---
title: Negatively Charged GeV Centers in Diamond
url: https://www.emergentmind.com/topics/negatively-charged-germanium-vacancy-centers-gev
type: topic
---

# Negatively Charged GeV Centers in Diamond

Negatively charged germanium-vacancy centers, denoted GeV⁻, are inversion-symmetric group-IV vacancy color centers in diamond in which a Ge atom occupies the bond-center position between two adjacent carbon vacancies along a \(\langle 111\rangle\) axis. In current quantum-optics and quantum-networking research, GeV⁻ is valued for its zero-phonon line near \(602\;\text{nm}\), large zero-phonon-line emission fraction, narrow and spectrally stable optical transitions, and an optically addressable spin-\(\tfrac12\) degree of freedom. The literature spans initial identification as a single-photon source, coherent spin control, waveguide and cavity integration, plasmonic interfacing, and millikelvin quantum-memory operation beyond \(20\;\text{ms}\) [1503.04938] [1612.02947] [2308.09666].

## 1. Split-vacancy structure and charge-state identification

GeV⁻ belongs to the same split-vacancy family as SiV⁻, SnV⁻, and PbV⁻. In this geometry, the Ge impurity sits between two adjacent vacant lattice sites, producing \(D_{3d}\) inversion symmetry rather than the polar symmetry of NV⁻. In the group-IV-vacancy literature summarized for GeV⁻, this inversion symmetry implies no permanent electric dipole moment and only second-order sensitivity to electric fields, charge noise, and local strain, which is the microscopic basis for the center’s reduced spectral diffusion relative to polar defects [1612.02947] [2606.28528].

The charge-state assignment developed progressively. Palyanov and co-workers introduced the GeV center as a new single-photon source in diamond and identified the split-vacancy structure by first-principles calculations, while noting that the experimental charge state was not yet fully resolved [1503.04938]. Subsequent optical and spin spectroscopy established that the \(602\;\text{nm}\) center is the negatively charged state: Siyushev and collaborators showed that the fourfold Zeeman splitting pattern of the optical lines is consistent with spin-\(\tfrac12\) splitting in both ground and excited manifolds, identifying the \(602\;\text{nm}\) band with GeV⁻ [1612.02947]. One membrane-growth study nevertheless explicitly noted that a secondary feature around \(640\;\text{nm}\) could plausibly indicate another charge state, while emphasizing that the \(602\;\text{nm}\) center integrated into photonic structures was the standard GeV emitter [1804.02130].

The electronic structure used across the GeV⁻ literature is the standard group-IV-vacancy picture built from \(e_g\) and \(e_u\) defect orbitals. Both the ground and excited manifolds are split by spin-orbit coupling and dynamic Jahn-Teller interactions, producing four fine-structure optical transitions in the zero-phonon line at cryogenic temperature. Bulk-like or weakly strained systems show ground-state splittings near \(152\)–\(165\;\text{GHz}\) and excited-state splittings near \(981\)–\(1159\;\text{GHz}\), whereas strain can drive much larger effective splittings: \(181\;\text{GHz}\) was measured for a cavity-coupled emitter at millikelvin temperature, and highly strained nanodiamonds reached ground-state splittings up to \(870\;\text{GHz}\) [1612.03036] [1804.02130] [2308.09666] [1903.05446]. This strain sensitivity does not remove the basic GeV⁻ level structure, but it strongly alters the quantitative spin-phonon landscape.

## 2. Optical spectroscopy and photophysics

The defining optical signature of GeV⁻ is a zero-phonon line at about \(602\;\text{nm}\), corresponding to roughly \(2.06\;\text{eV}\). Under hydrostatic pressure this line blue-shifts substantially; a room-temperature diamond-anvil-cell study observed a total shift of \(320\;\text{meV}\) \((78\;\text{THz})\) up to about \(170\)–\(180\;\text{GPa}\), with a low-pressure slope near \(2.7\;\text{meV/GPa}\) [2209.09792]. At low temperature, bulk and bulk-like emitters can approach the lifetime limit: resonant spectroscopy at \(2.2\;\text{K}\) showed linewidths as narrow as \(42\;\text{MHz}\), compared with a transform limit of about \(26\;\text{MHz}\) for a \(6\;\text{ns}\) lifetime [1612.02947]. In diamond nanowaveguides, resonant linewidths of \(73\pm1\;\text{MHz}\) were reported at \(5\;\text{K}\) [1612.03036]. By contrast, nanodiamond environments generally broaden the line: room-temperature single GeV⁻ centers in \(10\)–\(50\;\text{nm}\) nanodiamonds showed zero-phonon-line full widths at half maximum of \(4\)–\(7\;\text{nm}\) with an average near \(5.2\;\text{nm}\), and implanted commercial nanodiamonds yielded cryogenic photoluminescence-excitation linewidths of \(440\pm40\;\text{MHz}\), about four times the transform limit, together with visible spectral diffusion and blinking [2010.15475] [2503.19490].

