{"id":"de8308a2-9bb9-48ef-9c79-d34406cbf064","arxiv_id":"2502.07632","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The C center in silicon, which emits at 1571 nm, shows optically detected magnetic resonance, providing the first silicon defect with spin readout in the telecom L-band.","lead":"Researchers report the first optical readout of electron spin states in a carbon-oxygen defect in silicon, using light at telecommunications wavelengths. The result opens a route toward silicon-based quantum memories and networks that operate in the low-loss fiber L-band.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim hinges on the 965/1009 MHz ODMR dips being unambiguously from the C-center triplet; absent a control sample lacking C centers, assignment to another defect or to ODCR/heating artifacts is not excluded.","rationale":"The reader's weakest assumption exactly identifies the load-bearing concern: the ODMR dips may not come from the C-center spin triplet. My review of the manuscript confirms that no control sample, no single-defect measurement, and no direct lifetime measurement support the assignment. The spectral filtering to the 1571 nm line is helpful but not decisive because other radiation-induced defects in silicon can emit nearby, and the ODCR background is a known source of RF-dependent signals. The deviation of measured ZFS parameters from theory further weakens the link to the specific theoretical model. My proposed concrete test—a control sample without carbon implantation—directly tests whether the ODMR signal is specific to the C center. Since the reader already rendered a CONDITIONAL verdict based on this concern, my stress-test does not change the verdict; it reinforces the condition. The paper has several strengths: clear PL and PLE spectra identifying C0 and C1 lines, careful treatment of the ODCR background, and temperature/power dependencies that are internally consistent. But the central claim of a C-center spin-optical interface requires the defect assignment to be secure, and it is not yet. No additional structural or methodological flaws beyond this assignment issue were identified.","tokens_in":9893,"tokens_out":2401,"duration_ms":25224,"concrete_test":"Replicate the ODMR measurement on a control sample cut from the same SOI wafer and processed identically (same proton irradiation and thermal annealing) but without carbon implantation, or with a carbon fluence below the threshold for C-center formation. If the 965/1009 MHz dips persist in the control sample, the signal cannot be uniquely assigned to the C center and the central claim is not established. As a complementary check, sweep a narrow-band detector across 1540–1620 nm while holding RF at 987 MHz and verify that the ODMR contrast tracks only the C0 (1571 nm) and C1 (1560 nm) emission lines and not the ODCR background.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's strongest claim is the first demonstration of a spin-optical interface for the C center in the telecom L-band via ODMR. For this claim to hold, the two ODMR dips at 965 MHz and 1009 MHz in Fig. 1d must originate from spin transitions between the m_s=0 and m_s=±1 sublevels of the C_T triplet of the C center. The evidence provided is spectral filtering to the 1571 nm C0 line and consistency with the theoretical model of Ref. 17. However, the sample is a carbon-implanted, proton-irradiated SOI device layer, which is expected to contain a variety of radiation-induced and carbon-related defects, some of which emit in the 1540–1620 nm range. The paper explicitly acknowledges a strong ODCR background and uses a pulsed RF scheme to reduce it, but residual ODCR structure or microwave-heating effects could produce narrow dips unrelated to the C-center spin. No control sample without C centers was measured; no single-defect ODMR was performed; and the predicted spin-dependent lifetimes (m_s=0 > 1.4 ms, m_s=±1 > 10 ms) that form the basis of the proposed readout mechanism were not directly measured. Furthermore, the measured zero-field parameters (D/h=987 MHz, E/h=22 MHz) deviate from the theoretical predictions of Ref. 17, and the explanation given (suppression of spin-orbit coupling by vibrational modes) is plausible but speculative. The Discussion itself notes that 'coherence studies and the isolation of a single C center are still in progress,' underscoring that the defect identity has not been confirmed at the single-defect level. Therefore, while the observation of two RF-dependent dips in the 1571 nm PL is suggestive, the central claim is not uniquely established without excluding other origins.