Principles and advances in spin light-emitting diodes

Abstract

Spin light-emitting diodes (spin-LEDs), which converts carrier spin information to optical polarization, are important for modern display application and future technologies including spin-based electronic circuits and quantum information science. Spin-LEDs operate through spin-polarized carrier injection, directly converting spin angular momentum into photon helicity. Recent advances in materials and mechanism understanding have significantly improved spin injection efficiency and device performance. This review discusses the fundamental principles and operation mechanisms of spin-LEDs, where we examine conventional spin injection approaches and their constraints. We then highlight the emerging role of chiral-induced spin selectivity (CISS) in spin-LEDs, exploring how this novel mechanism could enhance circularly polarized electroluminescence efficiency and enable new device architectures. Finally, we provide perspectives on current challenges.

Introduction

Circularly polarized light (CPL) refers to photons in which the electric field vector rotates helically as it propagates, resulting in left- or right-handed circular polarization. Compared to other forms of light polarization, such as linear polarization, CPL offers greater potential due to the intrinsic angular momentum of photons, which enables the encoding of information beyond what is possible with linearly polarized light. The development of CPL sources plays an essential role in the modern display industry and future technologies including 3D display^1,2, quantum computing and sensing^3,4, information processing^5,6,7, and chiral catalysis^8,9.

Traditional approach for CPL generation mostly relies on a “top-down” method, where unpolarized photons are filtered and purified through optical components such as quarter-wave plates and polarizers. However, this method suffers from the inherent energy loss, complex infrastructure, and low sensitivity and resolution. Another strategy for CPL generation is through a “bottom-up” method, where materials or devices can directly produce CPL via photoexcitation or electrical excitation. Optical excitation gives rise to circularly polarized photoluminescence (CP-PL), while electrical excitation leads to circularly polarized electroluminescence (CP-EL). While substantial progress has been made in material design and mechanism understanding of CP-PL^10,11,12,13, the development of light-emitting diodes (LEDs) capable of efficiently generating CP-EL has lagged significantly.

In contrast to CP-PL, CP-EL enables direct electrical generation of CPL, eliminating the need for external optical components and reducing energy losses associated with intermediate conversion steps. By directly converting electrical energy into CPL, CP-EL can maximize the energy efficiency, with the potential to convert 100% excitons into usable CPL^14,15. Further, CP-EL allows for active electrical modulation of polarization states, making it a more practical choice for integrated photonic devices. These advantages make CP-EL a highly promising approach for real-world optoelectronic applications where tunability, scalability, and high efficiency are crucial.

There are generally two types of LEDs that generate CP-EL. One approach utilizes chiral materials as the emissive layer, where molecular chirality imprints the chirality into the excited state, leading to the CPL emission. LEDs that apply a chiral emissive layer to produce CP-EL are referred to as circularly polarized light-emitting diodes (CP-LEDs), or often circularly polarized organic light-emitting diodes (CP-OLEDs). This idea was initially explored by Peeters et al., who demonstrated the first CP-EL using a chiral polymer LED¹⁶, proving that molecular chirality alone could induce circular polarization in electroluminescence. Since then, significant progress has been made¹⁷.

The other approach employs spin-polarized carrier injection, where the angular momentum of injected carriers determines the polarization of electroluminescence. LEDs that rely on spin polarization injection to generate CP-EL are often referred to as spin-LEDs. These devices fall into the umbrella of spintronics as spin control and spin injection are crucial in their operation. In 1990, Datta and Das proposed a structure for injecting spin-polarized electrons into semiconductors¹⁸, laying the theoretical foundation for semiconductor spintronics. In 1999, Fiederling et al.¹⁹ demonstrated efficient spin injection from a dilute magnetic semiconductor (DMS) into a III-V semiconductor and successfully detected high circular polarization in the emitted light. Since then, significant progress has been made in optimizing spin injection efficiency, extending operational temperatures, and integrating new material platforms. Despite these advancements, conventional spin injection methods—such as ferromagnetic contacts—face challenges related to efficiency losses in spin injection²⁰.

More recently, Beard’s group pioneered a promising new direction of spin-LED research²¹, which relies on a new mechanism, i.e., chirality-induced spin selectivity (CISS), to enable spin-LED function. Unlike traditional spin injection by using ferromagnetic injectors, the CISS effect allows for efficient spin polarization and injection without the need for external magnetic fields or ferromagnetic electrodes²², showing the potential to revolutionize spin-based optoelectronics.

This review aims to discuss the fundamental mechanisms that operate spin-LED, highlighting the key difference between spin-LED and CP-LED. We also summarized recent developments of spin-LED, where we examine conventional spin injection approaches and their constraints. Finally, we highlight the emerging role of CISS in spin-LEDs, exploring how this novel mechanism could enhance CP-EL efficiency and enable new device architectures.

Principles of spin-LEDs

Although both CP-LED and spin-LED can generate CP-EL, they possess distinct fundamental principles and operational mechanisms. It is thus essential to first distinguish these two approaches, CP-LED vs. spin-LED.

