Solution-processed ferrimagnetic insulator thin film for the

Subscriber access provided by University of South Dakota

Energy, Environmental, and Catalysis Applications

Solution-processed ferrimagnetic insulator thin film for the microelectronic spin Seebeck energy conversion Inseon Oh, Jungmin Park, Junhyeon Jo, Mi-Jin Jin, Min-Sun Jang, Ki-Suk Lee, and Jung-Woo Yoo ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b08749 • Publication Date (Web): 06 Aug 2018 Downloaded from http://pubs.acs.org on August 16, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Solution-processed ferrimagnetic insulator thin film for the microelectronic spin Seebeck energy conversion Inseon Oh, Jungmin Park, Junhyeon Jo, Mi-Jin Jin, Min-Sun Jang, Ki-Suk Lee, and Jung-Woo Yoo* School of Materials Science and Engineering-Low dimensional Carbon Materials Center, Ulsan National Institute of Science and Technology, Ulsan, 44919, Korea.

Corresponding author email: [email protected]

KEYWORDS Yttrium iron garnet, Ferrimagnetic insulator, Solution process, Spin Seebeck effect, Inverse spin Hall effect, Magnon

1 Environment ACS Paragon Plus

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT

The longitudinal spin Seebeck effects with a ferro- or ferrimagnetic insulator provides a new architecture of a thermoelectric device that could significantly improve the energy conversion efficiency. Until now, epitaxial yttrium iron garnet (YIG) films grown on gadolinium gallium garnet (GGG) substrates by a pulsed laser deposition have been most widely used for spin thermoelectric energy conversion studies. In this work, we developed a simple route to obtain a highly uniform solution-processed YIG film and used it for the on-chip microelectronic spin Seebeck characterization. We improved the film roughness down to ~ 0.2 nm because the extraction of thermally induced spin voltage relies on the interfacial quality. The on-chip microelectronic device has a dimension of 200 m long and 20 m wide. The solution-processed 20 nm thick YIG film with a 10 nm Pt film was used for the spin Seebeck energy converter. For a temperature difference of T ~ 0.036 K applied on the thin YIG film, the obtained V ~ 28 V, which is equivalent to SLSSE ~ 80.4 nV/K, close to the typical reported values for thick epitaxial YIG films. The temperature and magnetic field dependent behaviors of spin Seebeck effects in our YIG films suggest active magnon excitations through the noncoherent precession channel. The effective SSE generation with the solution-processed thin YIG film provides versatile applications of the spin thermoelectric energy conversion.


Page 2 of 23

Page 3 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION Spin Seebeck effect (SSE) refers to a conversion between heat and spin flow, as the temperature gradient applied on a magnetic material induces a spin voltage driving a nonequilibrium spin flow.1-11 The observation of SSE in magnetic insulator2 suggested that unbalanced electron flow between up- and down- spin is irrelevant to the SSE. Instead, SSE is mediated by the nonequilibrium dynamics of localized spins, magnon.4, 6 The generated magnonic spin flow can be transferred to the attached conductor, where it can be further transformed into a charge current via the inverse spin Hall effect (ISHE).12-15 The SSE is of crucial importance in spin caloritronics16 which deal with entanglement between charge, spin, and heat flow, providing an alternative platform of energy generation from wasted heat.11 The spin caloritronic energy converter shares the same merits with the traditional thermoelectric device, e.g. silent, sustainable, scalable, solidstate, no moving parts, and simple to be ubiquitous applications.

The simplest and most straightforward setup of the SSE is the longitudinal configuration,3, 10 which consists of a ferromagnet/heavy-metal bilayer. In a longitudinal SSE (LSSE), the temperature gradient T applied to the perpendicular to the film (z direction). When the magnetization M of the ferromagnet is along the x direction, the spin current flows parallel to T with a polarization of x direction. Then, an ISHE-induced voltage is generated in the heavy-metal layer along the y direction according to the relation EISHE  DISHE J s   , where EISHE, Js, and  denote the electric field induced by the ISHE, thermally generated spin current vector, and spin-polarization vector of electrons, respectively. DISHE indicates the ISHE efficiency, which is associated with strength of a spin-orbit interaction (SOI) in a heavy metal film. Therefore, the applied vertical heat flux induces transverse electrical power in the ferromagnet/heavy-metal bilayer. Such a transverse



energy conversion has advantages in architecturing thermoelectric devices over traditional thermoelectric modules, where Wiedemann-Franz law strictly limits the improvements of thermoelectric figure of merit.11 Ferro- or ferrimagnetic insulator (FMI) has been most extensively employed for the pure spin current generator through LSSE3, 10-11 as well as the ferromagnetic resonance (FMR) induced spin pumping,17-19 because it eliminates artifacts caused by the electric conduction in the ferromagnetic layer. FMI also has advantage in terms of energy conversion because it assists maintaining a vertical temperature gradient more effectively than the metallic films.11

Most of LSSE studies used the yttrium iron garnet Y3F5O12 (YIG) as a source of spin flow, due mostly to its very low damping of magnetization dynamic and the resulting long spin-wave propagation lifetime.20-21 In its bulk crystalline form, YIG is a ferrimagnet with an electronic bandgap of approximately 2.8 eV, so that electronic excitations do not contribute to transport.22 The antiferromagnetic exchange interaction between Fe3+ ions located in two inequivalent sites (two in octahedral and three in tetrahedral environments) leads to a net moment becoming the ferrimagnet. The macroscopic properties of YIG are often described using the typical language of the ferromagnet.

