Co_2FeAl full Heusler compound based spintronic terahertz emitter
Rahul Gupta, Sajid Husain, Ankit Kumar, Rimantas Brucas, Anders Rydberg, Peter Svedlindh
CCo FeAl full Heusler compound based spintronic terahertz emitter
Rahul Gupta, ∗ Sajid Husain, Ankit Kumar, Rimantas Brucas,
1, 2
Anders Rydberg,
1, 3 and Peter Svedlindh
1, † Department of Materials Science and Engineering, Uppsala University, Box 35, SE-751 03 Uppsala, Sweden Ångström Microstructure Laboratory, Uppsala University, Box 35, SE-751 03 Uppsala, Sweden Department of Physics and Astronomy, FREIA, Box 516, SE-75120 Uppsala, Sweden (Dated: January 27, 2021)To achieve a large terahertz (THz) amplitude from a spintronic THz emitter (STE), materials with 100%spin polarisation such as Co-based Heusler compounds as the ferromagnetic layer are required. However, thesecompounds are known to loose their half-metallicity in the ultrathin film regime, as it is difficult to achieve L2 ordering, which has become a bottleneck for the film growth. Here, the successful deposition using room tem-perature DC sputtering of the L2 and B2 ordered phases of the Co FeAl full Heusler compound is reported.Co FeAl is used as ferromagnetic layer together with highly orientated Pt as non-ferromagnetic layer in theCo FeAl/Pt STE, where an MgO(10 nm) seed layer plays an important role to achieve the L2 and B2 orderingof Co FeAl. The generation of THz radiation in the CFA/Pt STE is presented, which has a bandwidth in therange of 0.1-4 THz. The THz electric field amplitude is optimized with respect to thickness, orientation, andgrowth parameters using a thickness dependent model considering the optically induced spin current, superdif-fusive spin current, inverse spin Hall effect and the attenuation of THz radiation in the layers. This study, basedon the full Heusler Co FeAl compound opens up a plethora possibilities in STE research involving full Heuslercompounds.
Due to several applications of terahertz (THz) radiation infundamental and applied sciences [1–5], it is of eminent in-terest to develop broadband, powerful and low-cost THz radi-ation sources. The THz radiation lies in the frequency rangefrom 0.1 to 30 THz (3.34 cm − − − ) and coin-cides with the energy scale of many collective excitations inthe materials [6, 7]. For example, the resonance frequency ofsecond-order magnetization dynamics in ferromagnetic mate-rials, governed by the inertia of the magnetic moment, liesin THz frequency range [8, 9]. For decades, many effortshave been made towards the development of broadband, pow-erful and low cost THz emitters. Most of the THz emit-ters nowadays are based on photo-carrier excitation of dif-ferent materials using femtosecond laser pulses [10]. Re-cently, it has been experimentally demonstrated that THz ra-diation can also be generated in ferromagnetic (FM) and non-ferromagnetic (NFM) magnetic heterostructures, where apartfrom the charge of the electron, the spin degree of freedom isutilized to generate the THz radiation [11, 12]. These mag-netic heterostructures are generally referred to as spintronicTHz emitters (STEs). In fact, for some of the STEs, such asCoFeB/(Pt,W,Pd) and Fe/Pt, a THz bandwidth in the range of0.3-30 THz have been reported [12, 13]. However, it shouldbe noticed that the the experimentally observed bandwidth de-pends on the pulse width of the femtosecond laser and thedetector used in the terahertz-time domain (THz-TDS) set-up[12].In