Effect of Co-Fe substitutions on the room-temperature spin polarization in Co_3-xFe_xSi Heusler-compound films

0.65 MB·English

by  K. Tanikawa

Effect of Co-Fe substitutions on the room-temperature spin polarization in Co Fe Si
3−x x
Heusler-compound films
K. Tanikawa,1 S. Oki,1 S. Yamada,1 K. Mibu,2 M. Miyao,1,3 and K. Hamaya1∗
1Department of Electronics, Kyushu University, 744 Motooka, Fukuoka 819-0395, Japan
2Nagoya Institute of Technology, Nagoya, Aichi 466-8555, Japan and
3CREST, Japan Science and Technology Agency, Sanbancho, Tokyo 102-0075, Japan
(Dated: January 15, 2013)
Using low-temperature molecular beam epitaxy, we study substitutions of Fe atoms for Co ones
3 in Co3−xFexSi (1.0 ≤ x ≤ 3.0) Heusler-compound films grown on Si and Ge. Even for the low-
1 temperature grown Heusler-compound films, the Co-Fe atomic substitution at A and C sites can
0 be confirmed by the conversion electron M¨ossbauer spectroscopy measurements. As a result, the
2 magnetic moment and room-temperature spin polarization estimated by nonlocal spin-valve mea-

n surements are systematically changed with the Co-Fe substitutions. This study experimentally
a verifiedthattheCo-FesubstitutioninCo3−xFexSiHeuslercompoundscandirectlyaffecttheroom-
J temperaturespin polarization.

2
1 PACSnumbers:

]
i I. INTRODUCTION trodes in lateral spin valves (LSVs),[18] where the CFS
c
films were grown by LT-MBE and highly L2 -ordered
s 1
- structures were also evaluated by conversion electron
l To demonstrate highly efficient spin injection and de-
r Mo¨ssbauer spectroscopy (CEMS) measurements.[19, 20]
tection in spintronic applications, ferromagnetic Heusler
t
m compounds with the chemical formula X YZ have been Contrary to the previous works with MTJs,[10–12] rel-
2 atively large P ∼ 0.8 was already obtained at room
. studied,[1–8] where X and Y are transition metals and
t temperature.[21] We infer that these results originate
a Z is a main group element such as Si and Ge. A
m schematic diagram of X YZ Heusler-compound struc- from the high-quality epitaxial growth with the precise
2
controloftheatomiccompositionusingLT-MBE,[19,20]
- tures is shown in Fig. 1(a), and four crystal sites are
d denoted as A(0,0,0), B(1,1,1), C(1,1,1), D(3,3,3) in and these growth techniques have already been estab-
n 4 4 4 2 2 2 4 4 4 lished for other Heusler compounds.[22, 23] Hence, if we
Wyckoff coordinates.
o tune the substitution of Fe for Co using our LT-MBE
c It is well known that Co2FeSi (CFS) is one of the techniques, the controlof P canbe demonstrated exper-
[ Co-based Heusler compounds with the highest Curie
imentally and one can clarify the influence of the Co-Fe
temperature and highly spin-polarized density of states
1 substitutions on P in CFS.
(DOS) at the Fermi level.[9] For ideal CFS with an
v In this study, we study the effect of the substitution
45 oLc2c1u-opriedderewditshtruCcoturaen,dthFee(Aat,Com)ss,itersesapnedctiBvelsyi.tes Aarle- ofFe for Co onP in low-temperaturegrownCo3−xFexSi
films. UtilizingourLT-MBEtechniques,wedemonstrate
6 though room-temperature spin-related functionality was
a substitution of Fe for Co occupying A and C sites
2 shown in magnetic tunnel junctions (MTJs) with CFS
1. electrodes,[8, 10–12] the spin polarization (P) estimated in Co3−xFexSi films even on group-IV semiconductors.
