betaFeedback

WorksPhysics and Astronomy

Comparative 181^{181}Ta-NQR Study of Weyl Monopnictides TaAs and TaP: Relevance of Weyl Fermion Excitations

Abstract

Based on our first detailed 181^{181}Ta nuclear quadrupole resonance (NQR) studies from 2017 on the Weyl semimetal TaP, we now extended our NQR studies to another Ta-based monopnictide TaAs. In the present work, we have determined the temperature-dependent 181^{181}Ta-NQR spectra, the spin-lattice relaxation time T1T_{1}, and the spin-spin relaxation time T2T_{2}. We found the following characteristic features that showed great contrast to what was found in TaP: (1) The quadrupole coupling constant and asymmetry parameter of EFG, extracted from three NQR frequencies, have a strong temperature dependence above ∼\sim80 K that cannot be explained by the density functional theory calculation incorporating the thermal expansion of the lattice. (2) The temperature dependence of the spin-lattice relaxation rate, 1/T1T1/T_{1} T, shows a T4T^{4} power law behavior above ∼\sim30 K. This is a great contrast with the 1/T1T∝T21/T_{1} T \propto T^{2} behavior found in TaP, which was ascribed to the magnetic excitations at the Weyl nodes with a temperature-dependent orbital hyperfine coupling. (3) Regarding the nuclear spin-spin interaction, we found the spin-echo signal decays with the pulse separation simply by a Lorentzian function in TaAs, but we have observed spin-echo modulations in TaP that is most likely due to the indirect nuclear spin-spin coupling via virtually excited Weyl fermions. From our experimental findings, we conclude that the present NQR results do not show dominant contributions from Weyl fermion excitations in TaAs.

Introduction

It is well known by now that the topological properties in materials open a new world in condensed matter physics. Particularly, topological semimetals, such as Dirac, Weyl, or line-node semimetals, are gapless states of matter characterized by their nodal band structures and surface states [1]. Weyl semimetals are realized in systems without spatial-inversion or time-reversal symmetry even more with the strong spin-orbit coupling. Quite interesting new phenomena such as ultra-high mobility [2], surface Fermi arcs [3], and chiral magnetic effect [4] are expected to emerge in these materials. Furthermore, in topological semimetals, we found new types of quasiparticles, Dirac- and Weyl-fermions, and those excitations have exhibited fascinating properties that have been the subject of many theoretical and experimental investigations. The first target materials in this field were taken among the monopnictides TMPnTMPn (TM=NbTM = \mathrm{Nb}, Ta; Pn=PPn = \mathrm{P}, As). Visualization of the nodal structure in topological semimetals has been realized by ARPES (angle-resolved photoemission spectroscopy) in the topologically protected surface states [5]. Indirectly, the large negative magnetoresistance [6], optical and resis-

tivity measurements [7], and chiral anomalies [8] are believed to be associated with those quasiparticles. In addition to the surface-sensitive probes like ARPES and electron spin resonance (ESR), the microscopic measurements which enable us to study the static and dynamical properties of quasiparticles as a bulk are highly expected. Along this line, we have succeeded for the first time to explore the Weyl fermion excitations in TaP through the temperature dependence of nuclear quadrupole resonance (NQR) relaxation rate, 1/T1T1/T_1T [9]. There, we have demonstrated that in addition to the T4T^4 power law of 1/T1T1/T_1T associated with the linear dispersion of Weyl nodes near the Fermi level, we have pointed out the importance of fluctuations in Dirac/Weyl-type orbital currents to the relaxation channel through the characteristic temperature dependence of the orbital hyperfine interaction [10]. This scenario is supported by the theory explicitly and the overall temperature dependence of 1/T1T∝T21/T_1T \propto T^2 has been interpreted properly [11].

In this paper we present an extended study of the sister compound, TaAs, using the same 181Ta{}^{181}\mathrm{Ta}-NQR technique. TaAs has been claimed to be a typical example of the Weyl nodal semimetal from band structure calculations. Regarding the nodal structure, both compounds have Weyl points near the Fermi level, EFE_F, 14 meV14\,\mathrm{meV} below EFE_F for the W2 Weyl points in TaAs [12], while 13 meV13\,\mathrm{meV} above EFE_F in TaP [13].

In the following, we will first briefly describe the experimental technique, then discuss the temperature dependence of NQR parameters, νQ\nu_Q and η\eta. This will be followed by the temperature dependence of nuclear magnetic relaxation time, T1T_1, and a comparison of spin-echo decay curves of TaAs and TaP, along with their interpretations.

Experimental

Basically, we followed the experimental procedure of previous 181Ta{}^{181}\mathrm{Ta}-NQR experiments and analysis [9]. Here, we briefly describe the essence of them.

Samples used in the present NQR experiments were prepared by the chemical transport reaction (CTR) method. Starting from microcrystalline powder synthesized by reacting 3-nine Tantalum and 6-nine Arsenic, single crystals of TaAs were grown in a temperature gradient from 900 °C (source) to 1000 °C (sink), and a transport agent concentration of 13 mg/cm3^{3} iodine. The crystals obtained by the CTR were characterized by electron-probe-microanalysis and powder X-ray diffraction (XRD) to ensure the single phase, tetragonal I41mdI4_{1}md (#109) structure.