a THz-TDS set-up, a femtosecond laser pulse createsa non-equilibrium spin polarization in the FM layer, whichsuperdiffuses through the FM/NFM interface into the NFMlayer. The spin current is converted to a charge current inthe NFM layer via the inverse spin Hall effect, which gen-erates the electromagnetic radiation in the THz range. The ∗ [email protected]; [email protected] † [email protected] output signal of the STE thus depends in three fundamen-tal processes; generation of optically induced spin currentin the FM (ultrafast dynamics in femtosecond region), spin-dependent interfacial transport across the FM/NFM interface,and the spin to charge conversion efficiency of the NFMlayer due to spin-orbit coupling (SOC). The ideal STE con-sists of a ∼ ∼ P );e.g. Co MnSi ( P = − MnGe ( P = FeAl ( P = FeAl (CFA) belongs to the Fm ˜3 m space group,exhibits half-metallicity and large ( ∼ ), partially ordered phase (B2) and disordered phase (A2).The L2 phase, being half metallic, is shown in Figure 1. TheB2 phase is formed when Fe and Al randomly share their lat-tice sites, whereas when Co, Fe, and Al are randomly dis-tributed on the lattice sites it corresponds to the A2 phase. Forthe L2 phase, half-metallicity results from hybridization ofthe transition-metal element d -orbitals forming bonding (2e g and 3t g ) and non-bonding states (2e u and 3t u ). The half-metallicity emerges from the separation of non-bonding hy-brids of Co as it cannot hybridize with the d -orbitals of Fe[17–19]. Depending upon the chemical ordering between Co,Fe and Al in CFA, the half-metallicity may be reduced dueto modification of the hybridization between the d -orbitals ofCo and Fe [18, 19]. The atomic ordering depends on thin filmgrowth parameters such as underlying substrate, buffer layer,growth temperature etc.To have a large amplitude of the THz-electric field, thethickness of the FM and NFM layers needs to be in the a r X i v : . [ c ond - m a t . m t r l - s c i ] J a n ( a )( b ) ( c ) FIG. 1. (a) In the first row, the colors of atoms, the perpendicular diffraction vector ( k f − k i ) for CFA(002)/MgO(002) Bragg peaks, and forCFA(111)/MgO(002) Bragg peaks in X-ray diffraction geometry are shown from the left to right, respectively. In the second row, the unitcell of CFA, the atomic orientations corresponding to CFA(002) and CFA(111) Bragg peaks are shown from left to right, respectively. In thethird row, the unit cell of MgO, the atomic orientations corresponding to MgO(002), and MgO(002) Bragg peaks are shown from left to right,respectively. In this model, the cross-sectional TEM zone axis is parallel to MgO [100] for the atomic arrangements of CFA(111), CFA(002),and MgO(002). (b) Cross-sectional TEM image of Pt(6 nm)/CFA(8 nm)/MgO(10 nm)/MgO, where 1, 2 and 3 indicate regions with L2 CFAphase, B2 CFA phase and MgO phase, respectively. (c) X-ray diffraction Bragg peaks for different CFA thickness with Pt(6 nm). nanometer range. However, it is difficult to achieve the L2 or-dered phase of CFA in ultralow thickness regime ( ∼ et al. [20] reported that B2 ordering of CFA canbe achieved for 20 nm thickness, which is too large for CFAbased STEs. Husain et al. [21, 22] also reported B2 orderingof CFA deposited on MgO and Si substrates for the relativelylarger thickness regime ( ∼
50 nm). Moreover, L2 orderinghas been achieved for Co FeSi Heusler compounds depositedon MgO, SrTiO , and MgAl O substrates having relativelylarger thickness ( ∼