A recurrent advantage over NV⁻ is the large zero-phonon-line fraction. Siyushev et al. reported that the phonon sideband contains about \(40\%\) of the emission, corresponding to a Debye-Waller factor of about \(0.6\) [1612.02947]. A later nanodiamond–microcavity experiment found a Debye-Waller factor \(>0.6\) for a selected single GeV⁻ center, again emphasizing that most emission is already in the useful coherent channel before any cavity engineering [2307.00916]. In room-temperature nanodiamonds, Huang–Rhys factors around \(0.5\)–\(0.8\) were extracted, with a mean \(S\approx0.65\), consistent with weak electron-phonon coupling relative to NV⁻ [2010.15475].

Brightness varies strongly with host environment and photonic loading. Bulk resonant excitation produced more than \(200\;\text{kcps}\) in the phonon sideband and up to \(0.6\;\text{Mcps}\) when the full emission band was collected [1612.02947]. In \(10\)–\(50\;\text{nm}\) nanodiamonds at room temperature, the maximum saturation count rate reached \(1.6\;\text{MHz}\) [2010.15475]. Cryogenic nanodiamond emitters coupled to plasmonic structures exceeded \(2\times10^6\;\text{counts/s}\) in the zero-phonon line under off-resonant excitation [1903.05446]. Lifetimes likewise span a wide range because the electromagnetic environment matters strongly: \(6.0\;\text{ns}\) is a representative bulk value; \(6.6\pm0.3\;\text{ns}\) was measured in a nanophotonic waveguide; \(2.53\pm0.20\;\text{ns}\) was reported for a selected nanodiamond emitter; and plasmonic integration reduced lifetimes from \(15.9\;\text{ns}\) on glass to \(8.1\;\text{ns}\) on silver and to \(3.8\;\text{ns}\) in a dielectric-loaded surface-plasmon-polariton waveguide [1612.03036] [2307.00916] [1801.10106].

The photophysics is often well described by a three-level model with a dark or shelving state. In one fiber-cavity membrane study, power-dependent \(g^{(2)}(\tau)\) fits implied that high \(532\;\text{nm}\) excitation shelved about \(96\%\) of the population into a dark state, plausibly another charge state, and led to an inferred quantum efficiency of \(17^{+8}_{-5}\%\) for the specific center under study [1912.05247]. This does not contradict reports of much higher effective optical performance in other devices; it instead underscores that charge-state dynamics and optical gating remain sample- and excitation-dependent.

## 3. Spin control and quantum-memory operation

The first complete demonstration of GeV⁻ spin control was provided by Siyushev et al. Using a magnetic field to lift the spin degeneracy, they showed optical spin pumping, optically detected magnetic resonance, Autler-Townes splitting in a microwave-optical double-resonance configuration, and coherent population trapping. In that work, optical pumping produced about \(59\%\) spin polarization, the spin relaxation time exceeded \(20\;\mu\text{s}\), and the pure dephasing time inferred from coherent population trapping was about \(19\;\text{ns}\) [1612.02947]. The same study also established the near-transform-limited optical linewidths that made GeV⁻ compelling as a spin-photon interface.

The major later advance was the realization of a GeV⁻ quantum memory at millikelvin temperature. In a dilution refrigerator below \(300\;\text{mK}\), coherent control of the lower orbital branch of the GeV⁻ ground state yielded a microwave transition near \(3.066\;\text{GHz}\), a Rabi frequency of \((2\pi)\,6.486\;\text{MHz}\), and \(98\%\) optical spin initialization within \(1\;\text{ms}\). The measured inhomogeneous dephasing time was \(T_2^*\approx1.43\;\mu\text{s}\); Hahn echo gave \(T_2\approx440\;\mu\text{s}\); Carr-Purcell-Meiboom-Gill decoupling extended the coherence to \(24.1\pm0.9\;\text{ms}\); and XY8 gave \(18\pm3\;\text{ms}\) [2308.09666]. In that regime the dominant noise was modeled as an Ornstein-Uhlenbeck process for magnetic detuning plus amplitude noise on the microwave drive, rather than phonon-driven orbital relaxation.