Temperature-dependent powder XRD was performed at the beamline ID22 at the European Synchrotron Research Facility (ESRF) in Grenoble in a temperature range between 80 and 300 K with a wavelength λ=0.39997\lambda= 0.39997 Å.

The NQR experiments were mostly carried out with high-quality polycrystals prepared by powdering several single crystals. The NQR spectra and the nuclear magnetic relaxation times were measured using a standard pulsed (spin-echo) NMR apparatus (Apollo, TecMag). The 181Ta{}^{181}\mathrm{Ta}-NQR spectra were taken using the frequency sweep method under zero applied magnetic field. In order to avoid any artificial broadening, fast Fourier-transformed (FFT) spin-echo signals were summed across the spectrum (FFT-summation), or the real part of spin-echoes was integrated after appropriate phase adjustments.

The quadrupole Hamiltonian can be written, using a set of principal axes [14], as

HQ=e2qQ4I(2I−1)[3IZ2−I(I+1)+η2(I+2−I−2)](1)\mathcal{H}_{\mathrm{Q}}=\frac{e^{2}qQ}{4I(2I-1)}\left[3I_{Z}^{2}-I(I+1)+\frac{\eta}{2}\left(I_{+}^{2}-I_{-}^{2}\right)\right] \tag*{(1)}

where eqeq is the largest component of the electric field gradient (EFG) tensor, VZZV_{ZZ}, and eQeQ the nuclear quadrupole moment. The EFG tensor is generally defined as ∣VXX∣≤∣VYY∣≤∣VZZ∣\lvert V_{XX}\rvert\leq\lvert V_{YY}\rvert\leq\lvert V_{ZZ}\rvert with the asymmetry parameter, η≡(VXX−VYY)/VZZ\eta\equiv(V_{XX}-V_{YY})/V_{ZZ}. The quadrupole-split nuclear energy levels, EmE_m, and the resultant transition frequencies can be readily calculated numerically by diagonalizing Eq. (1). For η=0\eta= 0, the energy levels can simply be expressed as,

Em=16hνQ[3m2−I(I+1)],νQ=3e2qQ2I(2I−1)h(2)E_m=\frac{1}{6}h\nu_{\mathrm{Q}}\left[3m^{2}-I(I+1)\right],\qquad\nu_{\mathrm{Q}}=\frac{3e^{2}qQ}{2I(2I-1)h} \tag*{(2)}

where νQ\nu_{\mathrm{Q}} is the quadrupole coupling constant. The NQR occurs for the transition between two levels mm and m+1m+1, and the resonance condition can be written as fQ=νQ(2∣m∣+1)/2f_{\mathrm{Q}}=\nu_{\mathrm{Q}}(2\lvert m\rvert+1)/2. Hence, three NQR lines for 181Ta{}^{181}\mathrm{Ta} with I=7/2I=7/2 are expected at 1νQ1\nu_{\mathrm{Q}}, 2νQ2\nu_{\mathrm{Q}}, and 3νQ3\nu_{\mathrm{Q}} with equal spacing. For the finite η\eta values, the calculated NQR frequencies with νQ=1\nu_{\mathrm{Q}}=1 MHz for respective transitions are shown in Figure 1.

Image not extracted

Figure 1. (Color online) EFG asymmetry parameter, η\eta, dependence of NQR frequencies for I=7/2I=7/2. Here, the quadrupole coupling parameter, νQ\nu_{\mathrm{Q}}, is set as 1 MHz. The cross point between 1νQ1\nu_{\mathrm{Q}} and 2νQ2\nu_{\mathrm{Q}} lines at η=0.5855\eta=0.5855 is called as the “magic eta”. η\eta values determined from the experimental NQR frequencies for TaAs (η=0.558\eta=0.558) and TaP (η=0.423\eta=0.423) are shown by dashed arrows.

Since the nuclear spin-lattice (longitudinal) relaxation time, T1T_{1}, is extremely long in TaAs at low temperatures (typically several hundred seconds), we mostly employed the progressive saturation method [15] to measure the recovery of nuclear magnetization below 130 K, as was in the case of TaP [9]. Above 130 K, a conventional inversion recovery method is employed. At 130 K, both methods yield essentially the same T1T_{1} value.

The recovery of nuclear magnetization was fitted to the theoretical function for the magnetic relaxation in NQR lines for 181Ta{}^{181}\mathrm{Ta} (I=7/2I=7/2) nucleus with finite η=0.558\eta=0.558 [16],

Mn(t)=M0[1−{Q1exp⁡(−3.03t/T1)+Q2exp⁡(−8.260t/T1)+Q3exp⁡(−17.074t/T1)}](3)M_{n}(t)=M_{0}\left[1-\left\{Q_{1}\exp(-3.03t/T_{1})+Q_{2}\exp(-8.260t/T_{1})+Q_{3}\exp(-17.074t/T_{1})\right\}\right] \tag*{(3)}

where QnQ_{n} are constants depending on which NQR transition is excited. For the preset T1T_{1} measurements, we typically used 2νQ2\nu_{\mathrm{Q}}-line corresponding to the ±3/2↔±5/2\pm3/2 \leftrightarrow\pm5/2 nuclear quadrupole transition. In this case, Q1Q_{1}, Q2Q_{2} and Q3Q_{3} are 0.076, 0.021, and 0.903, respectively.

