Diamond Valley Transistor Fundamentals

• Diamond-based valley transistors are devices that utilize the unequal occupation of diamond’s six conduction-band valleys to control electron transport.
• They exploit anisotropic effective-mass properties and directional drift characteristics to enable spatial separation and modulation of valley currents via dual-gate, two-drain architectures.
• Advanced implementations use low-temperature pulsed electrical injection and room-temperature ultrafast optical switching, balancing long relaxation times with sub-picosecond control.

Searching arXiv for recent and foundational papers on diamond valleytronics and related transistor concepts. A diamond-based valley transistor is a device class in which the operative state variable is the occupation of diamond’s conduction-band valleys rather than charge density alone. In diamond, six symmetry-equivalent conduction-band minima, strong effective-mass anisotropy, and unusually slow intervalley relaxation enable valley-polarized electron populations to be created, transported, steered, and read out over technologically relevant times and distances. The literature spans three complementary realizations: low-temperature bulk valley transport with microsecond-scale relaxation times (Hammersberg et al., 2014), dual-gate two-drain devices that electrostatically partition valley currents in space and time (Suntornwipat et al., 2020, Suntornwipat et al., 7 Apr 2026), and ultrafast optical generation and reversal of valley polarization at room temperature on subpicosecond timescales (Gindl et al., 2024).

1. Valley degree of freedom in diamond

Diamond is a multivalley semiconductor whose conduction band has six equivalent minima along the $\Delta$ lines in the crystallographic $(100)$ directions, located at approximately $76\%$ of the way to the $X$ point in the Brillouin zone. These valleys are oriented along the $\pm x$, $\pm y$, and $\pm z$ axes and are characterized by a longitudinal effective mass $m_l = 1.56\,m_0$ and a transverse effective mass $m_t = 0.28\,m_0$ (Hammersberg et al., 2014). For a valley with axis along $[001]$, the near-minimum dispersion is written in effective-mass form as

$(100)$0

with the obvious permutation of axes for the other valleys (Suntornwipat et al., 7 Apr 2026).

A valley-polarized state is a non-equilibrium distribution in which some valleys contain more electrons than others. Two notations recur in the literature. In low-field relaxation studies, the excess density in one valley is denoted $(100)$1 and its decay defines the valley relaxation time $(100)$2 through

$(100)$3

In ultrafast optical work, a polarization parameter is introduced as

$(100)$4

where $(100)$5 and $(100)$6 are the populations of the corresponding valley groups and $(100)$7 is the total conduction-band electron density [(Hammersberg et al., 2014); (Gindl et al., 2024)].

The physical utility of the valley index follows from anisotropic transport. Valleys whose longitudinal axis is parallel to an applied field have different drift characteristics from valleys whose axis is perpendicular to it. In the low-temperature transport literature this distinction appears as “cold” and “hot” valleys; in later device work it appears as distinct diffusion depths, transit times, and drain currents [(Hammersberg et al., 2014); (Suntornwipat et al., 7 Apr 2026)]. This anisotropy is the basis of electrical readout in bulk time-of-flight measurements and of current partitioning in transistor-like geometries.

2. Intervalley scattering and the stability of valley polarization

The central figure of merit for a diamond-based valley transistor is the valley polarization relaxation time. In diamond, depolarization is governed primarily by symmetry-allowed intervalley scattering. The relevant hierarchy is explicit: intravalley acoustic scattering is fast and controls mobility but does not change the valley index; intervalley $(100)$8-scattering transfers electrons between valleys on orthogonal axes and is the dominant depolarization channel; intervalley $(100)$9-scattering between opposite valleys on the same axis is strongly suppressed in the regimes of interest and is effectively irrelevant for depolarization (Hammersberg et al., 2014).

The long lifetime originates in the combination of crystal symmetry and high intervalley phonon energies. In the relaxation analysis, the characteristic phonon energies are reported as $76\%$0 and $76\%$1; in the dual-gate transport model, representative values are $76\%$2 and $76\%$3 [(Hammersberg et al., 2014); (Suntornwipat et al., 7 Apr 2026)]. In both descriptions, the consequence is the same: at low temperature the phonon occupation is very small, and at low electric field electrons seldom acquire enough excess energy to emit the relevant intervalley phonons. The relaxation time is therefore well approximated by the inverse $76\%$4-scattering rate, $76\%$5 (Hammersberg et al., 2014).

