We demonstrated the fabrication of a rare-earth-free ferromagnetic L10-type Fe–Ni alloy (L10-FeNi) by pulsed laser deposition (PLD). We deposited Fe and Ni on Cu(001) by alternating monoatomic deposition via automatically stabilized laser ablation. We examined the structural properties, magnetic properties, and surface morphology of the alloy specimens as the growth temperature (Ts) was varied. We adequately confirmed the construction of the most prominent L10-FeNi phase at 300 °C, which is significantly higher than previously reported growth temperatures, indicating that PLD followed by thermal treatment promoted two-dimensional growth of the adsorbent. The formation process of L10-FeNi was investigated from the standpoint of surface thermodynamics, and the results suggest that the surface free energy of PLD and its highly instantaneous deposition process by PLD played key roles. Our findings are expected to lead to advanced methods for the fabrication of L10-FeNi.

Spintronic applications have been found to greatly benefit from L10-type ordered thin films with large magnetic anisotropy.1 Structural ordering induces magnetic anisotropy,2 and it is directly related to the areal density and the precession speed of magnetic-recording devices.3 A number of studies have been carried out on L10 ferromagnetic structures, such as L10-FePt, L10-FeAu, and L10-FePd,4–6 and these structures were adequately employed in subsequent applications.7–9 However, these materials utilize large amounts of rare metals, and the demand for alternatives is rapidly increasing to address ecological concerns.10 

An L10-type ordered Fe–Ni alloy (L10-FeNi) is considered a potential candidate for replacing ferromagnetic L10-type rare metals11,12 because it shows remarkably high magnetic anisotropy and a large magnetic moment,13,14 and it is composed of environmentally abundant elements. Its magnetic anisotropy mainly originates from the spin–orbit interaction,15,16 and slight structural changes strongly affect macroscopic magnetic anisotropy. Kojima et al. reported that the uniaxial magnetic anisotropy constant (Ku) is monotonically proportional to the chemical order parameter (S).17,18,21 Shen et al. reported that in a comparison between molecular beam epitaxy (MBE) and pulsed laser deposition (PLD), the perpendicular magnetization of a bilayer system is correlated with the interfacial structure.19 These reports all emphasize the importance of structural properties and the need to develop growth techniques for the realization of L10-FeNi. However, the correlations between the lattice structure, the magnetic properties, and the surface and interfacial morphology have not been investigated satisfactorily, and a suitable technique for the fabrication of ferromagnetic L10-FeNi has yet to be developed.

Here, we report the useful fabrication of L10-FeNi by the PLD technique. In general, PLD enables the preparation of almost ideal films by layer-by-layer epitaxial growth,19 and the ordered structure can be improved. Shen et al. reported that an atomically flat surface was adequately grown, while three-dimensional island growth was suppressed by the PLD technique, thus providing great opportunities to design artificial thin-film structures. In our study, we adopted PLD to prepare L10-FeNi films. To avoid the formation of droplets, we utilized automatic control of the laser power to attain stable and satisfactory ablation on a small scale.

We investigated the relationship between the lattice structure, the magnetic properties, and the surface morphology of a PLD-grown L10-FeNi film. In particular, we systematically investigated the dependence of island formation on the growth temperature. We characterized the surface roughness by reflective high-energy electron diffraction (RHEED), and we used X-ray diffraction (XRD) with synchrotron radiation as a powerful tool to evaluate the structural properties. A superconducting quantum interference device (SQUID) magnetometer was used to evaluate the magnetic properties of the samples, and atomic force microscopy (AFM) was utilized to observe the surface morphology. Finally, we investigated the formation mechanism of the L10-FeNi film by PLD from the standpoint of surface thermodynamics.

