Skip to content

Lattice location of ion-implanted He 6 in diamond

FieldValue
JournalPhysical review. B./Physical review. B
Published2026-08-07
AuthorsUlrich Wahl, J. G. Correia, Ângelo Costa, Brecht Biesmans, Goele Magchiels
DOI10.1103/1286-zbbb

Low-fluence ⁶He implanted into single-crystal CVD diamond largely occupies tetrahedral interstitial T sites, based on beta emission channeling measurements from 30°C to 800°C. At 800°C, the channeling anisotropy fell by about 18%, with a best-fit T-site fraction of 92%. Fractions above 100% at lower temperatures are attributed to the background correction, whose absolute uncertainty is about ±15%. The authors associate the high-temperature loss with helium migration and estimate an activation energy of 1.63 to 2.89 eV.

CERN’s ISOLDE facility produced ⁶He through 1.4 GeV proton-induced spallation of a UC2 target. After release near 2000°C, the atoms were ionized to the 1+ charge state, accelerated to 30 keV, and mass separated. The isotope’s 806.7 ms half-life made on-line lattice-location measurements possible.

The host was an ElementSix <100>-oriented, single-crystal CVD plate measuring 3.0 × 3.0 × 0.25 mm³. Its nitrogen concentration was below 1 ppm. The beam passed through a 1 mm collimator at 0.9 pA, corresponding to 5 × 10⁶ atoms/s or 6.4 × 10⁸ atoms/s/cm². Implantations were performed at 30°C, 300°C, 600°C, and 800°C under vacuum better than 10⁻⁵ mbar.

The implanted fluence was approximately 0.72 × 10¹² atoms/cm² per pattern. The total fluence reached 2.9 × 10¹³ atoms/cm². SRIM simulations placed the projected range at 1375 Å, with 299 Å straggling. The short half-life prevents persistent accumulation. The maximum concentration was about 1.1 × 10¹⁴ atoms/cm³, leaving more than 2100 Å between implanted atoms on average.

⁶He decays by beta emission into stable ⁶Li. The endpoint energy is 3508 keV, and the average beta energy is 1569 keV. Electrons emitted from aligned lattice sites follow the crystal potential, producing direction-dependent yields around atomic rows. A two-dimensional silicon pad detector measured those patterns.

The detector was 3 × 3 cm² with 22 × 22 pixels of 1.3 × 1.3 mm². It sat 60 cm from the sample, giving a ±1.36° angular range and 0.05° standard-deviation resolution. Expected critical angles were about 0.40°, so the patterns required fine angular resolution.

The authors simulated about 250 lattice sites with the many-beam method and fitted linear combinations to the measurements. They included an rms thermal displacement of 0.044 Å for ⁶He. Electron and gamma backgrounds were corrected with a factor of 2.36. The correction had an absolute error of roughly ±10 to 15%, which permits fitted fractions above 100%.

One-site fits produced the lowest chi-square values for the ideal T site. The <111> and <100> patterns alone cannot separate substitutional S from T sites because their projections are identical. The <110> and <211> patterns can separate them and matched the axial channeling and planar blocking expected for T sites.

Displacing the emitter from the ideal T position produced minima only within about 0.1 Å. Two-site fits gave inconsistent improvements for most alternatives and often required negative fractions. Substitutional S sites could account for 2 to 10% in some fits. No relaxed substitutional position, including AB or SP→C 0.60, produced consistent gains. The authors therefore set other highly symmetric site fractions below 10%.

The data do not distinguish a genuinely random position from a low-symmetry position that produces little angular anisotropy.

The 30°C, 300°C, and 600°C data showed a slight upward trend rather than a loss of anisotropy. The best-fit fractions were:

Implantation temperatureBest-fit T-site fraction
30°C108%
600°C111%
800°C92%

Values above 100% are fit results after correction, not physical occupancies. A change in the background correction would move the entire curve. The relative decrease of about 18% at 800°C was consistent across all four crystallographic directions, so it remains statistically relevant despite the systematic uncertainty.

The authors considered two limiting effects during the 1.164 s radioactive lifetime. A few atomic jumps could move ⁶He to a low-symmetry site while preserving its depth. Many jumps could widen the implanted profile, send atoms to the surface, or move them deep enough to destroy channeling. Neither effect appeared as a new high-symmetry site in the two-site fits.