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports ensemble-level optical spectroscopy and optically detected magnetic resonance (ODMR) measurements of the C center (Ci-Oi pair) in carbon-implanted, proton-irradiated SOI silicon. Photoluminescence shows the C0 line at 1571 nm and C1 line at 1560 nm; PLE resolves C0 through C4 excited states. Under zero magnetic field, ODMR shows two dips at 965 and 1009 MHz, interpreted as the E-split transitions within an S=1 triplet excited state (CT) with D/h=987 MHz and E/h=22 MHz. The paper also reports ODMR on the C1 line, temperature/laser-power/RF-power dependencies, and anisotropic Zeeman splitting for three magnetic-field orientations. The central claim is the first demonstration of a spin-optical interface for the C center, with optical spin readout in the telecom L-band.","tokens_in":10266,"tokens_out":4844,"duration_ms":50544,"significance":"If the assignment of the ODMR signal to the C-center triplet is correct, this is an important result: it would be the first silicon defect with optically detected spin states in the telecom L-band, and it would provide experimental validation of the DFT predictions in Ref. 17. The paper has clear strengths: the PL and PLE spectra identify the known C-center lines; the ODMR signal is spectrally filtered to the 1571 nm C0 line; the interpretation is compared with an external, independently authored theory; and the authors explicitly acknowledge and partially mitigate the ODCR background. However, the claim rests on ensemble measurements with no control sample lacking C centers, no error bars or repeated statistics, no direct measurement of the spin-dependent lifetimes, and only a qualitative comparison with the spin Hamiltonian. These gaps are load-bearing for the 'first demonstration' claim and require experimental strengthening before publication.","major_comments":[{"comment":"The central assignment of the 965/1009 MHz ODMR dips to the C-center CT triplet is not uniquely established. The sample is carbon-implanted and also proton-irradiated, which is expected to produce multiple radiation-induced defects; the paper itself acknowledges a strong ODCR background that depends on RF power. No control sample without carbon (or with varied carbon fluence) is measured, and no single-defect ODMR is shown; the Discussion states that single-C-center isolation is still in progress. Without such a control or a clear correlation between the ODMR amplitude and the C0-line intensity, assignment to another defect or to residual ODCR/heating structure remains plausible. This issue is directly load-bearing for the claim of the first C-center spin-optical interface.","section":"Results, Fig. 1d; Methods, ODMR measurement"},{"comment":"No error bars, repeated traces, or statistical analysis are presented for any ODMR spectrum. The ODMR contrast is small and is superimposed on a large, RF-power-dependent background; single traces do not establish reproducibility of the 965/1009 MHz dips. At minimum, repeated independent measurements (e.g., multiple cooldowns and RF sweeps) with uncertainty intervals should be shown to demonstrate that the features are not noise or artifacts.","section":"Fig. 1d; Fig. 2"},{"comment":"The proposed readout mechanism assumes spin-dependent lifetimes (m_s=0 > 1.4 ms, m_s=±1 > 10 ms) and an ISC selection rule Δm_s=0 taken from Ref. 17, but these quantities are not measured here. Because the negative ODMR contrast is the central observable, the manuscript should either measure these lifetimes/ISC rates or explicitly label the mechanism as a hypothesis consistent with, but not established by, the data. This does not invalidate the ODMR observation itself, but it limits the strength of the 'optical spin readout' claim.","section":"Results, ODMR mechanism paragraph"},{"comment":"The measured ZFS parameters D/h=987 MHz and E/h=22 MHz deviate from the DFT predictions of Ref. 17, and the proposed explanation (vibrational suppression of spin-orbit coupling, with citation to Ham) is not quantitatively compared. In addition, Eq. (2) is labeled H_Z but defines the spin-spin term, while the Zeeman Hamiltonian is introduced later as Eq. (4). The authors should correct the labeling and either provide a quantitative comparison (e.g., calculated D/E including vibronic corrections) or clearly state the discrepancy as an open question. The anisotropic Zeeman data in Fig. 3 are also interpreted only qualitatively; a fit to the spin Hamiltonian would substantially strengthen the assignment.","section":"Eqs. (1)-(4) and following