Dysprosium substituted Ce:YIG thin films
with perpendicular magnetic anisotropy for
silicon integrated optical isolator applications
The MIT Faculty has made this article openly available.
Please share
how this access benefits you. Your story matters.
Citation
Zhang, Yan et al. "Dysprosium substituted Ce:YIG thin films with
perpendicular magnetic anisotropy for silicon integrated optical
isolator applications." APL Materials 7, 8 (August 2019): 081119 ©
2019 Author(s)
As Published
http://dx.doi.org/10.1063/1.5112827
Publisher
AIP Publishing
Version
Final published version
Citable link
https://hdl.handle.net/1721.1/129490
Terms of Use
Creative Commons Attribution 4.0 International license
for silicon integrated optical isolator
applications
Cite as: APL Mater. 7, 081119 (2019); https://doi.org/10.1063/1.5112827
Submitted: 04 June 2019 . Accepted: 05 August 2019 . Published Online: 22 August 2019
Yan Zhang, Qingyang Du, Chuangtang Wang, Wei Yan, Longjiang Deng, Juejun Hu, Caroline A. Ross, and Lei Bi
ARTICLES YOU MAY BE INTERESTED IN
Tunable perpendicular magnetic anisotropy in epitaxial Y3Fe5O12 films
APL Materials
7, 041104 (2019); https://doi.org/10.1063/1.5090292
Perpendicular magnetic anisotropy and spin mixing conductance in polycrystalline
europium iron garnet thin films
Applied Physics Letters
114, 052403 (2019); https://doi.org/10.1063/1.5074166
Systematic control of strain-induced perpendicular magnetic anisotropy in epitaxial
europium and terbium iron garnet thin films
APL Materials
ARTICLE scitation.org/journal/apmDysprosium substituted Ce:YIG thin films
with perpendicular magnetic anisotropy
for silicon integrated optical
isolator applications
Cite as: APL Mater. 7, 081119 (2019);doi: 10.1063/1.5112827 Submitted: 4 June 2019 • Accepted: 5 August 2019 • Published Online: 22 August 2019
Yan Zhang,1 Qingyang Du,2 Chuangtang Wang,1 Wei Yan,1 Longjiang Deng,1 Juejun Hu,2 Caroline A. Ross,2
and Lei Bi1,a)
AFFILIATIONS
1National Engineering Research Center of Electromagnetic Radiation Control Materials, University of Electronic Science and Technology of China, Chengdu 610054, People’s Republic of China
2Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA
a)Author to whom correspondence should be addressed:bilei@uestc.edu.cn
ABSTRACT
In this report, dysprosium substituted Ce1Y2Fe5O12 (Ce:YIG) thin films (Dy:CeYIG) with perpendicular magnetic anisotropy (PMA) are successfully deposited on silicon and silicon-on-insulator waveguides by pulsed laser deposition. The structural, magnetic, and magneto-optical properties of Dy:CeYIG films are investigated. We find that increasing dysprosium concentration leads to a decreased out-of-plane magnetic saturation field. Dy2Ce1Fe5O12and Dy2Ce1Al0.4Fe4.6O12thin films show PMA dominated by magnetoelastic effects due to thermal mismatch strain. These films further exhibit high Faraday rotation and low optical loss. Our work demonstrates that Dy substitution is an effective way to induce PMA in Ce:YIG thin films without compromising their magneto-optical figure of merit, making this material promising for self-biased transverse electric mode optical isolator applications.