CP-LEDs generate CPL using chiral emissive materials (Fig. 1a). The recombination of unpolarized charge carriers occurs within the chiral emissive layer (typically organic), where the intrinsic chirality of the material induces asymmetry in radiative transitions (Fig. 1b). This innate asymmetry gives rise to CPL. The degree of circular polarization of emitted light is quantified by the dissymmetry factor (g_CPL), which can be expressed as Eq.1²³.

$${g}_{{CPL}}=2\frac{I\left({\sigma }^{+}\right)-I\left({\sigma }^{-}\right)}{I\left({\sigma }^{+}\right)+I\left({\sigma }^{-}\right)}=\frac{4\cos {{\rm{\theta }}}_{{\rm{\mu }}m}\left|{{\boldsymbol{\mu }}}_{{\boldsymbol{ij}}}\right|\left|{m}_{{ij}}\right|}{{\left|{{\boldsymbol{\mu }}}_{{\boldsymbol{ij}}}\right|}^{2}+{\left|{{\boldsymbol{m}}}_{{\boldsymbol{ij}}}\right|}^{2}}$$

(1)

Fig. 1: Principles of circularly polarized LED (CP-LED) and spin-LEDs.

a Schematic of a CP-LED device. b Dissymmetry in intensity of R- L-CPL emitted during the transition between quantum state i and j. gCPL is determined by TEDM(μ') and TMDM (m') within the chiral emissive material. c Schematic of a Spin-LED device. Mechanism CPL generation under d HH-LH degeneracy and e HH-LH non-degeneracy in Spin-LEDs. The transition probability (oscillator strength) from conduction bands to HH bands are 3 times of that to LH bands. [b is adapted from ref. ²⁴. with permission].

Here, $I\left({\sigma }^{+}\right)$ and $I\left({{\rm{\sigma }}}^{-}\right)$ represents the intensities of right- and left- handed CPL. The value of g_CPL can further be related to the transition dipole moment of a molecule. The terms μ_ij and m_ij correspond to the transition electric dipole moment (TEDM) and transition magnetic dipole moment (TMDM) between states $\left|i\right\rangle$ (ground) and ${|j}{\rm{\rangle }}$ (excited). The term θ_μm is the angle between TEDM and TMDM. From Eq. 1, the CPL generation in CP-LEDs is entirely dependent on the transition moments (μ_ij and m_ij) and their spatial configuration within the chiral emissive layer (Fig. 1b)²⁴. Importantly, the handedness and g_CPL of CP-LEDs are independent of the spin polarization of charge carriers, relying solely on the chiral properties of the emissive material.

First reported in 1997 by Peeters et al.¹⁶, CP-LEDs have evolved into three main categories based on their emissive materials. Category 1: Chiral conjugated emissive polymers achieve CP-EL by incorporating chirality through side-chain modification^16,25,26, main-chain chirality²⁷, or chiral dopants²⁸, offering high g_CPL but limited to a 25% internal quantum efficiency (IQE) due to singlet-only emission. Category 2: Chiral metal complexes utilize heavy metals (e.g., Ir, Pt²⁹, Zn³⁰, Cu³¹) to enable near-unity intersystem crossing (ISC) via spin-orbit coupling (SOC), achieving 100% IQE but often exhibiting low g_CPL. Lanthanide metals (e.g., Eu) enhance g_CPL through f-f transitions but suffer from low quantum yields³². Category 3: Chiral thermally activated delayed fluorescence molecules achieve 100% IQE without heavy metals by converting triplet excitons to singlets via reverse ISC, offering a better balance between efficiency and g_CPL compared to the other two approaches^33,34,35.

Spin-LEDs, on the other hand, generate CPL by injecting spin-polarized charge carriers into an achiral emissive layer (Fig. 1c). In general, spin-LEDs offer greater tunability, most notably the ability to electronically modulate the polarization direction of CPL compared to CP-LEDs³⁶. Spin-LED normally consists of a ferromagnetic spin injector layer that can inject spin-polarized carriers and a semiconductor layer that can transfer the carrier spin angular momentum to photon angular momentum, i.e., CPL. Unlike CP-LEDs, spin-LEDs do not require chiral materials in the emissive layer. The helicity of emitted photons is directly determined by the spin state of recombining charge carriers. When electrons are injected from a ferromagnetic injector into a nonmagnetic semiconductor, they can retain their spin polarization over a certain distance (from 50 nm to 4 μm^37,38,39), provided that spin injection is efficient, spin transport occurs without significant depolarization, and spin lifetime is sufficiently long (from 0.1 to 20 ns^37,38,40).

It is important to note that not all semiconductors are capable of producing CP-EL when spin-polarized carriers are injected from spin injector contacts. In order to convert the spin angular momentum of carriers to photons, semiconductors in spin-LED need to possess spin-dependent optical transitions, which normally require a strong SOC and direct bandgap. Under the angular momentum conservation rule, the carrier spin information can then be transferred to the photon helicity. Below we use the traditional GaAs semiconductor as a model system to illustrate how spin polarized charge carriers can generate CP-EL. In GaAs with a zinc-blend structure, the conduction band is twofold degenerate at the center of the Brillouin zone, corresponding to spin-up and spin-down electrons (m_j = ±1/2, where m_j is magnetic angular momentum). The valence band is fourfold degenerate and consists of the heavy-hole (HH) and light-hole (LH) bands, which differ in their effective mass. Each valence band is twofold spin degenerate (m_j = ±3/2, ±1/2). Radiative interband transitions follow selection rule Δm_j = ±1⁴¹. The polarization and intensity of light emitted during these transitions can be understood using the correspondence principle, where the magnetic angular momentums associated with specific transitions determine the polarization characteristics of the emitted light. The resultant possible optical transistions⁴² are summarized below and in Fig. 1d.

• For transitions between the HH bands and the conduction bands, the recombination of a −1/2 electron with a −3/2 hole, and a +1/2 electron with a +3/2 hole, result in the light with opposite polarization directions (right hand σ⁺ and left hand σ^-, respectively) in the plane perpendicular to the wave vector k.

