United States

Ferromagnetic semiconductors (FMSs) which show both ferromagnetic and semiconducting characteristics are promising materials for spintronics applications. Recently, we have successfully grown Fe-doped FMSs; p-type (Ga,Fe)Sb with Curie temperature (TC) ≈ 340 K2 and n-type (In,Fe)Sb with TC ≈ 335 K3, which are promising for devices operating at room temperature. One of the important steps towards practical applications is to realize spin-valve effects in heterostructures containing Fe-doped FMSs. In this work, we demonstrate spin-valve magnetoresistance (MR) in all-semiconductor heterostructures containing Fe-doped FMS (Ga,Fe)Sb layers with TC ≈ 260 K. The samples examined here consist of (Ga0.75,Fe0.25)Sb (40 nm) / InAs (thickness tInAs = 0, 3, 6, 9 nm) / (Ga0.8,Fe0.2)Sb (40 nm) grown by low temperature molecular beam epitaxy (LT-MBE) [Fig. 1(a) and (b)]. In these structures, electrical current flows mainly in InAs due to the high resistivity of (Ga,Fe)Sb layers at low temperature (≤ 20 K)4. A clear spin-valve effect with an MR ratio of ~ 1.6% with an open minor loop is observed in the structure with tInAs = 3 nm at 3.7 K [Fig. 2], whose peaks (≈ ±0.1 T) correspond to the coercive forces in the magnetization characteristics of the (Ga,Fe)Sb layers obtained with superconducting quantum interference device (SQUID) magnetometry and anomalous Hall effect (AHE) measurements. We found that the spin-valve MR (MRSV) increases (from 0.03 to 1.6%) with decreasing tInAs (from 9 to 3 nm), which reflects the enhancement of spin-dependent scattering at the InAs/(Ga,Fe)Sb interfaces. This demonstration is the first important step towards device application of these promising high-TC FMSs.References: [1] L. Chen et al, Nano Lett., Vol. 11, p.2584 (2011). [2] N. T. Tu, et al, Appl. Phys. Lett., Vol. 108, p.192401(2016). [3] N. T. Tu, et al, Appl. Phys. Express, Vol. 11, p.063005(2018). [4] N. T. Tu, et al., Phys. Rev. B, Vol. 92, p.144403 (2015). [5] K. Sriharsha, et al., APL Mater., Vol. 7, p.021105 (2019).

MMM 2020

Current-in-plane spin-valve magnetoresistance in ferromagnetic semiconductor (Ga,Fe)Sb heterostructures with high Curie temperature

The 65th Annual Conference on Magnetism and Magnetic Materials (MMM 2020) was held in a virtual form. The COVID-19 pandemic, with its associated uncertainty about international travel and the ability to gather in large groups, meant that the Steering Committee decided to cancel the face-to-face meeting that had been planned in Palm Beach, Florida November 2-6, 2020.

We were reluctant to give up the benefits of meeting face-to-face, and know that shifting to a virtual conference is a significant change for this community. Nevertheless, we are first and foremost concerned about the health and safety of all our Conference participants and their loved ones on returning home after travelling.

The MMM 2020 Steering Committee was working exceptionally hard, along with our sponsors, the IEEE Magnetics Society and AIP Publishing, to make this event serve the needs of our participants, who live in time zones all around the globe, in the best possible way. 

You need a ticket to explore this event.

If you have already registered for MMM 2020, and you have already created an account, then you
just need to login (above).

If you have already registered for MMM 2020, but you have not yet created an account, then you
need to click on &quot;create an account&quot; (above) and then login.

Forgot your password? Go to
https://www.underline.io/log-in

If you have not yet registered for MMM 2020 but would like to do so, go to
https://magnetism.org/register

Still need help? Contact our team!


