Fe16N2 thin films possess high saturation magnetization (Ms) and high magnetic crystalline anisotropy (Ku) simultaneously. For application in magnetic recording and electrical machine, soft magnetic materials with high Ms and low Ku are desirable. In this report, carbon was doped into FeN thin films. Fe-CN martensite thin films, also named “Minnealloy,” were fabricated by a facing target sputtering system. The Fe-CN martensite phase was detected by X-ray diffraction (XRD). Vibrating sample magnetometry (VSM) measurements confirmed the high Ms property even though ordered Fe-CN was not formed. An X-ray photoelectron spectroscopy study was carried out to reveal the carbon and nitrogen electronic environment. The Ku value was obtained from the M-H loop and the law to approach saturation with reasonable consistency around 1–2 × 106 erg/cc, which is about five to ten times smaller than that of the ordered Fe16N2 thin film (1–2 × 107 erg/cc). The combination of high Ms and low saturation fields makes the Fe-CN martensite thin film a potential candidate for the magnetic write head and other applications.

The areal density of the hard drive increases with a smaller recording unit, which subsequently causes the thermal instability.1,2 The enhancement of crystalline anisotropy, e.g., using FePt media,3 stabilizes the grain but also requires a stronger writing field to switch the grains between up- and down-perpendicular magnetic states. However, the field produced by the magnetic writer is strictly restricted by the saturation magnetization of the writer material. The discovery4 and the recent experimental5 and theoretical validation6 of the high Ms property makes the Fe16N2 thin film a very promising candidate for replacing the FeCo alloy, which is considered to be the highest Ms material7 according to the Slate-Pauling curve.8 One issue about using Fe16N2 for a magnetic writer is its high crystalline anisotropy,1–5,9 which makes it harder to switch between up- and down-magnetization states. It is also noted that the high Ms soft magnetic material is very desirable in electric machines such as generators, transformers, inductors and wireless chargers. Investigation of new soft magnetic materials with a combination of high Ms and low coercivity could enable reduced machine size and energy efficiency.10 In this report, we use carbon as a dopant to engineer and lower the effective magnetic anisotropy of FeN thin films. Fe-CN martensite thin films, which were named Minnealloy,10 were obtained with low Ku and high Ms with optimum carbon and nitrogen concentrations in the film.

A facing-target sputtering system was used to prepare the films.5 To prepare the samples, a MgO(001) single crystal substrate was heated up to 250 °C for 2 h in vacuum. The base pressure of the sputtering chamber was 5 × 10−8 Torr. 15 nm of a Fe seed layer was grown first on the MgO substrate to induce the (001) texture. After cooling down to room temperature, the FeC target with 5 at. % carbon was sputtered in the Ar + N2 mixture gas atmosphere to obtain the Fe-NC martensite layer. Different nitrogen concentrations were obtained by changing the N2 partial pressure in the sputtering process. The film thickness was chosen to grow epitaxially with the MgO/Fe(001) texture to promote the Fe-CN growth. It also varies under different N2 partial pressures. Subsequent annealing was carried out in vacuum for various hours (5–36 h) at 120 °C, 150 °C, and 180 °C, respectively.

To characterize the thickness of the prepared films, X-ray reflectivity measurements were conducted. Figure 1(a) shows the reflectivity curve of the Fe-CN martensite thin film. Fitting the reflectivity curve yields the scattering length density (SLD) depth profile of the film in Fig. 1(b). The total thickness of the film is around 52 nm. The Fe-CN layer has a higher electron SLD compared to that of iron, and it remains constant in the z direction. This indicates the difference in the composition between Fe and the Fe-CN layer. The clear oscillation at the high angle indicates that the film has a relatively smooth texture. Also, in order to fit the reflectivity curve, an oxidation layer around 3 nm is needed and can be seen from Fig. 1(b) on top of the sample.

FIG. 1.

(a) X-ray reflectivity curve of the Fe-CN martensite thin film. (b) Scattering length density (SLD) of the film in the out-of-plane direction.

FIG. 1.

(a) X-ray reflectivity curve of the Fe-CN martensite thin film. (b) Scattering length density (SLD) of the film in the out-of-plane direction.

