Skip to content

Muonium as a probe of point defects in type-Ib diamond

FieldValue
JournalPhysical Review Research
Published2026-02-25
AuthorsK. Yokoyama, J. S. Lord, Hiroshi Abe, Takeshi Ohshima
DOI10.1103/jn9c-fprb

Diffusive tetrahedral muonium converted to a diamagnetic state after reaching negatively charged NV centers in type-Ib diamond. The fitted capture rate reached 9 ± 1 MHz at 20 K, compared with 2.8 ± 0.4 MHz at 290 K. Density matrix fits also identified electron spin exchange between muonium and neutral substitutional nitrogen.

Neutral tetrahedral muonium hops between interstitial sites and can reach point defects throughout the implanted volume. Prior work estimates its room-temperature diffusion coefficient at about 10⁻² cm²/s.

Interaction with neutral substitutional nitrogen exchanges the muonium electron spin with the paramagnetic donor electron. Interaction with negatively charged NV centers instead leaves muonium in a diamagnetic state.

The method detects both unpaired electrons and charged defects without magnetic moments. The authors identify the work as the first measurement of muonium state exchange dynamics in diamond.

Two single-crystal diamond samples had an average grain diameter of 1.2 mm and a mass of about 2 g each. The pristine sample was synthetically grown type-Ib diamond with approximately 100 ppm dispersed substitutional nitrogen, estimated by electron paramagnetic resonance.

The NV sample was produced by exposing type-Ib diamond to an electron beam and annealing it at 900 °C for 2 hours. The resulting NV concentration was approximately 4 × 10¹⁷ cm⁻³, or 2.5 ppm. More than 90% of its NV centers were considered negatively charged at room temperature.

Measurements used longitudinal field muon spin relaxation, or µSR, on the EMU spectrometer at the ISIS Pulsed Neutron and Muon Source. A 100% spin-polarized 4 MeV surface muon beam irradiated each sample. The beam range was approximately 200 mg/cm².

Longitudinal fields covered 0 to 4000 G and were parallel to the initial muon spin. Spectra were recorded at 290 K and 20 K. Each displayed spectrum averaged five million muon decay events.

At zero field, most muon spin polarization was lost, leaving integrated P near 0.1. A 400 G field produced rapid recovery, followed by slower recovery through 4000 G. Integrated polarization versus field produced the repolarization curves used in the fits.

Two neutral states dominate the model: tetrahedral MuT0 and bond-centered MuBC0. A constant diamagnetic background represents the remaining prompt muons.

Global fits to the full longitudinal field data set separate their contributions and return state transition rates. The Python program QUANTUM solves density matrix time evolution through eigenvalue decomposition. Transition rates connect block diagonal terms for each muonium state.

The code models four bond axes for MuBC0 and distributes their orientation over the sphere to match the granular sample. MuT0 has an isotropic hyperfine constant of 3711 MHz. MuBC0 has axial components of 168 MHz and −393 MHz.

Diffusive hopping is not modeled explicitly, so fitted rates can include motion and host interaction contributions. Prompt yields were about 60% for MuT0 and 30% for MuBC0 at both temperatures. These are close to the 0.7 and 0.2 fractions reported by Connell et al.

The global fit gives these fractions and defect interaction rates. Here $\Lambda_T^0$ is MuT0 relaxation and $\Lambda_{T/D^-}^0$ is MuT0 to Mu−D capture.

SampleTemperatureMuT0 fractionMuBC0 fractionPrincipal fitted rate
Pristine290 K0.600 ± 0.0050.315 ± 0.007$\Lambda_T^0$ = 58 ± 4 MHz
Pristine20 K0.586 ± 0.0040.324 ± 0.007$\Lambda_T^0$ = 53 ± 3 MHz
NV290 K0.632 ± 0.0040.317 ± 0.005$\Lambda_{T/D^-}^0$ = 2.8 ± 0.4 MHz
NV20 K0.595 ± 0.0040.355 ± 0.005$\Lambda_{T/D^-}^0$ = 9 ± 1 MHz

N0 donors substitute for carbon and carry one paramagnetic electron in a deep donor state. MuT0 scattering with N0 exchanges electron spins, then depolarizes the muon through its hyperfine interaction.

The fitted $\Lambda_T^0$ values were 58 ± 4 MHz at 290 K and 53 ± 3 MHz at 20 K. Their similarity supports the assumption that the donor cross section changed little between temperatures.

The collision relation is $\Lambda_T^0 = nv\sigma$. The donor density n was estimated as 2 × 10¹⁹ cm⁻³, but the Mu velocity v and cross section σ cannot be separated.

Forward site exchange from MuT0 to MuBC0 was much faster than the reverse. The forward rates were 1.19 ± 0.09 MHz at 290 K and 1.00 ± 0.06 MHz at 20 K. Both reverse rates were 0.034 ± 0.002 MHz.

These rates indicate a predominantly one-way transition toward the bond-centered state. This agrees with Odermatt et al.’s prediction that MuBC0 is the most stable diamond site.

MuBC0 relaxation was 0.02 ± 0.02 MHz at 290 K and 0.22 ± 0.04 MHz at 20 K. The low-temperature value is about an order of magnitude larger.