• In the case of the LH bands, transitions involving a −1/2 electron with a +1/2 hole, and a +1/2 electron with a −1/2 hole, also produce oppositely polarized light. However, transitions between electron and hole states with identical m_j (e.g., +1/2 with +1/2 or −1/2 with −1/2) are forbidden by optical selection rules.

For consistency, we also use the same parameter, g_CPEL, to define the quality of CP-EL. In spin-LEDs, we normally define the degree of circular polarization as DOCP, which relates with g_CPEL as,

$${g}_{{CPEL}}=2\times {|DOCP|}=2\frac{I\left({\sigma }^{+}\right)-I\left({\sigma }^{-}\right)}{I\left({\sigma }^{+}\right)+I\left({\sigma }^{-}\right)}$$

(2)

The DOCP value in spin-LED is directly related to the relative populations of the electron spin states and the degree of quantum confinement in semiconductors (if we exclude spin relaxation)⁴³.

$${|D}{OCP}|=\frac{I\left({\sigma }^{+}\right)-I\left({\sigma }^{-}\right)}{I\left({\sigma }^{+}\right)+I\left({\sigma }^{-}\right)}=\alpha \times \frac{n\downarrow -\,n\uparrow }{n\downarrow +\,n\uparrow }$$

(3)

Where, 0.5 ≤ α ≤ 1, $n\uparrow$ and $n\downarrow$ are the relative population of the electrons with different spin states m_j = +1/2 and m_j = −1/2, respectively. The spin populations satisfy the conditions $0\le {\rm{n}}\le 1$, and $n\uparrow +{n}\downarrow \,=\,1$. Therefore, the maximum |DOCP|_max equals to $\alpha$, if spin relaxation is not considered. In the case of weak confinement (Fig. 1d), where the HH and LH bands are degenerate in energy, α = 0.5 due to the 3:1 ratio of transition probabilities favoring the HH sub-band over the LH sub-band (Fig. 1d). The maximum circular polarization |DOCP|_max is 0.5 and the maximum attainable value of g_CPEL in this case is 1.

In situations with strong quantum confinement, for example quantum wells or quantum dots, the HH-LH band degeneracy is lifted (Fig. 1e)⁴⁴. The LH bands are at lower energy levels. In this case, only the HH bands participate in the radiative recombination process (Fig. 1e), and α is estimated to be 1. In this case, the maximum circular polarization |DOCP| is 1 and the maximum attainable value of g_CPEL is 2 under the HH-LH non-degenerate condition⁴³.

When spin relaxation is considered, the |DOCP| is also related to the spin injection efficiency (P_inj) through^45,46

$$|{\bf{DOCP}}|=\alpha \times F\times {P}_{{inj}}$$

(4)

where F is a renormalization factor that quantifies the fraction of spin-polarized carriers that recombine radiatively before losing their spin polarization. F is defined as F = 1/(1 + τ/τ_s), where τ is the total carrier lifetime and τ_s is the spin lifetime of the semiconductors. The spin injection efficiency P_inj depends on the interface between the spin injector and the semiconductor. To achieve a high |DOCP|, it is critical to obtain a short total carrier lifetime, a long spin lifetime and a high spin injection efficiency. Below we introduce the most common strategies for spin polarization injection in spin-LEDs.

Mechanism of spin polarization in spin injectors

Dilute magnetic semiconductor (DMS)

DMS are semiconductors containing a small amount of magnetic dopants, such as Mn²⁺ (Fig. 2a)⁴⁷. The low (dilute) concentration of magnetic dopant can introduce ferromagnetism in DMS⁴⁸, which arise from half-filled 3d⁵ shell of localized Mn²⁺ ions, and their interaction with the conduction electrons of the semiconductor through exchange interactions. Well-known examples of DMS include Ga_1-xMn_xAs⁴⁹, Cd_1-xMn_xSe⁵⁰, where x is generally in the range of 0.01–0.1. Under an applied magnetic field, the conduction and valence bands of DMS split into Landau levels, which further split into spin-dependent sublevels, resulting in a large energy difference between spin-up and spin-down states (i.e., Giant Zeeman Effect). At sufficiently low temperatures, the energy splitting is so significant that only one spin state is populated, leading to a spin-polarized band structure (Fig. 2b)⁵¹. By injecting the spin-polarized electrons into a LED, these spin-polarized electrons can be utilized to emit CP-EL^19,52. The highest reported DOCP was achieved using a ZnMnSe spin injector by Asshoff et al., who obtained a DOCP of 96% at 5 K under an external magnetic field of 6 T⁵³. One of the major limitations of DMS materials is their low Curie temperature (Tc), which makes them unsuitable for practical applications at room temperature. Below Tc, the DMS remains in a ferromagnetic phase, enabling efficient spin polarization and spin injection in spin-LEDs. However, near or above Tc, other magnetic phases such as paramagnetic, antiferromagnetic, or chiral spin states may emerge. These phases lead to the injection of carriers with non-uniform or random spin polarization, ultimately resulting in an average circular polarization that approaches zero. For most traditional DMS materials, the Curie temperature is typically below 50 K. Although recent advancements have pushed the Curie temperature as high as 260 K in a new type of DMS (e.g., (Ba, K)(Zn, Mn)₂As₂⁵⁴), this is still significantly lower than room temperature, limiting their usability in real-world devices.

Fig. 2: Spin polarization in dilute magnetic semiconductor (DMS).

a Schematic representation of a DMS. A non-magnetic semiconductor is doped with magnetic ions, introducing localized spin-polarized states. The exchange interactions between these spins induce ferromagnetic ordering, transforming the material into a DMS with tunable magnetic properties. b Representation of spin-polarized carrier transport in a DMS device, spin splitting in the band structure leads to asymmetric spin population under low temperature.