Close modal

X-ray diffraction (XRD) with the out-of-plane configuration is shown in Fig. 2. The Fe (002) peak can be observed, which indicated the (001) texture of the film. When no nitrogen was present in the FeC layer, a weak peak shows up on the left shoulder of the Fe (002) peak. Also, there is an obvious peak on the right shoulder of the Fe (002) peak, which indicates that the FeC layer contains grains whose lattice constant c is smaller than Fe. This shoulder does not exist in the Fe/MgO structure and is only found in FeC/MgO. In our epitaxially-grown thin films, it is common to experience strain/stress, but it usually collectively affects the whole layer rather than just some grains. However, one can find that Fe (002) still resides at 65°, so we exclude the possibility of the strain/stress effect. These observations indicate that carbon does not reside in the grain boundary of the thin film. XRD should not give any signal if carbon is totally diffused into the grain boundaries or in the amorphous state. With the introduction of nitrogen atoms in the system, both peaks shift to the left. The right shoulder peak merges with the Fe (002) peak at low concentrations (0.02 mTorr N2 partial pressure ) and the left shoulder eventually forms a Fe-NC martensite Fe8(N1-xCx) (002) peak. When the nitrogen pressure reached an optimum (0.08 mTorr N2 partial pressure), the lattice constant c becomes the nominal value for Fe16N2, indicating the formation of a Fe-CN martensite structure. Therefore, nitrogen atoms play an important role in lattice expansion in the c direction. Since carbon is already present in the film and it is located interstitially in the iron lattice, the required nitrogen partial pressure for stoichiometric Fe-CN martensite is lower than that of Fe16N2 (0.12–0.15 mTorr).11 

FIG. 2.

(a) The XRD pattern of the Fe-CN film under various nitrogen partial pressures during sputtering. The Fe-CN (002) peak progressively shifts to the left with increasing amount of nitrogen. (b) The lattice constant c calculated with respect to nitrogen partial pressure in (a).

FIG. 2.

(a) The XRD pattern of the Fe-CN film under various nitrogen partial pressures during sputtering. The Fe-CN (002) peak progressively shifts to the left with increasing amount of nitrogen. (b) The lattice constant c calculated with respect to nitrogen partial pressure in (a).

Close modal

After the sample was annealed, we did not find the ordered phase Fe16(N1-xCx)2 that corresponds to α′′-Fe16N2 at different annealing temperatures and for different annealing duration times. This may indicate that carbon and nitrogen atoms become resistant to diffuse in the Fe-CN layer after they formed a Fe-CN martensite phase. There might be an effect of carbon clustering in iron carbon martensite, which was first observed in Fe-C martensite by carburizing the iron foil.12 This phenomenon is also observed in the high resolution SEM image under a dark field with diffuse scattering.13 In this case, the short range ordering of carbon atoms in the octahedral interstitials takes place, which leads to an inhomogeneous distribution of carbon atoms in the Fe-CN layer. However, the clustering effect is not present in the Fe-N thin film as an ordered Fe16N2 phase can be obtained post-annealing. The diffusion of nitrogen atoms can possibly be more rapid than that of carbon, even though their activation energies are about the same and leads to an absence of significant clustering of nitrogen on aging.14 Since the clustering of carbon atoms happens at the early stage of tempering, it will subsequently prevent the diffusion of carbon and nitrogen into the ordered Fe-CN phase interstitials, and thus no ordered phase was detected from the XRD pattern.

X-ray Photoelectron Spectroscopy (XPS) was used to measure the chemical environment of carbon and nitrogen atoms as interstitial dopants. XPS utilizes the high energy output X-ray to bombard the sample surface to excite the electrons of various elements. Depending on the binding energy of each electron, they are collected in different energy ranges with its kinetic energy measured. Different chemical environments of carbon and nitrogen will yield different binding energies of the C 1s and N 1s electrons. In this experiment, Surface Science SSX-100 was used with bent quartz crystal monochromated Al Kα X-rays. Figure 3(a) shows the binding energy of the C1s electron in the Fe-CN martensite layer. It is obvious that the C 1s electron contains more than one electronic state. Decomposing the C 1s region yields three peaks. The major peak occurs at 248.6 eV, which describes the interaction between C atoms. The peak around 283 eV indicates the iron carbides existing in the system, which corresponds to the martensite dominant phase in the Fe-CN layer. The peak around 288.2 eV shows that carbonate also exists due to the oxygen impurity in the film. The N 1s electron binding energy is around 397.6 eV [Fig. 3(b)], which is a little smaller than that of Fe-N martensite (398 eV),15 which is mainly due to the addition of carbon in the system. These observations show that the Fe-CN martensite phase is formed in the sample and oxygen contributes to the electronic structure of the Fe-CN martensite phase. Auger Electron Spectroscopy (AES) was conducted to examine the elemental content of the film, and the result in Fig. 3(c) shows that the Fe-CN martensite thin film contains oxygen, including the seed layer of Fe. Therefore, the carbonate bond exists for the C 1s electron.

