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, H. Abe, T. Ohshima
DOI10.1103/jn9c-fprb

Diffusive tetrahedral muonium was trapped by negatively charged NV centers in type-Ib diamond at a fitted rate of 9 ± 1 MHz at 20 K. The corresponding rate was 2.8 ± 0.4 MHz at 290 K, so trapping was about three times stronger at low temperature. Electron spin exchange with substitutional nitrogen produced a separate relaxation channel for the same muonium state.

The experiment compared two synthetic single-crystal diamond samples in granular form with an average grain diameter of 1.2 mm. The pristine sample contained about 100 ppm substitutional neutral nitrogen, or $N_s^0$, based on electron paramagnetic resonance estimates. This P1 center replaces carbon and carries one deep-donor electron. Each sample weighed about 2 g.

The second sample began as the same material. Electron-beam irradiation created carbon vacancies, and heat treatment at 900 °C for 2 h formed NV centers. The estimated NV concentration was approximately $4 \times 10^{17}\ \mathrm{cm^{-3}}$, or 2.5 ppm. More than 90% of these centers were considered $\mathrm{NV^-}$ at room temperature.

Implanted muons can capture an electron during thermalization and form neutral muonium. The model treated tetrahedral $\mathrm{Mu}T^0$ and bond-centered $\mathrm{Mu}{BC}^0$ as active states. Their hyperfine constants were 3711 MHz, 168 MHz parallel, and -393 MHz perpendicular.

$\mathrm{Mu}T^0$ was expected to diffuse rapidly, with a prior room-temperature diffusion coefficient near $10^{-2}\ \mathrm{cm^2,s^{-1}}$. $\mathrm{Mu}{BC}^0$ was considered polaronic and immobile up to 1000 K.

The EMU spectrometer at ISIS used 100% polarized 4 MeV surface muons. Its muon range was about 200 mg/cm², and its beam diameter was about 1 cm FWHM. A closed-cycle refrigerator provided measurements at 290 and 20 K. Longitudinal fields covered 0 to 4000 G and were parallel to the initial muon spin polarization. Each plotted spectrum averaged 5 million muon decay events.

At zero field, the normalized polarization was about 0.1 because fast dynamics removed most of the signal. Polarization recovered by 400 G, then continued recovering toward 4000 G. The field response differed between pristine and NV samples, especially at 20 K.

Each $\mu$SR spectrum combined several muonium populations, so relaxation could not be assigned from a single trace. The model assumed that only $\mathrm{Mu}T^0$ and $\mathrm{Mu}{BC}^0$ were active. The remaining muons contributed a constant diamagnetic background.

QUANTUM solved the time evolution of the muon density matrix and calculated $P(t)$. The authors fitted all field scans for each sample and temperature as one dataset. This global fit separated the diffusing population from bond-centered muonium and resolved transition rates between states. Electron-spin relaxation was included for each state.

Diffusive motion was not calculated explicitly. Its effect was absorbed into fitted transition and relaxation rates. Because ${}^{12}\mathrm{C}$ has no nuclear spin and accounts for about 99% of natural carbon isotopes, the model omitted a nuclear-spin bath. Representative reduced chi-squared values ranged from 1.174 to 1.483.

Nitrogen spin exchange in pristine diamond

Section titled “Nitrogen spin exchange in pristine diamond”

The pristine data were consistent with nitrogen as the dominant interaction. Fitted prompt fractions were 0.600 ± 0.005 for $\mathrm{Mu}_T^0$ at 290 K and 0.586 ± 0.004 at 20 K. Bond-centered fractions were 0.315 ± 0.007 and 0.324 ± 0.007. These values agreed with earlier reported fractions of about 0.7 and 0.2.

The global fit returned these rates:

SampleTemperature$\mathrm{Mu}_T^0$ relaxation rate, MHz$\mathrm{Mu}T^0$ to $\mathrm{Mu}{BC}^0$ rate, MHz$\mathrm{NV^-}$ trapping rate, MHz
Pristine290 K58 ± 41.19 ± 0.09Not present
Pristine20 K53 ± 31.00 ± 0.06Not present
NV290 K73 ± 61.192.8 ± 0.4
NV20 K100 ± 121.009 ± 1

Values without uncertainties were fixed to the pristine results during the NV fits. The large $\mathrm{Mu}_T^0$ relaxation rate caused its fast early-time depolarization. The authors attributed this relaxation to electron spin exchange with paramagnetic $N_s^0$ centers.