Ferromagnetic metals (FM)

Transition metal ferromagnets such as Fe^46,55 (with ~45% spin polarization) and Co⁵⁶ (~35–50%) offer high Curie temperatures (Tc > 1000 K)⁵⁷, making them well-suited for practical applications. CoFe alloys^58,59, and CoFeB^{36,45,60,61,62,63,64,65,66}, which exhibit enhanced spin polarization (~50–70%)⁶⁷, are widely used as spin injectors in spin-LEDs. Heusler alloys such as Co₂MnGe⁶⁸, Co₂MnGa⁶⁹, and Co₂FeSi⁷⁰ are theoretically predicted to achieve 100% spin polarization. However, the experimentally measured DOCP from Heusler-based spin-LEDs remain lower than those achieved with transition metal-based counterparts. Up to now, the highest room temperature DOCP is reported to be 36% with CoFeB spin injector by Dainone et al.³⁶. The pursuit of high spin polarization in FM is also closely linked to the development of two fundamental phenomena in spintronics.

Giant magnetoresistance (GMR)

Giant Magnetoresistance (GMR) is a phenomenon observed in metallic multilayers composed of alternating ferromagnetic and nonmagnetic layers. The effect arises from spin-dependent scattering of conduction electrons, which is influenced by the relative alignment of the magnetizations of the ferromagnetic layers. When the magnetizations are parallel, the majority-spin electrons encounter fewer scattering events, leading to a lower resistance. Conversely, in the antiparallel alignment, scattering increases, resulting in a higher resistance (Fig. 3)⁷¹.

Fig. 3: Spin transport in multilayered structures.

Schematic illustration of spin-dependent electron transport in a ferromagnetic (F) and non-magnetic (NM) multilayered structure under parallel and antiparallel configurations. In the parallel configuration (left), the magnetizations of the F layers are aligned, allowing spin-polarized electrons to travel with minimal scattering, resulting in lower resistance. In the antiparallel configuration (right), alternating magnetization directions cause increased spin scattering, leading to higher resistance.

Spin-dependent scattering in ferromagnetic materials is governed by the asymmetry in the DOS at the Fermi level for spin-up and spin-down electrons. Majority-spin electrons, which correspond to the spin state with a higher DOS, experience a weaker scattering and show longer mean free paths. Minority-spin electrons, with lower DOS, encounter a stronger scattering, leading to shorter mean free paths. The spin dependent differential scattering effect enables significant changes in spin dependent electrical resistance.

Tunneling magnetoresistance (TMR)

Tunneling Magnetoresistance (TMR), first demonstrated by Julliere in 1975⁷² and observed at room temperature by Miyazaki and Moodera^71,73, occurs in magnetic tunnel junctions (MTJs). An MTJ consists of two ferromagnetic layers separated by a thin insulating barrier. The resistance of this structure depends on the relative alignment of the magnetizations in the ferromagnetic layers: resistance is lower when the magnetizations are aligned parallelly and higher when they are antiparallelly.

The mechanism of TMR is governed by spin polarization at barrier/ferromagnet interface, where spin-up and spin-down electrons at the Fermi level have unequal densities of states⁷⁴. This imbalance creates a preferential tunneling for one spin orientation. When a voltage is applied across the MTJ, electrons tunnel through the insulating barrier via quantum mechanical effects. In the parallel configuration, the alignment of the magnetizations of both ferromagnetic layers enhances the available density of states (DOS) for majority-spin electrons (occupied and unoccupied) in each layer, favoring efficient tunneling and resulting in a lower resistance (Fig. 4a). In the antiparallel configuration, majority-spin electrons in one layer correspond to minority-spin electrons in the other, leading to reduced tunneling probabilities and a higher resistance (Fig. 4b)^71,73,75,76.

Fig. 4: Schematic representation of spin-dependent tunneling in an magnetic tunnel junctions (MTJ).

a Parallel configuration, where the magnetizations of the two FM layers (FM1 and FM2) are aligned, leading to lower resistance and higher tunneling probability for majority spin carriers. b Antiparallel configuration, where the magnetizations are opposite, increasing resistance and reducing tunneling efficiency due to the density of states (DOS) mismatch. The energy band diagrams below depict the spin-dependent DOS and tunneling probabilities for each configuration.

Spin Hall effect (SHE)

Despite the success of FM in polarizing electron spins, these injectors face significant challenges for practical applications in thin, portable spin-LEDs. Their reliance on external magnetic fields to manipulate the magnetization of ferromagnetic layers limits the utility in applications. A promising alternative is the Spin Hall Effect (SHE), first proposed by Dyakonov and Perel in 1971⁷⁷. They predicted that in materials with strong relativistic SOC, an unpolarized electrical current could generate a transverse spin current, enabling spin polarization without the need for an external magnetic field (Fig. 5a). There are two main mechanisms of the SHE, i.e., Extrinsic SHE and Intrinsic SHE. The extrinsic SHE arises from Mott scattering, where electrons with opposite spins scatter in opposite directions upon colliding with impurities in the material⁷⁷. This spin-dependent scattering leads to the spin separation. The intrinsic SHE originates from the band structure of materials and is most prominent in systems with strong SOC and broken structural inversion symmetry⁷⁸. In such systems, the Rashba spin-orbit interaction arises due to the lack of structural inversion symmetry, and induces spin-momentum locking, where the electron’s spin orientation is always aligned perpendicular to its momentum.