The YIG films have been fabricated by several methods, for example, liquid phase epitaxy (LPE), 23-25

sputtering,26-27 pulsed laser deposition (PLD),28-30 and sol-gel method.31-35 The LSSE with

YIG has been studied mostly with the epitaxial YIG film on a gadolinium gallium garnet Gd3Ga5O12 (GGG) substrate grown by PLD, sputtering, and LPE, which requires highly expensive


Page 4 of 23

Page 5 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


materials and instruments. In general, the spin pumping with FMR require highly coherent precession motion in need of a monocrystalline FMI layer. But the SSE could largely depend on noncoherent precession motion of magnetic moments excited by thermal energy.36 Thus, the SSE could be effectively induced even with the polycrystalline film, which can be deposited on various substrates. Further, the magnonic excitation in the polycrystalline YIG could be significantly enhanced because magnetic stiffness will be reduced for higher degree of disorder. The LSSE for the solution processed FMI might have substantially different behavior in its dependence of field, temperature, and thickness of film. Another important factor we need to consider for the LSSE from a solution-processed polycrystalline YIG film is the interface quality. The transfer of a spin current to the neighboring heavy metal layer strongly relies on the spin mixing conductance at the interface.18-19, 37-38 Improving interface quality has been extensively studied for both the epitaxial film and the bulk crystal of YIG.18-19, 37, 39 Because the solution-processed film generally has higher degree of roughness, it is critically important to improve interface quality for effective extraction of EISHE from LSSE. Developing a well-defined solution-processing methodology for the FMI could significantly reduce the energy cost and extend further variety of applications. The solutionprocess can be easily applied to large-area manufacturing for versatile energy generation. On the other hand, it can also be directly employed on a micro-electronic chip to recycle heat dissipation because it is free from substrate.

In this work, we developed a highly uniform solution-processed YIG film and used it for the onchip microelectronic spin Seebeck devices. The solution-processed 20 nm thick YIG film with a 10 nm Pt film were used for the spin Seebeck energy converter. The top Au layer was used for a heater to induce a vertical temperature gradient as well as for a temperature sensor. The roughness



of a YIG film was reduced to 0.2 nm by simply diluting precursor solution, which sacrifices the thickness of a spin-coated film. In general, thicker YIG films induce higher LSSE voltage because more number of magnons can be involved in transferring the spin current at the interface of a YIG/Pt bilayer. Here, we show that improving surface roughness of a solution-processed YIG film can induce a LSSE voltage as high as that of the epitaxial YIG film although the thickness of the solution-processed YIG film is only 20 nm. The estimated spin Seebeck coefficient for our microelectronic device is SLSSE ~ 80.4 nV/K. This value is comparable to the typical reported values for thick epitaxial YIG films. Temperature dependence of LSSE voltage in our devices suggests effective phonon-drag invigorates thermal spin-pumping. And magnetic field dependent behavior of SSE reflects noncoherent magnon excitations due to higher degree of disorder. The demonstrated effective LSSE with solution-processed thin YIG films can be applied for a verity of substrates leading to ubiquitous spin thermoelectric power generations.

EXPERIMENTAL SECTION 1. Solution processed YIG film Yttrium nitrate hexahydrate [Y(NO3)36H2O, 99.99% purity] and iron(III) nitrate nonahydrate [Fe(NO3)3․9H2O, 99.99% purity] powder with a stoichiometric ratio of 3:5 was dissolved in a citric acid (C6H8O7․H2O). The precursor mixture was dissolved in water (100 mL) and stirred for 24 h at 80 ℃ to obtain a homogeneous solution. The solution was maintained at pH 1. The homogeneous solution was spin-coated on a SiO2 (500 um) substrate at 4000 rpm for 40 s. The film was then baked at 150 ℃ for 3 min to remove residual solvents. In order to obtain highly homogeneous solution-processed YIG films we used 1 M solution to reduce viscosity. Typical thickness of the spin-coated film was ~ 4 nm. The spin-coating and annealing procedure was 6 Environment ACS Paragon Plus

Page 6 of 23

Page 7 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


repeated for 5 times, which produces typical thickness of the film ~ 20 nm. Lastly, the film was annealed at 800 ℃ for 2 h in an ambient atmosphere. The crystallization of the film was probed by using a high resolution X-ray diffraction (D8 ADVANCE, Bruker AXS). The roughness of the film surface was determined by using an atomic force microscopy (DI-3100, Veeco). The thickness and detailed structure of the stacked film of our LSSE device were further investigated by using a transmission electron microscopy (JEM-2100, JEOL). Magnetic properties of our YIG films were measured by using a SQUID magnetometer (Quantum Design SQUID-VSM).