© 2019 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).https://doi.org/10.1063/1.5112827., s
I. INTRODUCTION
Integrated optical isolators or circulators are essential compo-nents of photonic integrated circuits (PICs) and optical communi-cation systems.1,2 Optical isolators allow one-way transmission of light, protecting sensitive components such as lasers and amplifiers from harmful reflections. Optical isolation can be achieved by break-ing the time-reversal symmetry of light propagation usbreak-ing nonrecip-rocal properties of magneto-optical materials. At the telecommuni-cation wavelengths, cerium doped yttrium iron garnet (Ce:YIG) is one of the best magneto-optical (MO) materials because of its large magneto-optical activity and low optical loss, which produces an exceptional magneto-optical figure of merit (FOM, defined as the ratio of Faraday rotation (FR) to optical loss).3The optical isolator
structure based on magneto-optical nonreciprocal phase shift (NRPS) effects was demonstrated using epitaxial Ce:YIG films as the guiding layer,4by bonding epitaxial Ce:YIG films onto semicon-ductor waveguides5–9or by monolithically depositing polycrystalline Ce:YIG with YIG seed layers on silicon photonic devices.2,10–12 Recent advances have led to waveguide optical isolators with per-formances approaching that of bulk optical isolators,11,12qualifying them as viable isolation solutions for PIC applications.
However, the vast majority of the waveguide magneto-optical isolators reported so far operate only for the transverse mag-netic (TM) polarization.13As most semiconductor lasers emit into the transverse electric (TE) mode, TE mode isolators are essen-tial for many applications. To achieve TE mode isolators without polarization rotators,9,14,15 the magneto-optical material must be
APL Mater. 7, 081119 (2019); doi: 10.1063/1.5112827 7, 081119-1
generate a NRPS. One approach to create such in-plane (IP) asym-metry is to use a magneto-optical rib waveguide made of a gar-net with an out-of-plane (OP) easy axis of maggar-netization, where a domain wall is placed at the center of the rib. A nonreciprocal TE-mode isolator based on the approach was theoretically analyzed16 and later experimentally demonstrated.17A TE-mode NRPS can also be produced by placing a magneto-optical cladding on one side of a waveguide. For example, a TE-mode Mach-Zehnder Interferome-ter (MZI) isolator was fabricated in a Si:H waveguide with a Ce:YIG garnet bonded on one sidewall. The device shows a maximum isola-tion of 17.9 dB at a wavelength of 1561.7 nm, although it requires an out-of-plane magnetic field to orient the magnetization of the garnet.18In our previous work, we experimentally demonstrated a monolithically integrated TE mode isolator with a maximum iso-lation of 30 dB, again necessitating an out-of-plane magnetic field to bias the garnet film.12For TE isolators based on the asymmet-ric magneto-optical cladding configuration, it is desirable to develop a magneto-optical material which exhibits perpendicular magnetic anisotropy (PMA) to facilitate operation with a low applied mag-netic field or even self-biased operation. Meanwhile, the material must exhibit a high magneto-optical FOM to ensure low-loss optical isolation.
Ce:YIG thin films grown on garnet or silicon substrates show an in-plane easy axis of magnetization and require magnetic fields as high as ∼2 kOe to saturate the film in the out-of-plane (OP) direction.10 For Ce:YIG thin films deposited on silicon substrates, the large difference in thermal expansion coefficients induces in-plane tensile strains in Ce:YIG and magnetic anisotropy due to magnetoelastic effects. Due to the positive magnetoelastic coeffi-cient of Ce:YIG, the tensile strain favors in-plane magnetization and increases the OP saturation field.10,19,20 On the other hand,
PMA may be achieved in a substituted garnet with negative mag-netostriction grown on Si. This has been observed in Bi-substituted Dy3Fe5O12 (DyIG) thin films whose OP hysteresis loops exhibit a squareness ratio (remanence/saturation, Mr/Ms) of 121,22 and a
tunable coercivity by incorporating nonmagnetic ions.23However, optical loss in the material was not reported. The negative magne-tostriction coefficient of DyIG suggests that Dy substitution may compensate or even overcome the positive magnetoelastic response of Ce ions in Ce:YIG films,24leading to PMA DyCeYIG thin films with large magneto-optical coefficients essential for TE mode device integration.