The PLD system was used to prepare the alloy samples. Target metals were ablated by a Nd:YAG laser with fourth-harmonic generation (wavelength: 266 nm; pulse width: 6 ns; repetition frequency: 10 Hz). In order to deposit a smooth thin film, the laser power was precisely controlled to maintain at 8 ± 0.08 mJ (the corresponding laser power density of the target: 1.5 J/cm2) to avoid the formation of droplets but allow satisfactory ablation of the target. Using a laser splitter via automatic control (LabVIEWTM, National Instruments), the laser power was controlled to remain constant; the deposition rate was stabilized by rotating and swinging the target metal. The typical deposition rate was 74.2, 70.6, and 78.1 s per monolayer (s/ML) for Fe, Ni, and Cu, respectively. During the sample preparation, an Fe seed layer (thickness: 1 nm) and a Au layer (thickness: 20 nm) were deposited at 70 °C on a MgO (100) substrate. A 50 nm Cu layer was then deposited at 300 °C. The temperature, thickness, and smoothness of the layers and the substrate were examined by measurements with an atomic force microscope (NaioAFM, Nanosurf) in the dynamic mode. Finally, 25 sets of Fe/Ni bilayers [each bilayer consisted of one Fe monolayer (ML) and one Ni ML] were deposited by the alternating monoatomic deposition technique on the Cu buffer layer [Fig. 1(a)]. The growth temperature (Ts) was varied from room temperature (RT) to 400 °C, and the structural properties and magnetic anisotropy (Ku) were evaluated as a function of Ts.

FIG. 1.

Schematic view of the PLD-grown FeNi film and RHEED patterns. (a) Structure of the multilayered FeNi/Cu/Au/MgO(100) film prepared by PLD. (b)–(e) RHEED patterns obtained on the FeNi surface for the growth temperatures (Ts) of RT, 150 °C, 300 °C, and 400 °C.

FIG. 1.

Schematic view of the PLD-grown FeNi film and RHEED patterns. (a) Structure of the multilayered FeNi/Cu/Au/MgO(100) film prepared by PLD. (b)–(e) RHEED patterns obtained on the FeNi surface for the growth temperatures (Ts) of RT, 150 °C, 300 °C, and 400 °C.

Close modal

RHEED observation was carried out at a beam energy of 20 kV, and the surface structure and roughness of the samples were evaluated in situ. Synchrotron-radiation XRD (SR-XRD) measurements were performed at SPring-8 BL46XU. We utilized grazing incidence to evaluate the superlattice (110) peak and the fundamental (220) peak of FeNi. The incident angle was set to 0.28° to carry out structural analysis of the in-plane geometry. The photon energy was set to 7.11 and 7.50 keV to execute X-ray anomalous scattering (AXS) using the Fe K adsorption edge. Measurements were carried out at RT. We estimated the degree of order (S), the magnetic anisotropy energy (Ku), and the surface roughness (RRMS) of the specimen. The details of evaluation methods are described in the supplementary material.

We obtained RHEED patterns of the FeNi films as a variation of growth temperature [Figs. 1(b)–1(e)]. The incident electron beam was set along the ⟨100⟩ azimuth of the MgO substrate, and we confirmed the flatness of the Cu buffer layer that was prepared in advance. Patterns with sharp streaks were clearly observed for the growth temperature Ts = RT, 150 °C, and 300 °C, as shown in Figs. 1(b)–1(d), respectively. On the other hand, the pattern completely disappeared for Ts = 400 °C, as shown in Fig. 1(e), indicating that the morphology of the deposited film deteriorated as a result of annealing. This suggests that structural transition occurred between 300 and 400 °C.

We used SR-XRD to characterize the crystallographic structure of the FeNi film as a variation of growth temperature (Fig. 2). The superlattice (110) peak of L10-FeNi was clearly observed for the growth temperatures Ts = 150, 300, and 400 °C, as shown in Fig. 2(a). A small superlattice signal remained for the growth temperature of Ts = RT, and it was confirmed by AXS. The estimated degree of order S was 0.25 ± 0.073 (for the growth temperature Ts = 150 °C), 0.38 ± 0.094 (Ts = 300 °C), and 0.25 ± 0.048 (Ts = 400 °C), showing the maximum value for Ts = 300 °C. The fundamental (220) peak of FeNi was clearly observed in the patterns of all specimens, as shown in Fig. 2(b). The in-plane and out-of-plane lattice parameters for Ts= 300 °C were calculated to be 0.358 and 0.353 nm, respectively, which are close to the previously reported values.16–18 As 300 °C is just below the order–disorder transition temperature (320 °C) of the FeNi system,20 these results indicate that the L10-FeNi phase was formed at 300 °C, and the superstructure formation was promoted by thermal treatment. It should also be noted that the optimized growth temperature of PLD was higher than that of MBE (187 °C).

FIG. 2.

XRD patterns measured using synchrotron radiation with a photon energy of 7.5 keV: (a) around the superlattice FeNi(110) peak and (b) around the fundamental FeNi(220) peak.

FIG. 2.