Withvaryingx(1.0≤x≤3.0),weclearlyobservemono-
0 was not so large (P ≤ 0.5). It has so far been ar-
tonicchangeinP atroomtemperature. Thisexperimen-
3 gued that the intentional Co-Fe substitutions between
1 tal study gives an important knowledge to obtain high-
the (A,C) and B sites can affect the decrease in P of
v: CFS electrodes.[8, 11, 13, 14] Theoretically, electronic performance Co3−xFexSi Heusler-compound electrodes.
i band structures were calculated and the effect of the
X
Co-Fe substitutions on P was examined for Co-Fe based
r Heusler-compound films.[15] However, there was no ex- II. SAMPLES AND MEASUREMENTS
a
perimental work on the influence of the Co-Fe substitu-
tions on P in the CFS films. Prior to the fabrication and characterizations of thin-
Recently, in spin light-emitting diodes (LED) with film samples, we briefly explain the crystal structure of
CFS spin injectors,[16, 17] highly efficient spin injec- general bulk samples by using Fig. 1(a). In the ideal
tion into III-V semiconductors was demonstrated. The L21-type CFS, there are two Co sites (A and C sites),
CFS films were grown on (Al,Ga)As by using low- oneFesite (Bsite), andoneSisite (Dsite),wherethe A
temperaturemolecularbeamepitaxy(LT-MBE).Wealso andC sitesaremagneticallyequivalent. With increasing
demonstrated relatively high performance for CFS elec- Fe concentration (x) from x = 1.0 to 3.0 in Co3−xFexSi,
Fe atoms can occupy the A and C sites, leading to the
replacement of Co atoms. When all the A and C sites
are occupied by Fe atoms, D0 -type Fe Si (FS) can be
3 3
∗E-mail: [email protected] obtained.[24, 25]
2
(a) X A,C site (b) (a)
224
Y B site 222 FeCoSi Exp.
Z D site 220 113 2
111
Total fit.
004
Fe CoSi 002
2 000
)
s
nit
u
b. site number
r
Si(111) (a 1
s
1nm nt
u 2
(0,0,0) o
C
3
FIG. 1: (Color online) (a) A schematic diagram of Heusler-
4
compounds with the chemical formula X2YZ. (b) Cross-
sectional TEM image of Fe2CoSi/Si(111). The inset shows 5
NED patterns for Fe2CoSi film near the interface. The zone -10 -5 0 5 10
axis is parallel tothe [110] direction. Velocity (mm/s)
(b) 50
Co3−xFexSi (1.0 ≤ x ≤ 3.0) layers with a thickness 40 Fe(I), (A,C) site
of ∼25 nm were grown directly on non-doped Si(111) Fe(II), B site
◦ Co, (A,C) site
and/or Ge(111) by LT-MBE at 100 C, where we co- %)30 Si, D site
evaporated Co, Fe and Si using Knudsen cells.[19, 20] (
a
In order to change the ratio of Fe to Co, the growth e
r20
A
rates of Co and Fe were tuned by adjusting the cell
temperatures. The x value in the films was confirmed
10
by measuring the energy dispersive x-ray spectra. In-
situreflectionhighenergyelectrondiffraction(RHEED)
0
patterns of Co3−xFexSi layersclearly exhibited symmet- 1 2 3 4 5
rical streaks, showing good two-dimensional epitaxial Site number n [t]
growth. Structural and magnetic properties were char-
acterized by transmission electron microscopy (TEM) FIG.2: (Coloronline)(a)Room-temperature57FeM¨ossbauer
images, nanobeam electron diffraction (NED) patterns, spectrum(solidcircles)ofanFe2CoSifilm,togetherwithfive
magnetization curves, and 57Fe Mo¨ssbauer spectra, as fitting subspectra as denoted in the text. (b) Fitted area
shown in our previous works.[19, 20, 26, 27] To enhance versussite numberfor Fe2CoSi film,estimated bytheCEMS
the detectability of the 57Fe Mo¨ssbauer spectra, we en- measurement and best fitting. The inset shows schematic
riched 57Fe nuclei to 20 % in the Knudsen cell of the Fe diagramofanorderedFe2CoSistructure. Therearetwotypes
ofmagnetically distinctFesites, i.e.,Fe(I),occupyingAorC
source. For the estimation of the spin polarization at
sitescoordinatedwithfourFeatomsandSiatoms,andFe(II),
room temperature, the Co3−xFexSi films were patterned occupyingtheBsitesurroundedbyeightFeatomsorCoones.