The pristine model failed the irradiated sample’s repolarization curves. Adding an NV− interaction channel allowed accurate fitting. The revised model converts MuT0 into a diamagnetic Mu−D state at an NV− center. Both directions were included, but capture from MuT0 dominated.

At 290 K, capture was 2.8 ± 0.4 MHz and the reverse rate was 0.065 ± 0.004 MHz. At 20 K, capture was 9 ± 1 MHz and the reverse rate was 0.034 ± 0.003 MHz. The threefold capture-rate increase from 290 to 20 K shows stronger trapping at lower temperature.

The authors suggest that the capture cross section may be larger at 20 K. Delocalization of the NV− excess electron through tunneling or zero-point motion is a possible reason. These are proposed explanations, not resolved measurements.

NV0 can accept an electron from nearby N0 donors and become NV−. More than 90% were NV− at room temperature, leaving little room for temperature-driven charge-state conversion. Without optical excitation, NV− centers stayed in their unpolarized ground state, whose 2.87 GHz splitting was negligible against thermal energy at 20 K.

Unresolved variables and measurement limits

Section titled “Unresolved variables and measurement limits”

The model treats only MuT0 and MuBC0 as active neutral states. The assumption of similar velocity at 20 K and 290 K remains untested. Diffusion can switch from phonon-assisted tunneling at high temperature to incoherent quantum tunneling at low temperature.

NV− density may also vary with temperature. Its effect is limited because the sample was nearly fully converted at room temperature. A detailed temperature series is needed to test velocity and capture cross section. Charge-state population should also be tracked.

Surface muons have limited depth resolution because their implantation depth and spread must be calculated. A low-energy source could move the probe toward subsurface layers, but this remains prospective.

High-field data can carry systematic offsets from positrons spiraling around field lines. Silver hardware caused background offsets of about 0.02. Reduced chi-squared values ranged from 1.174 to 1.483 for the four displayed datasets.

Granular averaging replaces one defined crystal orientation with an orientation distribution. Several NV fit parameters were fixed to pristine values, limiting independent tests of host interactions.

In this diamond system, muonium characterizes substitutional nitrogen and charged NV defects through different relaxation channels. Spin exchange reports access to unpaired donor electrons. Diamagnetic conversion reports capture by an electron-rich NV− center.

This gives defect metrology access to a charged center that magnetic relaxation alone would not reveal. For NV device engineering, the results support tracking nitrogen supply, vacancy generation, annealing, and charge conversion together.

The study links NV0 to NV− conversion to electron transfer from nearby N0 donors. The method could support process development by testing whether nitrogen content or vacancy treatment changes defect capture.

The authors propose muonium spectroscopy for stable MuT0 in silicon and silicon carbide. Those systems were not tested.

The diamond materials catalog from 6CCVD lists MPCVD-grown single-crystal diamond, with plates and wafers up to 125 mm and thickness from 0.1 to 500 micrometers. The muonium experiment needs type-Ib single-crystal diamond with substitutional nitrogen, plus electron irradiation and annealing at 900 °C for 2 hours to create NV centers. The engineering support page can be used to discuss specimen geometry and preparation.

View Original Abstract

Muonium (Mu), a bound state of a positively charged muon and an electron, can diffuse through crystal lattices and interact with defect centers in insulators and semiconductors. We demonstrate that this Mu’s diffusive property can be used to probe defects in a diamond crystal lattice, specifically, substitutional nitrogen atoms ( <a:math xmlns:a=“http://www.w3.org/1998/Math/MathML”> <a:msubsup> <a:mi mathvariant=“normal”>N</a:mi> <a:mrow> <a:mtext>s</a:mtext> </a:mrow> <a:mn>0</a:mn> </a:msubsup> </a:math> ) and nitrogen-vacancy (NV) centers in type-Ib diamond. Upon interaction with these defects, Mu can exchange its electron’s spin or change its charge state, which results in muon spin relaxation. However, muons in diamond (and semiconductors in general) can be in a few distinctive muonium states, with each state contributing to the muon signal. In addition, these states can undergo site and charge exchange interaction, forming a dynamic network. Hence, to study the Mu interaction with point defects, the muon data have to be deconvoluted to isolate signals from the diffusing species. To achieve this goal, we have modeled the Mu state exchange dynamics and numerically simulated the time evolution of muon spin polarization by the density matrix method. With a global curve fit to a set of longitudinal field scan data, one can extract the Mu transition rates that involve interaction with the defect centers. The diffusing tetrahedral interstitial Mu was found to interact with the paramagnetic <c:math xmlns:c=“http://www.w3.org/1998/Math/MathML”> <c:msubsup> <c:mi mathvariant=“normal”>N</c:mi> <c:mrow> <c:mtext>s</c:mtext> </c:mrow> <c:mn>0</c:mn> </c:msubsup> </c:math> center via electron spin exchange. In contrast, they are converted to form a diamagnetic center upon interaction with the negatively charged NV center.

  1. 2021 - Muon Spectroscopy: An Introduction
  2. 1998 - Semiconductors and Semimetals