In this paper, we report the fabrication of Dy-substituted Ce:YIG films with PMA on silicon via a single-step growth method. We also investigated Al substitution on the Fe sites, which low-ers the shape anisotropy and further promotes PMA.24The struc-tural, magnetic, and optical properties of the films were systemat-ically characterized with varying Dy concentrations as detailed in Secs.II–IV.
II. EXPERIMENTAL METHODS
74 nm thick Dy:CeYIG thin films with a 42 nm thick YIG top seed layer were grown on (100) Si and silicon-on-insulator (SOI) substrates by pulsed laser deposition (PLD). According to our prior study, the YIG seed layer crystallizes first and then serves as a tem-plate to promote the crystallization of the substituted Dy:CeYIG
a 3 μm SiO2cladding layer. Targets of Ce1Y2Fe5O12, Ce1Dy2Fe5O12, and Ce1Dy2Al1Fe4O12were ablated by using a 10 Hz, 248 nm Com-pex Pro KrF excimer laser. The laser fluence was 2.5 J/cm2and the distance between the target and the substrate was 55 mm. The sub-strate temperature was 400○C and oxygen pressure was 10 mTorr during the deposition process. The laser pulse ratios were chosen to yield Ce1DyxY2−xFe5O12 films with x = 2.00, 0.87, and 0, and Ce1Dy2Al0.42Fe4.58O12films. The stoichiometries were verified by X-ray photoelectron spectroscopy. The compositions are designated as Dy20, Dy08, Dy00, and Dy20-Al04. The films were subsequently rapid thermal annealed at 850○C at an oxygen gas partial pressure of 2 Torr for 3 min. The crystal structure was characterized by X-ray diffraction (XRD) (Shimadzu XRD-7000 X-X-ray diffractometer) with a Cu-Kαradiation source. Atomic force microscopy (AFM) was used to determine the film surface morphology. Room temperature in-plane (IP) or OP magnetic hysteresis loops were measured by vibrating sample magnetometry (VSM). The MO hysteresis loops were measured by using a custom-built Faraday effect characteri-zation system19with field and light propagation directions out of
plane.
To accurately measure the optical loss, 220 nm thick, 500 nm wide Si channel waveguides were fabricated on SOI wafers by stan-dard planar processes. The Si waveguide structures were first formed using an Elionix ELS-F125 electron beam lithography (EBL) system to expose a negative HSQ (hydrogen silsesquioxane) resist followed by reactive ion etching (RIE) with Cl2gases. A 600 nm thick SiO2 layer was deposited onto the patterned wafer to isolate the optical modes from interacting with the overlaid MO materials except at predefined window regions. The SiO2layer was first formed by spin-ning a 400 nm HSQ layer followed by rapid thermal annealing at 800○C for 5 min, then followed by PECVD of another SiO2layer of 200 nm. The process yields a planarized SiO2top surface facilitating subsequent window opening. Windows with different lengths were patterned into the SiO2cladding layer to expose the Si waveguide surface by RIE with a gas mixture of CHF3and Ar.12Ce:DyIG/YIG
thin films were then deposited on the waveguides. The transmission loss was characterized on a fiber butt coupled waveguide test station. The propagation loss of Ce:DyIG thin films was calculated from the measured propagation loss and simulated confinement factors in Si and Ce:DyIG.
