Spintronics stretches its arms to lasers

Three-dimensional TVs are becoming a hot new gadget in today’s electronics market. In order to create the 3D image, two separate images with orthogonal polarizations need to be delivered from a single display panel. Currently, this is achieved by dividing…

Three-dimensional TVs are becoming a hot new gadget in today’s electronics market. In order to create the 3D image, two separate images with orthogonal polarizations need to be delivered from a single display panel. Currently, this is achieved by dividing the pixels into two groups, so that each emits light for one of the images. As a result, each of the two images ends up with a diminished resolution. If we could use a single light source unit to emit light with two orthogonal polarizations, the image resolution would be greatly improved. An example of such a light source is the spin laser.

Our work is on the theory of spin lasers. The spin laser is enabled by the emerging field of spintronics, which interlinks electronic charge and spin properties. Current applications of spintronics are mostly centered around magnetic storage and information sensing, and focus on advances in metallic systems, such as tunnelling magnetoresistance and giant magnetoresistance.1 The spin laser offers an application for spintronics that is based in optics, and promises to spur cross-fertilization between the two fields.

There are three major differences between spin lasers and their conventional counterparts. First, in spin lasers, injected carriers are spin polarized. (As the hole spin decays much faster than electron spin,1 it is accurate to consider electrons as spin carriers.) We consider two magnetic contacts to achieve this (see Figure 1). The ratio of the currents through each contact determines the injection polarization. Alternatively, spin-polarized carriers can be injected optically, using circularly polarized light. The second difference is that the light emitted from spin lasers is circularly polarized due to the spin-polarized carriers. When an electron recombines with a hole in accordance with optical selection rules,3 the electron’s spin orientation determines the helicity (circular polarization) of the emitted photon, such that the total angular momentum is conserved. The output polarization can be controlled by adjusting either the injection polarization or the intensity.4 Third, there are two lasing thresholds in spin lasers: each spin feeds one corresponding mode (polarization) and the imbalance of spin-up and spin-down carrier injection leads to two separate thresholds for majority and minority spin carriers. According to our calculations, the two thresholds delimit three operational regimes (see Figure 2), while conventional lasers have only on and off regimes. Depending upon the strength of injection, a spin laser operates in off, full- or mixed-polarization regimes. It is especially interesting that between the two thresholds the emitted light is fully polarized, even with partially spin-polarized injection (0<|PJ|<1). This makes the spin laser a very good candidate for a spin-amplifier.
Figure 2. Emitted light intensity vs. injection intensity for spin lasers (PJ≠0.5) and conventional lasers (PJ=0, inset), where PJ is injection polarization. JT1 and JT2 indicate the two lasing thresholds of the spin laser, delineating the laser’s three operating regimes. +/−: Right/left circular polarization. Adapted from a figure published elsewhere.5

Operations of spin lasers in steady-state—with enhanced light emission, and spin filtering and amplification—have successfully been demonstrated.6–8 But the most interesting applications of spin lasers lie perhaps in their dynamic operation, where we expect that the spin lasers could outperform their conventional counterparts in two key categories: modulation bandwidth and chirp.5, 9 In spin lasers there are two basic methods of modulation: amplitude modulation (AM), and polarization modulation (PM). While AM changes injection intensity, PM modifies polarization at a fixed injection intensity. In both cases, bandwidths (the range where the frequency response is above -3dB) are enhanced with increased polarization (see Figure 3).
Figure 3. Frequency response with different polarizations. Infinite spin relaxation time (τs) is assumed except for the green curve. Adapted from a figure published elsewhere.9

Another advantage of spin lasers is that they can reduce frequency chirp. Generally, chirp is a parasitic frequency modification, and is due to dynamic change of the carrier-induced refractive index in the resonant cavity. In an ideal case, such as in a vacuum, that gives no chirp, there are two sidebands in the vicinity of the main peak corresponding to the resonance frequency for lasing (see Figure 4). Once the chirp is switched on, the spectrum is modified. The signal distortion can be reduced by injecting spin-polarized electrons. Spin lasers offer a relatively easy way to reduce chirp compared to other conventional methods,10 but the reduction depends on several factors such as injection intensity, modulation frequency and spin-resolved refractive index.5
Figure 4. Broadened spectra of electric fields, , of a laser with optical frequency ω0, where ωm is modulation frequency. Conventional lasers with (dotted line) and without chirp (solid line), and a spin laser (dashed line) with spin injection are shown. Adapted from a figure published elsewhere.5

What else may lie in store for spin lasers? Modern computers generate lots of excess heat. Contrary to popular belief, the main limitation in making them more energy efficient is not necessarily their microprocessors, but the way units communicate with each other.11 Conventional metallic interconnects are unable to keep up with the ever-growing demand for rapid data transfer within multi-core computer architectures of shrinking dimensions. Optical interconnects are a particularly promising platform for high-performance interconnects11 with VCSELs potentially playing an important role.12 We expect that the predicted advantages of spin lasers and recent experimental advances in their dynamic operation,13, 14 could be directly used to enable spin-enhanced optical interconnects. In fact, the polarization modulation discussed for spin lasers9 has already spurred novel schemes for spin interconnects.15 Currently, we are interested in studying spin dynamics, which is a very important (but largely missing) piece of the puzzle in understanding spin lasers and efficient spin modulation. Our focus is also on exploring schemes for robust electrical spin injection at room temperature that, combined with spin modulation, may bring higher resolution 3D TVs of the future one step closer to reality.