Fig. 5: Spin injection and CPL emission using SHE.

a Schematic representation of the SHE, where a charge current (J_c) induces a transverse spin current (J_s), leading to spin accumulation at the edges of the material. b Scanning electron microscope (SEM) image of the SHE spin-LED device, showing a 1.5 μm channel with two light-emitting diodes (LED 1 and LED 2) used for CPL emission. A polarization current (I_p) is applied to control spin injection. c CP-EL spectra from LED 1 and LED 2 for opposite polarization currents (+I_p and −I_p), demonstrating spin-dependent light emission. [b, c are adapted from ref. ⁸⁴. with permission].

When an electrical current is applied to the material, it drives the electrons through the system. As a result of spin-momentum locking, electrons with opposite spins are naturally deflected in the opposite directions as they move. This behavior is analogous to the Magnus force in classical mechanics, where the spin of a ball distorts its trajectory. In the SHE, this deflection causes spin-up electrons to accumulate on one edge of the material and spin-down electrons on the opposite edge, resulting in spin polarization along the lateral edges of the material (i.e., a spin current perpendicular to charge current). Traditionally, the SHE has been observed primarily in inorganic materials, as it typically requires strong SOC associated with heavy atoms. However, recent experimental and theoretical studies have predicted that the effective SOC in organic molecules can be enhanced through structural modifications, such as incorporating metal atoms like Cu and Al into the molecular framework^79,80,81,82.

A major advantage of the SHE is that it does not require a ferromagnetic material to generate spin polarization, making it particularly promising for practical applications in spintronic devices. In 2005, Wunderlich et al.⁶⁷ and Nomura et al.⁶⁶ demonstrated the intrinsic SHE by constructing a micro-LED with a coplanar p-n junction in (Al,Ga)As/GaAs heterostructure incorporating a two-dimensional hole gases (2DHG, Fig. 5b)^83,84. A p-channel current (I_p = 100 μA) applied along the x-direction deflected carriers with opposite spins to opposite edges of the 2DHG, creating spin polarization optically detected as CPL emitted from biased LEDs. The CPL’s polarization flipped when I_p was reversed. The device achieved 1% DOCP at room temperature without a magnetic field (Fig. 5c), but this remains the only reported use of SHE in spin-LEDs.

Electronically modulated polarization switching

One key advantage of spin-LEDs over CP-LEDs is the ability to electrically control the helicity of CPL. In spin-LEDs, polarization is determined by the spin states of injected carriers, rather than by the intrinsic chirality of the emissive layer, which is fixed and cannot be easily modulated externally. This capability is particularly significant for information processing, as conventional data transmission methods rely on intensity modulation—encoding information through the binary ON/OFF states of optical and electrical signals. In contrast, utilizing angular momentum modulation—switching between σ⁺/σ⁻ CPL states for information encoding presents the potential for substantially higher data transmission rates compared to intensity modulation. This advantage arises from the inherently faster transition dynamics of polarization states compared to intensity modulation⁸⁵. Notably, birefringent spin lasers have already demonstrated polarization oscillation at speeds 10 times faster than those achieved through intensity modulation⁸⁶.

The feasibility of dynamically controlling CPL polarization in spin-LEDs was first demonstrated by Nishizawa et al.⁸⁷, who designed a device incorporating two electrodes with opposite spin polarizations—one injecting spin-up carriers and the other spin-down carriers—into a common emissive layer (Fig. 6a). By applying an alternating current between these electrodes, they successfully achieved CPL polarization switching at frequencies of up to 1 kHz.

Fig. 6: Electronically modulated polarization switching in spin-LEDs.

a Schematic illustration of spin-LED device by Nishizawa et al. with dual spin-polarized electrodes enabling electrical control of CPL. By applying an alternating current between the two oppositely magnetized electrodes, the CPL helicity can be dynamically switched. b Schematic illustration of electrical control of CPL via the spin Hall effect. A pulsed current in the Hall-bar-shaped spin injector alters the magnetization direction, which in turn determines the polarization of the light emitted by the spin LED. [a is replotted from data in ref. ⁸⁷].

However, this approach is inherently complex due to the requirement for two distinct spin-polarized electrodes and an alternating current drive. More recently, in 2024, Dainone et al. demonstrated an alternative method for electrically switching CP-EL polarization via magnetization control³⁶. Their approach leverages the spin Hall effect, wherein a charge current pulse passing through a heavy metal generates a transverse spin current, with spin-up and spin-down carriers propagating in opposite directions perpendicular to the applied current. In their study, the spin current in tantalum (Ta) is sufficiently strong to induce magnetization switching in an adjacent ferromagnetic layer (CoFeB)⁸⁸, thereby modulating the spin polarization of injected charge carriers (Fig. 6b). As a result, they achieved reversible CPL polarization switching, with g_CPEL tunable between +0.23 and −0.25, all without the need for an external magnetic field. Under a small magnetic field, they can achieve an even larger CP-EL switching (g_CPEL = ±62%). Their study represents a significant advancement in electrically controlled spin-LEDs.

Fundamental challenges

Table 1 summarizes the key advantages, main limitations and highest achieved spin polarization and DOCP of different spin polarization mechanisms. Although DMS injectors demonstrate high spin injection efficiency due to the ohmic contact on semiconductors^{19,43,53,89,90}, their low Curie temperature prevents operation at room temperature. Despite the progress in spin polarization in FM and spin-dependent transport mechanisms such as GMR, TMR and SHE, a critical obstacle remains in their practical application through spin injection: the conductivity mismatch at the interface between ferromagnetic metals and semiconductors²⁰. The significant difference in electrical conductivity between the two materials severely limits the efficiency of spin injection. Consequently, the efficiency of spin-LEDs based on ferromagnetic metals remains relatively limited.