III. FILM STRUCTURE AND MAGNETIC PROPERTIES Figures 1(a)–1(c)show X-ray diffraction spectra for Dy:CeYIG films with a YIG top seed layer. A sharp peak located at 2θ = 32.96○ is attributed to the substrate Si (200). All films show characteristic peaks of polycrystalline garnet thin films for both the YIG seed layer and the Dy:CeYIG layers. The lattice constant of Dy:CeYIG increases with Dy content, from 12.37 Å in Dy00 to 12.41 Å in Dy20. This is because the ionic radius of Y3+ions (1.02 Å) is smaller than that of Dy3+ions (1.083 Å). Substituting Fe3+ions with Al3+decreases the lattice constant given the smaller radius of the Al3+ion (0.535 Å) compared to that of the Fe3+ion (0.645 Å). The lattice constant of the top YIG seed layer is 12.30 Å. A wide range θ−2θ scan inFig. 1(c) confirms that the film consists of a pure garnet phase. The average crystal grain size of the films was estimated using the Scherrer equa-tion to be 51–66 nm. The root mean square (RMS) surface roughness
APL Materials
ARTICLE scitation.org/journal/apmFIG. 1. X-ray diffraction patterns of dysprosium and aluminum substituted Ce:YIG with a top YIG seed layer in a 2θ range from (a) 25○to 40○and (c) 15○to 60○. (b) The
zoomed-in spectrum of the (420) diffraction peaks. AFM images showing (d) height and (e) phase maps.
of sample Dy20-Al04/YIG is 0.6 nm, as shown inFig. 1(d). The phase image inFig. 1(e)clearly shows the grain boundaries, from which an average grain size of 64 nm is inferred, consistent with the XRD results.
Figures 2(a)–2(e)show the room temperature magnetic hys-teresis loops of YIG and Dy:CeYIG/YIG films. By subtracting the
contribution of the top YIG layer (∼136 emu/cm3), saturation mag-netization (Ms) of the Dy00 film (i.e., the film without Dy sub-stitution) reaches 161 emu/cm3, slightly higher than that of YIG. The Ms decreases to 114 emu/cm3in Dy20 and further drops to 68 emu/cm3in Dy20-Al04 as shown inFig. 2(f). In the ferrimag-netic Dy:CeYIG lattice, the tetrahedral Fe3+ (3/formula unit) and
FIG. 2. Room temperature magnetization hysteresis loops for dysprosium and aluminum substituted Ce:YIG/YIG bilayers: (a) YIG; (b) Dy00; (c) Dy08; (d) Dy20; (e) Dy20-Al04.
(f) Saturation magnetization as a function of Dy3+concentration.
APL Mater. 7, 081119 (2019); doi: 10.1063/1.5112827 7, 081119-3
cally and the dodecahedral sites (Dy, Ce) couple with the octahe-dral Fe3+. Above the compensation temperature (220 K for bulk Dy3Fe5O12), the net magnetic momentM is given by M = Md− (MCe
c +MDyc +Ma), where McCe,McDy,Md, andMaare the magnetic moments of dodecahedral Ce3+, dodecahedral Dy3+, tetrahedral Fe3+, and octahedral Fe3+sites, respectively. The different temper-ature dependence of the Fe and Ce sublattices leads to a higher satu-ration magnetization at higher temperatures in Dy00.19The antipar-allel alignment betweenMDyc andMdcontributes to the reduction of Ms when Dy3+substitutes for nonmagnetic Y3+.25The decrease of Ms in Dy20-Al04 is due to nonmagnetic Al3+ions substituting for Fe3+ions in the tetrahedral sites.26
The magnetic hysteresis loops of the Dy:CeYIG/YIG films in Fig. 2indicate an out-of-plane easy axis for the Dy20 and Dy20-Al04 bilayer films and in-plane for YIG and Dy00. The Dy08/YIG bilayer shows isotropic behavior with both IP and OP satura-tion fields of ∼2.5 kOe. The Dy20/YIG film exhibits a coerciv-ity of 230 Oe, which increases to 380 Oe for Dy20-Al04. The net anisotropy in polycrystalline Dy:CeYIG thin films includes con-tributions from shape anisotropy, magnetoelastic anisotropy, and magnetocrystalline anisotropy. The last term is negligible due to ran-dom alignment of the polycrystalline grains as confirmed by XRD. For YIG thin films, the uniaxial anisotropyKu(YIG) is calculated to be (10.4 ± 0.3) × 104erg cm−3, based on the difference in areas between the OP and IP magnetic hysteresis loops,27which is com-parable to the calculated shape anisotropy of 2πMs2 = 11.6 × 104 erg cm−3, indicating that the dominant contribution to magnetic anisotropy in YIG is shape anisotropy.