XRD patterns measured using synchrotron radiation with a photon energy of 7.5 keV: (a) around the superlattice FeNi(110) peak and (b) around the fundamental FeNi(220) peak.

Close modal

The magnetic moment and magnetic anisotropy energy were investigated using a SQUID magnetometer at 100 K. Figure 3 shows the obtained magnetization curves for an in-plane and an out-of-plane external magnetic field. The values of Ms were estimated by hysteresis curves as 1050, 1030, 800, and 730 emu/cc, and the values of Ku were estimated using Eq. (S3) (supplementary material) as 1.60 × 105, 3.09 × 105, 1.30 × 106, and 1.45 × 104 erg/cc for the growth temperatures Ts = RT, 150 °C, 300 °C, and 400 °C, respectively. Unfortunately, the easy axis of magnetization was in-plane for all samples, and the obtained Ms was smaller than expected, which can be ascribed to the oxidation of the samples. However, we noted that Ku increased as S increased, and Ku showed the maximum value for the growth temperature Ts = 300 °C. With the exception of the growth temperature, this behavior agrees well with that observed in a previous study.18 It is also remarkable that there was a slight decrease in Ms for 300 °C.

FIG. 3.

Magnetization curves for an in-plane and an out-of-plane magnetic field, obtained for different growth temperatures Ts: (a) RT; (b) 150 °C; (c) 300 °C; and (d) 400 °C. The characterization temperature was 100 K.

FIG. 3.

Magnetization curves for an in-plane and an out-of-plane magnetic field, obtained for different growth temperatures Ts: (a) RT; (b) 150 °C; (c) 300 °C; and (d) 400 °C. The characterization temperature was 100 K.

Close modal

To analyze the detail of the surface structure as a variation of growth temperature, AFM observation was carried out for each specimen in various viewing fields (Fig. 4). We confirmed in advance that a flat surface was obtained on the Cu buffer layer. As shown in Figs. 4(a) and 4(c), rounded and small islands were distributed all over the surface. A typical island had a diameter of about 80 nm. In a large viewing field, a remarkable morphology was observed, as shown in Figs. 4(b) and 4(d). Islands with a square shape and directional facets oriented in the ⟨110⟩ direction appeared, suggesting that the FeNi adsorbent favorably grew in the ⟨110⟩ direction. The islands' shape resembled a typical Wulff construction, and it deteriorated as the temperature increased.22 The surface roughness (RRMS) was calculated to be 13.92 nm at RT and 15.10 nm for the growth temperature Ts = 150 °C.

FIG. 4.

AFM images showing the surface morphology of samples as a variation of the growth temperature: (a) and (b) RT; (c) and (d) 150 °C; (e) and (f) 300 °C; (g) and (h) 400 °C. Images were obtained in small (3 × 3 μm) and large (10 × 10 μm) viewing fields. Square and oriented islands were recognized for Ts = RT and 150 °C; smooth topography was clearly observed for Ts = 300 °C; and the surface morphology completely collapsed for Ts = 400 °C.

FIG. 4.

AFM images showing the surface morphology of samples as a variation of the growth temperature: (a) and (b) RT; (c) and (d) 150 °C; (e) and (f) 300 °C; (g) and (h) 400 °C. Images were obtained in small (3 × 3 μm) and large (10 × 10 μm) viewing fields. Square and oriented islands were recognized for Ts = RT and 150 °C; smooth topography was clearly observed for Ts = 300 °C; and the surface morphology completely collapsed for Ts = 400 °C.

Close modal

Surprisingly, a smooth surface was clearly observed for Ts = 300 °C, and the number of islands drastically decreased [Figs. 4(e) and 4(f)]. The value of RRMS obtained from the AFM image clearly decreased to 5.21 nm. These results suggest that the FeNi film grew in two dimensions, and a very smooth surface was realized by thermal treatment. A slight facet structure could be recognized, however, suggesting that the directional growth of the FeNi film continued for Ts = 300 °C. A small amount of dendritic hollow structures also appeared at this temperature. Since dendritic growth could progress even in three dimensions, the hollow structures might indicate the beginning of inter-diffusion in the underlayer, and it is consistent with the slight decrease in Ms for Ts = 300 °C. The surface morphology for Ts = 400 °C was significantly rough and large steps appeared, as shown in Figs. 4(g) and 4(h). The value of RRMS was 30.24 nm, representing an obvious increase from the value for Ts = 300 °C. This structural collapse simply corresponds to the decrease in S and Ku, as described earlier. This series of AFM images confirms that the smooth and two-dimensional growth was realized at the critical temperature of 300 °C, which is in good agreement with the XRD and SQUID results. It is useful to compare and discuss the microscopic AFM results and macroscopic XRD and SQUID results with support of surface thermodynamics for understanding the growth process of L10-FeNi in the PLD method.