Table 1 Summary of key advantages, main limitations and highest achieved spin polarization and DOCP of different spin polarization mechanisms

Several strategies have been proposed to mitigate the conductivity mismatch, most notably by introducing Schottky barriers^91,92,93 or thin tunneling barriers (Al₂O₃^56,94,95,96 or MgO^46,58,60,62) at the interface between the metal and the semiconductor. Schottky barriers rely on the creation of a depletion region at the interface, which facilitates spin injection through quantum tunneling. These approaches provide partial solutions to the conductivity mismatch issue by eliminating direct contact of metal on semiconductor but fail to fully overcome their intrinsic limitations.

In addition, to achieve CPL emission in Faraday geometry (surface emitting) without applying an external magnetic field, the spin injector must possess perpendicular magnetic anisotropy⁶³. This poses a challenge, as most ferromagnetic injectors exhibit in-plane magnetization. Special structure design should be taken into account to fabricate perpendicular magnetized spin injectors^45,64,65,66.

Additionally, the tunneling barrier often requires sophisticated and expensive ultrahigh vacuum (UHV) fabrication, such as molecular beam epitaxy or sputtering⁶², which can increase costs and potentially compromise device performance. All these constraints create a strong incentive to explore alternative mechanisms.

Emergence of chiral-induced spin selectivity (CISS)

Recently, Chiral-Induced Spin Selectivity (CISS), has emerged as a new paradigm for spin polarization at room temperature. CISS effect does not require the use of any ferromagnetic materials or external magnetic fields. More importantly, CISS materials can be semiconducting and form ohmic contact on semiconductors, thus eliminating the obstacle of conductivity mismatch.

The CISS effect enables spin manipulation through the interaction between electron spin and the chirality of materials. This effect is generally understood as a result from the electron motion in a chiral potential, which favors one spin orientation over the other^97,98,99. In a chiral system, the symmetry of electron motion is broken, and the electron’s spin state becomes correlated with its momentum and the applied electric field. This occurs because spin, electron momentum, and the electric field direction are mutually orthogonal, causing spin polarization to emerge naturally in chiral materials. The handedness of the material determines the preferred spin orientation, generating an effective magnetic field that stabilizes one spin state while suppressing the opposite one¹⁰⁰. This selective stabilization of spin is the hallmark of the CISS effect (Fig. 7a).

Fig. 7: The chiral-induced spin selectivity (CISS) effect and earlier experimental observations.

a Schematic illustration of CISS. When electrons move through a chiral potential field, they generate an effective magnetic field that acts on the electrons, stabilizing one spin direction. p is the pitch of chiral potential. b Energy distribution as photoelectrons injected (i) L, C-terminus connected, (ii) D, N-terminus connected, and (iii) D, C-terminus connected helical polyalanine films using a left (negative spin polarization; red, solid) or right circularly polarized laser (positive spin polarization; blue, dashed). Or linearly (no spin polarization; black, dotted) light. c Current–voltage curves of helical self-assembled polyalanine molecules monolayers (top) and its cluster (down) adsorbed on magnetic Au/Co/Au substrates. [a, b are adapted from ref. ¹¹³, and c is adapted from ref. ¹¹⁴. with permission].

The CISS effect has been widely observed in structured arrays of chiral organic molecules, including helicenes^101,102, DNA^103,104,105, oligopeptides^106,107, proteins^108,109 and polymers^110,111,112, with reported spin polarization values ranging from 30 to 60%, depending on the specific molecular system, structure, alignment, and degree of order.

The spin filtering efficiency via the CISS effect is highly sensitive to molecular orientation, structural organization and alignment. For example, Naaman et al. demonstrated that the spin selectivity of poly-D-alanine varies based their relative orientation to a substrate (attached via C- or N-terminal) (Fig. 7b)¹¹³. Nguyen et al. found that self-assembled polyalanine arrays exhibited higher spin polarization (75%) compared to chiral polyalanine clusters (60%), highlighting the importance of structural order and molecular alignment in optimizing spin transport (Fig. 7c)¹¹⁴.

Observation of CISS in chiral metal-halide semiconductors

CISS also occurs in a variety of chiral self-assembled materials, including hybrid thin films composed of organic and inorganic layers (>99%)¹¹⁵, metal-organic frameworks formed by metal ions coordinated to organic linkers (MOFs, >99%)¹¹⁶, and metal-halide semiconductors comprised of periodic arrangement of metal-halide octahedra (MHSs) (>90%)^22,117,118. Among these chiral materials, chiral MHSs, first reported by Billings et al. in 2017, have emerged as highly promising materials due to their unique chiroptical and spin-selective characters^21,119, and outstanding optoelectronic properties¹²⁰. Chiral MHSs typically adopt low-dimensional structures with a chemical formular of A₂BX₄ (2D) or ABX₃ (1D), where A represents a chiral organic cation, B is a divalent metal cation (e.g., Pb²⁺, Sn²⁺), and X is a halide anion (e.g., Cl⁻, Br⁻, I⁻)^121,122. By selecting combinations of cation and anion, chiroptical properties such as circular dichroism (CD), circularly polarized luminescence, and spin selectivity via the CISS effect can be precisely controlled⁹⁹. In 2D chiral MHS films, the metal–halide planes are oriented parallelly to the substrate, while the chiral organic molecules align perpendicularly, forming highly uniform, low-roughness layers. Their well-ordered structure enables efficient spin-polarized electron transport and can function as intrinsic spin filters through CISS without the need for magnetic components.