25 nm) [23]. X. Zhang et al. reportedthat the spin polarization varied between 50-60% varying thethickness of B2 ordered CFA films in the range of 5-20 unitcells ( ∼ and B2 orderingof CFA down to 3 nm thickness. Because of mixed L2 andB2 ordering, which will be discussed in the forthcoming sec-tion, we expect more than 50-60% spin polarization in ourCFA thin films. To achieve the L2 and B2 ordering, a MgO10 nm seed layer was used on MgO and Si substrates (cf. sup-plementary information section I). To confirm the L2 andB2 phases of CFA and the orientation of the Pt layer in ourCFA/Pt STEs, we have performed X-ray diffraction measure-ments in the Bragg-Brentano geometry using a Cu K α source.The presence of CFA(111), CFA(002) and Pt(002) diffractionpeaks indicate that the CFA film exhibits a mixed L2 and B2phase and that Pt is highly orientated following the CFA(002)orientation. However, the CFA(002) and CFA(004) Braggpeaks do not appear below 15 nm CFA thickness as shown ( a ) zxy ( b ) ( c ) FIG. 2. (a) A schematic of the experimental THz-TDS set-up. (b) Typical spectrum of THz electric field amplitude in time-domain forthe Pt(4.5)/CFA(3)/MgO(10)/MgO (blue color), Pt(3)/Fe(2)/MgO (green color) and Pt(6)/CFA(3)/MgO(10)/MgO STEs (black/red color).(c) Fast Fourier transform of time-domain signal for the Pt(4.5)/CFA(3)/MgO(10)/MgO (blue color), Pt(3)/Fe(2)/MgO (green color) andPt(6)/CFA(3)/MgO(10)/MgO STEs (black color). in Figure 1(c). This may be explained by the X-ray sourcenot having enough intensity to resolve these peaks. To con-firm our supposition, we utilized cross-sectional transmissionelectron microscopy (TEM) to visualize the atomic orienta-tion of CFA with respect to the MgO-substrate for the Pt(6nm)/CFA(8 nm)/MgO(10 nm)/MgO sample. As expected,two atomic orientations, CFA(111) and CFA(002), can beidentified in the TEM images. On the one hand, when welook at the structure along the MgO [100] substrate directionin cross-sectional TEM, MgO and CFA exhibit square latticescorresponding to the MgO(002) and CFA(002) crystal latticeplanes when the CFA(002) X-ray diffraction vector (corre-sponds to the B2 phase) is perpendicular to film plane. Onthe other hand, CFA exhibits a hexagonal lattice and MgO a square lattice corresponding to the CFA(111) and MgO(002)crystal lattice planes, respectively when the CFA(111) X-raydiffraction vector (corresponds to the L2 phase) is perpendic-ular to film plane as indicated in Figure 1(a). A cross-sectionalview of the CFA(8 nm) sample along the MgO [100] substratedirection is shown in Figure 1(b), where point 1 (hexagonalorientation) and point 2 (square orientation) correspond to theCFA(111) and CFA(002) crystal lattice planes, which con-firms our supposition that the CFA film exhibits a mixed L2 and B2 phase up to 3 nm thickness. The matching of the Ptand CFA layers to the MgO buffer and substrate also confirmsthat our STEs are highly oriented. As CFA is known for itslow damping, therefore, to further confirm our prediction, wehave performed ferromagnetic resonance spectroscopy to ob-tain the intrinsic Gilbert damping parameter, which is foundto be 5.58 (47) × − , which again indicates the mixed L2 and B2 phases of CFA. More details can be found in supple-mentary information section III.To compare the performance of our CFA/Pt STEs,we have also deposited the Fe(2 nm)/Pt(3 nm) refer-ence sample on MgO substrate. To