For $\mathrm{Mu}_T^0$ collisions, the rate was written as $nv\sigma$. The nitrogen density was estimated as $2 \times 10^{19}\ \mathrm{cm^{-3}}$. The measurement could not separate effective velocity $v$ from scattering cross section $\sigma$. The donor electron lies 2.2 eV below the conduction-band edge, so the authors treated $\sigma$ as nearly unchanged between 290 and 20 K. Comparable fitted rates left the effective velocities comparable in magnitude, but the authors left this interpretation unresolved.

$\mathrm{Mu}_{BC}^0$ relaxed at only 0.02 ± 0.02 MHz at 290 K and 0.22 ± 0.04 MHz at 20 K. Its site-exchange rate toward $\mathrm{Mu}_T^0$ was 0.034 ± 0.002 MHz in both fits. The authors described the $\mathrm{Mu}T^0$ to $\mathrm{Mu}{BC}^0$ rate as about two orders of magnitude larger. This agreed with prior work that predicted bond-centered muonium as the more stable site.

The pristine model failed to fit the NV spectra, which required an extra reaction path. Most NV centers were negatively charged, so the authors added a diamagnetic muonium state formed by interaction with $\mathrm{NV^-}$. The model assigned charge exchange with $\mathrm{NV^-}$ to the formation of $\mathrm{Mu}_D^-$.

The model allowed conversion in both directions between $\mathrm{Mu}_T^0$ and $\mathrm{Mu}_D^-$. Fitted trapping rates rose from 2.8 ± 0.4 MHz at 290 K to 9 ± 1 MHz at 20 K. The reverse rates were much smaller, 0.065 ± 0.004 MHz and 0.034 ± 0.003 MHz, respectively. The forward transition therefore behaved as an effective trap.

The authors suggested that diffusive $\mathrm{Mu}_T^0$ could bind to $\mathrm{NV^-}$ through a covalent bond. This remained a proposed interpretation rather than a directly observed bond. The experiment did not optically excite the sample, so all $\mathrm{NV^-}$ centers occupied their ground state. Their $m_S=0$ and $m_S=\pm1$ levels were split by 2.87 GHz at zero field. The applied field had little effect on their population at 20 K.

Surface muons have limited depth resolution, so this experiment probed a sample volume rather than a resolved subsurface layer. Lower-energy muons could improve depth profiling, but that application remained prospective. Granular material also averaged crystallographic orientation over the sample.

Several assumptions limit the extracted rates. Only two temperatures were measured, and the fit did not resolve velocity from capture cross section directly. The NV fits fixed three host-lattice rates to pristine values. High-field repolarization offsets of about 0.02 also came from silver backgrounds and incomplete beam coverage.

The threefold increase in NV trapping could not arise from a large rise in $\mathrm{NV^-}$ density. Conversion was already above 90% at room temperature, leaving little room for a density increase. The authors suggested a larger capture cross section at 20 K, perhaps from excess-electron delocalization by tunneling or zero-point motion.

They also reported a statistically significant rise in $\mathrm{Mu}_T^0$ relaxation in the NV sample. Its cause was unknown, so it could not be assigned to velocity or scattering.

The method can detect both unpaired-electron defects and charged defects without magnetic moments. Paramagnetic nitrogen causes electron spin exchange and fast $\mathrm{Mu}_T^0$ relaxation. Charged $\mathrm{NV^-}$ defects instead capture $\mathrm{Mu}_T^0$ and convert it into a diamagnetic state.

For diamond material selection, substitutional nitrogen and NV charge state both control the measured response. A high nitrogen concentration supplies the spin-exchange channel, while sufficient $\mathrm{NV^-}$ population supplies the trapping channel. The paper therefore provides a way to test both defect responses in one muon experiment.

The authors also propose muonium probing for Si and SiC. They suggest low-energy muons could extend defect measurements into subsurface layers. These ideas remain proposed applications, while the present measurements established the diamond reaction pathways.

The application requires nitrogen-rich type-Ib single-crystal diamond, with an electron-irradiated and annealed condition for NV formation. 6CCVD’s diamond materials catalog includes MPCVD-grown single-crystal plates up to 125 mm, which match the crystal family used for this defect-host study. The paper’s target nitrogen concentration and defect treatment can be discussed through engineering support for custom fabrication.

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