Time-of-Flight measurements combined with ensemble Monte Carlo simulation were used to extract the electron-phonon coupling constants over $76\%$6–$76\%$7 and $76\%$8–$76\%$9. The fitted parameters are $X$0 for acoustic intravalley scattering and $X$1 for intervalley $X$2-scattering, with acoustic velocities $X$3 and $X$4 (Hammersberg et al., 2014). Within that model, the valley polarization relaxation time can reach microseconds or longer for $X$5 and $X$6, whereas for “hot valleys” at $X$7 and $X$8 it is reduced to less than $X$9 (Hammersberg et al., 2014).

The same separation of timescales reappears in device form. In the dual-gate two-drain transistor operated at $\pm x$0, the intervalley scattering time is described as approximately $\pm x$1, while the transit times through the active region are only tens of nanoseconds (Suntornwipat et al., 7 Apr 2026). This disparity is the operational reason that valley populations remain largely conserved during transport. A diamond valley transistor therefore operates most naturally in two complementary regimes: a low-field transport regime that preserves valley information, and a higher-field regime in which intervalley scattering is deliberately accelerated to rewrite or redistribute the valley state. This suggests a direct transistor logic analogy, but that implication remains architectural rather than fully standardized in the current literature.

3. Device architectures and operating principles

The basic operational sequence of a valley transistor is conceptually consistent across the diamond literature: initialize carriers into a preferred subset of valleys, transport or manipulate them while preserving polarization, and read out the resulting valley-dependent transport or optical signature before relaxation erases the state (Hammersberg et al., 2014). Reported diamond implementations realize this sequence with field-effect geometries, but they are not conventional steady-state MOSFETs.

The first explicit electrical realization is a dual-gate field-effect transistor in ultra-pure single-crystalline diamond, in which electrons are generated by a short $\pm x$2 UV pulse and valley currents are controlled electrostatically (Suntornwipat et al., 2020). The device is fabricated on freestanding single-crystalline $\pm x$3 CVD diamond plates, $\pm x$4–$\pm x$5 thick, with nitrogen concentration below $\pm x$6 and ionized impurity concentration below $\pm x$7. Two top gates are isolated from diamond by a $\pm x$8 Al$\pm x$9O$\pm y$0 layer, two drains are laterally separated, and the backside carries a thin semitransparent gold contact. Under appropriate gate bias, electrons injected near the source edge are separated by anisotropic drift into valley-specific trajectories that preferentially feed different drains. In that work, the charge current and the valley current measured at the receiving electrodes are controlled separately by varying the gate voltages, and the same platform is used to modulate the charge state of nitrogen-vacancy centers (Suntornwipat et al., 2020).

A later realization refines this architecture into a dual-gate, two-drain, backplane-controlled diamond transistor designed explicitly for tunable valley-polarized transport (Suntornwipat et al., 7 Apr 2026). It uses $\pm y$1 high-purity CVD diamond plates, $\pm y$2–$\pm y$3 thick, oxygen termination, a $\pm y$4 ALD Al$\pm y$5O$\pm y$6 top dielectric, Ti/Al source-gate-drain metallization, and a $\pm y$7 back-surface Au backplane. The active transport length used in analysis is approximately $\pm y$8. Operation is pulsed Time-of-Flight rather than steady-state: a $\pm y$9 DPSS laser with FWHM approximately $\pm z$0 photo-generates carriers near the source, holes are rapidly collected, and only electrons drift toward the drains under a bias configuration with source voltage around $\pm z$1–$\pm z$2 and gate/backplane voltages in the range $\pm z$3–$\pm z$4 (Suntornwipat et al., 7 Apr 2026).

Within this geometry, the left gate acts primarily as a selector in depth. Low left-gate voltage favors the bulk-penetrating $\pm z$5 valleys, while higher left-gate voltage suppresses those carriers deeper into the bulk and allows the heavy-vertical-mass $\pm z$6 valley population to remain near the surface. The right gate then acts downstream as a lateral steering element that redirects the selected valley population between the two drains. The backplane supplies a competing vertical field from below and modifies surface confinement (Suntornwipat et al., 7 Apr 2026). In this sense, the diamond-based valley transistor is best described as an electrostatic valley router or valley-current partitioner implemented in a bulk field-effect structure.