optimize the pa-rameters of the CFA/Pt STE, we utilized the THz-TDSset-up schematically shown in Figure 2(a) (cf. sup-plementary information section VI). The THz emissionalong with its fast Fourier transform spectrum for thereference Pt(3)/Fe(2)/MgO, Pt(4.5)/CFA(3)/MgO(10)/MgO,and Pt(6)/CFA(3)/MgO(10)/MgO STEs are shown in Fig-ures 2(b,c). The peak-to-peak amplitude of THz E-field is found to be similar for the Pt(3)/Fe(2)/MgO andPt(4.5)/CFA(3)/MgO(10)/MgO STEs. Moreover, there issome ringing in the time domain signal after the main THzpulse due to multi-order reflections of the THz pulse withinthe sample, originating at the interface between the substrateand Si-lens and back side of the substrate (i.e., laser pumpside). The effect of the ringing is also visible in the FFT ofthe signal as shown in Figure (2c). This effect of the ringingcan be seen in several publications displaying the THz signalin the time and frequency domains [25–28]. The peak-to-peakamplitude of the THz signal ( E THz ) in time-domain is mea-sured for all CFA/Pt STEs and used as the dependent variablein equation (3) with the CFA thickness as an independent vari-able. The explanation of equation (3) is given in the followingsection.A relation between the THz electric field and the chargecurrent ( j c = − e (cid:82) t γ ( z ) j s ( z , ω ) dz ) in the NFM layer can beformulated using Ohm’s law by assuming that the electromag-netic radiation propagates as a plane-wave perpendicular tothe STE film plane ( i.e., z-direction as shown in Figure 2a),as the sample thickness is much smaller then the wavelengthof the THz radiation. The THz electric field after the STE isdescribed by E THz = Z ( ω ) e (cid:82) t γ ( z ) j s ( z , ω ) dz , where e is the charge of the electron, t ( = t CFA + t Pt ) is the total thickness of the STE, γ ( z ) is thespin to charge conversion efficiency, and j s is the spin currentdensity. Z ( ω ) is the impedance, which describes the conver-sion efficiency of charge current into THz electric field in theNFM layer, that can be expressed as [29],1 Z ( ω ) = ( n air + n MgO ) Z + (cid:90) t σ ( z , ω ) dz , (1)where ω /2 π is the frequency of the THz radiation, n air and n MgO are the refractive index of air and the MgO substrate,respectively, Z ( ∼ Ω ) is the impedance of free space, and σ is the electrical conductivity of the metal layers.Several mechanisms have been proposed for describinggeneration of the optically induced spin polarised current inthe FM layer, such as a superdiffusive spin current due to spinfiltering of hot electrons [30], the spin-Seebeck effect [31],and a demagnetization process based on magnon-phonon cou-pling [32]. Since the THz electric field amplitude is dependenton the spin polarised current generated in the CFA layer, onecan expect a larger amplitude due to multiple reflections of the pump pulse and therefore the Pt/CFA heterostructure canbe treated as an Fabry-Perot cavity. The shorter the cavity vol-ume is in the transverse direction ( i.e., z-direction as shown inFigure 2a), the more echos occur, which may cause an en-hancement of the spin current, and as a result the THz am-plitude increases. However, below a certain thickness of theFM layer, the possibility of having an out-of-plane componentof the spin polarization in the FM layer or even a superpara-magnetic behaviour in case of island growth cannot be ruledout, which will cause a decrease of the THz electric