-
COOD: Combined out-of-distribution detection using multiple measures for anomaly & novel class detection in large-scale hierarchical classification
Authors:
L. E. Hogeweg,
R. Gangireddy,
D. Brunink,
V. J. Kalkman,
L. Cornelissen,
J. W. Kamminga
Abstract:
High-performing out-of-distribution (OOD) detection, both anomaly and novel class, is an important prerequisite for the practical use of classification models. In this paper, we focus on the species recognition task in images concerned with large databases, a large number of fine-grained hierarchical classes, severe class imbalance, and varying image quality. We propose a framework for combining i…
▽ More
High-performing out-of-distribution (OOD) detection, both anomaly and novel class, is an important prerequisite for the practical use of classification models. In this paper, we focus on the species recognition task in images concerned with large databases, a large number of fine-grained hierarchical classes, severe class imbalance, and varying image quality. We propose a framework for combining individual OOD measures into one combined OOD (COOD) measure using a supervised model. The individual measures are several existing state-of-the-art measures and several novel OOD measures developed with novel class detection and hierarchical class structure in mind. COOD was extensively evaluated on three large-scale (500k+ images) biodiversity datasets in the context of anomaly and novel class detection. We show that COOD outperforms individual, including state-of-the-art, OOD measures by a large margin in terms of TPR@1% FPR in the majority of experiments, e.g., improving detecting ImageNet images (OOD) from 54.3% to 85.4% for the iNaturalist 2018 dataset. SHAP (feature contribution) analysis shows that different individual OOD measures are essential for various tasks, indicating that multiple OOD measures and combinations are needed to generalize. Additionally, we show that explicitly considering ID images that are incorrectly classified for the original (species) recognition task is important for constructing high-performing OOD detection methods and for practical applicability. The framework can easily be extended or adapted to other tasks and media modalities.
△ Less
Submitted 11 March, 2024;
originally announced March 2024.
-
Deep convolutional neural networks for multi-planar lung nodule detection: improvement in small nodule identification
Authors:
Sunyi Zheng,
Ludo J. Cornelissen,
Xiaonan Cui,
Xue** **g,
Raymond N. J. Veldhuis,
Matthijs Oudkerk,
Peter M. A. van Ooijen
Abstract:
Objective: In clinical practice, small lung nodules can be easily overlooked by radiologists. The paper aims to provide an efficient and accurate detection system for small lung nodules while kee** good performance for large nodules. Methods: We propose a multi-planar detection system using convolutional neural networks. The 2-D convolutional neural network model, U-net++, was trained by axial,…
▽ More
Objective: In clinical practice, small lung nodules can be easily overlooked by radiologists. The paper aims to provide an efficient and accurate detection system for small lung nodules while kee** good performance for large nodules. Methods: We propose a multi-planar detection system using convolutional neural networks. The 2-D convolutional neural network model, U-net++, was trained by axial, coronal, and sagittal slices for the candidate detection task. All possible nodule candidates from the three different planes are combined. For false positive reduction, we apply 3-D multi-scale dense convolutional neural networks to efficiently remove false positive candidates. We use the public LIDC-IDRI dataset which includes 888 CT scans with 1186 nodules annotated by four radiologists. Results: After ten-fold cross-validation, our proposed system achieves a sensitivity of 94.2% with 1.0 false positive/scan and a sensitivity of 96.0% with 2.0 false positives/scan. Although it is difficult to detect small nodules (i.e. < 6 mm), our designed CAD system reaches a sensitivity of 93.4% (95.0%) of these small nodules at an overall false positive rate of 1.0 (2.0) false positives/scan. At the nodule candidate detection stage, results show that a multi-planar method is capable to detect more nodules compared to using a single plane. Conclusion: Our approach achieves good performance not only for small nodules, but also for large lesions on this dataset. This demonstrates the effectiveness and efficiency of our developed CAD system for lung nodule detection. Significance: The proposed system could provide support for radiologists on early detection of lung cancer.
△ Less
Submitted 9 December, 2020; v1 submitted 13 January, 2020;
originally announced January 2020.
-
A Socio-Informatic Approach to Automated Account Classification on Social Media
Authors:
Laurenz A Cornelissen,
Petrus Schoonwinkel,
Richard J Barnett
Abstract:
Automated accounts on social media have become increasingly problematic. We propose a key feature in combination with existing methods to improve machine learning algorithms for bot detection. We successfully improve classification performance through including the proposed feature.