For the sample Dy08/YIG, the isotropic behavior indicates that the stress-induced anisotropy is balanced by the shape anisotropy in Dy08 and the top YIG layer. Assuming the two layers are mag-netically coupled, the total uniaxial anisotropy can be considered as the volume average of the top layerKu(YIG) and the bottom layer Ku(Dy08). The YIG and Dy08 have a similar shape anisotropy and the Dy08 is assumed to have an additional magnetoelastic anisotropy Kλ(Dy08) opposite in sign to the shape anisotropy. The measured net anisotropy of the bilayer isKυ= Kλ−2πMs2≅ 0, where Kλ(Dy08) = ζKλwith ζ being the ratio of the thickness of the bilayer:thickness of the Dy08 = 1.57. We estimate the stress induced anisotropy Kλ(Dy08) = 17.2 × 104erg cm−3. The stress induced anisotropyKλ is given by24
Kλ= −3 2λs
Y
1 − ν(αf− αs)ΔT,
where λs is the magnetostriction coefficient for a polycrystalline material, Y is the Young’s modulus of the film (YIG: 200 GPa = 2 × 1012dyn cm−2), ν is the Poisson’s ratio (YIG:0.29), αf and αs are the linear thermal expansion coefficients of the film and the sub-strate, respectively (YIG:10.4 × 10−6K−1, Si:3 × 10−6K−1), and ΔT is the temperature difference (equal to 830○C in our case). We esti-mate the magnetostriction coefficient of Ce1Dy0.87Y1.13Fe5O12to be
λs≈ −6.6 × 10−6.
IV. MAGNETO-OPTICAL AND OPTICAL PROPERTIES The room-temperature OP Faraday rotation hysteresis loops of Dy:CeYIG/YIG films are plotted inFig. 3. The Faraday rotation
vious reports of polycrystalline Ce:YIG thin films19,20 possibly due to the use of the top YIG seed layer in this work. With Dy sub-stitution, the films show higher Faraday rotations (−2830○/cm for Dy08 and −2840○/cm for Dy20 at 1550 nm), whereas Al substi-tution results in a reduction of Faraday rotation to −2080○/cm in Dy20-Al04/YIG compared to Dy20/YIG at 1550 nm. Wavelength-dependent Faraday rotation measurements [Fig. 3(e)] reveal that the Ce:DyIG films, similar to Ce:YIG, exhibit reduced Faraday rotation at longer wavelengths. The observation indicates that the enhanced Faraday rotation of Ce:DyIG in the near infrared region compared to YIG is also mainly due to the electric dipole transition of the Ce3+ 4f1electron to tetrahedral Fe3+in the Ce:DyIG.3,28A possible expla-nation of the enhanced Faraday rotation in Dy-substituted samples is summarized as follows. Dy introduces positive Faraday rotation in Dy3Fe5O12crystals at a wavelength of 1.06 μm of +310○
/cm29which is expected to counteract the negative Faraday rotation in Ce:YIG and lower the magnitude of Faraday rotation. However, substitut-ing Dy also increases the lattice constant due to the larger ionic radius of Dy3+ vs Y3+, which may stabilize more Ce ions in the 3+ valence state and thus increase Faraday rotation.28,30Moreover,
Al3+ions preferentially substitute into the tetrahedral sites in gar-net lattices,26thus diminishing the electric dipole transition between Ce3+and tetrahedral Fe3+and the Faraday rotation in the Dy20-Al04 sample.