Figure 2 shows the recovery of nuclear magnetization measured by the progressive saturation method at (a) 4.2 K and (b) 100 K for the 181Ta{}^{181}\mathrm{Ta}-NQR 2νQ2\nu_{\mathrm{Q}}-line in TaAs. Solid lines are the least-squares fitting to Eq. (3). For both temperatures, experimental data are perfectly fitted by the theoretical curve, verifying that the nuclear relaxation is governed by magnetic fluctuations as in the case of TaP [9].

Image not extracted

Figure 2. (Color online) The recovery curves of nuclear magnetization measured by the spin-echo amplitude for the 2νQ2\nu_Q line at (a) 4.2 and (b) 100 K. Solid lines are the least-squares fit to Eq. (3) for magnetic fluctuations. For details, see text.

For the measurements of the nuclear spin-spin (transverse) relaxation time, T2T_2, we simply measure spin-echo amplitude, EE, as a function of the time between the first exciting and the second refocusing pulses. The repetition time between spin-echo sequences was taken to be sufficiently longer than T1T_1 (typically 8–10 times longer than T1T_1 value) to avoid a saturation effect.

In order to extract the EFG theoretically, we performed band structure calculations using the density functional theory (DFT) solid-state code FPLO [17]. We used the Perdew-Wang parametrization of the local density approximation (LDA) for the exchange-correlation functional [18, 19]. The strong spin-orbit coupling in TaAs is taken into account by performing full-relativistic calculations, wherein the Dirac Hamiltonian with a general potential is solved. The quadrupole coupling νQ\nu_Q can be obtained by the calculated EFG at the Ta nuclear site which is defined as the second partial derivative of the electrostatic potential V(r⃗)V(\vec{r}) at the position of the nucleus

Vij=(∂i∂jV(0)−δijΔV(0)/3).V_{ij} = \left(\partial_i\partial_j V(0) - \delta_{ij}\Delta V(0)/3\right).

Experimental Results and Discussion

In this section, we present the static and dynamic properties revealed by the temperature dependence of the EFG parameters and nuclear magnetic relaxation times with phenomenological discussions.

NQR spectra and their temperature dependence

A typical example of 181Ta{}^{181}\mathrm{Ta}-NQR spectra in TaAs measured at 4.2 K is shown in Figure 3(a) for three NQR transitions. From the lowest frequency we define the lines as 1νQ1\nu_Q (±1/2↔±3/2\pm1/2 \leftrightarrow\pm3/2), 2νQ2\nu_Q (±3/2↔±5/2\pm3/2 \leftrightarrow\pm5/2) and 3νQ3\nu_Q (±5/2↔±7/2\pm5/2 \leftrightarrow\pm7/2) lines. For comparison, we also depict a similar spectrum of TaP in Figure 3(b). It is immediately seen that the frequency difference between 1νQ1\nu_Q and 2νQ2\nu_Q lines is smaller for TaAs than TaP. It means that the η\eta value is larger for TaAs than TaP as indicated in Figure 1. The EFG parameters, νQ\nu_Q and η\eta, were calculated using the same manner described in Ref. [9] and shown in Table I, together with values for TaP and NbP. For NbP, νQ\nu_Q and η\eta are extracted from the single-crystal 93Nb{}^{93}\mathrm{Nb}-NMR spectrum at ∼6.3\sim6.3 T, 4.2 K. The calculated values of νQ\nu_Q agrees with the experimental values of all compound within 4%.

Image not extracted

Figure 3. (Color online) A 181{}^{181}Ta-NQR spectrum in (a) TaAs in comparison with that in (b) TaP. Both spectra were taken at 4.2 K and curves are Lorentzian fit of the line profiles.

The temperature dependence of NQR frequencies is plotted in Figure 4. From the observed NQR frequencies we can extract νQ\nu_Q and η\eta by diagonalizing Eq. (1) and those temperature dependences, νQ(T)\nu_Q(T) and η(T)\eta(T), are shown in Figure 5(a) and (b) for TaAs (open triangles) and TaP (open circles), respectively. In general, νQ\nu_Q is expected to decrease with increasing temperature due to the thermal expansion of the lattice, and is often discussed by an empirical formula, νQ(T)=νQ(0)(1−αT3/2)\nu_Q(T)=\nu_Q(0)(1-\alpha T^{3/2}). Actually, νQ(T)\nu_Q(T) is well fitted to this empirical formula below ∼100\sim100 K for both TaAs and TaP, but the experimental data fall more rapidly above ∼100\sim100 K.

Image not extracted

Figure 4. (Color online) Temperature dependence of the NQR peak frequencies in (a) TaAs in comparison with those in (b) TaP. Here 1νQ1\nu_Q, 2νQ2\nu_Q, and 3νQ3\nu_Q line correspond to ±1/2↔±3/2\pm1/2 \leftrightarrow\pm3/2, ±3/2↔±5/2\pm3/2 \leftrightarrow\pm5/2, and ±5/2↔±7/2\pm5/2 \leftrightarrow\pm7/2 quadrupole transitions, respectively.