into the submicron-sized electrodes by using a conven-
tional electron-beam lithography and an Ar ion milling
technique. In order to form the LSVs, 100-nm-thick
and/or L2 -structures (solid circles) are clearly identi-
1
Cu strips bridging the two Co3−xFexSi electrodes were fied, consistent with our previous works on CFS (x= 1)
patterned by a conventional lift-off technique, together
[19,20]andFS(x=3).[26,27]Thesefeaturesimplythat
with bonding pads. Nonlocal spin-valve (NLSV) mea-
the Co atoms occupying the A and C sites are replaced
surementswerecarriedoutbyaconventionalcurrent-bias
lock-in technique (∼ 200 Hz). The detailed fabrication byFeonesinCo3−xFexSifilmswithincreasingxwithout
formingotherdisorderstructures. Forthefilmsgrownon
process and measurements are described elsewhere.[21]
Ge(111),wehavealreadyconfirmedalmostidenticalfea-
tures.
In order to verify the atomic substitution between Co
III. RESULTS AND DISCUSSION
and Fe, we examined the magnetic environments around
the Fe sites for a FCS film (x = 2.0) by 57Fe Mo¨ssbauer
Figure 1shows ahigh-resolutionTEMimageof oneof spectroscopy. Oneofthecrystalstructuresofanordered
the low-temperature grown Co3−xFexSi films, Fe2CoSi FCSisillustratedintheinsetofFig. 2(b). Therearetwo
(FCS, x = 2.0)/Si(111). We see no reaction layers near typesofmagneticallydistinctFesites,i.e.,Fe(I),occupy-
the interface. Also, lattice image and the NED pattern ing A or C site coordinated with four Fe atoms and four
in the inset of Fig. 1(b) of the FCS layer show single- Si atoms, and Fe(II), occupying the B site surrounded
crystalline characteristics. Superlattice reflections, (111) by eight Fe or Co atoms. In a previous work of bulk
and (113), originating from the presence of the D03- Co3−xFexSi samples reported by Jung et al.,[13] the two
3
6 values can be caused by the site occupation of surplus
Fe andSiatoms with(A,C) sites.[26, 27]Inother words,
on Si(111)
althoughthe presenceofabout 15%sites (sites 3and4)
5 on Ge(111) wasalsoarisingfromtheFe(I)-andFe(II)-likestructures,
local off-stoichiometry led to such disordered structures.
Finally, the site 5 with less than 15 % was regarded as
)
f.u. 4 6 thealmostnonmagneticsitewithasmallhyperfinemag-
/μB FeCoSi neticfieldoflessthan5T,whichmayoriginatefromthe
(S u.) 3 2 FCS/Ge interface.[27]
M 3 /f.μB 0 In our previous works for FS (x = 3),[26, 27] we have
M (-3 obtainedrelativelyhighlocaldegreeofthe structuralor-
dering(∼70%)forFS(x=3). Also,wehaveconfirmed
-6 that the local degree of the structural ordering exceeds
2 -100 -50 0 50 100
H (Oe) 70% is achievedandthe ratio ofsite 1 to site 2 is nearly
onetooneinFig. 2. Thus,evenifweconductthechange
1 1.5 2 2.5 3 in x in LT-MBE conditions, we can achieve the atomic
x (Co FeSi) substitutions without decreasing the structural ordering
3-x x
intheHeusler-compoundstructure. Consideringthefact
that there was no site occupation of Fe atoms with D
FIG. 3: (a) MS as a function of x for Co3−xFexSi films on
sites for CFS (x = 1),[19, 20] we can eventually judge
Si(111) and/orGe(111), measured at300K.Theinsetshows
that the Fe atoms are substituted for the Co ones occu-
theM-H curvefor Fe2CoSi, measured at 300 K.
pying with A and C sites, following the site preference
selectivity described for a bulk Co3−xFexSi system.[24]
sites with two different hyperfine magnetic fields were Takingthesestructuralcharacterizationsintoaccount,
completely distinguished by the 57Fe Mo¨ssbauer spec- wefurtherexaminedmagneticpropertieswithincreasing
troscopy. Thus, we can probe the local structures of the x inCo3−xFexSi films grownonSi(111)and/orGe(111).