An advantage of characterizing hysteresis via Faraday rotation is that the contribution of the top YIG layer is small (YIG:θf ∼ 200○/cm, Ce:YIG: θf ∼ −3000○/cm at 1550 nm) so that the Fara-day hysteresis loop originates primarily from the bottom layer Dy:CeYIG, unlike the VSM data where the magnetic signal of YIG is comparable to that of the Dy:CeYIG. The OP saturation field is ∼3 kOe in Dy00 and decreases to 600 Oe in Dy08. Dy20/YIG and Dy20-Al04/YIG exhibit square loops indicating perpendicular magnetic anisotropy.Figure 3(f)shows the ratio θr/θsand coerciv-ity obtained from the Faraday hysteresis loops. A squareness ratio of 1 is obtained in Dy20-Al04/YIG with the highest coercivity of 500 Oe.
The PMA of Ce:DyIG films is derived from the negative mag-netostriction contributed by the Dy3+combined with tensile strain induced during cooling after rapid thermal annealing (RTA). The tensile strain results from both volume changes during crystal-lization and the thermal mismatch between the film and the sub-strate. To illustrate the effects of annealing, Dy20-Al04/YIG films were grown using single-step and two-step deposition methods at two different RTA temperatures. The results are summarized in Fig. 4. For two-step deposition, a YIG seed layer was first deposited at 400○C followed by RTA at 800○C, and then the Dy20-Al04 film was deposited at 650○C with a cooling rate of 1○C/min to allow for strain relaxation. The OP Faraday loop has low square-ness and a saturation field of 1.4 kOe, much smaller than that of Ce:YIG (2–3 kOe) fabricated using a two-step deposition pro-cess.19 However, the film grown using the single-step deposition process (where Dy:CeYIG is grown first followed by YIG deposi-tion and annealing of the bilayer) shows a square OP hysteresis loop indicating PMA, and an OP coercivity of ∼1 kOe is obtained with an optimal RTA temperature of 900○C, as shown inFig. 4(b). Although the same concentration of Dy is present, the difference in magnetic anisotropy of the one-step and two-step deposited films
APL Materials
ARTICLE scitation.org/journal/apmFIG. 3. Faraday rotation (FR) hysteresis loops of dysprosium and aluminum substituted CeYIG/YIG bilayers: (a) Dy00; (b) Dy08; (c) Dy20; (d) Dy20-Al04. Curves of different
colors correspond to measurements taken at different wavelengths. (e) Faraday rotation of the films as functions of wavelength. (f) The ratio between the remanent and saturation Faraday rotation (black symbols) and coercivity (red symbols) as functions of composition.
reveals that the microstructure and strain state depend on the fabrication method.
In order to determine the optical transmission loss of Dy:CeYIG/YIG films, 500 nm wide straight SOI waveguides were fabricated with window lengths ranging from 50 μm to 4 mm and a fixed width of 10 μm [Fig. 5(a)]. By measuring the waveguide loss with different window lengths, the optical loss of Si/Dy:CeYIG/YIG can be determined. A cross-sectional SEM image of the win-dow area with the garnet cladding layer is shown in Fig. 5(b). The transmission loss as a function of window lengths is plot-ted in Fig. 5(c). The garnet/Si waveguide loss is thereby deter-mined to be 39.9 ± 0.9 dB/cm. The optical loss of Dy:CeYIG
films is inferred by accounting for the confinement factors in the garnet film. The waveguide modal profile of theHx field ampli-tude is simulated using the finite element software COMSOL Mul-tiphysics and plotted inFig. 5(d). The simulated confinement fac-tor in the Dy:CeYIG film is 41%, corresponding to an optical loss of ∼91 dB/cm in Dy20. The loss value gives a magneto-optical FOM of 31○/dB. This value is only slightly lower than that of polycrystalline Ce:YIG films on Si with in-plane anisotropy (38○/dB).12We note that strain relaxation at the edges of the gar-net film in Fig. 5(b) affects the local magnetic anisotropy and domain state, but this does not influence the optical absorption measurements.
