where the magnetic moments between the adjacent Gd honeycomb layers are antiferromagnetically aligned perpendicular to the cc axis.

GdNi3Ga9 and YbNi3Al9 have the same crystal structure and antiferromagnetic and ferromagnetic spiral magnetic structures, respectively. In the future, research that compares these materials is desired.

Acknowledgments

This study was supported by the JSPS KAKENHI Grant Numbers JP18K03539, JP20H01854, JP21K03467, JP21K13879, JP23H04870. The synchrotron radiation experiments at KEK were performed with the approval of the Photon Factory Program Advisory Committee (Proposal No. 2022G520).

References

  • Mühlbauer et al. (2009) S. Mühlbauer, B. Binz, F. Jonietz, C. Pfleiderer, A. Rosch, A. Neubauer, R. Georgii, and P. Böni, Science 323, 915 (2009).
  • Togawa et al. (2012) Y. Togawa, T. Koyama, K. Takayanagi, S. Mori, Y. Kousaka, J. Akimitsu, S. Nishihara, K. Inoue, A. Ovchinnikov, and J.-i. Kishine, Phys. Rev. Lett. 108, 107202 (2012).
  • Togawa et al. (2016) Y. Togawa, Y. Kousaka, K. Inoue, and J.-i. Kishine, J. Phys. Soc. Jpn. 85, 112001 (2016).
  • Momma and Izumi (2011) K. Momma and F. Izumi, Journal of applied crystallography 44, 1272 (2011).
  • Ohara et al. (2011) S. Ohara, T. Yamashita, Y. Mori, and I. Sakamoto, J. Phys.: Conf. Ser. 273, 012048 (2011).
  • Matsumura et al. (2017) T. Matsumura, Y. Kita, K. Kubo, Y. Yoshikawa, S. Michimura, T. Inami, Y. Kousaka, K. Inoue, and S. Ohara, J. Phys. Soc. Jpn. 86, 124702 (2017).
  • Ninomiya et al. (2017) H. Ninomiya, Y. Matsumoto, S. Nakamura, Y. Kono, S. Kittaka, T. Sakakibara, K. Inoue, and S. Ohara, J. Phys. Soc. Jpn. 86, 124704 (2017).
  • Ishii et al. (2018) I. Ishii, K. Takezawa, T. Mizuno, S. Kamikawa, H. Ninomiya, Y. Matsumoto, S. Ohara, K. Mitsumoto, and T. Suzuki, J. Phys. Soc. Jpn. 87, 013602 (2018).
  • Ishii et al. (2019) I. Ishii, K. Takezawa, T. Mizuno, S. Kumano, T. Suzuki, H. Ninomiya, K. Mitsumoto, K. Umeo, S. Nakamura, and S. Ohara, Phys. Rev. B 99, 075156 (2019).
  • Tsukagoshi et al. (2022) M. Tsukagoshi, T. Matsumura, S. Michimura, T. Inami, and S. Ohara, Phys. Rev. B 105, 014428 (2022).
  • Sato et al. (2022) Y. J. Sato, H. Manako, Y. Homma, D. Li, R. Okazaki, and D. Aoki, Physical Review Materials 6, 104412 (2022).
  • Kakihana et al. (2018) M. Kakihana, D. Aoki, A. Nakamura, F. Honda, M. Nakashima, Y. Amako, S. Nakamura, T. Sakakibara, M. Hedo, T. Nakama, et al., J. Phys. Soc. Jpn. 87, 023701 (2018).
  • Kakihana et al. (2019) M. Kakihana, D. Aoki, A. Nakamura, F. Honda, M. Nakashima, Y. Amako, T. Takeuchi, H. Harima, M. Hedo, T. Nakama, et al., J. Phys. Soc. Jpn. 88, 094705 (2019).