2. Spin thermoelectric device fabrication and characterization Our LSSE devices have a configuration of Au (30 nm)/Al2O3 (130 nm)/Pt (10 nm)/YIG (20 nm) on oxidized silicon substrates (500 um) as shown in Figure 1a. The device patterning was done by using the photo-lithography. The fabricated LSSE devices have a dimension of 200 m long and 20 m wide. The Pt film for the detection of LSSE was deposited by the e-beam evaporation under the base presser of ~ 10-7 torr. The top Au layer was used for a Joule heating wire to induce heat gradient over the underlying YIG film and also used for a temperature sensor to calibrate applied temperature gradient. The thick Al2O3 layer was inserted between Au and Pt layers for insulation. Both Al2O3 and Au layers were deposited successively by the e-beam evaporation. The current source for the Joule heating was induced by using a Keithley 2636 sourcemeter. The temperature of the Au layer was estimated based on the temperature dependence of the electrical resistivity of the Au layer. Temperature stabilization of the top Au layer depends on the heating current and usually took less than a few min (see Figure S-1b in the Supporting Information). LSSE measurements were performed 10 min after the temperature of the top Au layer was sufficiently stabilized. The detailed measurements and temperature calculations were further explained in the



section of the result and the Supporting Information. Under the applied temperature gradient, the induced voltage from SSE and ISHE was measured by using a Keithley 2182 nanovoltmeter. All measurements for LSSE were performed in a Quantum Design Physical Properties Measurement System (PPMS) with varying temperature and magnetic field.

RESULTS AND DISCUSSION

Figure 1. A schematic illustration of the LSSE device. (a) The micro LSSE device consists of a YIG/Pt

bilayer with a top on-chip Au heater and temperature sensor (left). A schematic illustration of the conversion between heat flow and electrical current (right). Temperature gradient induced by the Joule heating is applied to the vertical direction of the device. Magnon from the YIG film transfers a vertical spin current into the neighboring Pt film, where the spin current converts into the transverse charge current via the ISHE. The generated VLSSE (= EISHE×L) is measured at the both side ends of the Pt layer. (b) A TEM image of a Pt (10 nm)/YIG (20 nm) bilayer, which displays polycrystalline nature of the solution-processed YIG film. The inset displays the TEM image for the cross-section of the LSSE device.

Figure 1a shows a schematic illustration of our LSSE device. The device consists of a YIG/Pt bilayer for the LSSE spin thermoelectric conversion. The applied longitudinal heat gradient induces magnon excitation and propagation along the vertical direction, which transfers a vertical


Page 8 of 23

Page 9 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


spin current into the adjacent Pt line. Then, the ISHE generates a transverse electrical voltage along the Pt line. Figure 1b displays the TEM image of the fabricated devices. It clearly displays a polycrystalline structure of a 20 nm thick YIG film with a well-defined Pt/YIG interface. The inset of Figure 1b shows the TEM image for the cross-section of the LSSE device.

Figure 2. (a) The XRD patterns of thin YIG films on quartz substrates according to annealing temperatures.

(b) An AFM image of the YIG film annealed at 800 ℃ on the oxidized silicon substrate. The measured rms roughness scanned for 1×1 m2 area was 0.2 nm.

Figure 2a shows the XRD data for the samples with different annealing temperatures. The asgrown YIG film typically undergoes phase transformation with annealing temperature 600 – 1000 ℃



Page 10 of 23

depending on the details of environment.31, 36, 40 Our results showed spin-coated samples with annealing temperature over 800 ℃ become crystallized. The diffraction peaks correspond to (400), (420), and (422) of the YIG film.40-41 The strength of diffraction peaks is rather weak due to very thin thickness of the studied YIG film. Figure 2b shows AFM topography displaying surface roughness of the studied YIG film. One of the obvious disadvantage of the solution-processed YIG film for spin thermoelectric applications is a rough surface because the measured VLSSE (= EISHE×L) highly depends on an interfacial quality. The thicker YIG film generally induces the higher VLSSE, because more number of magnons can be involved in generating a spin current at the YIG/Pt interface42. Thus, keeping the thickness of the YIG film and a smooth surface is challenging for the solution-processed YIG film. Several polishing methods have been employed to improve the surface roughness of the solution-processed YIG films.11,

33

In this work, we developed the

solution-processed polycrystalline YIG thin films which have much lower value of root mean square (RMS) roughness ~ 0.2 nm (for 1×1 m area) than those of previous reports32-33 by using a dilute solution. But this process sacrifices the thickness of the YIG film. The obtained RMS of our YIG film is comparable to PLD and LPE deposited epitaxial YIG films.23-25, 28-30

Figure 3a shows magnetization vs magnetic field measured for our solution processed film. The measurements were done for the magnetic field parallel and perpendicular to the plane of the film. Results show ferromagnetic hysteresis loop with a coercivity about 25 Oe for in-plain magnetic field and 40 Oe for out-of plane magnetic field, respectively. The inset of Figure 3a is a close view of the hysteresis loop. It displays the gradual flip of the magnetization due to the presence of multiple domains in our polycrystalline YIG film. Figure 3b shows the temperature dependent magnetization curve measured for 5 – 300 K. The magnetization monotonically decreases as


Page 11 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


temperature increases, which could be attributed to the spin wave excitation as well as structural disorder. The inset of Figure 3b displays a plot for the Bloch T3/2 law, which describes temperature dependent population of magnons. Deviation from T3/2 law can be observed especially at below 100 K. As our polycrystalline YIG films have the discreteness of the lattice and various interaction mechanisms of the spin wave with each other,43 the magnetization as a function of temperature shows deviation from a standard magnon behavior.