FCS film by detecting the two different surroundings of The inset of Fig. 3 shows representative field-dependent
Featoms. The57FeMo¨ssbauerspectroscopyfortheFCS magnetization(M−H)curveat300Kforthe FCSfilm,
epilayergrownis presentedin Fig. 2(a). An evidentsex- where the applied field direction is parallel to the mag-
tet pattern, typical for magnetically ordered systems, is netic easy axis in the film plane. A clear ferromagnetic
obtained (black solid circles). The spectrum can be suc- hystereticcurveisobservedandthecoerciveforceisvery
cessfully fitted with five magnetic environments defined smallwithavalueof∼9Oe,whichisconsistentwithtyp-
assitenumbernfrom1to5,wherethefivefittingcurves ical characteristics of Heusler compounds.[1–7, 28] The
arealsopresentedinFig. 2(a). Notethatthequadrupole estimated saturation magnetization (MS) reaches more
shift was fixed to zero due to the cubic symmetry of the than ∼80 % of bulk MS.[24, 25] These features were
local Fe environment and the intensity ratio of the mag- seen for all the Co3−xFexSi films (1.0 ≤ x ≤ 3.0), im-
netically split sextet was also fixed to 3:4:1:1:4:3, since plying that the crystal quality of the Co3−xFexSi films
themagneticmomentwasalongthefilmplane,reflecting arevery good. Then, we summarize MS versusx for the
the magnetic shape anisotropy. The large two contribu- Co3−xFexSi films in the main panel of Fig. 3, where MS
tions, denoted by the site 1 and 2 with hyperfine mag- values of the Co3−xFexSi films grown on Si and Ge are
netic fields of 19.4and32.8T, respectively,arethe same plotted. We can find that MS is monotonically tuned
as the two sites, Fe(I) and Fe(II), in the bulk samples in withvaryingx. Thistendencyisingoodagreementwith
Ref.[13]. On the other hand, the other three subspectra thatofbulksamplesinapreviouswork,[24]inwhichthis
with smaller intensity can be recognized as disordered behaviorin MS versus x canbe explained by the change
sites. inthelocalmagneticmomentofFeatomsduetothesub-
The percentage of the fitted area versus site number stitutionofFeforCooccupyingAandCsites.[24]There-
for the FCS film is also presented in Fig. 2(b). If we fore, we conclude that the obtained monotonic change
obtain the perfectly substituted structure in FCS (x = in MS versus x results from the experimentally verified
2.0), there are Fe(I) and Fe(II) sites evenly, i.e., site 1 : atomicsubstitutioninCo3−xFexSifilms(1.0 ≤x≤ 3.0).
site2=50(%) :50(%). InactualourFCSepilayer,site Also, the electronic structures are determined pre-
1:site 2=37.3(%):36.6(%), indicatingthatthe local dominantly by the chemical surroundings of magnetic
degree of the structural ordering exceeds 70 % even for atomssuchasthe numberandtype ofthe nearestneigh-
the low-temperature grown FCS film. We note that the bors at given site.[11, 15, 22, 25, 29] Since the magnetic
ratio of site 1 to site 2 is nearly one to one. Since the properties in Co3−xFexSi films on Si or Ge were con-
hyperfine magnetic fields for the site 3 and site 4 were trolled by the Co-Fe atomic substitution, the control of
30.5 T and 25.1 T, respectively, the site 3 and 4 might P canalsobeexpectedbecauseofthesystematicchange
be very similar to Fe(II) and Fe(I), respectively. Such in the DOS at the Fermi level.[11, 13, 14] To confirm
deviationsofthehyperfinemagneticfieldsfromthe ideal this effect, we finally evaluated the room-temperature P
4
(a) (b) tion, interface spin polarization,resistivity, and spin dif-
fusionlengthoftheferromagneticelectrode,respectively.