FIG. 3.

The XPS result of carbon and nitrogen and the AES result of the Fe-CN martensite thin film: (a) Carbon 1s electron binding energy. Three peaks can be extracted from the raw data, which indicates iron carbonate, sp3 carbon and iron carbide. (b) Nitrogen 1s electron binding energy. (c) Depth profile of element concentration throughout the whole film.

FIG. 3.

The XPS result of carbon and nitrogen and the AES result of the Fe-CN martensite thin film: (a) Carbon 1s electron binding energy. Three peaks can be extracted from the raw data, which indicates iron carbonate, sp3 carbon and iron carbide. (b) Nitrogen 1s electron binding energy. (c) Depth profile of element concentration throughout the whole film.

Close modal

To characterize the magnetic properties of the films, VSM was used to measure the M-H loops of the samples under various nitrogen partial pressures with an external field applied in the film plane (Fig. 4). Figure 4(a) shows hysteresis loops for the samples prepared at a low nitrogen partial pressure. All samples show soft magnetic properties with a relatively low coercivity (15–40 Oe) and a low saturation field (300–500 Oe) as well as a low saturation magnetization (1.95–2.1 T). With the increase in the nitrogen partial pressure above 0.08 mTorr, the M-H loop as shown in Fig. 4(b) changes its shape with a higher 4πMs (∼2.4 T) and an increased saturation field (∼2000 Oe). This saturation field is still significantly lower than that of the FeN thin film (5000–7000 Oe).16 This indicates that the perpendicular magnetic anisotropy of the Fe-CN film is lower than of the FeN thin film. It can be noticed that the rotation of magnetically hard grains exists when the film approaches saturation, which is a good indication of the perpendicular easy-axis component of the Fe-CN layer. At lower nitrogen partial pressures, the Fe-CN martensite phase was not formed, and thus the whole film exhibited pure Fe-like behavior.

FIG. 4.

M-H loops of Fe-CN martensite films with the external field applied in plane: (a) Low nitrogen partial pressure sample exhibiting very soft behavior with low Ms (1.9–2.1 T). The Ms of nitrogen partial pressure at 0.072 mTorr starts to increase and the hysteresis can be seen. (b) At higher nitrogen partial pressures, 4πMs reaches 2.4 T and the M-H loop shows grain rotation when its moment approaches saturation.

FIG. 4.

M-H loops of Fe-CN martensite films with the external field applied in plane: (a) Low nitrogen partial pressure sample exhibiting very soft behavior with low Ms (1.9–2.1 T). The Ms of nitrogen partial pressure at 0.072 mTorr starts to increase and the hysteresis can be seen. (b) At higher nitrogen partial pressures, 4πMs reaches 2.4 T and the M-H loop shows grain rotation when its moment approaches saturation.

Close modal

In the Fe16N2 thin film, the tetragonality of the Fe16N2 lattice leads to high crystalline anisotropy. Therefore, the impact of the c/a ratio was investigated with various nitrogen concentrations in the Fe-CN thin film. In Fig. 4(a), the Fe-C thin film with zero amount of nitrogen in the film shows in-plane magnetic anisotropy. This result coincides with the estimation of the Fe-C thin film that the magnetic anisotropy constant Ku value for Fe-C is negative (in the base plane or perpendicular to the c axis).17 The lattice constant c progressively increases with the amount of nitrogen concentration as shown in Fig. 2(b), which increases the anisotropy. However, it is also notable that without nitrogen, the Fe-C thin film only possesses a low Ms as shown in Fig. 4(a). Therefore, an optimized condition should be satisfied to achieve both high Ms and low anisotropies. With a minimum amount of nitrogen flow in the gas chamber during sputtering, the Fe-CN layer does not possess high Ms, and its hysteresis loop also behaves like α-Fe, which indicates that the solubility of nitrogen and carbon does not reach the stoichiometric Fe-CN martensite value. Therefore, the film has properties of soft in-plane anisotropy and low Ms. With the increase of nitrogen flow, the saturation magnetization field (Hs) and the saturation magnetization (Ms) increase simultaneously and eventually reach an optimum situation where the Fe-CN martensite phase is formed.