Automated accounts on social media have become increasingly problematic. We propose a key feature in combination with existing methods to improve machine learning algorithms for bot detection. We successfully improve classification performance through including the proposed feature.
△ Less
Submitted 27 April, 2019;
originally announced April 2019.
-
Efficient spin transport in a paramagnetic insulator
Authors:
Koichi Oyanagi,
Saburo Takahashi,
Ludo J. Cornelissen,
Juan Shan,
Shunsuke Daimon,
Takashi Kikkawa,
Gerrit E. W. Bauer,
Bart J. van Wees,
Eiji Saitoh
Abstract:
The discovery of new materials that efficiently transmit spin currents has been important for spintronics and material science. The electric insulator $\mathrm{Gd}_3\mathrm{Ga}_5\mathrm{O}_{12}$ (GGG) is a superior substrate for growing magnetic films, but has never been considered as a conduit for spin currents. Here we report spin current propagation in paramagnetic GGG over several microns. Sur…
▽ More
The discovery of new materials that efficiently transmit spin currents has been important for spintronics and material science. The electric insulator $\mathrm{Gd}_3\mathrm{Ga}_5\mathrm{O}_{12}$ (GGG) is a superior substrate for growing magnetic films, but has never been considered as a conduit for spin currents. Here we report spin current propagation in paramagnetic GGG over several microns. Surprisingly, the spin transport persists up to temperatures of 100 K $\gg$ $T_{\mathrm{g}} = 180$ mK, GGG's magnetic glass-like transition temperature. At 5 K we find a spin diffusion length ${λ_{\mathrm{GGG}}} = 1.8 \pm 0.2 μ$m and a spin conductivity $σ_{\mathrm{GGG}} = (7.3 \pm 0.3) \times10^4$ $\mathrm{Sm}^{-1}$ that is larger than that of the record quality magnet $\mathrm{Y}_3\mathrm{Fe}_5\mathrm{O}_{12}$ (YIG). We conclude that exchange coupling is not required for efficient spin transport, which challenges conventional models and provides new material-design strategies for spintronic devices.
△ Less
Submitted 29 November, 2018;
originally announced November 2018.
-
Microwave control of thermal magnon spin transport
Authors:
**g Liu,
Frank Feringa,
Benedetta Flebus,
Ludo J. Cornelissen,
Johannes C. Leutenantsmeyer,
Rembert A. Duine,
Bart J. van Wees
Abstract:
We observe that an rf microwave field strongly influences the transport of incoherent thermal magnons in yttrium iron garnet. Ferromagnetic resonance in the nonlinear regime suppresses thermal magnon transport by 95%. The transport is also modulated at non-resonant conditions in two cases, both related to the magnon band minimum. Firstly, a strong enhancement of the nonlocal signal appears at a st…
▽ More
We observe that an rf microwave field strongly influences the transport of incoherent thermal magnons in yttrium iron garnet. Ferromagnetic resonance in the nonlinear regime suppresses thermal magnon transport by 95%. The transport is also modulated at non-resonant conditions in two cases, both related to the magnon band minimum. Firstly, a strong enhancement of the nonlocal signal appears at a static magnetic field below the resonance condition. This increase only occurs at one field polarity and can be as large as 800%. We attribute this effect to magnon kinetic processes, which give rise to band-minimum magnons and high-energy chiral surface modes. Secondly, the signal increases at a static field above the resonance condition, where the rf frequency coincides with the magnon band minimum. Our study gives insight into the interplay between coherent and incoherent spin dynamics: The rf field modifies the occupation of relevant magnon states and, via kinetic processes, the magnon spin transport.
△ Less
Submitted 27 October, 2018;
originally announced October 2018.
-
A Network Topology Approach to Bot Classification
Authors:
Laurenz A Cornelissen,
Richard J Barnett,
Petrus Schoonwinkel,
Brent D. Eichstadt,
Hluma B. Magodla
Abstract:
Automated social agents, or bots, are increasingly becoming a problem on social media platforms. There is a growing body of literature and multiple tools to aid in the detection of such agents on online social networking platforms. We propose that the social network topology of a user would be sufficient to determine whether the user is a automated agent or a human. To test this, we use a publicly…
▽ More
Automated social agents, or bots, are increasingly becoming a problem on social media platforms. There is a growing body of literature and multiple tools to aid in the detection of such agents on online social networking platforms. We propose that the social network topology of a user would be sufficient to determine whether the user is a automated agent or a human. To test this, we use a publicly available dataset containing users on Twitter labelled as either automated social agent or human. Using an unsupervised machine learning approach, we obtain a detection accuracy rate of 70%.