FIG. 4. Faraday rotation (FR) hysteresis loops at 1310 nm
of (a) Dy20-Al04 deposited by a two-step method (black) and a single-step method (red). (b) The effect of annealing at 850○C (black) and 900○C (red) on Faraday rotation of
Dy20-Al04.
APL Mater. 7, 081119 (2019); doi: 10.1063/1.5112827 7, 081119-5
FIG. 5. (a) Optical microscope image, (b) SEM
cross-sectional image, and (d) simulated Hx field distribution
of Si/Dy20 waveguides. (c) The transmission loss for a 500 nm wide Si/Dy20 waveguide.
V. CONCLUSION
A magneto-optical thin film with perpendicular magnetic anisotropy, Dy-substituted Ce:YIG, was grown on silicon and SOI substrates by PLD using a YIG top seed layer. By replacing Y ions with Dy, perpendicular magnetic anisotropy is obtained as a result of magnetoelastic anisotropy, while maintaining a Faraday rotation up to −2840○
/cm in the film and a high magneto-optical figure of merit of 31○
/dB, comparable to that of CeYIG with in-plane anisotropy. Moreover, the top seed layer process enables direct contact between Dy:CeYIG and the Si waveguide core to enhance the device nonre-ciprocal phase shift. The tunable anisotropy and excellent magneto-optical performance of Dy:CeYIG films provide opportunities for new magneto-optical devices. In particular, the high remanence and coercivity in the out of plane direction make Dy:CeYIG a promis-ing material candidate for magnet-free, self-biased TE mode optical isolators.
ACKNOWLEDGMENTS
The authors are grateful for support by the National Natural Science Foundation of China (NSFC) (Grant Nos. 51522204 and 61475031) and Ministry of Science and Technology of the People’s Republic of China (MOST) (Grant No. 2016YFA0300802). Q.D., J.H., and C.A.R. acknowledge support from the National Science Foundation (NSF) (Grant No. 1607865).
REFERENCES 1
D. Jalas, A. Petrov, M. Eich, W. Freude, S. Fan, Z. Yu, R. Baets, M. Popovi´c, A. Melloni, J. D. Joannopoulos, M. Vanwolleghem, C. R. Doerr, and H. Renner,
Nat. Photonics7(8), 579–582 (2013).
2
L. Bi, J. Hu, P. Jiang, D. H. Kim, G. F. Dionne, L. C. Kimerling, and C. A. Ross,
Nat. Photonics5(12), 758–762 (2011).
3M. Gomi, H. Furuyama, and M. Abe, J. Appl. Phys. 70(11), 7065–7067
(1991).
4J. Fujita, M. Levy, R. M. Osgood, and L. W. H. Dötsch,Appl. Phys. Lett.76(16),
2158–2160 (2000).
5
H. Yokoi, T. Mizumoto, and Y. Shoji,Appl. Opt.42(33), 6605–6612 (2003).
6
Y. Shoji, T. Mizumoto, H. Yokoi, I.-W. Hsieh, and R. M. Osgood,Appl. Phys. Lett.92, 071117 (2008).
7M. Tien, T. Mizumoto, P. Pintus, H. Kromer, and J. E. Bowers,Opt. Express 19(12), 11740–11745 (2011).
8
S. Ghosh, S. Keyvavinia, W. V. Roy, T. Mizumoto, G. Roelkens, and R. Baets,
Opt. Express20(2), 1839–1848 (2012).
9S. Ghosh, S. Keyvaninia, Y. Shirato, T. Mizumoto, G. Roelkens, and R. Baets, IEEE Photonics J.5(3), 6601108 (2013).