Image not extracted

Figure 5. (Color online) Temperature dependences of quadrupole coupling parameter, νQ\nu_Q, and asymmetry parameter of the EFG, η\eta, for TaAs and TaP are shown in panels (a) and (b), respectively. Both values were extracted from experimental data shown in Figure 4 by diagonalizing Eq. (1). The dashed curves with open stars are DFT calculated values of νQ\nu_Q and η\eta using the lattice parameters of respective temperatures shown in Table II. The calculated νQ(T)\nu_Q(T) follows the empirical form with νQ(T)=νQ0(1−αT3/2)\nu_Q(T) = \nu_{Q0}(1-\alpha T^{3/2}), νQ0=20.24\nu_{Q0} = 20.24 MHz, α=3.92×10−7\alpha= 3.92 \times10^{-7} K−3/2^{-3/2}.

In particular, for TaAs, we have made the DFT calculation for EFG parameters by putting the measured lattice parameters measured by synchrotron XRD at selected temperatures above 80 K shown in Table II. As the temperature decreases, the lattice parameters exhibit a slight, monotonous reduction while maintaining a constant c/ac/a ratio, indicative of isotropic behavior. The calcalculated νQ(T)\nu_Q(T) and η\eta are shown by stars in Figure 5(a). As can be seen from the figure, the thermal expansion of the lattice cannot account for the temperature dependence of the observed NQR parameters.

DFT calculationExperimental
νQ\nu_Q (MHz)η\etaνQ\nu_Q (MHz)η\eta
TaAs20.8300.62520.2490.558
TaP20.0570.33019.2500.423
NbP0.6670.19120.650.15

Table 1. DFT calculated EFG parameters νQ\nu_Q and η\eta. The values of νQ\nu_Q and η\eta are in good agreement with those extracted from the NQR frequencies in TaAs and TaP [9]. Also, those obtained by the quadrupole split 93Nb{}^{93}\mathrm{Nb}-NMR spectrum in a single-crystalline NbP are in good agreement with the DFT calculation.

TT (K)aa (Å)cc (Å)c/ac/aVV (Å3^3)
3003.4375211.647623.38838137.63
2603.4364211.643743.38833137.50
2203.4356311.640893.38828137.40
1803.4349211.638263.38822137.32
1403.4343711.636273.38818137.25
1003.4337511.633973.38812137.17
803.4334911.632993.38809137.14

Table 2. Lattice parameters measured by synchrotron XRD for TaAs at selected temperatures above 80 K.

It should be noted here that the fractional decrease of νQ(T)\nu_Q(T) and η(T)\eta(T) is very similar as shown by scaling TaAs results (shown by open open diamonds) to TaP by factors of 0.95 and 0.76 for νQ(T)\nu_Q(T) and η(T)\eta(T), respectively.

EFG generally has contributions from lattice symmetry and asymmetrical charge distribution around the nucleus in concern,

eq(T)=eqlattice(T)+eqel(T).(4)e_q(T) = e_{q\mathrm{lattice}}(T) + e_{q\mathrm{el}}(T). \tag*{(4)}

The first term was literally calculated by measured thermal lattice expansion and agrees well with the experimental values below 80 K. On the other hand, experimental data deviate from the first term, showing an almost linear dependence on temperature above 80 K. This fact suggests that the second term becomes dominant above 80 K, indicating an unusual electronic contribution to the EFG is induced by a reason that is yet to be identified.

Nuclear spin-lattice relaxation

The temperature dependence of the 181Ta{}^{181}\mathrm{Ta} nuclear spin-lattice relaxation rate divided by TT, 1/T1T1/T_1T, in TaAs is shown in Figure 6 together with previous data of TaP [9]. Also, we reproduced data of 75As{}^{75}\mathrm{As}-NQR taken by Wang et al. [20]. There is a general tendency for the 1/T1T1/T_1T exhibiting that from high-temperature power law type relaxation process crosses over to temperature-dependent Korringa type around T∗∼20T^* \sim20–40 K.

Image not extracted

Figure 6. (Color online) Temperature dependence of 1/T1T1/T_1T measured for the 2νQ2\nu_Q line in TaAs is shown by filled circles. For comparison, the similar data taken by 75^{75}As-NQR are shown in cross squares [20]. Also, data in TaP are shown in open circles where the T2T^2 temperature dependence of 1/T1T1/T_1T which is characteristic to the Weyl fermion excitations is seen above ∼30\sim30 K [9].