field am-plitude as indicated for the minimum CFA thickness used inthis study (cf. Figure 3)). Therefore, due to this assumption[27, 33], E THz is proportional to tanh [ t CFA − t c λ pol ] , where t c is thecritical layer thickness and λ pol is a characteristic length thatdescribes the saturation of the spin current when the thicknessof the FM layer is larger than the critical thickness. t c is thethickness where the magnetization has an out-of-plane com-ponent, and therefore the applied magnetic field is not suffi-cient to saturate all the magnetic spins in film plane direction,as a result the lowest amplitude of THz-electric field is foundas shown in Figure (3). Thickness dependent magnetizationmeasurements are discussed in the forthcoming section.The generated spin current super diffuses from CFA to Ptthrough the CFA/Pt interface. Considering reflection of thespin current at the interface, the spatial dependence of the spinpolarised current inside in the Pt layer is defined as [34], j s ( z ) = j s ( CFA ) sinh (( z − t Pt ) / λ Pt ) sinh ( t Pt / λ Pt ) , (2)where j s ( CFA ) is the spin current density at the CFA/Pt in-terface and λ Pt is the spin diffusion length in Pt. The emittedTHz electric field is linearly dependent on the pump fluency,which indicates that the spin current is proportional to theenergy density of the pump pulse. The measurements wereperformed at ∼
40 mW laser pump power corresponding toa pump fluency of about 0.15 mJ/cm calculated for an esti-mated laser spot size of 20 µ m in diameter. This power levelis well below the region, starting at about 1mJ/cm , where theTHz signal starts to deviate from a linear relationship betweenTHz power and the laser pump fluency. The external magneticfield was fixed at 85 mT (0.67 × A/m) during the measure-ments. Therefore, j s ( CFA ) will scale as A / t , where A is theabsorbance of the incident pump pulse by the STE. To knowthe absorbance, we have measured the reflected ( P re f ) andtransmitted ( P trans ) powers in the optical region as a functionof CFA thickness and calculated using A = − T − R as shownin Figure 4(d), where T ( = P trans / P inc ) and R ( = P re f / P inc ) arethe transmittance and reflectance in the optical region, and P inc is the incident power. The absorbance for all STEs was foundto be larger than ∼ E THz (arb. units)
P t ( t n m ) / C F A ( 3 n m )
P t ( 6 n m ) / C F A ( t n m ) F i t t e d t C F A (n m ) t Pt (nm ) R e f e r e n c e P t ( 3 ) / F e ( 2 )
FIG. 3. CFA (red) and Pt (black) thickness dependent THz electric field peak-to-peak amplitude. The blue curve is fitted with a CFA thicknessdependent THz electric field amplitude using equation (3). Green color corresponds to THz electric field peak-to-peak amplitude for thePt(3)/Fe(2)/MgO STE. factor exp − t / ζ THz for the THz electric field, where ζ THz is theeffective inverse attenuation coefficient. This factor is only de-pendent upon the total thickness of the layers [27]. Therefore,making use of all assumptions mentioned above in Ohm’s law,the final form of the THz electric field can be expressed as, E THz ∝ At tanh (cid:18) t CFA − t c λ pol (cid:19) · tanh (cid:18) t Pt λ Pt (cid:19) (3) · n air + n MgO + Z ( σ CFA t CFA + σ Pt t Pt ) e − t / ζ THz , where σ CFA and σ Pt are the electrical conductivities of CFAand Pt, respectively. The experimental data for the THz elec-tric field have been fitted using equation (3) as shown in Fig-ure 3. We have used the experimentally measured A ∼ n air =1, n MgO =3.1 (cf. supplementary information sectionVII), σ CFA = .