Figure 3. Magnetic characteristics of the solution-processed YIG films. (a) Magnetic hysteresis loops

measured at 300 K for the applied magnetic fields perpendicular and parallel to the plane of the YIG film.



Page 12 of 23

(b) Temperature dependence of the magnetization measured for B = 0.1 T. Inset displays a plot for the Block’s T3/2 law of the spin wave.

Figure 4a shows the obtained LSSE signal for different heating currents in a top Au layer. The onchip Joule heating can provide a direct thermal contact to the YIG/Pt LSSE device44-46 and the precise estimation of the temperature gradient. Schematic at the top-left corner of Figure 4a displays the VLSSE measurement configuration. The temperature gradient was applied from the top gold heating layer to the bottom of the oxidized silicon substrate, which was in thermal contact with a gold plate of a PPMS sample puck. The measurement of VLSSE was performed while an external magnetic field was swept in the plane of the device. As increasing the heating current through the Au line, significant enhancement of VLSSE can be observed. The obtained VLSSE for 40 mA of heating current was around 28 V. Figure 4b shows variation of resistance of Au, estimated TAu, and measured VLSSE upon applying currents in the Au heating layer. Because Joule heating power is proportional to I2, temperature of Au layer and the obtained VLSSE are proportional to I2. To compare the SSE coefficient (SLSSE = EISHE/∇𝑧 T = (△VLSSE/L)/(△T/d)) quantitatively, we first 𝑑𝑇

need to estimate the applied T on the YIG film. We employed the Fourier’s law (𝑞𝑥 = −𝜅 𝑑𝑥 ) of a heat conduction to calculate the temperature gradient, where 𝜅 is the thermal conductivity of the material. Based on assumption that the heat flux mainly flows through a normal direction, the 𝑥

temperature gradient applied on each layer in the device is proportional to 𝜅. For example, when the applied current was 20 mA, the estimated temperature of the gold layer was 315.43 K based on the change of the resistance of the Au layer (see Supporting Information). Then, the temperature gradient applied on each layer can be estimated as long as thermal conductivity (𝜅) and thickness (x) are given. The reference values used for thermal conductivities of Si, SiO2, YIG, Pt, and Al2O3, 12 Environment ACS Paragon Plus

Page 13 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


were 150, 1.4, 8, 72, and 39 W/mK, respectively. Then, the temperature gradient and the spin Seebeck coefficient of the YIG thin film was estimated to be 0.011 K and SLSSE ~ 72.8 nV/K, respectively. Figure 4c displays the measured VLSSE (= EISHE×L) as a function of T in YIG for the precise estimation of SLSSE. The linear fit produces VLSSE /T ~ 804 V/K, which corresponds to SLSSE ~ 80.4 nV/K. The measured spin Seebeck coefficient of our YIG thin film is comparable to what have been reported for thick epitaxial YIG films,3, 10, 42, 47-51 although recent work showed significant improvement in SLSSE for epitaxial YIG film52. We also note that the obtained SLSSE in our study is much higher than those of YIG films made by solution-process.11, 33 In order to clarify the possible artifacts in measured VLSSE due to leakage or capacitive contributions through the Al2O3, we tested the control sample of Au (30 nm)/ Al2O3 (130 nm)/Pt (10 nm) on oxidized Si substrate, which has the same configuration without the YIG film (Figure S2 in the Supporting Information). Results showed that the measured VLSSE in our solution-processed YIG/Pt devices is originated from LSSE.



Figure 4. LSSE characterization for our solution-processed polycrystalline YIG film. (a) Magnetic field

dependent VLSSE loops at different heating currents. (b) The obtained VLSSE (■), resistance of the Au layer (■), and the estimated temperature of the Au layer (■) as a function of the heating current. As the heating current of the Au layer increases, resistance of the Au layer, calculated temperature of the Au layer, and measured VLSSE all increase by ~ I2. (c) Measured VLSSE as a function of the estimated T in YIG.

The VLSSE depending on a system temperature was measured by sweeping the magnetic field at each fixed temperature as shown in the Figure 5a. The applied current to the Au heating line was 20 mA. The measured VLSSE = V(+H) -V(-H) is about 7.9 V at 300 K. The VLSSE increases