technical paper

We investigate the effect of the distribution of pinning sites on the magnetization behavior in systems with Dzyaloshinskii–Moriya interaction (DMI). We consider the standard classical spin Hamiltonian with Heisenberg exchange, DMI, uniaxial anisotropy statistically varying across the lattice sites, and the Zeeman energy. The statistical variation of the anisotropy constants follows from a spatial Gaussian distribution with non-zero mean and standard deviation σ, and serves as a way to model the structural disorder in the sample. First, we discuss the reduction of the model to a lattice resolved mean-field theory and the development of energy minimisation algorithm for solving it. This approach allows to compute systematically and efficiently the magnetisation versus field, M(H), dependence for variable σ and temperature, and can be used for computing qualitative thermodynamic phase diagrams to explore material behaviour and to guide the computationally costly Monte-Carlo simulations. We show that as the σ increases, relative to the strength of exchange interaction and DMI, the nature of the reversal modes observed along a typical M(H) hysteresis loop changes in a certain temperature window. Namely, in ‘clean’ systems with small σ, the reversal proceeds first through the appearance of skyrmion lattices at low external fields, while in ‘dirty’ systems with large σ the reversal is through the nucleation of individual or small groups of skyrmions. We systematically quantify this effect and discuss its broader implications for applications.

The effects of disorder on hysteresis loops in chiral magnets

Threshold logic [1] efficiently implements complex logical functions within a single gate, but its widespread adoption has been impeded by insufficient nonlinearity in conventional transistors. The recent experimental demonstration of cascaded logic gates with current-driven magnetic domain walls [2] operates based on the nonlinearity intrinsic to anisotropic systems, thus naturally enabling threshold logic gates. However, these logic gates are described in [2] as reconfigurable NAND and NOR gates, rather than as three-input minority threshold logic gates. We therefore propose a novel domain wall threshold logic (DWTL) paradigm that enables circuits that are simpler and more efficient than can be achieved with the reconfigurable gates of [2]. Similar to [2], magnetic domains magnetized in the +z (-z) direction represent logic ‘1’ (‘0’), and each in-plane region causes magnetization reversal due to chiral coupling resulting from the interfacial DMI. As demonstrated in [2], DWTL gates produce an output domain that is the inversion of the majority of the inputs. That is, this logic gate performs the minority threshold function, which is significantly more logically expressive than NAND or NOR. By leveraging this minority gate behavior, the one-bit full adder of Fig. 1 can be realized with three logic gates, 80% fewer than the 15 required in [2]. We also propose DWTL functions with logical expressivity increased further by using more than three inputs, and with inputs that have non-uniform impacts on the output magnetization. As shown in Fig. 2a, a four-input DWTL gate can be realized in which one input has twice the width of the other three; therefore, this wider input will have roughly twice the impact on the output magnetization. DWTL gates with greater logical expressivity further reduce the number of devices required to perform logical operations, such as in the three-bit DWTL full adder of Fig. 2b that requires only two gates per bit (87% reduction).References: [1] D. Hampel and R. O. Winder, IEEE Spectrum, Vol. 8, pp. 32-39 (1971) [2] Z. Luo, A. Hrabec, T. P. Dao, et al., Nature, Vol 579, pp. 214-218 (2020)

Threshold Logic with Current-Driven Magnetic Domain Walls

The understanding of magnetization reversal in STTMRAM cells is of considerable interest. We use time-resolved conductance measurement, micromagnetic simulations and modeling to study it. The samples are fabricated from proprietary supersoft free layers of high volume and low moment. The experiments reveal the oscillatory nature of the conductivity signature of the switching (see figure). Modeling indicates that the conductance oscillations arise from the domain wall motion that happens in the so-called precessional regime above the Walker breakdown, when the domain wall propagation involves back-and-forth motions on top of a global drift. We first extend a collective coordinate model, taking into account the position of the wall and the tilt of the in-plane component of the magnetization within the wall. The mean wall velocity and its oscillations can be understood from the continuous exchange of energy between these two degrees of freedom, and the work supplied by the STT. In discs of diameters 70-100nm, experiments identify several distinct switching paths that occur stochastically and lead to distinct switching durations, while in disks of diameter 40 nm much less variability is found. Micromagnetics is used to understand these switching paths. For disks with 40 nm of diameter, the wall sweeps through the disk in an almost monotonous manner regardless of the wall state (Bloch, Néel or more complex). Conversely at larger diameters, the manner in which the wall crosses the disk depends on the wall state, and sets the variability of the transition time during the switching. The wall oscillation and its advance are essentially stopped if and only if a Bloch line is nucleated from the edge of the wall. This leads to switching times that are usually long compared to the standard deviation; this is one of the main mechanisms determining the write error rate in the ultrafast switching regime. Our model is finally used to design sample specifications with shall lead to rather reproducible propagation times, of direct interest for magnetic memories.