414 MS/m (cf. supplementary informationsection V), σ Pt = λ Pt = . t c , λ pol and ζ THz as global fitting parameters, yielding t c = . ( ) nm, λ pol = . ( ) nm and ζ THz = . ( ) nm for the CFA/Pt STE.To understand the THz electric field amplitude below thecritical thickness, we have studied the thickness (2, 3 and 15 nm) dependent static magnetic properties of the CFA filmscapped with Pt(6nm). It can be seen that the samples with3 and 15 nm thickness exhibit in-plane magnetization withsquare-like hysteresis loops following the expected behav-ior of epitaxially/highly oriented grown CFA thin films [20].However, in case of the 2nm thin CFA film, the magnetiza-tion saturates at a relatively higher magnetic field as shown inFigure 4a. Moreover, the hysteresis curve is not square-like,which indicates that the magnetization is dominated by an out-of-plane or easy-cone anisotropy. Thus, the applied magneticfield used in the THz measurements is not sufficient to satu-rate the magnetization, which results into a low value for theTHz electric field amplitude.In the above model (equation 3), we did not consider theeffect of the Si lens, which is used to focus the emitted THzradiation towards the detector. Moreover, the amplitude ofthe THz electric field depends upon the interface roughness;the higher the interface roughness, the lower the THz elec-tric field amplitude [35]. This may be due to interfacial spinmemory loss (SML) at the FM/Pt interface as the SML de-pends on the interface mixing/roughness [36]. The roughnessof our CFA/Pt STEs lies in the range of 0.5-1.1 nm. It meansthat there is still room for improvements with respect to theTHz electric field amplitude by reducing the interface rough- (a) (d)(c) (b) FIG. 4. Magnetization versus applied magnetic field for (a) Pt(6)/CFA(2)/MgO(10)/MgO, (b) Pt(6)/CFA(3)/MgO(10)/MgO, and (c)Pt(6)/CFA(15)/MgO(10)/MgO samples. (d) Tranmittance ( T ), reflectance ( R ) and absorbance ( A ) of the CFA/Pt STEs. The laser wavelengthwas 800 nm. ness of the CFA/Pt STE. Importantly, the bandwidth is foundto be 0.1-4 THz for CFA/Pt STEs, which is limited by thelow temperature (LT)-GaAs detector used in our system. Bychanging the LT-GaAs detector to a ZnTe detector and using ashorter pulse width of the femtosecond laser, we would expecta bandwidth of our STEs upto 30 THz [12].In summary, the Co FeAl full Heusler compound basedspintronic terahertz emitter (STE) is reported, which has abandwidth in the range of 0.1-4 THz. We have deposited anordered L2 and B2 mixed phase of Co FeAl in the ultra-thin film regime at room temperature using MgO(10 nm) asseed layer. We utilized the thickness-dependent THz electricfield model to extract the THz parameters for our Co FeAl/PtSTEs. The THz electric field amplitude was optimized with respect to the thickness, growth parameters, and orienta-tion. The THz electric field is found to be maximum forPt(6)/CFA(3) STE and minimum for Pt(6)/CFA(2) STE, re-spectively. Our study reveals that below 3 nm CFA thickness,the THz electric field amplitude decreases, which is explainedby the out-of-plane component of the magnetization. Thisstudy provides a unique direction for STE research, utilizingthe Co-based full Heusler compounds.
ACKNOWLEDGMENTS
