🔬 materials science

Fine structure of the M-center in Si

This study characterizes the fine structure of the M-center in silicon by identifying multiple temperature-dependent photoluminescence lines, distinguishing a second excited state from other nearby emission features that likely originate from distinct defects or perturbed configurations.

Original authors: Aurora Teien, David R. Gongora, Arnulf Johannes Snedker-Nielsen, Viktor Bobal, Augustinas Galeckas, Peter Granum, Stefano Paesani, Marianne Etzelmüller Bathen, Lasse Vines

Published 2026-09-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Aurora Teien, David R. Gongora, Arnulf Johannes Snedker-Nielsen, Viktor Bobal, Augustinas Galeckas, Peter Granum, Stefano Paesani, Marianne Etzelmüller Bathen, Lasse Vines

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Inside the solid world of silicon, the same material that forms the backbone of our computers and smartphones, scientists are hunting for tiny flaws that behave like artificial atoms. These flaws, known as color centers, are missing pieces or extra atoms trapped within the crystal lattice that can trap light and emit it again at very specific colors. While diamonds and silicon carbide have long been the favorites for this work, silicon offers a distinct advantage: it is the language of the modern electronics industry, meaning any technology built from it could be manufactured on a massive scale with existing tools. Among the various defects found in silicon, one called the M-center has recently drawn attention. It glows with a color that matches the wavelengths used for long-distance fiber-optic communication, and it possesses a magnetic property that could allow it to store information. However, to use it as a building block for future quantum networks, researchers must first understand exactly what it is made of and how it behaves, because the M-center is not just a single, simple point of light but a complex system with hidden layers.

A team of researchers from the University of Oslo and the University of Copenhagen set out to map the fine details of this M-center. They began by creating the defect in high-purity silicon wafers. The process involved shooting carbon ions into the silicon at high speeds, followed by a brief, intense heating to rearrange the atoms. They then introduced hydrogen atoms, which are believed to be a necessary partner for the carbon to form the M-center structure. To test the stability and origin of the light, they created many samples, varying the amount of carbon they implanted and using two different versions of the carbon atom: the common type and a slightly heavier version. After a final heating step at either a moderate or high temperature, they placed the samples in a cryostat, a device that cools them to near absolute zero, and shone a blue laser on them to make the defects glow.

When the researchers examined the light emitted by the M-center, they found that the main glow was not alone. Surrounding the primary beam of light were several fainter, neighboring lines, some shifted slightly toward the blue end of the spectrum and others toward the red. These neighboring lines were so close in energy that they appeared as a fine structure around the main signal. The team then subjected these samples to a series of tests to see how these lines reacted to changes in their environment. They warmed the samples from near absolute zero up to 70 Kelvin and observed how the brightness of each line changed. They also swapped the carbon isotope in some samples and varied the intensity of the initial ion bombardment.

The results revealed a clear distinction between the different lines. The main glow and the red-shifted lines behaved in a standard way: as the temperature rose, they became dimmer, a process known as thermal quenching. These lines also remained unchanged regardless of whether the carbon was the heavy or light version, suggesting they were not directly tied to the vibration of the carbon atom itself. However, one specific blue-shifted line behaved differently. As the temperature rose from the coldest point, this line actually got brighter before it began to fade. This unusual behavior, known as negative thermal quenching, indicates that this specific line of light comes from a higher energy state within the M-center itself, acting like a second step on a ladder that the defect climbs as it warms up. The researchers calculated that the energy required to reach this state is very small, just 1.5 millielectronvolts.

The other blue-shifted lines and the red-shifted lines did not show this unique temperature behavior, nor did they shift when the carbon isotope was changed. This lack of sensitivity to the carbon mass suggests that these extra lines do not come from a different vibration of the M-center itself. Instead, the researchers propose that these lines likely arise from other defects that happen to glow at a similar color, or perhaps from the M-center existing in a slightly different, disturbed configuration. The study also found that the formation of the M-center is delicate; if the initial bombardment of carbon ions is too strong, the defect fails to form or is overwhelmed by other non-glowing defects. Furthermore, a higher final heating temperature seemed to erase the M-center in some samples, leaving only the main glow in others, hinting that the defect is fragile and can be altered or destroyed by heat.

By carefully separating these different signals, the team has provided a clearer picture of the M-center's electronic structure. They confirmed that the defect has a specific magnetic orientation that responds to the direction of the light used to excite it, a property essential for controlling it with lasers. While the exact atomic arrangement of the M-center remains a subject of theoretical debate, the experimental evidence now points to a specific excited state that could be useful for reading and writing quantum information. The work establishes that the M-center is a viable candidate for quantum technologies, provided that the surrounding noise from other defects can be managed. The researchers have mapped the landscape of this defect, showing where the true signal lies and where the background noise hides, laying the groundwork for future engineers to build quantum devices that operate within the familiar world of silicon.