Walker-like domain wall motion in STTMRAM cells: evidence and consequences for the write error rate

Magnetoresistive random access memories (MRAM) based on spin-transfer torque are a scalable low-power nonvolatile memory solution being actively developed by industry, where a perpendicularly magnetized thin disk is reversed by a spin polarized current. During the switching process domain walls (DWs) can form [1-4] and understanding their dynamics in a disk geometry becomes of great interest. Here we show that DW surface tension always leads to oscillatory DW motion and instabilities that affect the switching dynamics. In a disk, a DW generally takes a circular form to minimize the sum of the surface and volume energies of the reversed domain, with the domain boundary perpendicular to the element boundary [1]. In a collective coordinate model the domain wall has two degrees of freedom, its position relative to the center of the disk q and the angle of the spins relative to the normal to the DW Φ. Here, Φ=0 is a Neel and Φ=90 deg. is a Bloch DW. We analyze the dynamics with micromagnetic and the collective coordinate model. Fig. 1a shows micromagnetic simulations of the evolution of the average perpendicular magnetization (m_z) with fixed initial DW position q, such that m_z<0, and different initial angles Φ, with no field or current applied. Whereas both Neel and Bloch DWs states relax towards a reversed state (m_z=-1), an intermediate initial angle DWs (Φ=45 deg) switches. Fig. 1b shows the (q,Φ) phase diagram indicating that the dynamics of domains close to the disk center is sensitive to small changes in the initial phase. Fig. 2 shows the same phase diagram as in Fig. 1b, now driven by a fixed spin-polarized J/Jc= 0.1. The results from both relaxation and switching simulations indicate that the DW phase Φ plays a critical role in the dynamics. A collective coordinate model will be presented which captures aspects of these instabilities and illustrates their physical origin.References: [1] G. Chaves-O'Flynn et al., Phys. Rev. Appl. 4, 024010 (2015) [2] T. Devolder et al., Phys. Rev. B 93, 024420 (2016) [3] P. Bouquin et al., Appl. Phys. Lett. 113, 222408 (2018) [4] J. B. Mohammadi et al., arXiv:2003.13875 (2020)

Micromagnetic instabilities associated with domain wall dynamics in perpendicular magnetized magnetic tunnel junction nanopillars

Magnetostatically coupled monodomain patterns have been studied for use as cellular automata and logic gates [1, 2]. Here the external magnetic field clocked switching is deterministic, and the magneto-optic Kerr (MOKE) effect microscopy or magnetic force microscopy (MFM) characterization only detects slow dynamics. We investigate similar systems, but with the use of spintronics for fast, all-electronic control and measurement. This is relevant both for low power magnetic logic, and for detection of fast dynamics in coupled superparamagnets that could be used in reservoir computing. We have demonstrated the ability to deterministically switch the state of a magnetic tunnel junction (MTJ) using spin orbit torque (SOT) [3], and this is used to switch the control dot at the beginning of the coupled patterns (Fig.1a). We have also detected magnetization changes in patterned nanodot MTJs using tunnel magnetoresistance (TMR) and a conductive atomic force microscopy (C-AFM) probe tip [4], and this is used to detect the dot magnetization at different locations. The dots were patterned by E-beam lithography from an MTJ thin film stack with a 3 nm thick CoFeB free layer. Fig.1b shows the schematic for a control dot and a neighboring secondary dot. Mumax3 simulations were conducted to match the magnetic field generated by the control dot with the coercive field of the secondary dot to facilitate the signal processing in magnetically coupled dots [5]. Fig.2 shows the magnetic field profile along +x direction. Inside the secondary dot, the magnetic field ranged from 2 to 15 mT, comparable to the 8 mT coercive field of the secondary dot, as shown in the Fig.2 inset. This ensures us that we could propagate a signal inside magnetically coupled dots. C-AFM measurements test the predicted switching and propagation along the nanodot chain of Fig.1.References: 1. R. P. Cowburn and M. E. Welland, Science 287 1466 (2000). 2. A. Imre, et al., Science 311 205 (2006). 3. M. Bapna, et al., Phys. Rev. Applied 10, 024013 (2018). 4. B.Parks, et al., Phys. Rev. Applied 13, 014063 (2020). 5. A. Vansteenkiste et al., AIP Adv. 4, 107133 (2014).