This work is supported by the Swedish Research Council(grant no. 2017–03799) and Olle Engkvists Stiftelse (grantno. 182–0365). RG acknowledges Vassilios Kapaklis for helpin the XRD measurments and Lars Rieker for TEM measure-ments. [1] Z. Jin, A. Tkach, F. Casper, V. Spetter, H. Grimm, A. Thomas,T. Kampfrath, M. Bonn, M. Kläui, and D. Turchinovich, NaturePhysics , 761 (2015).[2] M. Hangyo, M. Tani, and T. Nagashima, International journalof infrared and millimeter waves , 1661 (2005).[3] X. Yang, X. Zhao, K. Yang, Y. Liu, Y. Liu, W. Fu, and Y. Luo, Trends in biotechnology , 810 (2016).[4] J. A. Zeitler, P. F. Taday, D. A. Newnham, M. Pepper, K. C.Gordon, and T. Rades, Journal of Pharmacy and Pharmacology , 209 (2007).[5] E. P. J. Parrott, Y. Sun, and E. Pickwell-MacPherson, Journal ofMolecular Structure , 66 (2011). [6] W. Zhang, P. Maldonado, Z. Jin, T. S. Seifert, J. Arab-ski, G. Schmerber, E. Beaurepaire, M. Bonn, T. Kampfrath,P. M. Oppeneer, and D. Turchinovich, Nature Communication (2020).[7] P. Uhd Jepsen, D. G. Cooke, and M. Koch, Laser Photonics Rev , 124 (2011).[8] M.-C. Ciornei, J. M. Rubí, and J.-E. Wegrowe, Phys. Rev. B ,020410 (2011).[9] K. Neeraj, N. Awari, S. Kovalev, D. Polley, N. Z. Hagström,S. S. P. K. Arekapudi, A. Semisalova, K. Lenz, B. Green, J.-C.Deinert, et al. , Nature Physics , 1 (2020).[10] E. A. Prieto, A. D. L. Reyes, V. D. A. Vistro, N. I. Cabello,M. A. Faustino, J. P. Ferrolino, J. D. Vasquez, H. Bardolaza,J. P. Afalla, V. K. Mag-usara, H. Kitahara, M. Tani, A. Som-intac, A. Salvador, and E. Estacio, Applied Physics Express ,082012 (2020).[11] T. Kampfrath, M. Battiato, P. Maldonado, G. Eilers, J. Nötzold,S. Mährlein, V. Zbarsky, F. Freimuth, Y. Mokrousov, S. Blügel, et al. , Nature nanotechnology , 256 (2013).[12] T. Seifert, S. Jaiswal, U. Martens, J. Hannegan, L. Braun,P. Maldonado, F. Freimuth, A. Kronenberg, J. Henrizi, I. Radu,E. Beaurepaire, Y. Mokrousov, P. M. Oppeneer, M. Jour-dan, G. Jakob, D. Turchinovich, L. M. Hayden, M. Wolf,M. Münzenberg, M. Kläui, and T. Kampfrath, Nature Photonics (2016).[13] D. Yang, J. Liang, C. Zhou, L. Sun, R. Zheng, S. Luo, Y. Wu,and J. Qi, Advanced Optical Materials , 1944 (2016).[14] S. Husain, R. Gupta, A. Kumar, P. Kumar, N. Behera, R. Bru-cas, S. Chaudhary, and P. Svedlindh, Applied Physics Reviews, In Press (2020).[15] A. Mann, J. Walowski, M. Münzenberg, S. Maat, M. J. Carey,J. R. Childress, C. Mewes, D. Ebke, V. Drewello, G. Reiss, andA. Thomas, Phys. Rev. X , 041008 (2012).[16] G. M. Müller, J. Walowski, M. Djordjevic, G.-X. Miao,A. Gupta, A. V. Ramos, K. Gehrke, V. Moshnyaga, K. Samwer,J. Schmalhorst, et al. , Nature Materials , 56 (2009).[17] I. Galanakis, P. H. Dederichs, and N. Papanikolaou, Phys. Rev.B , 174429 (2002).[18] G. H. Fecher and C. Felser, Journal of Physics D: AppliedPhysics , 1582 (2007).[19] S. Skaftouros, K. Özdo˘gan, E. ¸Sa¸sıo˘glu, and I. Galanakis, Phys.Rev. B , 024420 (2013).[20] M. Belmeguenai, H. Tuzcuoglu, M. S. Gabor, T. Petrisor,C. Tiusan, D. Berling, F. Zighem, T. Chauveau, S. M. Chérif,and P. Moch, Phys. Rev. B , 184431 (2013).[21] S. Husain, S. Akansel, A. Kumar, P. Svedlindh, and S. Chaud-hary, Scientific reports , 28692 (2016).[22] S. Husain, V. Barwal, A. Kumar, R. Gupta, N. Behera, S. Hait,N. K. Gupta, P. Svedlindh, and S. Chaudhary, Journal of Mag-netism and Magnetic Materials , 165258 (2019).[23] K. Kudo, Y. Hamazaki, S. Yamada, S. Abo, Y. Gohda, andK. Hamaya, ACS Applied Electronic Materials , 2371 (2019).