The magnetic anisotropy constant of the Fe-CN thin film can be estimated from the in-plane and out-of-plane M-H loops. The out-of-plane loop is dominated by the strong demagnetizing field of the film, which corresponds to a typical hard axis loop scan with little hysteresis. The crossing point of in-plane and out-of-plane loops yields the anisotropy field Hk of the whole film which includes the perpendicular crystalline anisotropy of the Fe-CN layer. This effective anisotropy energy is defined as the difference between the parallel and perpendicular magnetization energy densities. Therefore, the anisotropy constant (Ku) of the Fe-CN layer can be obtained by

Ku=2πMs2HkMs/2.

For our sample, Ms and Hk can be obtained from the M-H loop as shown in Fig. 5. The estimated Ku value for this Fe-CN martensite thin film is 2.4 × 106 erg/cc, which is about four to five times lower than that of the Fe16N2 thin film (1 × 107 erg/cc).

FIG. 5.

In-plane and out-of-plane M-H loops of the Fe-CN martensite thin film. The perpendicular anisotropy of the Fe-CN layer is dominated by the large shape anisotropy. Anisotropy constant can be obtained from the difference between shape anisotropy and effective anisotropy, as determined from the anisotropy field at the crossing point of in-plane and out-of-plane loops.

FIG. 5.

In-plane and out-of-plane M-H loops of the Fe-CN martensite thin film. The perpendicular anisotropy of the Fe-CN layer is dominated by the large shape anisotropy. Anisotropy constant can be obtained from the difference between shape anisotropy and effective anisotropy, as determined from the anisotropy field at the crossing point of in-plane and out-of-plane loops.

Close modal

Another way to obtain the Ku value is to use the Law to Approach Saturation (LAS), which describes the behavior of magnetic moment when it is approaching the saturation as follows:18–20 

M(H)/Ms=1βK2/Ms2H2.

Considering the Fe-CN layer which possesses the perpendicular anisotropy, the rotation of magnetically hard grains happens when the field is applied in the film plane. Therefore, this formula can be applied when the Fe-CN perpendicular grains rotate when its approaching saturation is around 2000 Oe. To satisfy the condition for LAS, the sample is demagnetized first at a field strength of 15 kOe and gradually decreased to zero in positive and negative directions alternatively in the film plane in Fig. 6(a). After demagnetization, the initial magnetization curve was obtained as shown in Fig. 6(b). To better visualize the saturation magnetization field (Hs) for the sample, M vs dM/dH is plotted and Hs is labeled in Fig. 6(c), where the dM/dH value reaches zero. For a uniaxial material, β = 4/15.21 Figure 6(d) shows the M(H)/Ms vs 1/H2 plot. The negative slope of the linear fitting of the experiment data gives the anisotropy constant of the Fe-CN layer. The obtained value for Ku is 1 × 106 erg/cc.

FIG. 6.

(a) Demagnetized Fe-CN martensite thin film M-H loop. (b) Fe-CN layer approaching saturation. (c) Derivative of M over H, where the saturation field of the Fe-CN layer is determined. (d) Fitting M/Ms vs 1/H2 yields the anisotropy constant of the Fe-CN layer.

FIG. 6.

(a) Demagnetized Fe-CN martensite thin film M-H loop. (b) Fe-CN layer approaching saturation. (c) Derivative of M over H, where the saturation field of the Fe-CN layer is determined. (d) Fitting M/Ms vs 1/H2 yields the anisotropy constant of the Fe-CN layer.

Close modal

Although both carbon and nitrogen atoms occupy the interstitials of the Fe lattice in the c direction and expand the lattice, they do not seem to influence the anisotropy in the same manner. According to Takahashi, the crystalline anisotropy of α′-FeC does not even change with respect to carbon concentration. The difference of one valence electron makes carbon more “isotropic” when it is sitting in the interstitials, which leads to a situation that localization of electrons is not likely to happen compared to nitrogen. In our study, since the primary interstitial element of the Fe-CN thin film with high Ms and low anisotropy is nitrogen, it shows that the addition of carbon in the film reduces the anisotropy faster than saturation magnetization.

In summary, we have prepared textured Fe-CN martensite thin films by a facing-target sputtering system. The addition of carbon lowers the magnetic anisotropy in the Fe-N films. The saturation magnetization field of the Fe-CN thin films is greatly reduced, and its anisotropy is calculated using M-H loops, which is one order of magnitude smaller than that of the Fe16N2 thin film. Also, the Ms of the Fe-CN layer is reasonably high. Therefore, this provides a good opportunity for Fe-CN martensite to be used in a magnetic writer and other possible applications. In applications where a substantial amount of Fe-CN is needed, bulk synthesis10,22,23 can be further investigated and applied instead.