In 2019, Lu et al. first demonstrated that charge transport in 2D chiral (MBA)₂PbI₄ perovskites (MBA = methylbenzyl ammonium) (Fig. 8a) is highly spin dependent²². The spin transport property is directly dictated by the handedness of the organic ligands, achieving a spin polarization of up to 86% as shown by the difference in current in Fig. 8b, in which much higher current is measured when the tip is magnetized in the “up” direction versus that in the “down” direction. That is, carriers with their spin oriented parallel to the tip magnetization direction are preferentially transferred from the (R-MBA)₂PbI₄ film to the magnetized tip over those with antiparallel spin orientation. The opposite behavior is observed for the (S-MBA)₂PbI₄ film, where higher current is observed when the tip is magnetized in the down direction as compared to the up direction (Fig. 8c). Notably, this spin selectivity surpasses that observed in previously reported chiral self-assembled molecular arrays. Building on this work, Lu et al. reported in 2021 that 2D chiral (MBA)₂SnI₄ perovskites achieved a record 94% spin polarization¹²³. Interestingly, the minimal variation in spin-polarized current between these two systems suggests that spin dephasing does not play a significant role in the MHSs. Instead, the exceptionally high spin polarization is primarily attributed to the spin-filtering effect induced by the oriented chiral organic molecules.

Fig. 8: Observation of chiral-induced spin selectivity (CISS) in chiral metal-halide semiconductors.

a Structure of 2D MHSs, (R/S-)methylbenzylammonium lead iodide (R/S-MBA)₂PbI₄. Current–voltage curves of chiral b (R-MBA)2PbI₄ and c (S-MBA)₂PbI₄ thin films on a fluorine-doped tin oxide (FTO) substrate measured by a conductive-probe atomic force microscope. The difference in bias-dependent current indicate the highly spin selective for the chiral (MBA)₂PbI₄ thin films. [Adapted from ref. ²². with permission].

Demonstration of CISS in spin-LEDs

spin-LEDs based on a bilayer structure

Recognizing the potential of chiral MHSs for spintronic applications, Beard and coworkers developed the first CISS-enabled spin-LED in 2021, incorporating a 2D chiral (R/S-MBA)₂PbI₄ layer as the spin-injection layer (Fig. 9a)²¹. Specifically, the chiral MHS layer acts as a spin-polarization layer to generate spin-polarized holes. These holes are then transported into the non-chiral, light-emitting perovskite nanocrystal emissive layer, where they recombine with electrons. Notably, the handedness of the chiral MHS layer fully determines the polarization of the emitted CP-EL. Their spin injection efficiency of 86% significantly surpasses conventional spintronic injection methods. A key finding of their study is the direct correlation between ${g}_{{CPEL}}$ and the spin lifetime ${\tau }_{s}$ of the non-chiral emitting layer. When the CsPbI₃ nanocrystal emission layer was replaced with CsPb(Br₀.₁I₀.₉)₃ nanocrystals, the spin lifetime increased from 4 ps to 14 ps, leading to a 10 fold enhancement of ${g}_{{CPEL}}$ at room temperature.

Fig. 9: Spin-LEDs based on a bilayer structure.

a Schematic illustration of a spin LED device. using 2D chiral perovskite (R- or S-MBA)₂PbI₄ as spin filtering, non-chiral CsPbI₃ nanocrystals as light-emitting layer; spin-polarized holes are injected at the interface. bP_CP-EL of the CISS layer/mixed halide perovskite NC heterostructure. c The same 2D perovskite acts as a spin filter, allowing only spin-polarized holes (blue circles) to move through the LED and recombine in the AlGaInP multiple quantum well emitting CP-EL (yellow helix). d Circularly polarized transient absorbance determines the spin lifetime of carriers in the AlGaInP multiple quantum well. e DOCP of the CISS layer/Multiple quantum well structure. [a, b are adapted from ref. ²¹, c–e are adapted from ref. ¹¹⁹. with permission].

Subsequent demonstration were shown by Wang and coworkers, where 2D chiral perovskites were used to inject spin-polarized carriers to the CdSe/ZnS quantum dots layer, generating CP-EL¹²⁴. Chiral metal-organic frameworks (MOFs) can also lead to spin injection in spin-LEDs¹²⁵. Mustaqeem et al. integrated a chiral MOF layer with CdSe/ZnS quantum dots, where spin-polarized holes influence both the polarization direction and intensity of CP-EL, achieving a g_CPEL of 12.24%.

Recently, Beard and coworkers utilized (R/S-MBA)₂PbI₄ to inject spin-polarized carriers to a commercially relevant (Al₀.₃₂Ga₀.₆₈)₀.₅In₀.₅P multiple quantum wells and successfully demonstrated efficient spin-LEDs at room temperature (Fig. 9b). A key advantage of this design is that it eliminates tunneling barriers at the perovskite/III–V semiconductor interface, ensuring efficient spin injection. In their device, spin-polarized carriers were successfully injected across the chiral perovskite/III–V interface, leading to the spin accumulation in the emitting layer. This results in a g_CPEL of up to 0.3 ± 0.08. The enhanced g_CPEL was attributed to the longer spin lifetime of (Al₀.₃₂Ga₀.₆₈)₀.₅In₀.₅P (τ_s = 100 ps) compared to their previous CsPb(Br₀.₁I₀.₉)₃ (τ_s = 14 ps) (Fig. 9c). Their study demonstrates that chiral hybrid perovskites can be directly integrated with traditional semiconductors, functioning as an integral part of the device stack. By incorporating chiral MHSs, conventional III–V LEDs, which are typically designed for light-charge conversion, can be transformed into devices that also enable spin-to-light conversion, significantly expanding their functionality¹¹⁹.