△ Less
Submitted 17 September, 2018;
originally announced September 2018.
-
Deploying South African Social Honeypots on Twitter
Authors:
Laurenz A Cornelissen,
Richard J Barnett,
Morakane AM Kepa,
Daniel Loebenberg-Novitzkas,
Jacques Jordaan
Abstract:
Inspired by the simple, yet effective, method of tweeting gibberish to attract automated social agents (bots), we attempt to create localised honeypots in the South African political context. We produce a series of defined techniques and combine them to generate interactions from users on Twitter. The paper offers two key contributions. Conceptually, an argument is made that honeypots should not b…
▽ More
Inspired by the simple, yet effective, method of tweeting gibberish to attract automated social agents (bots), we attempt to create localised honeypots in the South African political context. We produce a series of defined techniques and combine them to generate interactions from users on Twitter. The paper offers two key contributions. Conceptually, an argument is made that honeypots should not be confused for bot detection methods, but are rather methods to capture low-quality users. Secondly, we successfully generate a list of 288 local low quality users active in the political context.
△ Less
Submitted 17 September, 2018;
originally announced September 2018.
-
Enhanced magnon spin transport in NiFe$_2$O$_4$ thin films on a lattice-matched substrate
Authors:
J. Shan,
A. V. Singh,
L. Liang,
L. Cornelissen,
A. Gupta,
B. J. van Wees,
T. Kuschel
Abstract:
We investigate magnon spin transport in epitaxial nickel ferrite (NiFe$_2$O$_4$, NFO) films grown on magnesium gallate spinel (MgGa$_2$O$_4$, MGO) substrates, which have a lattice mismatch with NFO as small as 0.78%, resulting in the reduction of antiphase boundary defects and thus in improved magnetic properties in the NFO films. In the nonlocal transport experiments, enhanced signals are observe…
▽ More
We investigate magnon spin transport in epitaxial nickel ferrite (NiFe$_2$O$_4$, NFO) films grown on magnesium gallate spinel (MgGa$_2$O$_4$, MGO) substrates, which have a lattice mismatch with NFO as small as 0.78%, resulting in the reduction of antiphase boundary defects and thus in improved magnetic properties in the NFO films. In the nonlocal transport experiments, enhanced signals are observed for both electrically and thermally excited magnons, and the magnon relaxation length ($λ_m$) of NFO is found to be around 2.5 $μ$m at room temperature. Moreover, at both room and low temperatures, we present distinct features from the nonlocal spin Seebeck signals which arise from magnon polaron formation. Our results demonstrate excellent magnon transport properties (magnon spin conductivity, $λ_m$ and spin mixing conductance at the interface between Pt) of NFO films grown on a lattice-matched substrate that are comparable with those of yttrium iron garnet.
△ Less
Submitted 24 July, 2018;
originally announced July 2018.
-
Nonlocal magnon spin transport in yttrium iron garnet with tantalum and platinum spin injection/detection electrodes
Authors:
**g Liu,
Ludo J. Cornelissen,
Juan Shan,
Bart J. van Wees,
Timo Kuschel
Abstract:
We study the magnon spin transport in the magnetic insulator yttrium iron garnet (YIG) in a nonlocal experiment and compare the magnon spin excitation and detection for the heavy metal paramagnetic electrodes platinum (Pt|YIG|Pt) and tantalum (Ta|YIG|Ta). The electrical injection and detection processes rely on the (inverse) spin Hall effect in the heavy metals and the conversion between the elect…
▽ More
We study the magnon spin transport in the magnetic insulator yttrium iron garnet (YIG) in a nonlocal experiment and compare the magnon spin excitation and detection for the heavy metal paramagnetic electrodes platinum (Pt|YIG|Pt) and tantalum (Ta|YIG|Ta). The electrical injection and detection processes rely on the (inverse) spin Hall effect in the heavy metals and the conversion between the electron spin and magnon spin at the heavy metal|YIG interface. Pt and Ta possess opposite signs of the spin Hall angle. Furthermore, their heterostructures with YIG have different interface properties, i.e. spin mixing conductances. By varying the distance between injector and detector, the magnon spin transport is studied. Using a circuit model based on the diffusion-relaxation transport theory, a similar magnon relaxation length of ~ 10 μm was extracted from both Pt and Ta devices. By changing the injector and detector material from Pt to Ta, the influence of interface properties on the magnon spin transport has been observed. For Ta devices on YIG the spin mixing conductance is reduced compared with Pt devices, which is quantitatively consistent when comparing the dependence of the nonlocal signal on the injector-detector distance with the prediction from the circuit model.