[24] X. Zhang, H. Xu, B. Lai, Q. Lu, X. Lu, Y. Chen, W. Niu, C. Gu,W. Liu, X. Wang, et al. , Scientific reports , 1 (2018).[25] J. Klier, G. Torosyan, N. S. Schreiner, D. Molter, F. Ell-rich, W. Zouaghi, E. Peytavit, J.-F. Lampin, R. Beigang,J. Jonuscheit, et al. , International Journal of Antennas and Prop-agation , 10.1155/2015/540175 (2015).[26] Y. Kinoshita, N. Kida, M. Sotome, T. Miyamoto, Y. Iguchi,Y. Onose, and H. Okamoto, ACS Photonics , 1170 (2016).[27] G. Torosyan, S. Keller, L. Scheuer, R. Beigang, and E. T. Pa- paioannou, Scientific reports , 1 (2018).[28] U. Nandi, M. Abdelaziz, S. Jaiswal, G. Jakob, O. Gueckstock,S. M. Rouzegar, T. Seifert, M. Kläui, T. Kampfrath, and S. Preu,Applied Physics Letters , 022405 (2019).[29] T. S. Seifert, Spintronics with Terahertz Radiation: Probing anddriving spins at highest frequencies , Ph.D. thesis (2018).[30] M. Battiato, K. Carva, and P. M. Oppeneer, Phys. Rev. Lett. , 027203 (2010).[31] G.-M. Choi, C.-H. Moon, B.-C. Min, K.-J. Lee, and D. G.Cahill, Nature physics , 576 (2015).[32] G.-M. Choi, B.-C. Min, K.-J. Lee, and D. G. Cahill, Naturecommunications , 1 (2014).[33] R. Allenspach and A. Bischof, Phys. Rev. Lett. , 3385 (1992).[34] O. Mosendz, J. E. Pearson, F. Y. Fradin, G. E. W. Bauer, S. D.Bader, and A. Hoffmann, Phys. Rev. Lett. , 046601 (2010).[35] G. Li, R. Medapalli, R. V. Mikhaylovskiy, F. E. Spada, T. Ras-ing, E. E. Fullerton, and A. V. Kimel, Phys. Rev. Materials ,084415 (2019).[36] K. Gupta, R. J. H. Wesselink, R. Liu, Z. Yuan, and P. J. Kelly,Phys. Rev. Lett. , 087702 (2020).[37] J. Neu and C. A. Schmuttenmaer, Journal of Applied Physics , 231101 (2018).[38] M. S. Gabor, T. Petrisor, C. Tiusan, M. Hehn, and T. Petrisor,Phys. Rev. B , 134413 (2011).[39] M. Belmeguenai, M. S. Gabor, F. Zighem, Y. Roussigné,D. Faurie, and C. Tiusan, Phys. Rev. B , 104424 (2016).[40] M. Scheller, C. Jansen, and M. Koch, Optics Communications , 1304 (2009).[41] J. C. Lagarias, J. A. Reeds, M. H. Wright, and P. E. Wright,SIAM Journal on optimization , 112 (1998).[42] Y. Zhang, L. Zhang, P. Sun, Y. He, Y. Zou, and Y. Deng, Appliedspectroscopy , 1102 (2016).[43] M. Scheller, Optics express , 10647 (2011).[44] T. Kampfrath, J. Nötzold, and M. Wolf, Applied physics letters , 231113 (2007).[45] H. T. Nembach, T. J. Silva, J. M. Shaw, M. L. Schneider, M. J.Carey, S. Maat, and J. R. Childress, Phys. Rev. B , 054424(2011).[46] R. Gupta, N. Behera, V. A. Venugopal, S. Basu, A. K. Puri,P. Ström, M. A. Gubbins, L. Bergqvist, R. Brucas, P. Svedlindh,and A. Kumar, Phys. Rev. B , 024401 (2020).[47] M. A. W. Schoen, J. M. Shaw, H. T. Nembach, M. Weiler, andT. J. Silva, Phys. Rev. B , 184417 (2015).[48] M. A. Schoen, D. Thonig, M. L. Schneider, T. Silva, H. T. Nem-bach, O. Eriksson, O. Karis, and J. M. Shaw, Nat. Phys. , 839(2016).[49] J. M. Shaw, H. T. Nembach, T. J. Silva, and C. T. Boone, Journalof Applied Physics , 243906 (2013).[50] M. Björck and G. Andersson, Journal of Applied Crystallogra-phy , 1174 (2007).[51] S. Husain, A. Kumar, P. Kumar, A. Kumar, V. Barwal, N. Be-hera, S. Choudhary, P. Svedlindh, and S. Chaudhary, PhysicalReview B , 180404(R) (2018).[52] L. Zhu, D. C. Ralph, and R. A. Buhrman, Phys. Rev. Lett. ,077201 (2019).[53] S. Zhang, Z. Jin, X. Liu, W. Zhao, X. Lin, C. Jing, and G. Ma,Optics letters , 3080 (2017).[54] S. Akansel, V. A. Venugopal, A. Kumar, R. Gupta, R. Brucas,S. George, A. Neagu, C.-W. Tai, M. Gubbins, G. Andersson, et al. , Journal of Physics D: Applied Physics51