4. Readout mechanisms, transport signatures, and ultrafast control

Electrical readout in bulk diamond valley devices relies on the Shockley–Ramo theorem. In the Time-of-Flight geometry, the motion of electrons induces a displacement current at the contacts,

$\pm z$7

where $\pm z$8 is the carrier velocity component normal to the contacts and $\pm z$9 is the electrode spacing (Hammersberg et al., 2014). Because different valleys have different drift velocities and trajectories, the induced current develops characteristic multistep or multi-peak structure. In early bulk measurements, step-like current transients were used to fit the deformation potentials and valley relaxation rates; in the dual-gate transistor, two distinct temporal peaks, commonly interpreted as a fast surface-associated component and a slower bulk-penetrating component, shift systematically with gate voltage and drain selection [(Hammersberg et al., 2014); (Suntornwipat et al., 7 Apr 2026)].

The robustness of this transport signature has been characterized directly. In the dual-gate two-drain device, temperature sweeps from $m_l = 1.56\,m_0$0 to $m_l = 1.56\,m_0$1 at fixed bias show that increased acoustic-phonon scattering broadens the current pulse and shifts its arrival time from $m_l = 1.56\,m_0$2 at $m_l = 1.56\,m_0$3 to $m_l = 1.56\,m_0$4 at $m_l = 1.56\,m_0$5, while the valley-selective steering remains observable over the full range (Suntornwipat et al., 7 Apr 2026). Repeated measurements performed one week apart show negligible change in signal shapes and peak positions, indicating stability against experimental drift (Suntornwipat et al., 7 Apr 2026).

Optical generation and readout add a distinct room-temperature control modality. In bulk diamond, linearly polarized $m_l = 1.56\,m_0$6 infrared pulses of $m_l = 1.56\,m_0$7 duration and field amplitude up to $m_l = 1.56\,m_0$8 generate valley polarization through unidirectional intervalley scattering: valleys with light effective mass along the field are accelerated more strongly, cross phonon-emission thresholds more readily, and scatter preferentially into valleys with heavy effective mass along the polarization axis (Gindl et al., 2024). Monte Carlo simulations give a room-temperature valley polarization of approximately $m_l = 1.56\,m_0$9 for $m_t = 0.28\,m_0$0, corresponding to up to about $m_t = 0.28\,m_0$1 of electrons in the two valleys whose longitudinal axis is aligned with the pump polarization (Gindl et al., 2024).

The optical readout is based on polarization anisotropy of Drude absorption. The measured quantity is

$m_t = 0.28\,m_0$2

which directly links the absorption anisotropy to the valley population imbalance (Gindl et al., 2024). In that framework, the room-temperature valley polarization relaxation time is $m_t = 0.28\,m_0$3 at low carrier density, and two orthogonally polarized pump pulses separated by $m_t = 0.28\,m_0$4 reverse the sign of the polarization at room temperature (Gindl et al., 2024). This demonstrates sub-picosecond write and switch operations, although the presently demonstrated readout is optical rather than all-electrical. A plausible implication is that THz or IR gate fields could serve as dynamic valley-control elements in future transistor geometries, but that step remains extrapolative.

5. Related platforms and enabling diamond transistor technologies

The broader valleytronics literature supplies two kinds of context for the diamond-based valley transistor: analogues that demonstrate what a valley-controlled transistor function can look like, and enabling diamond FET platforms that address transport quality rather than valley manipulation itself.

A direct conceptual analogue is the observation of spontaneous valley polarization in an AlAs two-dimensional electron system (Hossain et al., 2020). There, a small gate-induced reduction in density at zero magnetic field and zero applied valley-splitting field drives an abrupt transition from equal two-valley occupation to full occupation of a single valley. In the main sample the critical density is $m_t = 0.28\,m_0$5, corresponding to $m_t = 0.28\,m_0$6, the transition produces an approximately two-fold resistance change and up to about $m_t = 0.28\,m_0$7 anisotropy in $m_t = 0.28\,m_0$8, and the ordered phase disappears above $m_t = 0.28\,m_0$9 (Hossain et al., 2020). This is not a diamond result, but it establishes that gate-tuned transistor-like switching can be mediated by valley occupancy itself, not merely by total density. Applied to diamond, the literature presents this as an analogy rather than a demonstrated mechanism.

A distinct enabling strand concerns high-quality diamond FETs. In hydrogen-terminated diamond/h-BN devices, a clean gate-tunable two-dimensional hole gas was shown to exhibit quantum oscillations, with low-temperature Hall mobilities of approximately $[001]$0 and $[001]$1 in two devices, and a cyclotron effective mass $[001]$2 (Sasama et al., 2019). The carriers in that work are holes near $[001]$3, not electrons in the six conduction valleys, so it is not a valley transistor platform in the strict sense. It is nevertheless important because it shows that high-quality gate-controlled two-dimensional transport can be realized on diamond with a low-disorder dielectric interface.