Figure 5 shows a compilation of the structural, magnetic, and morphological properties of the samples as a variation of growth temperature. Both S and Ku exhibited their respective maximum values at 300 °C [Figs. 5(a) and 5(b)], and a smooth surface was realized at the same temperature [Fig. 5(c)]. The microscopic surface smoothness is believed to account for the macroscopic structural and magnetic properties, which is in good agreement with the results of previous studies, except for the critical temperature of 300 °C.17,18

FIG. 5.

Summary of the structural, magnetic, surface properties of a PLD-grown film. (a) Order parameter (S) vs. growth temperature (Ts). (b) Magnetic anisotropy energy (Ku) vs. Ts. (c) Surface roughness (RRMS) vs. Ts. (d) Proposed growth scenario when the PLD technique is used.

FIG. 5.

Summary of the structural, magnetic, surface properties of a PLD-grown film. (a) Order parameter (S) vs. growth temperature (Ts). (b) Magnetic anisotropy energy (Ku) vs. Ts. (c) Surface roughness (RRMS) vs. Ts. (d) Proposed growth scenario when the PLD technique is used.

Close modal

Next, let us qualitatively consider how the smooth surface was formed based on simple surface thermodynamics22–24 [Fig. 5(d)]. Originally, the surface morphology was determined to minimize the surface free energy (G), which was characterized by the enthalpy and island area, and its shape was given according to the following equation:23 

ΔG*=γoϕo/αo,
(1)

where γo is the surface free enthalpy, ϕo is the surface area of the island, and αo is the anisotropic term depending on the growth direction. Although all terms depended on the temperature, a smooth surface and a L10 structure were experimentally realized at 300 °C; so, we considered two-dimensional growth for simplicity. The anisotropy term αo is expressed as a linear combination of the nearest and next-nearest neighbor interactions by Wulff's construction22 

α01=εb(1)2a2+2εb(2)2a2,
(2)
α11=22εb(1)2a2+22εb(2)2a2,
(3)

where εb(1) is the nearest neighbor term and εb(2) is the next-nearest neighbor term. The difference between α01 and α11 determined the shape anisotropy of the island. The AFM images clearly show directional growth at RT, and it continued at 300 °C. The contribution of α11 should be satisfactorily larger than α10, yielding the relationship εb(1)εb2 and indicating that the nearest-neighbor interaction was significantly larger than the next-nearest neighbor interaction. Such anisotropic behavior suggests that the growth temperature could be much lower than the roughening temperature, and the thermal stability of the constructed island would be significantly high.24 It can also mean that the increase in αo contributed to the decrease in surface free energy, ΔG*, thus allowing a thermally stable island to be constructed.

It is known that the area of the island (ϕo) is proportional to the surface free energy (ΔG*). As reported by Shen et al., the size of an island produced by PLD is significantly smaller than that produced by MBE.19 It is difficult to directly compare our results with those in other reports18,19 because of the differences in resolution, temperature, and substrate. The observed diameter of a PLD-grown island was about 80 nm for the growth temperature Ts = 150 °C, and that of an MBE-grown island was about 100 nm.18 The size effect in the present study is considered to be comparable to that of MBE, and it is assumed that the contribution to ΔG* might be small.

A decisive factor of island growth is the nucleation density during PLD. Shen et al. noted that the nucleation density and interlayer mass transport characterize the island growth, and they are experimentally related to the instantaneous deposition rate and kinetic energy. The instantaneous deposition rate of PLD can be estimated from the average deposition rate (∼70 s/ML), pulse width (6 ns), and laser repetition (10 Hz). The instantaneous deposition rate of PLD is about 1 × 104 ML/min during a very short time of 6 ns, and it leads to a small critical nucleus with a higher density on the surface. During the much longer waiting period of 100 ms, the nucleus remained almost unchanged until the next ablation event. This resulted in the formation of a large number of small islands on the surface, as shown in Figs. 4(a) and 4(c). The kinetic energy is also a remarkable factor for characterization of the surface morphology. In our experiment, the ablating laser power was controlled to remain as small as possible (8 mJ), and it automatically stabilized. The thermal contribution by annealing was included in the Maxwell–Boltzmann distribution, but the PLD-grown islands were observed to show clear directional facets when compared to MBE-grown islands.18 Therefore, the kinetic energy of the adatoms might have been suppressed. Shen et al. also suggested that even if the kinetic energy increases, it is still difficult to exceed the step-edge-barrier (Ehrlich–Schwoebel barrier), which makes it unfavorable for adatoms to hop to the top of an island.19 This would offer a possible explanation for the thermally stable islands and the high critical temperature of the PLD-grown films in our study.