Page 14 of 23

Page 15 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


gradually with lowering temperature and exhibits a peak at around 110 K. Then, it starts to decrease as temperature decreases further. In general, the temperature dependence of VLSSE highly rely on the magnon excitation and propagation47. The excitation of magnon and its population follows Bloch T3/2 law. Thus, less number of magnons involve in SSE at lower temperature. However, the propagation length of magnon increases as temperature decreases. Thus, LSSE is often described by an atomistic spin model, VLSSE ∝ 1-exp(-L/ξ),42, 53 where ξ is the characteristic length of the magnon and L is the thickness of the sample. Increasing thickness of a YIG film allows more number of magnons involved in LSSE induced spin current generation at the Pt/YIG interface until the thickness of the YIG film becomes comparable to the propagation length of magnon. The estimated value of ξ for the epitaxial YIG film from the literatures is around ~ 1 m at room temperature and it monotonically increases as temperature is lowered.47 Thus, VLSSE from the epitaxial YIG film with thickness around several m increases as temperature is lowered from room temperature and shows maximum at particular temperature, where the size effect start to dominate magnon scattering. Then, it decreases as temperature is lowered further due to the reduced excitation of magnon at very low temperature. The temperature dependence of our YIG film also displays similar behavior. But it should not be correlated with the characteristic length of magnon because our solution-processed YIG film has thickness much less than ξ at all temperature. Thus, the mechanism of LSSE in the solution-processed thin YIG film would be affected by other factors. The propagation of phonon can also significantly affect SSE through two different mechanisms. One is the phonon-mediated process in the YIG film. The phonon-mediated non-equilibrium drives thermal spin pumping, which leads to strong enhancement of SSE at low temperature following the temperature dependent thermal conductivity of YIG.2, 10, 49 However, it was reported that the polycrystalline YIG film displayed negligible low temperature enhancement



of SSE.10 Another mechanism that can induce strong enhancement of SSE at low temperature is phonon-drag at the interface between the YIG film and the substrate.54 This effect is effective for the transverse configuration of SSE producing extremely long magnon propagation length55. But such phonon-drag effects become negligible for the longitudinal configuration of SSE.56 It is due to the fact that the phonon transport in the substrate and magnon transport in the YIG film occurs in series in the longitudinal configuration while the two processes take place in parallel for the transverse configuration.56 The substrate phonon-drag effect could be effective in our studies because our solution-processed YIG films are extremely thin compared to YIG films used for other LSSE studies. The thermal non-equilibrium from the substrate could continuously drive magnon excitation and propagation in our thin YIG film leading to a relatively high value of SLSSE. Figure 5c displays magnetic field dependence of VLSSE measured at 300 K with 20 mA heating current. For a thick YIG film, high magnetic field generally suppresses VLSSE due to the suppression of long wavelength magnon excitations.47 If the thickness of the YIG film is less than the characteristic length ξ, high field suppression disappears.47 The high magnetic field dependence of VLSSE in our device is shown in Figure 5c. The observed VLSSE even increases as the magnetic field increases. This behavior might be attributed to the polycrystalline nature of our solution-processed film because magnetization of our YIG film gradually increases with increasing field due to multidomain structure and magnetic anisotropy. Behavior of LSSE in a polycrystalline YIG is substantially different from that of epitaxial film,10, 42, 47, 49 which requires further exploration.


Page 16 of 23

Page 17 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. (a) Temperature dependence of LSSE signals for a Joule heating current of the top gold layer

Iheating = 20 mA. Measurements were done with sweeping magnetic field between -500 and 500 Oe. (b) Temperature dependent saturated LSSE signals which is calculated by VLSSE = [VLSSE (+H) - VLSSE (-H)]. (c) Magnetic field dependent LSSE signals measured at 300 K for Iheating = 20 mA.

CONCLUSIONS We developed a highly uniform solution-processed YIG film for the on-chip LSSE energy converter, which can be directly applied on various substrates with expanded functionalities. 17 Environment ACS Paragon Plus


Improving the surface roughness of the thin solution-processed YIG film allows us to improve SLSSE as high as that of the thick epitaxial YIG film. The temperature and magnetic field dependent behavior of LSSE for the polycrystalline YIG film is substantially different from that of the epitaxial YIG. In particular, the temperature dependence of LSSE reflects the substrate phonondrag effect enhances LSSE from the thin FMI layer. The solution-processed thin YIG film with low roughness will provide an alternative route for thermal energy harvesting and lead to versatile spin thermoelectric power generation.

ASSOCIATED CONTENT s Supporting Information ○

The supporting information is available free of charge on the ACS Publications website at DOI: XXXXX Details about calibration of temperature in the on-chip LSSE devices, Measurement on the control device of Au (30 nm)/ Al2O3 (130 nm)/Pt (10 nm) on oxidized Si substrate (PDF)

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] Notes The authors declare no competing financial interest.