paragraph"}],"minor_comments":[{"comment":"The label H_Z should be H_SS, since the equation defines the spin-spin interaction; the Zeeman term is defined in Eq. (4).","section":"Eq. (2)"},{"comment":"The phrase 'two sharp resonant peaks centered at 987 MHz' is ambiguous; it should read 'a pair of peaks centered at 987 MHz' to clarify that the two peaks are at 965 and 1009 MHz.","section":"Results, Fig. 1d caption and text"},{"comment":"The displayed equation for ODMR contrast is missing from the text: the sentence 'The ODMR contrast was deduced as follows:' is followed immediately by 'where PL represents...' with no equation. Please insert the formula.","section":"Methods, ODMR measurement"},{"comment":"Reference 9 appears to contain an incorrect author name ('Inc, P.'); please verify and correct the citation.","section":"References"},{"comment":"The statement that carriers are transported from C1-C4 to C0 is an inference from the PLE spectrum; please phrase this as an inference rather than a direct observation.","section":"Results, PLE paragraph"},{"comment":"The sentence 'its lack of dark spin sublevels implies a potentially more stable spin state, offering longer coherence times' is speculative and not supported by the measurements presented; please soften or provide supporting evidence.","section":"Discussion"}],"recommendation":"major_revision","confidential_remarks":"The manuscript addresses a timely topic and the basic experimental observation is plausible, but the central claim of a first C-center spin-optical interface requires stronger evidence for the defect assignment. A control sample without carbon centers, or single-defect ODMR/correlation of ODMR with the C0 PL intensity, would resolve the main concern. I did not see evidence of inappropriate citation practices; the reference list is appropriate. The paper may become publishable after the major experimental gaps are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper reports the first ODMR on the C center in silicon, a defect with a 1571 nm zero-phonon line in the telecom L-band. If the assignment holds, this closes an obvious gap—T center is O-band, so an L-band spin-photon interface is genuinely new. The experiment is straightforward but well executed: they spectrally filter to the C0 line, see a clean ZFS doublet at 965/1009 MHz, and the same 965 MHz dip shows up when reading out the C1 line. The anisotropic splitting under external field looks like what you'd expect from an ensemble of oriented triplets, similar to NV centers in diamond. That is real evidence, not just a single peak.\n\nThe soft spots are real but not fatal. The sample is carbon-implanted plus proton-irradiated, so other defects are present; there is no control sample without C centers. The authors acknowledge a strong ODCR background and use pulsed RF to suppress it, but they do not fully rule out RF artifacts. No error bars or repeated statistics are shown, and the readout mechanism leans on theoretical lifetimes from Ref. 17 that are not directly measured. The measured ZFS differs from that theory, and the explanation (vibrational suppression of spin-orbit) is plausible speculation.\n\nThere are minor labelling slips in the Hamiltonian section—Eq. (2) is the spin-spin term, not the Zeeman term, and D/E get called longitudinal/transverse inconsistently. These are fixable.\n\nOverall, the central claim is likely right but not uniquely established. A referee should ask for a control sample, error bars, and ideally single-defect ODMR. That is addressable in revision. The paper deserves serious peer review. I would cite it once the assignment is confirmed; even now it's the only L-band spin readout candidate in silicon.","headline":"First ODMR of the C center in silicon—a plausible L-band spin-photon interface, but the defect assignment needs a control sample before I'd call it airtight.","tokens_in":10818,"tokens_out":2608,"would_cite":true,"duration_ms":26612,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that the C center, a carbon-oxygen defect in silicon with a zero-phonon line at 1571 nm, provides the first optically detected spin readout in the telecom L-band, driven by microwave transitions in its excited triplet…","keywords":["C center","silicon color center","telecom L-band","optically detected magnetic resonance","spin-photon interface","carbon-oxygen defect","quantum memory"],"falsifier":"Run an identically annealed and proton-irradiated control sample that received no carbon implantation through the same ODMR sequence: if the 965 and 1009 MHz dips persist, the assignment to C-center spins