Quite generally, 1/T1T1/T_1T can be expressed by using the wave vector (qq) and frequency (ω\omega) dependent magnetic susceptibility, χ(q,ω)\chi(q,\omega), characterizing the magnetic excitations in a system as,

1T1T=2γN2kBg2μB2∑qAq2χ⊥′′(q,ωN)ωN,(5)\frac{1}{T_1T} = \frac{2\gamma_N^2 k_B}{g^2\mu_B^2}\sum_q A_q^2 \frac{\chi''_\perp(q,\omega_N)}{\omega_N}, \tag*{(5)}

where γN\gamma_N is the nuclear gyromagnetic ratio, kBk_B the Boltzmann constant, gg the electron gg-factor, μB\mu_B the Bohr magneton, AqA_q the qq-dependent hyperfine coupling constant, χ⊥′′(q,ω)\chi''_\perp(q,\omega) the transverse component of imaginary part of χ(q,ω)\chi(q,\omega), and ωN\omega_N the NQR frequency. Since at present we do not have any plausible microscopic theory to calculate χ(q,ω)\chi(q,\omega) in multiband systems like TaAs, we have adopted the theoretical 1/T1T1/T_1T for non-interacting itinerant electrons based on the band structure calculation with random phase approximation (RPA). Also, since we do not have the information about AqA_q, we cannot perform a quantitative analysis of 1/T1T1/T_1T. So, we try to interpret the data qualitatively only using the shape of temperature dependence. In what follows we will discuss it by setting three cases: [Case-1] simple calculation from the density of states (DOS), [Case-2] in-gap states near the Fermi level, and [Case-3] Weyl fermion excitations.

Case-1: Simple 1/T1T1/T_1T (TT) from DOS

Here, we have simply adopted the theoretical 1/T1T1/T_1T for non-interacting itinerant electrons based on the band structure calculation with RPA. For such a system 1/T1T1/T_1T may be expressed using density of state near the Fermi level as [14] [21],

1T1T∝Ahf2T∫f(E−μc)[1−f(E−μc)]D(E)2 dE(6)\frac{1}{T_1T} \propto\frac{A_{\mathrm{hf}}^2}{T}\int f(E-\mu_c)[1-f(E-\mu_c)]D(E)^2\,\mathrm{d}E \tag*{(6)}

where f(E)f(E) is a fermi distribution function, D(E)D(E) the energy dependent DOS, and μc\mu_c the temperature-dependent chemical potential. If AhfA_{\mathrm{hf}} does not change with temperature and is set to one, the calculation of 1/T1T1/T_1T is straightforward from calculated D(E)D(E) based on the band structure shown as Figure 7(a) for TaAs and TaP. The calculated 1/T1T1/T_1T is shown by curved lines in Figure 7(b). Also, we show the experimental 1/T1T1/T_1T data of TaAs and TaP. We can clearly observe that, aside from the absolute value of 1/T1T1/T_1T, the temperature dependencies for experimental and calculated 1/T1T1/T_1T do not match each other. This shows that the simple Korringa-type relaxation process cannot account for the experimental results observed.

Image not extracted

Figure 7. (Color online) (a) The calculated D(E)D(E) curves within ΔE∼±1000\Delta E \sim\pm1000 K based on the band structure are shown for TaAs and TaP. (b) Temperature dependence of 1/T1T1/T_1T calculated from DOS (solid lines) are compared with the experimental data for TaAs (filled circles) and TaP (open circles). Here, the hyperfine coupling constant AhfA_{\mathrm{hf}} is set to one and is assumed to be temperature-independent.

Case-2: 1/T1T1/T_1T from in-gap states

Simple D(E)D(E) calculations predict a fairly high energy scale for the excitation of the valence band. To reconcile this with the experimental data, here we assume the existence of rather narrow bands crossing the Fermi energy shown inset of Figure 8. Following the common phenomenological treatment of 1/T1T1/T_1T at the high-temperature region, an activated-type temperature dependence of 1/T1T1/T_1T has been assumed. Including the low temperature upturn the data have been fitted to the following empirical form,

Image not extracted

Figure 8. (Color online) Fitting for 1/T1T1/T_1T assuming a rectangular in-gap state shown as inset a least-sure fit of the data has adapted to Eq. (7). The result is shown by a solid curve with low-temperature power law exponent β=0.55\beta= 0.55 and high-temperature activation energy of Δ=283\Delta= 283 K (24.4 meV).

1T1T=αT−β+(1T1T)0exp⁡(−ΔkBT)(7)\frac{1}{T_1T}=\alpha T^{-\beta}+\left(\frac{1}{T_1T}\right)_0\exp\left(-\frac{\Delta}{k_{\mathrm{B}}T}\right) \tag*{(7)}

where the first term is associated with the local moment type fluctuations of the in-gap state, and the second term is due to an activation process in high temperatures with the energy of Δ\Delta. The solid line is a least-squares fit of the data to Eq. (7). We found α=8×10−4\alpha= 8 \times10^{-4} sec. K and β=0.55\beta= 0.55 for the first term, and (1/T1T)0=0.18(1/T_1T)_0 = 0.18 sec. K and Δ/kB=283\Delta/k_{\mathrm{B}} = 283 K (24.4 meV) for the second term. The energy scale in the activation process found in TaAs is nearly one order of magnitude larger than those observed in similar materials, SmB6_6 (Δ=4.3\Delta= 4.3 meV) [22], FeGa3_3 (Δ=1.1\Delta= 1.1 meV) [23], and PuB4_4 (Δ=1.8\Delta= 1.8 meV) [24]. There also exists a common feature of 1/T1T1/T_1T in low-temperature region, where the exponential decrease of 1/T1T1/T_1T with decreasing temperature crosses over to other excitations which give rise to an upturn of 1/T1T1/T_1T. Within the present model, the low-temperature behavior must be due to the local moment-type fluctuations in the occupied narrow band. If this is the case, 1/T1T1/T_1T should be treated by the exchange narrowed theory and β\beta should be one. The fit value, β=0.55\beta= 0.55, may indicate that the assumed in-gap state is rather spatially extended so that the exchange narrowed theory may not be applicable. The origin of the in-gap state is not clear, but it may be associated with Anderson localization [25] or impurity states.