Spintronic control of magnetostatically coupled dot assemblies

In the racetrack or similar technologies, information (bit) stored in the direction of the magnetic moment of domains placed next to each other in a wire is shifted when electrical current passes along or parallel to the wire. While using electrons as ‘carrier’, these technologies sometimes lack the deterministic motion of the bits causing data integrity issues [1]. In this work, we are proposing a more deterministic, ‘carrier’ free transmission of data in an insulating ferromagnetic wire where the data is encoded in the oscillation frequencies of magnetic moments. Fig. 1 shows the schematic view of the device, including a thin transmission strip made of soft magnetic material. An Antiferromagnet (AF)/Ferromagnet (F) coupled perpendicular spin torque oscillator (STO) [2] is used to generate the full 360o in-plane magnetization rotation in a small confined region of the transmission strip that is also the free layer of the STO. This precession at STO end generates a traveling rotational magnetization wave along the wire, as seen in Fig. 1. While the spatial wavelength is increasing with distance, the rotational frequency along the wire remains constant, which is a key for our data transmission method. By changing the current amplitude according to the transmission bits, STO frequency can be tuned. Even at 800 nm away from the STO end (Fig. 2), the variation of the rotational frequency with time matches the variation of input STO frequency (or the input current pulse), affirming reliable data transmission. For this micromagnetic simulation, archetypal values for F such as α = 0.01 and Ms = 800 emu/cc are considered. 1 μm wire with a width of 10 nm and thickness of 2 nm is discretized using a grid size of 4nm x 2nm x 2nm.References: [1] Foerster M. et al., Domain Wall Memory Device (2016). In: Xu Y., Awschalom D., Nitta J. (eds) Handbook of Spintronics. Springer, Dordrecht [2] J.-G. Zhu, A. Shadman, IEEE Trans. Magn., vol. 55, no. 3, 2019.

Data Transmission in an Insulator utilizing Rotational Magnetization Wave

It has been shown both theoretically [1] and experimentally [2] that a spectrum analyzer based on an externally driven spin torque nano-oscillator (STNOs) can perform high-speed, wide-band spectrum analysis in a microwave frequency band. In such a spectrum analyzer, an STNO generates a signal whose frequency is tuned in a linear manner over a scanning bandwidth Δf. Theoretical estimations show that when employing an STNO in spectrum analysis, the following can be achieved: a Δf of 10 GHz, a linear scanning rate ρ=Δf/T (where T is the scanning period) of up to 2 GHz/ns, all in a frequency range of operation between 200 MHz and 70 GHz [3]. The fast speed of frequency analysis in a STNO-based spectrum analyzer can be achieved mainly due to the nano-scale size of the STNO, and the consequent ns characteristic scanning times. At a sufficiently large scanning rate ρ, the accuracy of the frequency determination ΔF (defined as the minimum separation required to distinguish two neighboring frequencies) of an STNO-based spectrum analyzer is fundamentally limited by the bandwidth theorem. In contrast, at low scanning rates, this accuracy is limited by the STNO generation linewidth [4], which is determined by the thermal (or random-walk) noise that is causing the jitter of the STNO generation frequency. Fig. 1 shows the frequency accuracy ΔF as a function of the scanning rate ρ for the STNO-based spectrum analyzer [2], where the gray solid line is the result of numerical modeling, dash-dotted line is a result of the experimental measurement [2], and dashed line shows the limit due to the “bandwidth” theorem. It is evident that at sufficiently large scanning rate, ΔF approaches the fundamental limit ΔF =1/T, whereas at smaller ρ the ΔF is limited by the magnitude of the STNO linewidth.References: [1] S. Louis, et al., Appl. Phys. Lett., 113, p. 112401, (2018). [2] A. Litvinenko, et al., “Experimental demonstration of a rapid sweep-tuned spectrum analyzer with temporal resolution based on a spin-torque nano-oscillator,” arXiv preprint arXiv:2004.03508, (2020). [3] S. Louis, “Development of a spectrum analyzer using a spin torque nano-oscillator,” Ph.D. dissertation, Department of Electrical and Computer Engineering, Oakland University, (2020). [4] Joo-Von Kim, et al., Phys. Rev. Lett. 100, 017207 (2008).