is wrong.","tokens_in":9720,"feed_emoji":"📡","tokens_out":19686,"duration_ms":164932,"temperature":0.7,"pith_summary":"The paper claims to have demonstrated optical readout of spin states in the C center, a carbon-oxygen defect in silicon whose sharp emission line sits at 1571 nm in the telecom L-band. Exciting the defect with light near or below the silicon bandgap while sweeping microwaves, the authors see two dips at 965 and 1009 MHz in the 1571 nm photoluminescence, which they attribute to zero-field transitions between spin sublevels of an excited triplet state of the defect. If that assignment holds, the C center is the first silicon color center whose spin state can be read out through telecom L-band light, the wavelength range where silica fibers lose least signal. A spin-photon interface in this band matters because it would let quantum information stored in a silicon defect couple to light that can travel long distances through fiber, a practical ingredient for quantum communication and quantum memory.","feed_headline":"Silicon defect reads out spins at 1571 nm","feed_subtitle":"A microwave-driven carbon-oxygen defect links spin states to telecom L-band light, a step toward silicon quantum memories.","key_machinery":"The central mechanism is optically detected magnetic resonance through the $C_T$ spin-triplet excited state of the C center, an interstitial carbon-oxygen pair defect whose zero-field splitting parameters are $D/h = 987$ MHz and $E/h = 22$ MHz. The triplet's $m_s=0$ sublevel decays non-radiatively to the ground state in more than 1.4 ms, while the $m_s=\\pm1$ sublevels take more than 10 ms; resonant microwaves transfer population from $m_s=0$ to the longer-lived $m_s=\\pm1$ states, trapping carriers and reducing the 1571 nm photoluminescence. The ODMR readout is the difference in photoluminescence intensity with and without resonant microwaves, measured by lock-in detection, and pulsed RF excitation mitigates a background from microwave-induced free-carrier heating.","core_discovery":"The paper's central claim is that the C center, a carbon-oxygen interstitial pair made by carbon implantation, annealing, and proton irradiation, has an optically readable spin state in an ensemble. Under 1064 nm or 1310 nm excitation, spectrally filtering the photoluminescence to the C0 line at 1571 nm reveals two microwave resonances at 965 and 1009 MHz at zero magnetic field, interpreted as the longitudinal and transverse zero-field splittings $D/h = 987$ MHz and $E/h = 22$ MHz of an $S=1$ triplet excited state ($C_T$) that lies 2.64 meV below the C0 level. The mechanism is spin-dependent non-radiative decay: the $m_s=0$ sublevel returns to the ground state in more than 1.4 ms, whereas the $m_s=\\pm1$ sublevels take more than 10 ms, so resonant microwaves that pump population into $m_s=\\pm1$ trap carriers and lower the 1571 nm emission. Magnetic-field-dependent measurements show anisotropic splitting consistent with multiple quantization axes and a tensorial g-factor. The paper concludes that the C center is the first spin-optical interface in the telecom L-band, validating the theoretical prediction of an L-band emitter with quantum memory in silicon.","pith_inferences":["With single-defect confirmation, the same pulsed-RF protocol would likely transfer to other silicon defects that have spin-dependent non-radiative decay paths, expanding the set of telecom-band spin-photon interfaces.","If the non-magnetic singlet ground state indeed shields nearby nuclear spins from magnetic noise, the practical quantum memory would live in those nuclei rather than in the defect's own electron spin; a nuclear-spin coherence measurement under repeated optical pumping would test that directly.","The measured transverse splitting of 22 MHz means the zero-field ODMR frequency shifts with magnetic-field orientation, so an oriented ensemble of C centers could serve as a vector field sensor once the g-tensor axes are mapped.","Because 1310 nm excitation already produces the C0 emission, a silicon microcavity tuned to 1571 nm could in principle combine resonant spin preparation with single-photon emission entirely inside telecom bands, provided the emission rate is enhanced by the cavity."],"forward_implications":["The C center can be read out at 1571 nm, so spin-state information is available directly in the L-band without frequency conversion.","Both the C0 (1571 nm) and C1 (1560 nm) lines show the same microwave resonances, giving two telecom channels for spin readout.","The anisotropy of the magnetic-field splitting provides a way to identify