Page 18 of 23

Page 19 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ACKNOWLEDGMENT I. O. and J. P. contributed equally to this work. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (2017M3A7B4049172 and 2017R1A2B4008286)

Reference Uchida, K.; Takahashi, S.; Harii, K.; Ieda, J.; Koshibae, W.; Ando, K.; Maekawa, S.; Saitoh, 1. E. Observation of the Spin Seebeck Effect. Nature 2008, 455, 778-781. 2. Uchida, K.; Xiao, J.; Adachi, H.; Ohe, J.; Takahashi, S.; Ieda, J.; Ota, T.; Kajiwara, Y.; Umezawa, H.; Kawai, H.; Bauer, G. E. W.; Maekawa, S.; Saitoh, E. Spin Seebeck Insulator. Nat. Mater. 2010, 9, 894-897. 3. Uchida, K.; Adachi, H.; Ota, T.; Nakayama, H.; Maekawa, S.; Saitoh, E. Observation of Longitudinal Spin-Seebeck Effect in Magnetic Insulators. Appl. Phys. Lett. 2010, 97, 172505. 4. Xiao, J.; Bauer, G. E. W.; Uchida, K.; Saitoh, E.; Maekawa, S. Theory of Magnon-Driven Spin Seebeck Effect. Phys. Rev. B 2010, 81, 214418. 5. Jaworski, C. M.; Yang, J.; Mack, S.; Awschalom, D. D.; Heremans, J. P.; Myers, R. C. Observation of the Spin-Seebeck Effect in a Ferromagnetic Semiconductor. Nat. Mater. 2010, 9, 898-903. 6. Adachi, H.; Ohe, J.; Takahashi, S.; Maekawa, S. Linear-Response Theory of Spin Seebeck Effect in Ferromagnetic Insulators. Phys. Rev. B 2011, 83, 094410. 7. Bosu, S.; Sakuraba, Y.; Uchida, K.; Saito, K.; Ota, T.; Saitoh, E.; Takanashi, K. Spin Seebeck Effect in Thin Films of the Heusler Compound Co2MnSi. Phys. Rev. B 2011, 83, 224401. 8. Uchida, K.; Adachi, H.; An, T.; Ota, T.; Toda, M.; Hillebrands, B.; Maekawa, S.; Saitoh, E. Long-Range Spin Seebeck Effect and Acoustic Spin Pumping. Nat. Mater. 2011, 10, 737-741. Jaworski, C. M.; Yang, J.; Mack, S.; Awschalom, D. D.; Myers, R. C.; Heremans, J. P. 9. Spin-Seebeck Effect: A Phonon Driven Spin Distribution. Phys. Rev. Lett. 2011, 106, 186601. 10. Uchida, K.; Ota, T.; Adachi, H.; Xiao, J.; Nonaka, T.; Kajiwara, Y.; Bauer, G. E. W.; Maekawa, S.; Saitoh, E. Thermal Spin Pumping and Magnon-Phonon-Mediated Spin-Seebeck Effect. J. Appl. Phys. 2012, 111, 103903. 11. Kirihara, A.; Uchida, K.; Kajiwara, Y.; Ishida, M.; Nakamura, Y.; Manako, T.; Saitoh, E.; Yorozu, S. Spin-Current-Driven Thermoelectric Coating. Nat. Mater. 2012, 11, 686-689. 12. Saitoh, E.; Ueda, M.; Miyajima, H.; Tatara, G. Conversion of Spin Current into Charge Current at Room Temperature: Inverse Spin-Hall Effect. Appl. Phys. Lett. 2006, 88, 182509. 13. Valenzuela, S. O.; Tinkham, M. Direct Electronic Measurement of the Spin Hall Effect. Nature 2006, 442, 176-179.