  • Tabata et al. (2019) C. Tabata, T. Matsumura, H. Nakao, S. Michimura, M. Kakihana, T. Inami, K. Kaneko, M. Hedo, T. Nakama, and Y. Ōnuki, J. Phys. Soc. Jpn. 88, 093704 (2019).
  • Sakakibara et al. (2019) T. Sakakibara, S. Nakamura, S. Kittaka, M. Kakihana, M. Hedo, T. Nakama, and Y. Ōnuki, J. Phys. Soc. Jpn. 88, 093701 (2019).
  • Topertser et al. (2019) V. Topertser, R.-I. Martyniak, N. Muts, Y. Tokaychuk, and R. Gladyshevskii, Chem. Met. Alloys 12, 21 (2019).
  • Nakamura et al. (2020) S. Nakamura, J. Inukai, T. Asaka, J.-i. Yamaura, and S. Ohara, J. Phys. Soc. Jpn. 89, 104005 (2020).
  • Ishida et al. (1985) M. Ishida, Y. Endoh, S. Mitsuda, Y. Ishikawa, and M. Tanaka, J. Phys. Soc. Jpn. 54, 2975 (1985).
  • Yamasaki et al. (2007) Y. Yamasaki, H. Sagayama, T. Goto, M. Matsuura, K. Hirota, T. Arima, and Y. Tokura, Phys. Rev. Lett. 98, 147204 (2007).
  • Sutter et al. (1997) C. Sutter, G. Grübel, C. Vettier, F. De Bergevin, A. Stunault, D. Gibbs, and C. Giles, Phys. Rev. B 55, 954 (1997).
  • Fabrizi et al. (2009) F. Fabrizi, H. C. Walker, L. Paolasini, F. de Bergevin, A. T. Boothroyd, D. Prabhakaran, and D. F. McMorrow, Phys. Rev. Lett. 102, 237205 (2009).
  • Sagayama et al. (2010) H. Sagayama, N. Abe, K. Taniguchi, T.-h. Arima, Y. Yamasaki, D. Okuyama, Y. Tokura, S. Sakai, T. Morita, T. Komesu, et al., J. Phys. Soc. Jpn. 79, 043711 (2010).
  • Kishine et al. (2005) J.-i. Kishine, K. Inoue, and Y. Yoshida, Progress of Theoretical Physics Supplement 159, 82 (2005).
  • Kousaka et al. (2007) Y. Kousaka, S.-i. Yano, J.-i. Kishine, Y. Yoshida, K. Inoue, K. Kikuchi, and J. Akimitsu, J. Phys. Soc. Jpn. 76, 123709 (2007).
  • Hidaka et al. (2020) H. Hidaka, K. Mizuuchi, E. Hayasaka, T. Yanagisawa, J. Ohara, and H. Amitsuka, Phys. Rev. B 102, 174408 (2020).
  • Bouvier et al. (1991) M. Bouvier, P. Lethuillier, and D. Schmitt, Phys. Rev. B 43, 13137 (1991).
  • Blanco et al. (1991) J. A. Blanco, D. Gignoux, and D. Schmitt, Phys. Rev. B 43, 13145 (1991).
  • Hannon et al. (1988) J. P. Hannon, G. T. Trammell, M. Blume, and D. Gibbs, Phys. Rev. Lett. 61, 1245 (1988).
  • Lovesey et al. (2005) S. W. Lovesey, E. Balcar, K. Knight, and J. F. Rodríguez, Physics Reports 411, 233 (2005).
  • (30) S. W. Lovesey and S. P. Collins, X-ray Scattering and Absorption by Magnetic Materials (Oxford University Press, New York, 1996).
  • Zheludev et al. (1997) A. Zheludev, S. Maslov, G. Shirane, Y. Sasago, N. Koide, and K. Uchinokura, Physical review letters 78, 4857 (1997).
  • Ninomiya et al. (2018) H. Ninomiya, T. Sato, Y. Matsumoto, T. Moyoshi, A. Nakao, K. Ohishi, Y. Kousaka, J. Akimitsu, K. Inoue, and S. Ohara, Physica B: Condensed Matter 536, 392 (2018).