Problem and Context
Color centers in solid-state materials are emerging as candidates for quantum technologies, including quantum networks and photonic quantum computers. While diamond and silicon carbide are established hosts, silicon (Si) offers superior scalability and integration potential due to its mature semiconductor infrastructure. Among various Si defects, the M-center is a candidate of interest due to its telecommunications-band emission and paramagnetic ground state, which shares characteristics with the well-studied T-center. The M-center was first observed in 1973 in electron-irradiated p-type Czochralski silicon, with a zero-phonon line (ZPL) at 0.760 eV (1631 nm). While recent first-principles calculations propose an atomic structure consisting of a carbon-hydrogen complex ( $C_sC_iC_sH_i$ ), direct experimental verification of this structure and a detailed understanding of its electronic fine structure remain incomplete. Specifically, the origin of emission lines in the vicinity of the main M-center ZPL requires further investigation to distinguish between excited states of the M-center, perturbations of its configuration, or entirely separate defects.

Methodology
The authors fabricated samples using float-zone (Fz) and silicon-on-insulator (SOI) wafers. The fabrication process involved:

Ion Implantation: Samples were implanted with $^{13}$ C ions at 33 keV. Three fluences were tested ( $5\times10^{13}$ , $1\times10^{14}$ , and $5\times10^{14}$ cm $^{-2}$ ). A series of samples was also produced using $^{12}$ C ions to investigate isotope effects.
Annealing and Hydrogenation: Following carbon implantation, samples underwent rapid thermal annealing (1000 °C, 20 s). A sacrificial aluminum layer was deposited to align carbon and hydrogen implantation profiles. Hydrogen was implanted at 25 keV, followed by aluminum removal.
Final Annealing: Samples were annealed at either 300 °C or 575 °C for 3 minutes in an $N_2$ atmosphere to optimize defect formation.
Optical Characterization: Photoluminescence (PL) measurements were conducted using a 400 nm continuous-wave laser. Spectra were acquired with a high-resolution imaging spectrograph and an InGaAs detector array at cryogenic temperatures (starting at ~3.7 K). Temperature-dependent sweeps (3.7 K to 70 K) and polarization-dependent measurements were performed to analyze thermal quenching and symmetry.

Key Results
The study identified a rich fine structure surrounding the main M-center ZPL (labeled M1 at ~761.2 meV):

Satellite Peaks: Five additional emission lines were observed. Three are blue-shifted relative to M1 by 1.1 meV ( $M_{b1}$ ), 2.8 meV ( $M_{b2}$ ), and 3.6 meV ( $M_{b3}$ ). Two are red-shifted by 4.7 meV ( $M_{r1}$ ) and 6.4 meV ( $M_{r2}$ ).
Fluence Dependence: The intensity of the blue-shifted peaks ( $M_{b1}$ , $M_{b2}$ , $M_{b3}$ ) is significantly higher in samples with lower implantation fluence ( $5\times10^{13}$ cm $^{-2}$ ) compared to higher fluences. At the highest fluence ( $5\times10^{14}$ cm $^{-2}$ ), M-center emission was undetectable, suggesting that high fluence inhibits formation or promotes competing non-radiative centers. The red-shifted peaks showed no significant fluence dependence.
Isotope Independence: No shift in the satellite peaks was observed between samples implanted with $^{12}$ C versus $^{13}$ C. This lack of isotope shift suggests these peaks do not arise from local vibrational modes dependent on the carbon mass, but rather from perturbations or distinct defect configurations.
Thermal Behavior:
- Most peaks exhibit normal thermal quenching above 30 K.
- The $M_{b2}$ peak (2.8 meV blue-shift) exhibits negative thermal quenching at low temperatures (<30 K), where intensity increases with temperature before quenching occurs at higher temperatures.
- Arrhenius analysis of the $M_{b2}$ intensity yields a thermal activation energy of 1.5 meV.
- The $M_{b2}$ peak is the only satellite line detectable in samples annealed at 575 °C, whereas the main M-center signal is largely suppressed or absent in $^{13}$ C samples at this temperature.
Polarization: Both the T-center and M-center ensembles exhibit a four-fold rotation anisotropy in their polarization dependence, consistent with the low symmetry expected for these defects in the Si lattice.

Conclusions and Significance
The paper reports the observation of fine structure in the M-center emission spectrum. Based on the experimental data, the authors propose the following interpretations:

The $M_{b2}$ peak (2.8 meV blue-shift) likely originates from a second excited state of the M-center. This is supported by its negative thermal quenching behavior and the extracted 1.5 meV activation energy, which is comparable to the energy spacing between bound-exciton levels in the T-center.
The remaining blue-shifted ( $M_{b1}, M_{b3}$ ) and red-shifted ( $M_{r1}, M_{r2}$ ) peaks display normal thermal quenching and are unaffected by carbon isotope substitution or implantation fluence. Consequently, the authors suggest these likely arise from other defects with similar emission energies or from perturbed configurations of the M-center, rather than being intrinsic excited states or vibrational sidebands of the primary M-center structure.

The significance of this work lies in providing novel experimental insights into the electronic structure of the M-center. By characterizing the thermal and fluence dependence of these satellite lines, the study helps distinguish between intrinsic excited states and extrinsic defects. This understanding is essential for developing reliable fabrication and defect-engineering protocols for the M-center, which is a potential candidate for spin-photon interfaces and quantum networking applications due to its telecom emission and paramagnetic properties.

More like this