Kimura, T.; Otani, Y.; Sato, T.; Takahashi, S.; Maekawa, S. Room-Temperature Reversible 14. Spin Hall Effect. Phys. Rev. Lett. 2007, 98, 156601. 15. Ando, K.; Takahashi, S.; Ieda, J.; Kajiwara, Y.; Nakayama, H.; Yoshino, T.; Harii, K.; Fujikawa, Y.; Matsuo, M.; Maekawa, S.; Saitoh, E. Inverse Spin-Hall Effect Induced by Spin Pumping in Metallic System. J. Appl. Phys. 2011, 109, 103913. 16. Bauer, G. E. W.; Saitoh, E.; van Wees, B. J. Spin Caloritronics. Nat. Mater. 2012, 11, 391399. 17. Jia, X. T.; Liu, K.; Xia, K.; Bauer, G. E. W. Spin Transfer Torque on Magnetic Insulators. Europhys. Lett. 2011, 96, 17005. 18. Jungfleisch, M. B.; Lauer, V.; Neb, R.; Chumak, A. V.; Hillebrands, B. Improvement of the Yttrium Iron Garnet/Platinum Interface for Spin Pumping-Based Applications. Appl. Phys. Lett. 2013, 103, 022411. 19. Qiu, Z.; Ando, K.; Uchida, K.; Kajiwara, Y.; Takahashi, R.; Nakayama, H.; An, T.; Fujikawa, Y.; Saitoh, E. Spin Mixing Conductance at a Well-Controlled Platinum/Yttrium Iron Garnet Interface. Appl. Phys. Lett. 2013, 103, 092404. Serga, A. A.; Chumak, A. V.; Hillebrands, B. YIG Magnonics. J. Phys. D: Appl. Phys. 20. 2010, 43, 264002. 21. Chumak, A. V.; Vasyuchka, V. I.; Serga, A. A.; Hillebrands, B. Magnon Spintronics. Nat. Phys. 2015, 11, 453-461. 22. Jakubisova, E. L.; Visnovsky, S.; Chang, H. C.; Wu, M. Z. Interface Effects in NanometerThick Yttrium Iron Garnet Films Studied by Magneto-Optical Spectroscopy. Appl. Phys. Lett. 2016, 108, 082403. 23. Pirro, P.; Bracher, T.; Chumak, A. V.; Lagel, B.; Dubs, C.; Surzhenko, O.; Gornert, P.; Leven, B.; Hillebrands, B. Spin-Wave Excitation and Propagation in Microstructured Waveguides of Yttrium Iron Garnet/Pt Bilayers. Appl. Phys. Lett. 2014, 104, 012402. Cornelissen, L. J.; van Wees, B. J. Magnetic Field Dependence of the Magnon Spin 24. Diffusion Length in the Magnetic Insulator Yttrium Iron Garnet. Phys. Rev. B 2016, 93, 020403(R). 25. Dubs, C.; Surzhenko, O.; Linke, R.; Danilewsky, A.; Bruckner, U.; Dellith, J. SubMicrometer Yttrium Iron Garnet LPE films with Low Ferromagnetic Resonance Losses. J. Phys. D: Appl. Phys. 2017, 50, 204005. Houchen, C.; Peng, L.; Wei, Z.; Tao, L.; Hoffmann, A.; Longjiang, D.; Mingzhong, W. 26. Nanometer-Thick Yttrium Iron Garnet Films with Extremely Low Damping. IEEE Magn. Lett. 2014, 5, 6700104. 27. Lustikova, J.; Shiomi, Y.; Qiu, Z.; Kikkawa, T.; Iguchi, R.; Uchida, K.; Saitoh, E. Spin Current Generation from Sputtered Y3Fe5O12 Films. J. Appl. Phys. 2014, 116, 153902. 28. Sun, Y. Y.; Song, Y. Y.; Chang, H. C.; Kabatek, M.; Jantz, M.; Schneider, W.; Wu, M. Z.; Schultheiss, H.; Hoffmann, A. Growth and Ferromagnetic Resonance Properties of NanometerThick Yttrium Iron Garnet Films. Appl. Phys. Lett. 2012, 101, 152405.. 29. Onbasli, M. C.; Kehlberger, A.; Kim, D. H.; Jakob, G.; Klaui, M.; Chumak, A. V.; Hillebrands, B.; Ross, C. A. Pulsed Laser Deposition of Epitaxial Yttrium Iron Garnet Films with Low Gilbert Damping and Bulk-Like Magnetization. APL Mater. 2014, 2, 106102. 30. d'Allivy Kelly, O.; Anane, A.; Bernard, R.; Ben Youssef, J.; Hahn, C.; Molpeceres, A. H.; Carretero, C.; Jacquet, E.; Deranlot, C.; Bortolotti, P.; Lebourgeois, R.; Mage, J. C.; de Loubens, G.; Klein, O.; Cros, V.; Fert, A. Inverse Spin Hall Effect in Nanometer-Thick Yttrium Iron Garnet/Pt System. Appl. Phys. Lett. 2013, 103, 082408.