0 ρ and λ are those for the nonmagnetic wire, and d
V N N
is the separation distance between the injector and de-
)
Ω ΔRs tector (d = 400 nm). We determined ρ by the four-
Cu m F
( -1 terminal resistance measurements. Also, we know that
I
I Fe2C oSi H V/Δ ΔRs = 1.3mΩ ρRNAaFn/NdλiNs athree2r.5esµisΩtacnmceanodf5t0h0enmint,errefsapceectbiveetwlye.[e1n8,t2h1e]
-2 FeCoSi/Cu LSV ferromagnet and nonmagnet, but this term can be ig-
2
500 nm nored since we have already confirmed that RA <
F/N
-1000 -500 0 500 1000
0.1 fΩm2. Note that the area parameter A is defined
H (Oe)
as (S S /S ), where S , S , and S are the areas
(c) inj det N inj det N
1 of the junctions in the spin injector and spin detector,
CoFeSi (Ref. 21)
2 (x = 1) and the cross section of the nonmagnetic strip, respec-
0.8 tivelyasdescribedinourpreviousworks.[18,21,30]The
Fe2CoSi A values for the Co3−xFexSi/Cu junctions were directly
0.6 (x = 2.0) measured by SEM observations. Hence, using Eq. (1),
|F
P we can obtain P by making an assumption about the
F
| 0.4 value of λ . In this study, we used the λ value of 4 ±
F F
0.2 Fe1(.x5C =o 11..55S)i Fe3Si(R(xe f=. 23.10)) e1rinnmg ofourr xpr=evi2o.u0s(wFoCrSk)s.[a1n8d, 211.5] T(Fhee1.r5eCsuo1lt.5inSgi)Pconvsiadl--
F
ues estimated as a function of ρF for the Co3−xFexSi
0 20 40 60 80 100 120 140 films (x = 1.0, 1.5, 2.0, and 3.0) are shown in Fig. 4(c).
ρ (μΩm) There is an almost monotonic change in the magnitude
F of P with varyingρ , which implies that the structural
F F
ordering crucially affects the PF values for Co3−xFexSi
FIG. 4: (Color online) (a) Scanning electron microscope im- films, consistent with general interpretations in Heusler
age of fabricated Fe2CoSi/Cu LSV. (b) Room-temperature compounds.[16, 17, 21, 34–36]
NLSVsignalforFe2CoSi/CuLSV.(c)Room-temperaturePF
We note that the magnitude of P values for the
as a function of ρF for Co3−xFexSi/Cu LSVs. For x = 1.0 F
and3.0,wereplottedthedatapresentedinRef. [21]. TheλF Co3−xFexSi/Cu LSVs are also changed monotonically
values used are 5 ± 1 nm (x= 3.0), 4 ± 1 nm (x= 2.0 and with varying x. Since the resistivity shown in horizontal
1.5), and 3 ± 1 nm (x= 1.0). axis depended on the microscopic structural ordering in
the Heusler wires, the scattering of the horizontal axis
couldnotbe avoided.[21]Also,the signofP depending
F
for Co3−xFexSi films by using NLSV measurements and on the structural ordering should be considered.[17, 37]
the analysis based on the one-dimensional spin diffusion However, Fig. 4(c) apparently reveals that the room-
model.[18, 21, 30] Since we have so far evaluated CFS temperature P can be tuned by systematically intro-
F
(x = 1.0) and FS (x = 3.0),[21] we hereafter focus on ducing Co-Fe substitution in Co3−xFexSi films even on
Fe1.5Co1.5Si (x = 1.5) and FCS (x = 2.0). A scanning group-IV semiconductor platforms.
electron micrograph (SEM) of a representative FCS/Cu
LSV is shown in Fig. 4(a). The center-to-center sepa-
ration between FCS injector and detector (d) is 400 nm.
Figure 4(b) shows a nonlocal magnetresistance curve of
IV. SUMMARY
aFCS/CuLSV,measuredatroomtemperature. We ob-
serveclearNLSV signaldepending onthe magnetization
configuration between the spin injector and detector. It We have studied the effect of the Co-Fe substitu-
shouldbe notedthatarelativelylargespinsignal(∆RS) tion in Co3−xFexSi on P by using LT-MBE techniques
of∼1.3mΩisseen,whichisroughlyfivetimeslargerthan and nonlocal spin valve studies. Even for the low-
thatforPy/CuLSVsofthe samesizeandlow-resistance temperature grown Heusler-compound films, the Co-Fe
ohmic interfaces.[18] atomic substitution at A and C sites can be confirmed
Inourpreviousworks,[18,21,30]wesimplifiedthespin by the conversionelectronMo¨ssbauerspectroscopymea-
signals on the basis of a one-dimensional spin diffusion surements. Furthermore, the magnetic moment is sys-
model:[31–33] tematically changed with the Co-Fe substitutions. The
room-temperature P was monotonically tuned by sys-
2
∆RSA≈ (cid:16)(1−PPFF2)ρρFλλFs+inh(1(−PdPI/I2λ)R)AF/N(cid:17) , (1) tifieemdattihcaatllythcehCanog-Finegsuxb.sTtithuistiostnudinyCexop3−erxiFmeexnStiaHllyeuvselerr-
N N N
compounds can sensitively affect the room-temperature
where P , P , ρ , and λ are the bulk spin polariza- spin polarization.