For the above reasons, we propose the following qualitative mechanism for film growth by PLD. A high-density nucleus is formed on the surface by impulsive laser ablation, after which small islands grow all over the surface. As the deposition progresses, a square and oriented large island is formed. The flat square island grows two-dimensionally until it encounters a neighboring island, and a smooth surface is eventually formed. The surface free energy, shape of the island, and impulsive deposition could all contribute to the surface morphology, and they account for the stable and smooth surface produced by the PLD technique. Such PLD growth behavior is significantly different from that reported in previous studies, and PLD is expected to be an advanced method for fabricating the L10-FeNi system.

In summary, we demonstrated the fabrication of the L10-FeNi phase by PLD, where an automatically controlled YAG laser was used for alternating monoatomic deposition of Fe and Ni on a Cu(001) substrate. We characterized the magnetic properties, lattice structure, and surface morphology of the samples as the growth temperature was varied. Our results show that the L10-FeNi structure was adequately fabricated, and Ku shows remarkable correspondence with S and the surface smoothness. The formation of L10-FeNi was most promoted at 300 °C, which is significantly higher than the growth temperatures reported in previous works. It is suggested that PLD and thermal treatment promoted two-dimensional growth of the adsorbent, FeNi. The growth mode of PLD was explained by simple surface thermodynamics, and the impulsive deposition and the low kinetic energy were shown to play key roles in the construction of a smooth surface. Based on our findings, PLD could be an advanced method for the fabrication of L10-FeNi.

See supplementary material for the estimation methods of the order parameter S and the magnetic anisotropy energy Ku.

This work was partly supported by KAKENHI, JSPS [Scientific Research (B): 16H03873; Young Scientists (A): 24684029] and the Inter-University Cooperative Research Program of Institute for Materials Research, Tohoku University (Proposal Nos. 16K0103, 17K0007, and 18K0050). The SR-XPD measurements were performed at BL46XU of SPring-8 (Proposal Nos. 2012A1076, 2012B1214, 2013A1366, 2013B1222, 2014A1280, 2014B1503, 2015A1516, 2015B1304, 2016A1156, 2016B1185, 2017A1078, 2017A1600, 2017A1784, 2017B1588, and 2017B1814). The magnetization measurements by SQUID were carried out under the Visiting Researcher's Program of the Institute for Solid State Physics, University of Tokyo. The authors also thank Dr. T. Kojima (IMRAM, Tohoku University) for our valuable discussions on estimation of the order parameter.