Page 20 of 23

Page 21 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Wang, M. Q.; Zhu, X. Y.; Wei, X. Y.; Zhang, L. Y.; Yao, X. Preparation and Annealing 31. Process of Y3Fe5O12 by Sol-Gel Method. Ferroelectrics 2001, 264, 249-254. 32. Guo, X.; Chen, Y.; Wang, G. S.; Zhang, Y. Y.; Ge, J.; Tang, X. D.; Ponchel, F.; Remiens, D.; Dong, X. L. Growth and Characterization of Yttrium Iron Garnet Films on Si Substrates by Chemical Solution Deposition (CSD) Technique. J. Alloys. Compd. 2016, 671, 234-237. 33. Hirata, S.; Ono, T.; Amemiya, Y.; Tabei, T.; Yokoyama, S. Influence of Surface Smoothing on Spin Seebeck Effect of Ce1Y2Fe5O12 Deposited by Metal Organic Decomposition. Jpn. J. Appl. Phys. 2017, 56, 04CN04. 34. Sharma, V.; Saha, J.; Patnaik, S.; Kuanr, B. K. Synthesis and Characterization of Yttrium Iron Garnet (YIG) Nanoparticles - Microwave Material. AIP Adv. 2017, 7, 056405. 35. Musa, M. A.; Azis, R. a. S.; Osman, N. H.; Hassan, J.; Zangina, T. Structural and Magnetic Properties of Yttrium Iron Garnet (YIG) and Yttrium Aluminum Iron Garnet (YAlG) Nanoferrite via Sol-Gel Synthesis. Results Phys. 2017, 7, 1135-1142. 36. Chang, F. J.; Lin, J. G.; Huang, S. Y. Robust Spin Current Generated by the Spin Seebeck Effect. Phy. Rev. Mater. 2017, 1, 031401R. Jia, X.; Liu, K.; Xia, K.; Bauer, G. E. W. Spin Transfer Torque on Magnetic Insulators. 37. Europhys. Lett. 2011, 96, 17005. 38. Qiu, Z.; Hou, D.; Uchida, K.; Saitoh, E. Influence of Interface Condition on Spin-Seebeck Effects. J. Phys. D: Appl. Phys. 2015, 48, 164013. 39. Aqeel, A.; Vera-Marun, I. J.; van Wees, B. J.; Palstra, T. T. M. Surface Sensitivity of the Spin Seebeck Effect. J. Appl. Phys. 2014, 116, 153705. 40. Jang, M. S.; Roh, I. J.; Park, J.; Kang, C. Y.; Choi, W. J.; Baek, S. H.; Park, S. S.; Yoo, J. W.; Lee, K. S. Dramatic Enhancement of the Saturation Magnetization of a Sol-Gel Synthesized Y3Fe5O12 by a Mechanical Pressing Process. J. Alloys. Compd. 2017, 711, 693-697. 41. Dongquoc, V.; Kuchi, R.; Van, P. C.; Yoon, S. G.; Jeong, J. R. Effects of Heating Rate on the Magneto-Optical Properties of Bismuth-Substituted Yttrium Iron Garnet Films Prepared via Modified Metal-Organic Decomposition. Curr. Appl. Phys. 2018, 18, 241-245. 42. Kehlberger, A.; Ritzmann, U.; Hinzke, D.; Guo, E. J.; Cramer, J.; Jakob, G.; Onbasli, M. C.; Kim, D. H.; Ross, C. A.; Jungfleisch, M. B.; Hillebrands, B.; Nowak, U.; Klaui, M. Length Scale of the Spin Seebeck Effect. Phys. Rev. Lett. 2015, 115, 096602. Solt, I. H. Temperature Dependence of YIG Magnetization. J. Appl. Phys. 1962, 33, 118943. 1191. 44. Wu, S. M.; Fradin, F. Y.; Hoffman, J.; Hoffmann, A.; Bhattacharya, A. Spin Seebeck Devices Ssing Local on-Chip Heating. J. Appl. Phys. 2015, 117, 17C509. Jiang, Z.; Chang, C. Z.; Masir, M. R.; Tang, C.; Xu, Y.; Moodera, J. S.; MacDonald, A. H.; 45. Shi, J. Enhanced Spin Seebeck Effect Signal due to Spin-Momentum Locked Topological Surface States. Nat. Commun. 2016, 7, 11458. 46. Wu, S. M.; Zhang, W.; Amit, K. C.; Borisov, P.; Pearson, J. E.; Jiang, J. S.; Lederman, D.; Hoffmann, A.; Bhattacharya, A. Antiferromagnetic Spin Seebeck Effect. Phys. Rev. Lett. 2016, 116, 097204. 47. Guo, E. J.; Cramer, J.; Kehlberger, A.; Ferguson, C. A.; MacLaren, D. A.; Jakob, G.; Klaui, M. Influence of Thickness and Interface on the Low-Temperature Enhancement of the Spin Seebeck Effect in YIG Films. Phys. Rev. X 2016, 6, 031012. 48. Sola, A.; Bougiatioti, P.; Kuepferling, M.; Meier, D.; Reiss, G.; Pasquale, M.; Kuschel, T.; Basso, V. Longitudinal Spin Seebeck Coefficient: Heat flux vs. Temperature Difference Method. Sci. Rep. 2017, 7, 46752.



Iguchi, R.; Uchida, K.; Daimon, S.; Saitoh, E. Concomitant Enhancement of the 49. Longitudinal Spin Seebeck Effect and the Thermal Conductivity in a Pt/YIG/Pt System at Low Temperatures. Phys. Rev. B 2017, 95, 174401. 50. Yuasa, H.; Nakata, F.; Nakamura, R.; Kurokawa, Y. Spin Seebeck Coefficient Enhancement by Using Ta50W50 Alloy and YIG/Ru Interface. J. Phys. D: Appl.Phys. 2018, 51, 134002. 51. Yuasa, H.; Tamae, K.; Onizuka, N. Spin Mixing Conductance Enhancement by Increasing Magnetic Density. AIP Adv. 2017, 7, 055928. 52. Prakash, A.; Brangham, J.; Yang, F. Y.; Heremans, J. P. Spin Seebeck Effect Through Antiferromagnetic NiO. Phys. Rev. B 2016, 94, 014427. 53. Ritzmann, U.; Hinzke, D.; Nowak, U. Propagation of Thermally Induced Magnonic Spin Currents. Phys. Rev. B 2014, 89, 024409. 54. Adachi, H.; Uchida, K.; Saitoh, E.; Ohe, J.; Takahashi, S.; Maekawa, S. Gigantic Enhancement of Spin Seebeck Effect by Phonon Drag. Appl. Phys. Lett. 2010, 97, 252506. 55. Cornelissen, L. J.; Liu, J.; Duine, R. A.; Ben Youssef, J.; van Wees, B. J. Long-Distance Transport of Magnon Spin Information in a Magnetic Insulator at Room Temperature. Nat. Phys. 2015, 11, 1022-1026. 56. Siegel, G.; Prestgard, M. C.; Teng, S.; Tiwari, A. Robust Longitudinal Spin-Seebeck Effect in Bi-YIG Thin Films. Sci. Rep. 2014, 4, 4429.


Page 22 of 23

Page 23 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TOC graphic


Solution-processed ferrimagnetic insulator thin film for the

Recommend Documents