The influence of a thermal noise on the accuracy of a spectrum analyzer based on a spin torque nano-oscillator

The Meyer-Neldel (MN) rule, also known as entropy-enthalpy compensation, is a peculiar phenomenon that has been reported across diverse fields of the natural sciences, in semiconductors, chemical reactions, biology, etc [1]. In magnetism, it was very recently observed in the decay of the skyrmion lattice [2]. Consider a family of thermally activated processes whose rates k obey an Arrhenius-type law, i.e., k = f0 exp(-βΔE), in which ΔE is the internal energy barrier, β = (kBT)-1 is Boltzmann's factor, and f0 is the Arrhenius prefactor — the so-called "attempt frequency". If the Meyer-Neldel rule applies, the prefactor scales like an exponential of the energy barrier. In this work, we compute the mean waiting time between thermally activated magnetization reversals in a nanodisk, with parameters that resemble that of a free CoFeB layer, as used in magnetic random access memories [3]. We use two independent approaches, namely Langer's theory [4] and forward flux sampling [5]. By varying the perpendicular anisotropy, and the interfacial Dzyaloshinkii-Moriya interaction (DMI), we find that the Arrhenius prefactor can take extreme, seemingly non-physical values up to 1021 Hz, which is orders of magnitude beyond the typically assumed value of 109 Hz, and vary drastically as a function of material parameters. We show that it behaves like an exponential of the energy barrier, ΔE, which stems from a linear variation of the activation entropy, ΔS, with the energy barrier, thereby demonstrating a new case of the Meyer-Neldel rule (Fig. 1). Our results show that the Arrhenius prefactor is not an attempt frequency, because it also contains an important entropic contribution which cannot be neglected in complex systems. This suggests that modelling information retention times with a barrier-independent prefactor in magnetic storage elements is not justified [6].References: [1] A. Yelon, B. Movaghar, and R. Crandall, Reports on Progress in Physics, Vol. 69, p.1145 (2006) [2] J. Wild, T. N. Meier, S. Pollath et al., Science Advances, Vol. 3, p. e1701704 (2017) [3] J. Sampaio, A. Khvalkovskiy, M. Kuteifan et al., Applied Physics Letters, Vol. 108, p.112403 (2016) [4] J. S. Langer, Annals of Physics, Vol. 54, p. 258 (1969) [5] R. J. Allen, P. B. Warren, and P. R. ten Wolde, Physical Review Letters, Vol. 94, p. 018104 (2005) [6] L. Desplat and J.-V. Kim, submitted (2020)

Entropy-Reduced Retention Times in Magnetic Memory Elements: A Case of the Meyer-Neldel Compensation Rule