quantization axes in ensembles, and the complexity should reduce when scaling to single defects.","Because the C center ground state is a singlet, nearby nuclear spins are magnetically isolated while the defect is in the ground state, which supports using 13C nuclei as long-lived quantum memories.","The measured zero-field splittings deviate from theoretical values, indicating that vibrational effects must be included in a complete model of the C center's spin physics."],"supporting_citations":[{"why":"Supplies the theoretical prediction of the spin-triplet excited state, its zero-field splitting, and the spin-dependent lifetimes that the experiment validates.","marker":"17"},{"why":"Provides the T-center precedent of an optically active spin-photon interface in silicon, the benchmark the C center extends from the O-band to the L-band.","marker":"7"},{"why":"Identifies the C center as a carbon-related radiation damage center in silicon, grounding the defect assignment.","marker":"10"},{"why":"Provides the pseudodonor excited-state model and the spin Hamiltonian used to describe the C center's triplet levels.","marker":"12"},{"why":"Documents the low-temperature optical activity of the triplet state of excitons bound to carbon-oxygen defects, supporting the proposed intersystem crossing path.","marker":"16"},{"why":"Explains how free-carrier heating and cyclotron resonance create a background that obscures ODMR in silicon, motivating the pulsed RF scheme.","marker":"22"},{"why":"Introduces delayed optical detection of magnetic resonance for defects in silicon and gallium arsenide, the background-rejection approach adapted here.","marker":"23"}],"fun_headline_variants":["Telecom L-band spin readout in silicon defect","First spin readout in telecom L-band for silicon","Silicon color center reads spins via 1571 nm light","C center: spin readout at telecom wavelength","Microwave-driven C center enables L-band spin readout"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central claim stands or falls on whether the two microwave dips come from spin transitions in the C center's own triplet state rather than from some other defect, from the cyclotron-resonance background, or from microwave heating.","fun_headline_variants_meta":{"raw":{"variants":["Telecom L-band spin readout in silicon defect","First spin readout in telecom L-band for silicon","Silicon color center reads spins via 1571 nm light","C center: spin readout at telecom wavelength","Microwave-driven C center enables L-band spin readout"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000223,"raw_usage":{"total_tokens":1474,"prompt_tokens":981,"completion_tokens":493,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":415}},"tokens_in":597,"tokens_out":493,"duration_ms":4865,"temperature":1.0,"reasoning_tokens":415,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T12:05:11.937713+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run an identically annealed and proton-irradiated control sample that received no carbon implantation through the same ODMR sequence: if the 965 and 1009 MHz dips persist, the assignment to C-center spins is wrong.","supporting_citations":[{"cited_title":"& Gali, A","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical prediction of the spin-triplet excited state, its zero-field splitting, and the spin-dependent lifetimes that the experiment validates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the T-center precedent of an optically active spin-photon interface in silicon, the benchmark the C center extends from the O-band to the L-band."},{"cited_title":"C., Woolley, R., Newman, R","cited_arxiv_id":null,"evidence_quote":"Identifies the C center as a carbon-related radiation damage center in silicon, grounding the defect assignment."},{"cited_title":"H., Monemar, B","cited_arxiv_id":null,"evidence_quote":"Provides the pseudodonor excited-state model and the spin Hamiltonian used to describe the C center's triplet levels."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the low-temperature optical activity of the triplet state of excitons bound to carbon-oxygen defects, supporting the proposed intersystem crossing path."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Explains how free-carrier heating and cyclotron resonance create a background that obscures ODMR in silicon, motivating the pulsed RF scheme."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces delayed optical detection of magnetic resonance for defects in silicon and gallium arsenide, the background-rejection approach adapted here."}],"review_version":1}