This behavior reflects a central advantage of GeV⁻ over SiV⁻. The larger ground-state splitting, measured as about \(181\;\text{GHz}\) for the memory device and strain-enhanced in some nanostructures to much larger values, suppresses phonon-mediated orbital transitions more strongly at a few hundred millikelvin. The 2026 focused-ion-beam implantation study situates these results in a broader node context, noting record electron-spin coherence up to \(20\;\text{ms}\) and nuclear-spin coherence up to \(2.5\;\text{s}\) in a GeV-based multi-qubit node [2308.09666] [2606.28528]. This suggests a hierarchy in which the GeV electron spin acts as the fast optical interface and nearby nuclear spins provide longer-lived storage.

## 4. Formation routes and material platforms

The GeV center entered the diamond-defect literature as both an implanted and a grown-in defect. Palyanov et al. showed that GeV centers could be fabricated by ion implantation and by chemical-vapor deposition, and that both routes yielded the \(602\;\text{nm}\) single-photon source associated with the split-vacancy defect [1503.04938]. This dual accessibility has remained a defining practical advantage.

Single-crystal membrane platforms were developed soon afterward. A bottom-up lift-off plus microwave-plasma-CVD process using GeO\(_2\) and metallic Ge produced single-crystal diamond membranes about \(300\;\text{nm}\) thick with GeV emission distributed over the membrane area. These membranes retained surface roughness around \(1\;\text{nm}\) and were patterned into microring and microdisk resonators; whispering-gallery modes with quality factors up to \(1500\) were measured, and collected GeV emission from microrings was enhanced by about \(14\times\) relative to bare membrane regions [1804.02130].

Nanodiamond hosts introduced additional flexibility and additional disorder. Very small HPHT nanodiamonds, \(10\)–\(50\;\text{nm}\) in size with mean size around \(20\;\text{nm}\), were shown to host single GeV⁻ centers with bulk-like room-temperature spectral stability, high polarization visibility, and MHz-level brightness [2010.15475]. Commercial \(180\;\text{nm}\) nanodiamonds were later implanted with \(^{74}\mathrm{Ge}^-\) at \(57\;\text{keV}\) and annealed at \(1200^\circ\text{C}\) for \(1\;\text{h}\), yielding GeV-related emission in about \(90\%\) of surveyed nanodiamonds, but at the cost of substantial lattice damage, multiple emitters per nanodiamond, and many NV centers generated by the high implantation fluence \(10^{14}\;\text{ions/cm}^2\) [2503.19490]. That work explicitly recommended lowering the Ge fluence toward \(10^{12}\;\text{ions/cm}^2\) to reduce strain, charge noise, and clustering.

Deterministic near-surface creation advanced further with focused-ion-beam implantation. In type-IIa electronic-grade diamond, Ge ions implanted at \(5\), \(10\), \(35\), and \(70\;\text{keV}\), followed by \(1200^\circ\text{C}\) and \(1500^\circ\text{C}\) vacuum anneals, produced GeV centers at projected depths from \(5.5\) to \(30\;\text{nm}\). Formation yield depended strongly on energy and fluence, reaching up to \(33\%\) at \(35\) and \(70\;\text{keV}\), while low-fluence implantation generated single GeV centers verified by \(g^{(2)}(0)<0.5\) [2606.28528]. For nanophotonics this is important because near-surface placement with tens-of-nanometers precision is the regime needed for high field overlap in waveguides and cavities.

## 5. Nanophotonic, cavity, and plasmonic integration

Bhaskar et al. established waveguide quantum electrodynamics with a single GeV⁻ in a nanoscale diamond waveguide. They reported nearly lifetime-broadened optical transitions in the nanostructure, \(73\pm1\;\text{MHz}\) linewidth at \(5\;\text{K}\), and an \(18\pm1\%\) reduction of waveguide transmission by a single GeV⁻ in one pass, corresponding to a cooperativity \(C\ge 0.10\pm0.01\). By homodyne detection of the resonance fluorescence, they showed that the GeV-waveguide system is nonlinear at the single-photon level [1612.03036]. This result was notable because it did not require a cavity or slow-light enhancement.