△ Less
Submitted 30 March, 2018;
originally announced March 2018.
-
Criteria for accurate determination of the magnon relaxation length from the nonlocal spin Seebeck effect
Authors:
J. Shan,
L. J. Cornelissen,
J. Liu,
J. Ben Youssef,
L. Liang,
B. J. van Wees
Abstract:
The nonlocal transport of thermally generated magnons not only unveils the underlying mechanism of the spin Seebeck effect, but also allows for the extraction of the magnon relaxation length ($λ_m$) in a magnetic material, the average distance over which thermal magnons can propagate. In this study, we experimentally explore in yttrium iron garnet (YIG)/platinum systems much further ranges compare…
▽ More
The nonlocal transport of thermally generated magnons not only unveils the underlying mechanism of the spin Seebeck effect, but also allows for the extraction of the magnon relaxation length ($λ_m$) in a magnetic material, the average distance over which thermal magnons can propagate. In this study, we experimentally explore in yttrium iron garnet (YIG)/platinum systems much further ranges compared with previous investigations. We observe that the nonlocal SSE signals at long distances ($d$) clearly deviate from a typical exponential decay. Instead, they can be dominated by the nonlocal generation of magnon accumulation as a result of the temperature gradient present away from the heater, and decay geometrically as $1/d^2$. We emphasize the importance of looking only into the exponential regime (i.e., the intermediate distance regime) to extract $λ_m$. With this principle, we study $λ_m$ as a function of temperature in two YIG films which are 2.7 and 50 $μ$m in thickness, respectively. We find $λ_m$ to be around 15 $μ$m at room temperature and it increases to 40 $μ$m at $T=$ 3.5 K. Finite element modeling results agree with experimental studies qualitatively, showing also a geometrical decay beyond the exponential regime. Based on both experimental and modeling results we put forward a general guideline for extracting $λ_m$ from the nonlocal spin Seebeck effect.
△ Less
Submitted 19 September, 2017;
originally announced September 2017.
-
Nonlocal magnon-polaron transport in yttrium iron garnet
Authors:
L. J. Cornelissen,
K. Oyanagi,
T. Kikkawa,
Z. Qiu,
T. Kuschel,
G. E. W. Bauer,
B. J. van Wees,
E. Saitoh
Abstract:
The spin Seebeck effect (SSE) is observed in magnetic insulator|heavy metal bilayers as an inverse spin Hall effect voltage under a temperature gradient. The SSE can be detected nonlocally as well, viz. in terms of the voltage in a second metallic contact (detector) on the magnetic film, spatially separated from the first contact that is used to apply the temperature bias (injector). Magnon-polaro…
▽ More
The spin Seebeck effect (SSE) is observed in magnetic insulator|heavy metal bilayers as an inverse spin Hall effect voltage under a temperature gradient. The SSE can be detected nonlocally as well, viz. in terms of the voltage in a second metallic contact (detector) on the magnetic film, spatially separated from the first contact that is used to apply the temperature bias (injector). Magnon-polarons are hybridized lattice and spin waves in magnetic materials, generated by the magnetoelastic interaction. Kikkawa et al. [Phys. Rev. Lett. \textbf{117}, 207203 (2016)] interpreted a resonant enhancement of the local SSE in yttrium iron garnet (YIG) as a function of the magnetic field in terms of magnon-polaron formation. Here we report the observation of magnon-polarons in \emph{nonlocal} magnon spin injection/detection devices for various injector-detector spacings and sample temperatures. Unexpectedly, we find that the magnon-polaron resonances can suppress rather than enhance the nonlocal SSE. Using finite element modelling we explain our observations as a competition between the SSE and spin diffusion in YIG. These results give unprecedented insights into the magnon-phonon interaction in a key magnetic material.
△ Less
Submitted 14 June, 2017;
originally announced June 2017.