An Arrhenius model used a 1.545 Å jump between neighboring T sites and four nearest-neighbor sites. The assumed ⁶He attempt frequency was 23.9 THz, giving a diffusion prefactor of 3.80 × 10⁻³ cm²/s. Applying the short lifetime to the two limiting scenarios produced an activation-energy range of 1.63 to 2.89 eV. This is a model-dependent estimate because the experiment did not separate the two migration scenarios.

Comparison with theory and earlier studies

Section titled “Comparison with theory and earlier studies”

Most theoretical studies cited by the paper predict T as the most stable helium site. They place the hexagonal H site at the migration saddle point in a T-H-T path. Their predicted interstitial migration energies span 1.41 to 2.36 eV, consistent with the experimental estimate of about 2 eV.

Earlier channeling work implanted ³He into natural diamond at a fluence of 4.2 × 10¹⁵ cm⁻² and estimated only about 33% on T sites. That study used fewer angular scans and no quantitative comparison with simulated site patterns. The current low-fluence CVD result identifies T sites as the majority location. The two results used different materials, isotopes, energies, and fluences, so they are not direct replicates.

Experimental diffusion studies often report 1.0 to 1.6 eV but much smaller entropy constants than expected for simple interstitial migration. The authors use this mismatch to argue that many earlier release measurements included defect-assisted or other release mechanisms.

Measurement limits and device implications

Section titled “Measurement limits and device implications”

Several limits constrain the assignment. The maximum implantation temperature was 800°C, so the experiments could not track helium beyond the onset of migration. They could not distinguish loss of anisotropy caused by long-range diffusion from movement into a low-symmetry site. The fitted fractions carry about ±15% absolute uncertainty, and the deep 1375 Å implant profile increases dechanneling uncertainty.

HR1 and HR2 have zero-phonon lines at 536 nm and 560 nm, but their structures were not identified in this experiment. Reported He-related color centers were produced at fluences of 5 × 10¹⁴ to 5 × 10¹⁶ atoms/cm², 10 to 1000 times higher than this study. The authors note that a few-percent production yield would be below the experiment’s detection power.

For implantation processes, the data show that temperature can change the dominant lattice state during implantation itself. At 800°C, some ⁶He can migrate before decay, so a process using this probe cannot treat the as-implanted T-site fraction as temperature-independent. The low concentration also makes helium-bubble accumulation unlikely under these conditions.

For geological retention, the authors state that an activation energy near 2 eV makes simple interstitial helium unstable over geological times. Long-term retention would require binding to a defect, incorporation in mineral or liquid inclusions, or formation of small helium bubbles.

The study required a <100>-oriented, single-crystal CVD diamond plate, 3.0 × 3.0 × 0.25 mm³, with nitrogen below 1 ppm, as the host for 30 keV ⁶He implantation and beta emission channeling. 6CCVD’s diamond materials catalog lists single-crystal MPCVD plates and wafers up to 125 mm, with thicknesses from 0.1 to 500 micrometers, so the plate thickness falls within the listed range. The <100> orientation and low-nitrogen requirement remain separate selection criteria that should be confirmed through engineering support.

View Original Abstract

The properties of He in diamond are of interest with respect to geological dating and the creation of color centers with possible quantum applications. We report on the lattice location of the short-lived ion-implanted nuclear probe He 6 ( t 1 / 2 = 807 ms ) in diamond, which was performed using the beta emission channeling method at CERN’s ISOLDE facility. He 6 was implanted with 30 keV into a single-crystalline chemical-vapor deposited diamond sample kept at a temperature ranging from 30 °C up to 800 °C. By means of comparing the measured emission channeling patterns along different crystallographic directions with simulated yields for a variety of possible sites, we conclude that the implanted He 6 occupies tetrahedral (T) interstitial sites, in agreement with theoretical predictions that T sites should be the preferred position of He in diamond. Implantation at 800 °C resulted in a drop in the tetrahedral interstitial fraction by ∼18%, which we interpret as the onset of diffusion, He 6 thus being able to change to lattice sites of low crystallographic symmetry, or reach the surface of the sample or escape to the bulk during its lifetime. We estimate the activation energy for interstitial migration of He to be around 1.63-2.89 eV, which agrees with theoretical predictions of 1.41-2.36 eV from the literature. Activation energies around 2 eV would mean that simple interstitial He cannot be stable in diamond on geological time scales, thus to remain inside, it should be bound to some defect in the material or exist in another form such as within inclusions of other minerals or liquids, or possibly small He bubbles.