F I F F
5
Acknowledgments and Grant-in-Aid for challenging Exploratory Research
from JSPS. The Mo¨ssbauer measurements were sup-
The authors would like to thank Prof. T. Kimura ported by the Nanotechnology Network Platform of
of Kyushu University for their useful discussions. MEXT, Japan.
This work was partially supported by CREST-JST,
STARC, SCOPE-MIC, TEPCO Memorial Foundation, References
[1] R.A.deGroot, F.M.Mueller, P.G.vanEngen,andK. [21] K.Hamaya,N.Hashimoto,S.Oki,S.Yamada,M.Miyao,
H.J. Buschow, Phys. Rev.Lett. 50, 2024 (1983). and T. Kimura, Phys. Rev.B 85, 100404(R) (2012).
[2] I. Galanakis, P. H. Dederichs, and N. Papanikolaou, [22] K. Hamaya, H. Itoh, O. Nakatsuka, K. Ueda, K. Ya-
Phys.Rev.B 66, 174429 (2002). mamoto, M. Itakura, T. Taniyama, T. Ono, and M.
[3] K. Inomata, S. Okamura, R. Goto, and N. Tezuka, Jpn. Miyao, Phys. Rev.Lett. 102, 137204 (2009).
J. Appl.Phys. 42, L419 (2003). [23] S.Yamada,K.Hamaya,T.Murakami,B.Varaprasad,Y.
[4] T. Marukame, T. Kasahara, K. Matsuda, T. Uemura, K. Takahashi, A. Rajanikanth, K. Hono, and M. Miyao,
andM.Yamamoto,Jpn.J.Appl.Phys.,Part244,L521 J. Appl.Phys. 109, 07B113 (2011).
(2005). [24] V. Niculescu, J. I. Budnick, W. A. Hines, K. Raj, S.
[5] Y.Sakuraba,M.Hattori,M.Oogane,Y.Ando,H.Kato, Pickart, and S.Skalski, Phys. Rev.B 19, 452 (1979).
A. Sakuma, T. Miyazaki, and H. Kubota, Appl. Phys. [25] L. Hongzhi, Z. Zhiyong,M. Li, X.Shifeng, L. Heyan,Q.
Lett.88, 192508 (2006). Jingping, L. Yangxian, and W. Guangheng, J. Phys. D:
[6] G. H. Fecher and C. Felser, J. Phys. D: Appl. Phys. 40, Appl. Phys.40, 7121 (2007).
1582 (2007). [26] K.Hamaya,T.Murakami, S.Yamada, K.Mibu,andM.
[7] M.I.Katsnelson,V.Yu.Irkhin,L.Chioncel,A.I.Licht- Miyao, Phys. Rev.B 83, 144411 (2011).
enstein,R.A.deGroot,Rev.Mod.Phys.80,315(2008). [27] S.Yamada,J.Sagar,S.Honda,L.Lari,G.Takemoto,H.
[8] N. Tezuka, N. Ikeda, F. Mitsuhashi, and S. Sugimoto, Itoh, A.Hirohata, K. Mibu, M. Miyao, and K. Hamaya,
Appl.Phys.Lett. 94, 162504 (2009). Phys. Rev.B 86, 174406 (2012).
[9] S. Wurmehl, G. H. Fecher, H. C. Kandpal, V. Kseno- [28] J.Sagar,H.Sukegawa,L.Lair,V.K.Lazarov,S.Mitani,
fontov,C.Felser, H.-J.Lin,andJ.Morais, Phys.Rev.B K. O’Grady, and A. Hirohata, Appl. Phys. Lett. 101,
72, 184434 (2005). 102410 (2012).