-
Magnon Planar Hall Effect and Anisotropic Magnetoresistance in a Magnetic Insulator
Authors:
J. Liu,
L. J. Cornelissen,
J. Shan,
T. Kuschel,
B. J. van Wees
Abstract:
Electrical resistivities can be different for charge currents travelling parallel or perpendicular to the magnetization in magnetically ordered conductors or semiconductors, resulting in the well-known planar Hall effect and anisotropic magnetoresistance. Here, we study the analogous anisotropic magnetotransport behavior for magnons in a magnetic insulator Y$_{3}$Fe$_{5}$O$_{12}$. Electrical and t…
▽ More
Electrical resistivities can be different for charge currents travelling parallel or perpendicular to the magnetization in magnetically ordered conductors or semiconductors, resulting in the well-known planar Hall effect and anisotropic magnetoresistance. Here, we study the analogous anisotropic magnetotransport behavior for magnons in a magnetic insulator Y$_{3}$Fe$_{5}$O$_{12}$. Electrical and thermal magnon injection, and electrical detection methods are used at room temperature with transverse and longitudinal geometries to measure the magnon planar Hall effect and anisotropic magnetoresistance, respectively. We observe that the relative difference between magnon current conductivities parallel and perpendicular to the magnetization, with respect to the average magnon conductivity, i.e. $|(σ_{\parallel}^{\textrm{m}}-σ_{\perp}^{\textrm{m}})/σ_{0}^{\textrm{m}}|$ , is approximately 5% with the majority of the measured devices showing $σ_{\perp}^{\textrm{m}}>σ_{\parallel}^{\textrm{m}}$.
△ Less
Submitted 24 December, 2016;
originally announced December 2016.
-
Influence of yttrium iron garnet thickness and heater opacity on the nonlocal transport of electrically and thermally excited magnons
Authors:
Juan Shan,
Ludo J. Cornelissen,
Nynke Vlietstra,
Jamal Ben Youssef,
Timo Kuschel,
Rembert A. Duine,
Bart J. van Wees
Abstract:
We studied the nonlocal transport behavior of both electrically and thermally excited magnons in yttrium iron garnet (YIG) as a function of its thickness. For electrically injected magnons, the nonlocal signals decrease monotonically as the YIG thickness increases. For the nonlocal behavior of the thermally generated magnons, or the nonlocal spin Seebeck effect (SSE), we observed a sign reversal w…
▽ More
We studied the nonlocal transport behavior of both electrically and thermally excited magnons in yttrium iron garnet (YIG) as a function of its thickness. For electrically injected magnons, the nonlocal signals decrease monotonically as the YIG thickness increases. For the nonlocal behavior of the thermally generated magnons, or the nonlocal spin Seebeck effect (SSE), we observed a sign reversal which occurs at a certain heater-detector distance, and it is influenced by both the opacity of the YIG/heater interface and the YIG thickness. Our nonlocal SSE results can be qualitatively explained by the bulk-driven SSE mechanism together with the magnon diffusion model. Using a two-dimensional finite element model (2D-FEM), we estimated the bulk spin Seebeck coefficient of YIG at room temperature. The quantitative disagreement between the experimental and modeled results indicates more complex processes going on in addition to magnon diffusion and relaxation, especially close to the contacts.
△ Less
Submitted 3 August, 2016;
originally announced August 2016.
-
Temperature dependence of the magnon spin diffusion length and magnon spin conductivity in the magnetic insulator yttrium iron garnet
Authors:
Ludo J. Cornelissen,
Bart J. van Wees
Abstract:
We present a systematic study of the temperature dependence of diffusive magnon spin transport, using a non-local device geometry. In our measurements, we detect spin signals arising from electrical and thermal magnon generation, and we directly extract the magnon spin diffusion length $λ_m$ for temperatures from 2 to 293 K. Values of $λ_m$ obtained from electrical and thermal generation agree wit…
▽ More
We present a systematic study of the temperature dependence of diffusive magnon spin transport, using a non-local device geometry. In our measurements, we detect spin signals arising from electrical and thermal magnon generation, and we directly extract the magnon spin diffusion length $λ_m$ for temperatures from 2 to 293 K. Values of $λ_m$ obtained from electrical and thermal generation agree within the experimental error, with $λ_m=9.6\pm0.9$ $μ$m at room temperature to a minimum of $λ_m=5.5\pm0.7$ $μ$m at 30 K. Using a 2D finite element model to fit the data obtained for electrical magnon generation we extract the magnon spin conductivity $σ_m$ as a function of temperature, which is reduced from $σ_m=5.1\pm0.2\times10^5$ S/m at room temperature to $σ_m=0.7\pm0.4\times10^5$ S/m at 5 K. Finally, we observe an enhancement of the signal originating from thermally generated magnons for low temperatures, where a maximum is observed around $T=7$ K. An explanation for this low temperature enhancement is however still missing and requires additional investigations.
△ Less
Submitted 6 July, 2016;
originally announced July 2016.