spin-LEDs based on a spin bulk-heterojunction (spin-BHJ)

In addition to the classical bilayer structure enabled by CISS effect, spin-LEDs can also be fabricated by a single-layer structure where spin transfer occurs within the emitting layer. For instance, Xiao and co-workers used core-shell MAPbBr₃ nanocrystals¹²⁶ with a chiral perovskite shell as the emitting layer to fabricate green spin-LEDs. They proposed that the chiral shell enabled the injection of spin-polarized electrons into the achiral core, yielding a g_CPEL of 2.16 × 10⁻³. Moon’s group improved this approach with CsPbBr₃ nanocrystals, reaching a higher g_CPEL of 0.24¹²⁷. Since the emissive layers in these reports are also chiral, it is unclear whether the circularly polarized luminescence stems from spin polarization within the emissive layer or is simply due to the chiral optical transitions or light propagation. That is, it is under debate if these should be categorized as “CP-LED” or “spin-LED”.

Recently, Lu and coworkers showed that an ultrafast spin transfer process occurs in quasi-2D chiral perovskites¹²⁸. They proposed a spin bulk-heterojunction (spin-BHJ) concept when spin polarization and spin transfer occur within a chiral 2D/3D perovskite layer. They showed that an ultrafast energy and spin funneling process from 2D to 3D perovskites can generate spin-polarized excitons in 3D phase (Fig. 10b), giving rise to circularly polarized luminescence. The fabricated green spin-LEDs exhibited a g_CPEL of 7.8×10⁻² at room temperature. By tuning the halide composition, the authors extended the emission wavelength to the near-infrared (NIR) region, achieving a maximum g_CPEL of 1.48 × 10⁻² at room temperature¹²⁹. Their works show that the performance of CP-EL is directly related to the spin transfer and spin lifetime of chiral perovskites, indicating the important role of spin states. Along this direction, several groups have further optimized quasi-2D perovskites for spin-LED applications. Xiao’s group developed mixed-ligand quasi-2D perovskites, achieving room-temperature CP-EL in green to deep blue region, and a maximum g_CPEL value of 0.175¹³⁰. Similarly, Yang’s group fabricated red-emissive spin-LED with engineered passivation, achieving a maximum g_CPEL of around 4 × 10⁻³¹³¹. Wang and coworkers introduced chiral ionic liquids as additives in lead bromide-based achiral hybrid perovskites, resulting in spin-LEDs with a g_CPEL value of 0.26¹³². While these LEDs devices can successfully generate high performance CP-EL, the exact mechanism and the underlying role of spin polarization and CISS effect require further investigation. Whether describing these LEDs as CP-LED or spin-LED remains to be further addressed.

Fig. 10: Spin-LEDs based on a spin bulk-heterojunction (spin-BHJ).

a Schematic illustration of the structure of quasi-2D perovskite with n = 1-$\infty$. b Illustration of the energy and spin funneling process, which transfers spin-polarized excitons from a 2D to a 3D perovskite phase. The CISS effect generates spin polarization in excitons within chiral 2D perovskites. These spin-polarized excitons are then sequentially transferred to thicker quasi-2D perovskite layers, acting as a funnel for both energy and spin between 2D and 3D phases. This process results in a long-lived spin polarization in the final state. [Adapted from ref. ¹²⁸. with permission].

Future perspective

The development of spin-LEDs has advanced significantly in recent years, particularly with the integration of spin-Hall effect and CISS effect. The use of solution-processable chiral perovskites has enabled an efficient spin injection without the need for ferromagnetic electrodes, addressing a major limitation in conventional spin-LED devices. While the spin polarization in CISS spin-LEDs is determined by the fixed handedness of the chiral spin-polarizing layer, tunability can be realized by different chiral layers, giving rise to various degrees of spin polarization. In addition, recent studies show that the tailored chiral organic components can be used to construct chiral systems that can be controlled through external stimuli, such as light and/or temperature^133,134, which may be further used to realize bi-polarization spin-LEDs.

However, critical challenges remain that must be overcome before spin-LEDs can be used in practical applications.

One of the primary challenges is maintaining spin polarization throughout the device. While chiral materials have demonstrated efficient spin injection, minimizing spin relaxation losses remains a key hurdle. Notably, despite the high spin polarization and injection efficiency, the g_CPEL of the device is ultimately constrained by the spin lifetime in the emissive layer. Extending this spin lifetime through improved emissive layer design will be crucial for enhancing overall device performance.

Another major challenge is the stability and scalability of chiral spintronic materials. Many chiral perovskites, despite their promising optical and spin properties, suffer from environmental instability, which limits their long-term performance. Developing more stable material compositions and protective encapsulation strategies will be essential for integrating spin-LEDs into commercial devices. Additionally, current fabrication methods for chiral spin-LEDs are often limited to small-scale short-term demonstrations. Scalable production techniques must be developed to ensure compatibility with conventional semiconductor manufacturing.

In conclusion, while spin-LEDs based on CISS materials represent a breakthrough in spintronic optoelectronics, further research is needed to enhance device stability and ultimately increase both the efficiency and g_CPEL of these devices. By addressing these challenges, the field can move closer to realize practical, high-performance spintronic light sources with real world applications.