Current-perpendicular-to-plane (CPP) giant magnetoresistance (GMR) devices using Co-based Heusler alloys (Co2YZ, where Y is usually a transition metal and Z is a main group element) have attracted much interest as reading heads for next-generation hard disk drives [1-5]. To take full advantage of the half-metallic character of Co-based Heusler alloys, it is important to suppress the antisite of Co atoms, where Co atoms replace Y or Z site [6]. To do this increasing Y or Z composition in Co2YZ is highly effective, and we recently demonstrated the enhancement of MR ratios in off-stoichiometric Mn-rich Co2MnSi-based MTJs [7] and CPP-GMR devices [4]. In this study we systematically investigated the Ge-composition dependence of MR ratios in Co2Fe(Ga,Ge) (CFGG)-based CPP-GMR devices, and achieved the highest MR ratios of 88% at room temperature. We prepared CFGG-based CPP-GMR devices with various Ge composition α in Co2Fe1.03Ga0.41Geα electrodes ranging from 0.56 to 1.25. The layer structures of the fabricated CPP spin valves were as follows: (from the substrate side) MgO (10 nm)/Co50Fe50 (10)/Ag (100)/CFGG lower electrode (10)/NiAl (0.21)/Ag (5)/NiAl (0.21)/CFGG upper electrode (8)/Ru (5), grown on MgO(001) substrates. Ultrathin (0.21 nm) NiAl layers were aimed to enhancing the MR ratio [3]. Just after the deposition of CFGG upper electrodes, the layer structure was in-situ annealed at 550°C. We fabricated CPP spin valves with the nominal junction sizes ranging from 70 × 120 nm to 400 × 640 nm by using EB lithography and Ar ion milling. The MR characteristics were measured using a dc four-probe method at 290 K. We observed the MR ratio enhanced from 21% to 88% as α increased from 0.56 (Ge-stoichiometric CFGG) to 1.25 (Ge-rich CFGG). To the best of our knowledge, the MR ratio of 88% is the highest value ever reported among CPP-CMR devices. This result suggests that Ge-rich CFGG is effective in improving the half-metallic character of CFGG possibly due to the suppression of Co antisite with the increase of Ge composition.References: [1] Y. Sakuraba et al., Appl. Phys. Lett. 101, 252408 (2012). [2] H. Narisawa et al., Appl. Phys. Express 8, 063008 (2015). [3] Y. Du et al., Appl. Phys. Lett. 107, 112405 (2015). [4] M. Inoue et al., Appl. Phys. Lett. 111, 082403 (2017). [5] J. W. Jung et al., Appl. Phys. Lett. 108, 102408 (2016). [6] S. Picozzi et al., Phys. Rev. B 69, 094423 (2004). [7] H.-x. Liu et al., Appl. Phys. Lett. 101, 132418 (2012).

Enhancement of magnetoresistance characteristics of Ge-rich Co2Fe(Ga,Ge)-based current-perpendicular-to-plane giant magnetoresistance devices

Development of complementary metal–oxide–semiconductor (CMOS) logic is expected to approach its fundamental limit as the scaling down of the CMOS technology is reaching an end. As a route to extend the technology roadmap beyond traditional CMOS logic, novel spin-based logic architectures are being developed to provide nonvolatile data retention, near-zero leakage, and scalability. In particular, architectures based on magnetic domain walls (DWs) take the advantage of the fast motion, high density, non-volatility and flexible design of DWs to process and store information [1,2]. Such schemes have so far relied on DW manipulation and clocking using an external magnetic field, which hinders their implementation in dense, large-scale chips. Here we demonstrate a method for performing all-electric logic operations and cascading using DW racetracks [3]. Our concept for the magnetic DW logic is based on efficient DW motion induced by spin-orbit torques (SOTs) [4] and recently developed chiral coupling between adjacent magnets where the magnetic anisotropy competes with the interfacial Dzyaloshinskii–Moriya interaction (DMI) [5,6]. First we created a DW inverter, the essential basic building block for all implementations of Boolean logic, by embedding a narrow in-plane magnetized region in a racetrack with out-of-plane magnetization and demonstrate the operation of the DW inversion on application of an electric current (Fig. 1a). Building on the principles used to construct the DW inverter, we then fabricated reconfigurable NAND and NOR logic gates, making our concept for current-driven DW logic functionally complete. Finally, we cascaded several NAND gates to build XOR and full adder gates, demonstrating electrical control of magnetic data and device interconnection in logic circuits (Fig. 1b). Our work provides a viable platform for scalable all-electric magnetic logic, paving the way for memory-in-logic applications.References: [1] Allwood, D. A. et al. Science 309, 1688–1692 (2005). [2] Parkin, S. & Yang, S.-H. Nature Nanotechnol. 10, 195–198 (2015). [3] Luo, Z. et al. Nature 579, 214-218 (2020). [4] Manchon, A. et al. Rev. Mod. Phys. 91, 035004 (2019). [5] Luo, Z. et al. Science 363, 1435 (2019). [6] Dao, P. D. et al. Nano Lett. 19, 5930-5937 (2019).

Premium content

Downloads

Next from MMM 2020

The effects of disorder on hysteresis loops in chiral magnets

Similar lecture

Properties of quaternary-alloy ferromagnetic semiconductor (In,Ga,Fe)Sb