-
Magnon spin transport driven by the magnon chemical potential in a magnetic insulator
Authors:
Ludo J. Cornelissen,
Kevin J. H. Peters,
Rembert A. Duine,
Gerrit E. W. Bauer,
Bart J. van Wees
Abstract:
We develop a linear-response transport theory of diffusive spin and heat transport by magnons in magnetic insulators with metallic contacts. The magnons are described by a position dependent temperature and chemical potential that are governed by diffusion equations with characteristic relaxation lengths. Proceeding from a linearized Boltzmann equation, we derive expressions for length scales and…
▽ More
We develop a linear-response transport theory of diffusive spin and heat transport by magnons in magnetic insulators with metallic contacts. The magnons are described by a position dependent temperature and chemical potential that are governed by diffusion equations with characteristic relaxation lengths. Proceeding from a linearized Boltzmann equation, we derive expressions for length scales and transport coefficients. For yttrium iron garnet (YIG) at room temperature we find that long-range transport is dominated by the magnon chemical potential. We compare the model's results with recent experiments on YIG with Pt contacts [L.J. Cornelissen, et al., Nat. Phys. 11, 1022 (2015)] and extract a magnon spin conductivity of $σ_{m}=5\times10^{5}$ S/m. Our results for the spin Seebeck coefficient in YIG agree with published experiments. We conclude that the magnon chemical potential is an essential ingredient for energy and spin transport in magnetic insulators.
△ Less
Submitted 13 April, 2016;
originally announced April 2016.
-
Magnetic field dependence of the magnon spin diffusion length in the magnetic insulator yttrium iron garnet
Authors:
L. J. Cornelissen,
B. J. van Wees
Abstract:
We investigated the effect of an external magnetic field on the diffusive spin transport by magnons in the magnetic insulator yttrium iron garnet (YIG), using a non-local magnon transport measurement geometry. We observed a decrease in magnon spin diffusion length $λ_m$ for increasing field strengths, where $λ_m$ is reduced from 9.6$\pm1.2$ $μ$m at 10 mT to 4.2$\pm0.6$ $μ$m at 3.5 T at room temper…
▽ More
We investigated the effect of an external magnetic field on the diffusive spin transport by magnons in the magnetic insulator yttrium iron garnet (YIG), using a non-local magnon transport measurement geometry. We observed a decrease in magnon spin diffusion length $λ_m$ for increasing field strengths, where $λ_m$ is reduced from 9.6$\pm1.2$ $μ$m at 10 mT to 4.2$\pm0.6$ $μ$m at 3.5 T at room temperature. In addition, we find that there must be at least one additional transport parameter that depends on the external magnetic field. Our results do not allow us to unambiguously determine whether this is the magnon equilibrium density or the magnon diffusion constant. These results are significant for experiments in the more conventional longitudinal spin Seebeck geometry, since the magnon spin diffusion length sets the length scale for the spin Seebeck effect as well and is relevant for its understanding.
△ Less
Submitted 4 December, 2015;
originally announced December 2015.
-
Long distance transport of magnon spin information in a magnetic insulator at room temperature
Authors:
L. J. Cornelissen,
J. Liu,
R. A. Duine,
J. Ben Youssef,
B. J. Van Wees
Abstract:
The transport of spin information has been studied in various materials, such as metals, semiconductors and graphene. In these materials, spin is transported by diffusion of conduction electrons. Here we study the diffusion and relaxation of spin in a magnetic insulator, where the large bandgap prohibits the motion of electrons. Spin can still be transported, however, through the diffusion of non-…
▽ More
The transport of spin information has been studied in various materials, such as metals, semiconductors and graphene. In these materials, spin is transported by diffusion of conduction electrons. Here we study the diffusion and relaxation of spin in a magnetic insulator, where the large bandgap prohibits the motion of electrons. Spin can still be transported, however, through the diffusion of non-equilibrium magnons, the quanta of spin wave excitations in magnetically ordered materials. Here we show experimentally that these magnons can be excited and detected fully electrically in linear response, and can transport spin angular momentum through the magnetic insulator yttrium iron garnet (YIG) over distances as large as 40 micrometer. We identify two transport regimes: the diffusion limited regime for distances shorter than the magnon relaxation length, and the relaxation limited regime for larger distances. With a model similar to the diffusion-relaxation model for electron spin transport in (semi)conducting materials, we extract the magnon relaxation length lambda = 9.4 micrometer in a 200 nm thin YIG film at room temperature.
△ Less
Submitted 23 May, 2015;
originally announced May 2015.