Case-3: 1/T1 T (T) from Weyl fermion excitations

The first successful observation of the Weyl fermion excitations in topological materials has been achieved in the temperature dependence of 1/T1T1/T_1T in TaP where the T2T^2 power law dependence has observed in high temperatures [9]. This T2T^2 dependence was interpreted as competing relaxation channels between the spin and orbital. The spin channel is due to the Weyl fermion excitations associated with the linear dispersion around the Weyl points and has T4T^4 dependence of 1/T1T1/T_1T. The orbital channel is the relaxation process associated with the fluctuations of the orbital hyperfine field which leads to the T−2T^{-2} dependence [10, 11]. In high temperatures, both contributions are equally acting in TaP, then the T2T^2 dependence has been observed which is associated with the excitation of Weyl points located 13 meV above Fermi level. The same scenario was applied to the T2T^2 dependence observed in 75^{75}As-NQR measurements in TaAs [20]. However, our 181^{181}Ta-NQR measurements in TaAs revealed T4T^4 dependence, meaning the orbital contribution is negligibly small. Following the previous calculation for TaP, 1/T1T1/T_1T from the spin channel for TaAs has been calculated using,

1T1T=α[4μ(T)4+8π2μ(T)2T2+(28π4/15)T4],μ(T)=μ(0)1+c[T/μ(0)]2,(8)\frac{1}{T_1T} = \alpha\left[4\mu(T)^4 + 8\pi^2\mu(T)^2T^2 + \left(28\pi^4/15\right)T^4\right], \qquad \mu(T) = \frac{\mu(0)}{1+c[T/\mu(0)]^2}, \tag*{(8)}

where μ(T)\mu(T) is the temperature-dependent chemical potential [11] in the unit of K. Then we have obtained a reasonably good fit to the data with α=6.14×10−13\alpha= 6.14 \times10^{-13} sec.−1^{-1} K−5^{-5}, μ(0)=120\mu(0) = 120 K, c=35c = 35 as shown by a solid curve in Figure 9. It should be noted that the deviation above ∼100\sim100 K may be due to a cutoff effect of the Weyl fermion excitation toward the Korringa process. We also note that we currently have no explanation as to why the orbital relaxation channel is not visible compared with the case of TaP.

Image not extracted

Figure 9. (Color online) Experimental temperature dependence of 1/T1T1/T_1T are fitted to theoretical Weyl fermion excitations by a solid curve in TaAs. The normalized temperature dependence of chemical potential, μ(T)/μ(0)\mu(T)/\mu(0), is shown in the inset.

Based on the given information, it is difficult to draw a definitive conclusion about the temperature dependence of 1/T1T1/T_1T. While Case 2 appears to be the most plausible scenario, the lack of information regarding the hyperfine coupling constants of both spin and charge relaxation channels prevents us from making a conclusive statement.

Nuclear spin-spin relaxation

The nuclear spin-spin (transverse) relaxation time, T2T_2, was obtained by measuring the spin-echo amplitude, EE, as a function of the time tt between the first exciting pulse and the spin-echo position. The amplitude E(t)E(t) can generally be expressed as,

E(t)=E0e−Δ2t2[c0+c1cos⁡(Jt+ϕ)e−t/τc],(9)E(t) = E_0 e^{-\Delta^2 t^2}\left[c_0+c_1\cos(Jt+\phi)e^{-t/\tau_c}\right], \tag*{(9)}

where Δ\Delta is the second moment of the direct nuclear spin-spin interaction due to the classical dipolar coupling in the first place, c0c_0 and c1c_1 are constants, and the cosine term is the oscillatory term due to the indirect coupling or nuclear quadrupole coupling with their characteristic decay constant, τc\tau_c. The spin-echo modulation is well known for the case that the interaction is given by the formula J(I⃗i⋅I⃗j)=J/2(Ii+Ij−+Ii−Ij+)+JIz2J(\vec{I}_i\cdot\vec{I}_j)=J/2(I_{i+}I_{j-}+I_{i-}I_{j+})+JI_{z}^{2}. Here, the Iz2I_z^2 term is responsible to the oscillation because this term is invariant for the refocusing pulse (π\pi-rotation in the rotating frame) making an oscillatory behavior of the formation of spin-echo as a function of tt. The clear evidences for this oscillated spin-echo decay have been documented for the nuclear quadrupole interaction [26, 27] and the indirect nuclear spin-spin coupling via conduction electrons (the Ruderman-Kittel interaction) [28]. The direct nuclear spin-spin coupling includes the same term, but the coupling constant JJ is so small (the oscillation has a range of several milliseconds at least) that one could not see this effect except for a special case. It should be noted that the direct nuclear spin-spin interaction is easily detuned by inhomogeneous broadening (either external field, sample inhomogeneity, or both) making the decay longer with an exponential function,

E(t)=E0exp⁡(−αt).E(t)=E_0\exp(-\alpha t).