[10] Z.Gercsi,A.Rajanikanth,Y.K.Takahashi,K.Hono,M. [29] T. M. Nakatani, A. Rajanikanth, Z. Gercsi, Y. K. Taka-
Kikuchi, N. Tezuka, and K. Inomata, Appl. Phys. Lett. hashi, K. Inomata, and K. Hono, J. Appl. Phys. 102,
89, 082512 (2006). 033916 (2007).
[11] K.Inomata,N.Ikeda,N.Tezuka,R.Goto, S.Sugimoto, [30] S.Oki,S.Yamada,N.Hashimoto,M.Miyao,T.Kimura,
M.Wojcik,andE.Jedryka,Sci.Technol.Adv.Mater.9, and K.Hamaya, Appl. Phys.Exp.5, 063004 (2012).
014101 (2008). [31] T. Valet, and A. Fert,Phys. Rev.B 48, 7099 (1993).
[12] M. Oogane, M. Shinano, Y. Sakuraba, and Y. Ando, J. [32] S. Takahashi and S. Maekawa, Phys. Rev. B 67, 052409
Appl.Phys.105, 07C903 (2009). (2003).
[13] V. Jung, B. Balke, G. H. Fecher, and C. Felser, Appl. [33] T. Kimura, J. Hamrle, and Y. Otani, Phys. Rev. B 72,
Phys.Lett. 93, 042507 (2008). 014461 (2005); T. Kimura, and Y. Otani, J. Phys. Con-
[14] V. Ksenofontov, M. W´ojcik, S. Wurmehl, H. Schneider, dens. Matter. 19, 165216 (2007).
B. Balke, G. Jakob, and C. Felser, J. Appl. Phys. 107, [34] J. Herfort, H.-P. Sch¨onherr, K.-J. Friedland, and K.
09B106 (2010). H. Ploog, J. Vac. Sci. Technol. B 22, 2073 (2004); M.
[15] H. Mori, Y. Odahara, D. Shigyo, T. Yoshitake, E. Hashimoto,J.Herfort,H.-P.Sch¨onherr,andK.H.Ploog,
Miyoshi, Thin Solid Films 520, 4979 (2012). J.Appl.Phys.98,104902(2005);J.Herfort,B.Jenichen,
[16] M.Ramsteiner,O.Brandt,T.Flissikowski,H.T.Grahn, V. Kaganer, A. Trampert, H.-P. Sch¨onherr, and K. H.
M. Hashimoto, J. Herfort, and H. Kostial, Phys. Rev. B Ploog, Physica E 32, 371 (2006).
78, 121303(R) (2008). [35] H.Schneider,G. Jakob,M. Kallmayer, H.J. Elmers, M.
[17] P.Bruski,S.C.Erwin,M.Ramsteiner,O.Brandt,K.-J. Cinchetti, B. Balke, S. Wurmehl, C. Felser, M. Aeschli-
Friedland,R.Farshchi,J.Herfort,andH.Riechert,Phys. mann, and H. Adrian, Phys.Rev.B 74, 174426 (2006).
Rev.B 83, 140409(R) (2011). [36] C. G. F. Blum, C. A. Jenkins, J. Barth, C. Felser, S.
[18] T. Kimura, N. Hashimoto, S. Yamada, M. Miyao, K. Wurmehl, G. Friemel, C. Hess, G. Behr, B. Bu¨chner, A.
Hamaya, NPGasia materials 4, e9 (2012). Reller,S.Riegg,S.G.Ebbinghaus,T.Ellis,P.J.Jacobs,
[19] S. Yamada, K. Hamaya, K. Yamamoto, T. Murakami, J.T.Kohlhepp,andH.J.M.Swagten,Appl.Phys.Lett.
K. Mibu, and M. Miyao, Appl. Phys. Lett. 96, 082511 95, 161903 (2009).
(2010). [37] S.Oki,K.Masaki,N.Hashimoto,S.Yamada,M.Miyata,
[20] K. Kasahara, K. Yamamoto, S. Yamada, T. Murakami, M.Miyao,T.Kimura,andK.Hamaya,Phys.Rev.B74,
K.Hamaya,K.Mibu,andM.Miyao,J.Appl.Phys.107, 174412 (2012).
09B105 (2010).