10X. Y. Sun, Q. Du, T. Goto, M. C. Onbasli, D. H. Kim, N. M. Aimon, J. Hu, and
C. A. Ross,ACS Photonics2(7), 856–863 (2015).
11Q. Du, C. Wang, Y. Zhang, Y. Zhang, T. Fakhrul, W. Zhang, C. Gonçalves,
C. Blanco, K. Richardson, L. Deng, C. A. Ross, L. Bi, and J. Hu,ACS Photonics 5(12), 5010–5016 (2018).
12Y. Zhang, Q. Du, C. Wang, T. Fakhrul, S. Liu, L. Deng, D. Huang, P. Pintus,
J. Bowers, C. A. Ross, J. Hu, and L. Bi,Optica6(4), 473–478 (2019).
13
C. Zhang, P. Dulal, B. J. H. Stadler, and D. C. Hutchings,Sci. Rep.7(1), 5820 (2017).
14Y. Shoji, A. Fujie, and T. Mizumoto,IEEE J. Sel. Top. Quantum Electron. 22(6), 264–270 (2016).
15
D. Huang, P. Pintus, and J. E. Bowers,Opt. Mater. Express8(9), 2471–2483 (2018).
16A. F. Popkov, M. Fehndrich, M. Lohmeyer, and H. Dötsch,Appl. Phys. Lett. 72(20), 2508–2510 (1998).
17
M. Fehndrich, A. Josef, L. Wilkens, J. Kleine–Börger, N. Bahlmann, M. Lohmeyer, P. Hertel, and H. Dötscha,Appl. Phys. Lett.74(20), 2918–2920 (1999).
18
E. Ishida, K. Miura, Y. Shoji, H. Yokoi, T. Mizumoto, N. Nishiyama, and S. Arai,
Opt. Express25(1), 452–462 (2017).
19Y. Zhang, C. T. Wang, X. Liang, B. Peng, H. P. Lu, P. H. Zhou, L. Zhang,
J. X. Xie, L. J. Deng, M. Zahradnik, L. Beran, M. Kucera, M. Veis, C. A. Ross, and L. Bi,J. Alloys Compd.703, 591–599 (2017).
20T. Goto, M. C. Onbasli, D. H. Kim, V. Singh, M. Inoue, L. C. Kimerling, and
C. A. Ross,Opt. Express22(16), 19047–19054 (2014).
21
M. Gomi, K. Utsugi, and M. Abe, IEEE Trans. Magn. 22(5), 1233–1235 (1986).
APL Materials
ARTICLE scitation.org/journal/apm22
A. N. Shaposhnikov, A. R. Prokopov, V. N. Berzhansky, A. V. Karavainikov, Y. E. Vysokikh, N. N. Gerasimenko, and D. I. Smirnov,Mater. Res. Bull.95, 115–122 (2017).
23
M. Gomi, K. Satoh, and M. Abe, J. Appl. Phys. 63(8), 3642–3644 (1988).
24K. Maeto, A. Itoh, S. Roike, F. Inoue, and R. Rawanishi,IEEE Transl. J. Magn. Jpn.2(9), 784–790 (1987).
25
A. Z. Arsad and N. B. Ibrahim, J. Magn. Magn. Mater. 410, 128–136 (2016).
26
B. G. Ravi, X. Z. Guo, Q. Y. Yan, R. J. Gambino, S. Sampath, and J. B. Parise,
Surf. Coat. Technol.201(16-17), 7597–7605 (2007).
27
A. S. Tang, M. C. Onbasli, X. Sun, and C. A. Ross,ACS Appl. Mater. Interfaces 10(8), 7469–7475 (2018).
28X. Liang, J. Xie, L. Deng, and L. Bi,Appl. Phys. Lett.
106(5), 052401 (2015).
29
S. H. Wemple, J. F. Dillon, Jr., L. G. V. Uitert, and W. H. Grodkiewicz,Appl. Phys. Lett.22(7), 331–333 (1973).
30
X. Liang, Y. Zhu, B. Peng, L. Deng, J. Xie, H. Lu, M. Wu, and L. Bi,ACS Appl. Mater. Interfaces8(12), 8175–8183 (2016).
APL Mater. 7, 081119 (2019); doi: 10.1063/1.5112827 7, 081119-7