The experimental spin-echo decay curves taken at 2νQ2\nu_Q line of TaAs and TaP are shown in Figure 10(a), where spin-echo decays basically exponential for both compounds, but the oscillation was seen only for TaP. As shown in Figure 10(b), the oscillatory part of the decay in TaP can be fitted to ΔE(t)=c1cos⁡(ωpt+ϕ)exp⁡(−t/τc)\Delta E(t)=c_1\cos(\omega_p t+\phi)\exp(-t/\tau_c) with c1=0.30c_1=0.30, ωp/2π=3.58\omega_p/2\pi=3.58 kHz, and τc=150\tau_c=150 μ\musec. with ωp\omega_p is an energy scale of indirect coupling. If the oscillation is caused by an indirect nuclear spin-spin coupling via virtual excitation of Weyl fermions as illustrated in Figure 10(c), the absence of oscillation in TaAs indicates an absence of Weyl fermion excitations. This may be consistent with the 1/T1T1/T_1T behavior discussed in Case 2.

Image not extracted

Figure 10. (Color online) (a) Typical spin echo decay for 2νQ2\nu_Q lines in TaAs (filled circles) and TaP (open circles) at 4.2 K. A strong spin-echo modulation has been observed in TaP while in TaAs spin-echo decays exponentially without any modulation. The oscillatory part of the spin-echo decay in TaP (open circles) is shown in (b) with the data fit to Eq. (9) (solid curve). A cartoon of an indirect nuclear spin-spin coupling via virtual excitation of Weyl fermions is illustrated in (c).

Concluding Remarks

We presented an extended comparative microscopic study of one of the typical Weyl semimetals, TaAs, beyond previous work on TaP, utilizing the 181{}^{181}Ta-NQR technique. The experimental results are contrasted between the above two monopnictides. The NQR parameters, νQ\nu_Q and η\eta, are in good agreement with the ab initio calculations for both compounds. However, their temperature dependence above approximately 100 K shows

distinct characteristics, in the sense that νQ(T)\nu_Q(T) deviates considerably from calculated values using simple thermal expansion. This discrepancy is likely due to a manifestation of the change of the electronic structure above 100 K.

Likewise, nuclear spin-lattice relaxation rate 1/T1T1/T_1T and nuclear spin-echo decay have great contrast between TaP and TaAs. In TaP, 1/T1T1/T_1T are well documented by the Weyl fermion excitations with a temperature-dependent orbital hyperfine interaction. However, in TaAs, we observed a T4T^4 power law dependence of 1/T1T1/T_1T, which could potentially be associated with the linear dispersion of the Weyl fermions within a certain temperature range. Despite this observation, we were unable to draw a conclusive picture for 1/T1T1/T_1T. The lack of information regarding the hyperfine coupling constants of both spin and charge relaxation channels prevents us from making a conclusive statement.

The work on TaAs shows that there are still many open questions in the field of Weyl semimetals, and this is even more true when trying to understand local measurement methods such as the 181{}^{181}Ta-NQR. There is an urgent need to use more local methods like NQR and NMR but also muon spin spectroscopy (μ\muSR) or ESR to fully understand electronic excitations near the Fermi level in detail.

Acknowledgements

We thank G. Auffermann, U. Burkhardt, and V. Süß for help with the synthesis and characterization of the TaAs crystals. B. D. was supported by the Ministry of Culture and Innovation and the National Research, Development and Innovation Office within the Quantum Information National Laboratory of Hungary (Grant No. 2022-2.1.1-NL-2022-00004) K134437, K142179 and by a grant of the Ministry of Research, Innovation and Digitization, CNCS/CCCDI-UEFISCDI, under projects number PN-III-P4-ID-PCE-2020-0277. We thank U. Nitzsche (IFW Dresden) for technical support. We thank the ESRF (ID22) for providing beamtime.