1.
D.
Weller
,
A.
Moser
,
L.
Folks
,
M. E.
Best
,
W.
Lee
,
M. F.
Toney
,
M.
Schwickert
,
J.-U.
Thiele
, and
M. F.
Doerner
,
IEEE Trans. Magn.
36
,
10
(
2000
).
2.
S.
Okamoto
,
N.
Kikuchi
,
O.
Kitakami
,
T.
Miyazaki
, and
Y.
Shimada
,
Phys. Rev. B
66
,
024413
(
2002
).
3.
A.
Moser
,
K.
Takano
,
D. T.
Margulies
,
M.
Albrecht
,
Y.
Sonobe
,
Y.
Ikeda
,
S.
Sun
, and
E. E.
Fullerton
,
J. Phys. D: Appl. Phys.
35
,
R157
(
2002
).
4.
T.
Shima
,
T.
Moriguchi
,
S.
Mitani
, and
K.
Takanashi
,
Appl. Phys. Lett.
80
,
288
(
2002
).
5.
K.
Takanashi
,
S.
Mitani
,
M.
Sano
,
H.
Fujimori
,
H.
Nakajima
, and
A.
Osawa
,
Appl. Phys. Lett.
67
,
1016
(
1995
).
6.
M.
Ohtake
,
S.
Ouchi
,
F.
Kirino
, and
M.
Futamoto
,
J. Appl. Phys.
111
,
07A708
(
2012
).
7.
M.
Yoshikawa
,
E.
Kitagawa
,
T.
Nagase
,
T.
Daibou
,
M.
Nagamine
,
K.
Nishiyama
,
T.
Kishi
, and
H.
Yoda
,
IEEE Trans. Magn.
44
,
2573
(
2008
).
8.
N.
Inami
,
G.
Kim
,
T.
Hiratsuka
,
H.
Naganuma
,
M.
Oogane
, and
Y.
Ando
,
J. Phys.: Conf. Ser.
200
,
052008
(
2010
).
9.
T.
Seki
,
S.
Mitani
,
K.
Yakushiji
, and
K.
Takanashi
,
Appl. Phys. Lett.
88
,
172504
(
2006
).
10.
M.
Mizuguchi
,
T.
Kojima
,
M.
Kotsugi
,
T.
Koganezawa
,
K.
Osaka
, and
K.
Takanashi
,
J. Magn. Soc. Jpn.
35
,
370
(
2011
).
11.
M.
Kotsugi
,
C.
Mitsumata
,
H.
Maruyama
,
T.
Wakita
,
T.
Taniuchi
,
K.
Ono
,
M.
Suzuki
,
N.
Kawamura
,
N.
Ishimatsu
,
M.
Oshima
,
Y.
Watanabe
,
M.
Taniguchi
 et al,
Appl. Phys. Express
3
,
013001
(
2010
).
12.
T.
Shima
,
M.
Okamura
,
S.
Mitani
, and
K.
Takanashi
,
J. Magn. Magn. Mater.
310
,
2213
(
2007
).
13.
L.
Néel
,
J.
Pauleve
,
R.
Pauthenet
,
J.
Laugier
, and
D.
Dautreppe
,
J. Appl. Phys.
35
,
873
(
1964
).
14.
Y.
Miura
,
S.
Ozaki
,
Y.
Kuwahara
,
M.
Tsujikawa
,
K.
Abe
, and
M.
Shirai
,
J. Phys.: Condens. Matter
25
,
106005
(
2013
).
15.
M.
Kotsugi
,
M.
Mizuguchi
,
S.
Sekiya
,
M.
Mizumaki
,
T.
Kojima
,
T.
Nakamura
,
H.
Osawa
,
K.
Kodama
,
T.
Ohtsuki
,
T.
Ohkochi
,
K.
Takanashi
,
Y.
Watanabe
 et al,
J. Magn. Magn. Mater.
326
,
235
(
2013
).
16.
M.
Kotsugi
,
H.
Maruyama
,
N.
Ishimatsu
,
N.
Kawamura
,
M.
Suzuki
,
M.
Mizumaki
,
K.
Osaka
,
T.
Matsumoto
,
T.
Ohkochi
,
T.
Ohtsuki
,
T.
Kojima
,
M.
Mizuguchi
,
K.
Takanashi
,
Y.
Watanabe
 et al,
J. Phys.: Condens. Matter
26
,
064206
(
2014
).
17.
T.
Kojima
,
M.
Mizuguchi
,
T.
Koganezawa
,
K.
Osaka
,
M.
Kotsugi
, and
K.
Takanashi
,
Jpn. J. Appl. Phys.
51
,
010204
(
2012
).
18.
T.
Kojima
,
M.
Ogiwara
,
M.
Mizuguchi
,
M.
Kotsugi
,
T.
Koganezawa
,
T.
Ohtsuki
,
T.
Tashiro
, and
K.
Takanashi
,
J. Phys.: Condens. Matter
26
,
064207
(
2014
).
19.
J.
Shen
,
Z.
Gai
, and
J.
Kirschner
,
Surf. Sci. Rep.
52
,
163
(
2004
).
20.
J. F. J.
Bryson
,
N. S.
Church
,
T.
Kasama
, and
R. J.
Harrison
,
Earth Planet. Sci. Lett.
388
,
237
(
2014
).
21.
T.
Kojima
,
M.
Mizuguchi
, and
K.
Takanashi
,
Surf. Sci.
619
,
44
(
2014
).
22.
23.
L. Z.
Mezey
and
J.
Giber
,
Jpn. J. Appl. Phys. Part 1
21
,
1569
(
1982
).
24.
C.
Rottman
and
M.
Wortis
,
Phys. Rev. B
29
,
328
(
1984
).

Supplementary Material