This work was partially supported by Seagate Technology, Western Digital and ARPA-E (Advanced Research Projects Agency-Energy) project under Contract No. 0472-1595. Parts of this work were carried out in the Characterization Facility through the NSF MRSEC Program at University of Minnesota. The authors thank the useful discussion with Dr. Yanfeng Jiang.

1.
S. H.
Charap
,
P.-L.
Lu
, and
Y.
He
,
IEEE Trans. Magn.
33
,
978
(
1997
).
2.
D.
Weller
and
A.
Moser
,
IEEE Trans. Magn.
35
,
4423
(
1999
).
3.
T.
Suzuki
,
K.
Harada
,
N.
Honda
, and
K.
Ouchi
,
J. Magn. Magn. Mater.
193
,
85
(
1999
).
4.
T. K.
Kim
and
M.
Takahashi
,
Appl. Phys. Lett.
20
,
492
(
1972
).
5.
J.-P.
Wang
,
N.
Ji
,
X.
Liu
,
Y.
Xu
,
C.
Sanchez-Hanke
,
Y.
Wu
,
F. M. F. D.
Groot
,
L. F.
Allard
, and
E.
Lara-Curzio
,
IEEE Trans. Magn.
48
,
1710
(
2012
).
6.
N.
Ji
,
X.
Liu
, and
J. P.
Wang
,
New J. Phys.
12
,
6
(
2010
).
7.
R. S.
Sundar
and
S. C.
Deevi
,
Int. Mater. Rev.
50
,
157
(
2005
).
8.
R. M.
Bozorth
,
Ferromagnetism
849
(
1951
).
9.
H.
Takahashi
,
K.
Mitsuoka
,
M.
Komuro
, and
Y.
Sugita
,
J. Appl. Phys.
73
,
6060
(
1993
).
10.
M.
Mehedi
,
Y.
Jiang
,
P. K.
Suri
,
D. J.
Flannigan
, and
J.-P.
Wang
,
J. Phys. D: Appl. Phys.
50
,
37LT01
(
2017
).
11.
N.
Ji
,
Y.
Wu
, and
J. P.
Wang
,
J. Appl. Phys.
109
,
07B767
(
2011
).
12.
J. M.
Genin
and
P. A.
Flinn
,
Phys. Lett.
22
,
392
(
1966
).
13.
S.
Nagakura
,
K.
Shiraishi
, and
Y.
Hirotsu
,
Trans. Jpn. Inst. Met.
16
,
601
(
1975
).
14.
E. J.
Mittemeher
,
L.
Cheng
,
P. J.
der Schaaf
,
C. M.
Brakman
, and
B. M.
Korevaar
,
Metall. Trans. A
19
,
925
(
1988
).
15.
Y.
Jiang
,
X.
Zhang
,
A.
Al Mehedi
,
M.
Yang
, and
J.-P.
Wang
,
Mater. Res. Express
2
,
116103
(
2015
).
16.
N.
Ji
,
M. S.
Osofsky
,
V.
Lauter
,
L. F.
Allard
,
X.
Li
,
K. L.
Jensen
,
H.
Ambaye
,
E.
Lara-Curzio
, and
J. P.
Wang
,
Phys. Rev. B-Condens. Matter Mater. Phys.
84
,
245310
(
2011
).
17.
M.
Takahashi
,
Y.
Takahashi
,
K.
Sunaga
, and
H.
Shoji
,
J. Magn. Magn. Mater.
239
,
479
(
2002
).
18.
B. D.
Cullity
and
C. D.
Graham
,
Introduction to Magnetic Materials
(
John Wiley & Sons
,
2011
).
19.
J. P.
Wang
,
D. H.
Han
,
H. L.
Luo
,
Q. X.
Lu
, and
Y. W.
Sun
,
Appl. Phys. A: Mater. Sci. Process.
61
,
407
(
1995
).
20.
Z. L.
Zhao
,
J. P.
Wang
,
J. S.
Chen
, and
J.
Ding
,
Appl. Phys. Lett.
81
,
3612
(
2002
).
21.
J. F.
Herbst
and
F. E.
Pinkerton
,
Phys. Rev. B
57
,
10733
(
1998
).
22.
K. H.
Jack
,
Proc. R. Soc. London A. Math. Phys. Sci.
41
,
41
55
(
1948
).
23.
L.
Cheng
,
A.
Böttger
, and
E. J.
Mittemeijer
,
Metall. Trans. A
23
,
1129
(
1992
).