Plasmonic integration exploited the short wavelength and high zero-phonon-line fraction differently. In dielectric-loaded surface-plasmon-polariton waveguides on silver, a single GeV center embedded in a nanodiamond was remotely excited on-chip and coupled back into guided plasmons with an effective \(\sim6\)-fold Purcell enhancement, \(\beta\approx56\%\), and propagation length \(33\pm3\;\mu\text{m}\) on single-crystal silver, yielding a figure of merit about \(180\pm25\) [1801.10106]. In a related architecture, deterministic placement of nanodiamonds at plasmonic hot spots combined with a Bragg reflector produced unidirectional single-photon emission, a roughly tenfold decay-rate enhancement, and \(g^{(2)}(0)=0.06\); the same work also documented ground-state splittings up to \(870\;\text{GHz}\) in highly strained nanodiamonds [1903.05446].

Open microcavities have now covered both room-temperature spectral funneling and cryogenic Purcell enhancement. A single GeV in a diamond membrane coupled to a fiber Fabry-Pérot cavity with finesse about \(11{,}000\) showed a \(31^{+11}_{-15}\)-fold enhancement of spectral density at room temperature, a regime the authors explicitly distinguished from large spontaneous-emission-rate enhancement because thermal broadening still dominated [1912.05247]. An AFM-transferred nanodiamond GeV⁻ in an open Fabry-Pérot microcavity under ambient conditions preserved a finesse of \(7{,}700\) and yielded a \(48\)-fold spectral-density enhancement, again with an effective Purcell factor below unity because the room-temperature emitter linewidth exceeded the cavity linewidth [2307.00916]. The decisive cryogenic step came with a tunable open microcavity that produced a direct Purcell-effect-induced lifetime reduction of up to \(4.5\pm0.3\) and coherent coupling rates up to \(360\pm20\;\text{MHz}\), while extracting a GeV quantum efficiency of at least \(0.34\pm0.05\) and likely much higher [2312.14313].

## 6. Applications, distinctions, and unresolved issues

The application space follows directly from the combination of optical and spin performance. GeV⁻ is repeatedly positioned as a promising node for quantum communication and quantum networking because it combines a strong spin-photon interface with long-lived spin coherence [2308.09666] [2606.28528]. On-chip waveguide and plasmonic demonstrations show that the defect can serve as a nanoscale guided single-photon source, while cavity experiments indicate a route to large Purcell enhancement and high zero-phonon-line collection. Independently of networking, high-pressure spectroscopy established GeV⁻ as a viable optical probe for extreme-pressure sensing, with a predictable pressure-induced blue shift up to megabar conditions [2209.09792].

Several limitations remain active research topics. Early studies did not always identify the charge state unambiguously, and even in later cavity work dark-state shelving or optical gating was frequently attributed to another charge state [1804.02130] [1912.05247]. Nanodiamond and implantation-based platforms can suffer from strong strain, multiple emitters, NV co-generation, and spectral diffusion; one implanted-nanodiamond study measured a \(428\;\text{GHz}\) ground-state splitting for a single emitter, far above the \(150\;\text{GHz}\) zero-strain value, and directly linked the excess splitting to implantation-induced strain [2503.19490]. Room-temperature cavity enhancements must also be interpreted carefully: in broad-linewidth emitters, spectral-density enhancement does not imply a proportionate increase in the spontaneous-emission rate [2307.00916].

A persistent source of confusion is nomenclature. In diamond, GeV⁻ denotes the inversion-symmetric color center discussed above. In silicon, however, “GeV” is also used for Ge-vacancy complexes that are deep electronic defects rather than optical color centers. For the single-vacancy GeV complex in Si, first-principles calculations placed the \(\mathrm{D}^0/\mathrm{D}^-\) and \(\mathrm{D}^-/\mathrm{D}^{2-}\) charge-transition levels at \(-0.51\;\text{eV}\) and \(-0.35\;\text{eV}\) below the conduction-band minimum, with localization length about \(0.45\)–\(0.46\;\text{nm}\); subsequent transistor experiments used arrays of such GeV complexes as strongly correlated deep centers in silicon nanoelectronics [1803.01493] [2102.01390]. Those silicon defects are physically distinct from the diamond GeV⁻ center even though the acronym is shared.

Taken together, the GeV⁻ literature shows a defect system that has moved from initial discovery and structural assignment to coherent spin control, open-system waveguide QED, plasmonic interfacing, deterministic near-surface creation, and millikelvin quantum-memory performance. The remaining agenda is correspondingly specific: stabilize the charge state under realistic pumping, reduce implantation- and surface-induced strain where needed, and combine the now-demonstrated \(>20\;\text{ms}\) spin coherence with high-cooperativity nanophotonics to realize fully integrated spin-photon nodes.

Source: https://www.emergentmind.com/topics/negatively-charged-germanium-vacancy-centers-gev