High-frequency diamond FET modeling provides a second enabling perspective. Hydrodynamic simulations of p-diamond TeraFETs identify a relatively low minimum resonant mobility, with transition mobility $[001]$4 for $[001]$5, $[001]$6, and $[001]$7, and emphasize a favorable resonant window of $[001]$8 to approximately $[001]$9 (Zhang et al., 2020). When the channel length is reduced to $(100)$00, the simulated p-diamond TeraFET exhibits the highest DC response among the compared materials in a large frequency window (Zhang et al., 2020). These results are not valleytronic demonstrations, but they suggest that short-gate diamond electrostatics and THz-scale field control are compatible with the dynamical conditions relevant to ultrafast valley manipulation.

6. Limitations, misconceptions, and open directions

Several limitations structure the present state of the field. First, long valley lifetime is not equivalent to coherent valley quantum control. The low-temperature bulk and transistor studies predominantly address valley occupancy and relaxation, while the room-temperature ultrafast work also concerns population transfer and decay; coherent superpositions of valley states and their decoherence mechanisms are not treated as demonstrated device functionality [(Hammersberg et al., 2014); (Suntornwipat et al., 7 Apr 2026); (Gindl et al., 2024)]. A common misconception is therefore to identify “valley transistor” directly with a valley qubit. The present literature supports the former more clearly than the latter.

Second, the electrical diamond valley-transistor demonstrations are pulsed and photo-injected. In both the 2020 and 2026 architectures, electrons are generated by short UV pulses near the source, and the output is a time-resolved induced drain current rather than a conventional DC transfer curve (Suntornwipat et al., 2020, Suntornwipat et al., 7 Apr 2026). This does not diminish their device character, but it does mean that the reported operation is closer to pulsed field-effect valley routing than to steady-state CMOS-style switching. The literature itself notes that DC characteristics, GHz switching behavior, device scaling to integrated arrays, and purely electrical injection schemes remain unresolved (Suntornwipat et al., 7 Apr 2026).

Third, operational windows differ sharply between low-temperature transport and room-temperature ultrafast control. For electrically transported valley states in bulk or dual-gate devices, the favorable regime is below approximately $(100)$01 and at moderate fields, where microsecond or longer relaxation times are available (Hammersberg et al., 2014). By contrast, room-temperature optical manipulation reaches only picosecond retention, albeit with sub-picosecond switching (Gindl et al., 2024). These are complementary rather than contradictory results: one emphasizes memory time under weak-field transport, the other emphasizes write speed under strong AC fields. The literature therefore supports both cryogenic long-lived valley transport and room-temperature ultrafast valley rewriting, but not yet a single architecture combining long retention with all-electrical room-temperature operation.

Material purity and interface quality remain decisive. The long-lifetime analyses assume very high-purity CVD diamond, with nitrogen concentration below $(100)$02, charged impurity concentration below $(100)$03, and free electron density below $(100)$04 to suppress electron-electron scattering (Hammersberg et al., 2014). Device studies additionally identify Al$(100)$05O$(100)$06/diamond interface traps and surface roughness as practical limitations that shift effective thresholds and degrade signal quality when carriers are forced too near the surface (Suntornwipat et al., 7 Apr 2026). These constraints indicate that scalable valley-transistor technology in diamond will depend as much on interface engineering and carrier injection as on intrinsic valley physics.

The most direct future directions are already named in the device literature: improved dielectrics and interfaces, quantitative extraction of a numerical valley polarization metric $(100)$07, detailed Monte Carlo modeling with full anisotropic band structure, systematic extension to higher temperatures, and deeper integration with NV-center functionality (Suntornwipat et al., 7 Apr 2026, Suntornwipat et al., 2020). Ultrafast work adds the prospect of THz or IR field-driven valley gates and shows that dynamic room-temperature control is physically viable (Gindl et al., 2024). Taken together, these studies define the diamond-based valley transistor not as a single settled device archetype, but as a convergent research program linking long-lived multivalley transport, electrostatic current steering, and ultrafast valley-state manipulation in a material distinguished by a stiff lattice, high phonon energies, and unusually favorable intervalley stability.