Nov 29 2019
Researchers from MIT have developed an innovative circuit design that allows precise control of computing using magnetic waves, instead of using electricity. This achievement is a step toward viable magnetic-based devices, with the ability to compute far more efficiently compared to electronics.
An MIT-invented circuit uses only a nanometer-wide “magnetic domain wall” to modulate the phase and magnitude of a spin wave, which could enable practical magnetic-based computing—using little to no electricity. Image Credit: Image courtesy of the researchers, edited by MIT News.
Classical computers are dependent on enormous amounts of electricity to perform computing and store data. They also produce a lot of waste heat. In a quest for more efficient substitutes, scientists have started designing magnetic-based “spintronic” devices, which consume comparatively less electricity and produce almost no heat.
Spintronic devices make use of the “spin wave”—a quantum property of electrons—in magnetic materials that have a lattice structure. This method involves the modulation of the spin wave properties to generate some measurable output that can be correlated to computation.
To date, modulation of spin waves has necessitated injected electrical currents using heavy components that can lead to signal noise and effectively invalidate any intrinsic performance gains.
The researchers at MIT created a circuit architecture that modulates a passing spin wave by making use of only a nanometer-wide domain wall in layered nanofilms of magnetic material, without the need for any electrical current or extra components. The spin wave can in turn be adjusted to regulate the location of the wall, as required. This offers precise control of two varying spin wave states, corresponding to the 0s and 1s used in classical computing.
Going forward, spin wave pairs could be injected into the circuit via dual channels, tuned for different properties, and integrated to produce some measurable quantum interference—analogous to the use of photon wave interference for quantum computing. Scientists propose that interference-based spintronic devices, such as quantum computers, could perform highly complicated tasks that traditional computers find hard to perform.
People are beginning to look for computing beyond silicon. Wave computing is a promising alternative. By using this narrow domain wall, we can modulate the spin wave and create these two separate states, without any real energy costs. We just rely on spin waves and intrinsic magnetic material.
Luqiao Liu, Professor, Department of Electrical Engineering and Computer Science (EECS), MIT
Liu is the principal investigator of the Spintronic Material and Device Group in the Research Laboratory of Electronics. Collaborating with Liu on the study are Jiahao Han, Pengxiang Zhang, and Justin T. Hou, three graduate students in the Spintronic Material and Device Group; and EECS postdoc Saima A. Siddiqui.
Flipping Magnons
Spin waves are ripples of energy that have small wavelengths. Portions of the spin wave, which are typically the collective spin of several electrons, are known as magnons. Although magnons are not real particles, similar to individual electrons, it is possible to measure them similarly for computing applications.
In their study, the scientists made use of a tailored “magnetic domain wall,” a nanometer-sized barrier between two adjacent magnetic structures. A pattern of nickel/cobalt nanofilms—each with a thickness of a few atoms—was layered, which had some favorable magnetic properties with the potential to deal with a high volume of spin waves.
The researchers then positioned the wall in the middle of a magnetic material with a unique lattice structure, and integrated the system into a circuit.
Constant spin waves were excited in the material on one side of the circuit. When the wave passes through the wall, its magnons instantly spin in the opposite direction: Magnons located in the first region spin north, whereas those in the second region—past the wall—spin south. This leads to a dramatic shift in the phase (angle) of the wave and a slight drop in magnitude (power).
During the experiments, a separate antenna was positioned on the opposite side of the circuit, to detect and transmit an output signal. Results showed that at the output state, the input wave’s phase flipped by 180°. The magnitude of the wave—measured from the highest peak to the lowest—had also dropped considerably.
Adding Some Torque
Subsequently, the researchers found a mutual interaction between the domain wall and the spin wave, enabling them to efficiently switch between two states. If the domain wall is absent, the circuit would be magnetized evenly; in the presence of the domain wall, the circuit has a split, modulated wave.
They discovered that by controlling the spin wave, it would be possible to control the domain wall position. This is dependent on a phenomenon known as “spin-transfer torque,” which is when spinning electrons typically agitate a magnetic material to flip its magnetic orientation.
In this study, the power of the injected spin waves was increased to activate a specific spin of the magnons. This actually pulls the wall toward the increased wave source. Thus, the wall gets jammed beneath the antenna—effectively rendering it unable to modulate waves and ensuring even magnetization in this state.
The researchers used a special magnetic microscope to demonstrate that this technique leads to a micrometer-sized shift in the wall, which is sufficient to position it anywhere along the material block. Notably, a few years ago, the mechanism of magnon spin-transfer torque was hypothesized, but not proven.
There was good reason to think this would happen, But our experiments prove what will actually occur under these conditions.
Luqiao Liu, Professor, Department of Electrical Engineering and Computer Science (EECS), MIT
According to Liu, the entire circuit is similar to a water pipe. The valve (domain wall) regulates the way water (spin wave) flows through the pipe (material).
But you can also imagine making water pressure so high, it breaks the valve off and pushes it downstream. If we apply a strong enough spin wave, we can move the position of domain wall—except it moves slightly upstream, not downstream.
Luqiao Liu, Professor, Department of Electrical Engineering and Computer Science (EECS), MIT
Innovations such as these could facilitate viable wave-based computing for particular tasks, such as the signal-processing method known as “fast Fourier transform.” Going forward, the scientists intend to develop a working wave circuit that can perform basic computations.
Besides other things, they have to enhance materials, minimize possible signal noise, and further analyze how fast they can toggle between states by moving around the domain wall. “That’s next on our to-do list,” stated Liu.
Source: http://www.mit.edu/