References

References

  1. [1]M. Franz and L. Molenkamp (Eds.), Topological Insulators, Contemporary Concept of Condensed Materials Science Vol. 6, Elsevier 2013, ISBN 978-0-444-63314-9.
  2. [2]C. Shekhar, A. K. Nayak, Y. Sun, M. Schmidt, M. Nicklas, I. Leermakers, U. Zeitler, Y. Skourski, J. Wosnitza, Z. Liu, Y. Chen, W. Schnelle, H. Borrmann, Y. Grin, C. Felser, and B. Yan, Nat. Phys. 11, 645 (2015).
  3. [3]X. Wan, A. M. Turner, A. Vishwanath, and S. Y. Savrasov, Phys. Rev. B 83, 205101 (2011).
  4. [4]A. A. Zyuzin, A. A. Burkov, Phys. Rev. B 86, 115133 (2012).
  5. [5]S.-M. Huang, S.-Y. Xu, I. Belopolski, C.-C. Lee, G. Chang, B. Wang, N. Alidoust, G. Bian, M. Neupane, C. Zhang, S. Jia, A. Bansil, H. Lin, and M. Zahid Hasan, Nat. Commun. 1, 8373 (2015).
  6. [6]X. Huang, L. Zhao, Y. Long, P. Wang, D. Chen, Z. Yang, H. Liang, M. Xue, H. Weng, Z. Fang, X. Dai, and G. Chen, Phys. Rev. X 5, 031023 (2015).
  7. [7]B. Xu, Y. M. Dai, L. X. Zhao, K. Wang, R. Yang, W. Zhang, J. Y. Liu, H. Xiao, G. F. Chen, A. J. Taylor, D. A. Yarotski, R. P. Prasankumar, and X. G. Qiu, Phys. Rev. B 93, 121110(R) (2016).
  8. [8]C.-L. Zhang, S.-Y. Xu, I. Belopolski, Z. Yuan, Z. Lin, B. Tong, G. Bian, N. Alidoust, C.-C. Lee, S.-M. Huang, T.-R. Chang, G. Chang, C.-H. Hsu, H.-T. Jeng, M. Neupane, D. S. Sanchez, H. Zheng, J. Wang, H. Lin, C. Zhang, H.-Z. Lu, S.-Q. Shen, T. Neupert, M. Zahid Hasan, and S. Jia, Nat. Commun. 7, 10735–1–9 (2015).
  9. [9]H. Yasuoka, T. Kubo, Y. Kishimoto, D. Kasinathan, M. Schmidt, B. Yan, Y. Zhang, H. Tou, C. Felser, A. P. Mackenzie, and M. Baenitz, Phys. Rev. Lett. 118, 236403 (2017).
  10. [10]Z. Okvátovity, F. Simon, and B. Dóra, Phys. Rev. B 94, 245141 (2016).
  11. [11]Z. Okvátovity, H. Yasuoka, M. Baenitz, F. Simon, and B. Dóra, Phys. Rev. B 99, 115107 (2019).
  12. [12]F. Arnold, M. Naumann, S.-C. Wu, Y. Sun, M. Schmidt, H. Borrmann, C. Felser, B. Yan, and E. Hassinger, Phys. Rev. Lett. 117, 146401 (2016).
  13. [13]F. Arnold, C. Shekhar, S.-C. Wu, Y. Sun, R. D. dos Reis, N. Kumar, M. Naumann, M. O. Ajeesh, M. Schmidt, A. G. Grushin, J. H. Bardarson, M. Baenitz, D. Sokolov, H. Borrmann, M. Nicklas, C. Felser, E. Hassinger, and B. Yan, Nat. Commun. 7, 11615 (2016).
  14. [14]C. P. Slichter, Principles of Magnetic Resonance 3rd. Ed., Springer -Verlag 1990, ISBN: 978-3-662-09441-9.
  15. [15]V. F. Mitrović, E. E. Sigmund, and W. P. Halperin, Phys. Rev. B 64, 024520 (2001).
  16. [16]J. Chepin, and J. H. Ross Jr., J. Phys.: Condens. Matter, 3 8103 (1991).
  17. [17]K. Koepernik, and H. Eschrig, Phys. Rev. B 59, 1743 (1999).
  18. [18]J. P. Perdew, and Y. Wang, Phys. Rev. B 45, 13244 (1992).
  19. [19]A. A. Mosto, J. R. Yates, Y. S. Lee, I. Souza, D. Vanderbilt, and N. Marzari, Comput. Phys. Commun. 178, 685 (2008).
  20. [20]C. G. Wang, Y. Honjo, L. X. Zhao, G. F. Chen, K. Matano, R. Zhou, and G.-q. Zheng, Phys. Rev. B 101, 241110(R) (2020).
  21. [22]M. Takigawa, H. Yasuoka, Y. Kitaoka, T. Tanaka, H. Nozaki, and Y. Ishizawa, J. Phys. Soc. Jpn. 50, 2525 (1981).
  22. [23]A. A. Gippius, V. Yu. Verchenko, A. V. Tkachev, N. E. Gervits, C. S. Lue, A. A. Tsirlin, N. Büttgen, W. Krätschmer, M. Baenitz, M. Shatruk, and A. V. Shevelkov, Phys. Rev. B 89, 104426 (2014).
  23. [24]A. P. Dioguardi, H. Yasuoka, S. M. Thomas, H. Sakai, S. K. Cary, S. A. Kozimor, T. E. Albrecht-Schmitt, H. C. Choi, J.-X. Zhu, J. D. Thompson, E. D. Bauer, and F. Ronning, Phys. Rev. B 99, 035104 (2019).
  24. [25]S. Takagi, H. Yasuoka, S. Ogawa, and J. H. Wernick, J. Phys. Soc. Jpn. 50, 2539 (1981).
  25. [26]H. Abe, H. Yasuoka, and A. Hirai, J. Phys. Soc. Jpn. 21, 77 (1966).
  26. [27]H. Yasuoka, H. Nishihara, Y. Nakamura, and J. P. Remeika, Phys. Lett. 37A, 299 (1971).
  27. [28]C. Froidevaux, and M. Weger, Phys. Rev. Lett. 12, 123 (1964).

Paper details

vlnc:W_kup79kxk
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

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