{"id":"2a2aae04-91cb-408b-baa7-9de1273c3e9a","arxiv_id":"2504.19875","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A fiber laser based stimulated Raman photothermal microscope with low-NA, long-working-distance optics achieves high-sensitivity chemical imaging with improved SNR over fiber-laser SRS and works for multi-well plates and thick cleared tissue.","lead":"This paper demonstrates a fiber-laser stimulated Raman photothermal microscope that avoids the bulky solid-state lasers and high-numerical-aperture optics that normally constrain Raman imaging. The system enables label-free chemical imaging of cells, multi-well plates, and thick cleared tissue from a compact source.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 105x/12x SNR comparison behind the two-order-of-magnitude claim lacks a documented protocol; if SRS was measured at a lower modulation frequency or different lock-in bandwidth, the headline improvement could be inflated.","rationale":"The reader correctly identified the urea thermal-enhancement premise (missing dn/dT and lack of a water-versus-urea comparison) as a weakness in a secondary component of the claims. However, the more load-bearing issue for the central abstract claim is the quantitative SNR improvement, because the entire novelty of 'two-order-of-magnitude improvement' rests on the ~105x/12x numbers in Section 2.3 and Fig. S3. That comparison is under-specified: no modulation frequencies, lock-in settings, detection bandwidth, or error bars are reported, and the paper elsewhere shows strong frequency dependence of SRP SNR. This is a missing-control problem rather than an internal inconsistency. The reader did flag the missing error bars/protocol in the rationale, so there is partial agreement, but the reader's formal weakest assumption was urea. Since the recommended verdict remains conditional pending the same class of verification, I keep the verdict unchanged. A single controlled re-measurement would settle whether the headline SNR ratio is robust; if it collapses, the paper's main quantitative claim weakens substantially, while the low-NA and tissue-clearing demonstrations would still stand as qualitative advances.","tokens_in":16792,"tokens_out":4335,"duration_ms":47130,"concrete_test":"Reproduce the Fig. S3 SNR comparison on DMSO with all three detection modes at the same modulation frequency (e.g., 600 kHz), same lock-in time constant and acquisition time, same pump/Stokes powers, and report N≥5 replicates with error bars. Also repeat at 125 kHz. If the SRP-to-un-balanced-SRS SNR ratio changes by more than a factor of about 3 between the two frequencies, or if the 95% confidence intervals overlap with a 10x ratio, the two-order-of-magnitude claim should be downgraded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2.3 reports the central quantitative claim: fiber-laser SRP gives ~105-fold SNR over un-balanced fiber-laser SRS and ~12-fold over autobalanced SRS on DMSO (Fig. S3). This ratio is load-bearing for the abstract's 'two-order-of-magnitude improvement' claim. The supporting comparison is under-specified. Fig. S3 gives powers (28 mW pump, 90 mW Stokes, 23 mW probe) but not the modulation frequencies, lock-in time constant or filter slope, detection bandwidth, or number of replicates and error bars. Since §2.3 shows SRP SNR is strongly frequency-dependent, peaking at 600 kHz, while the earlier thermal-lens characterization and SRS modulation-depth measurement (Fig. S2) were made at 125 kHz, a comparison run at different frequencies would conflate modality advantage with operating-point choice. No error bars are provided for any SNR value, and no statement that all three modes used identical lock-in settings and acquisition times. If the un-balanced SRS measurement was taken at a lower frequency where fiber-laser 1/f noise is larger, the 105x ratio would be inflated. This concern is about the evidence base, not the underlying physics: SRP's independent probe beam is a plausible noise-isolation mechanism, but the quantitative headline requires a controlled, fully specified comparison.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a stimulated Raman photothermal (SRP) microscope driven by a compact picosecond fiber laser, using a third CW probe beam and low-NA long-working-distance collection in an inverted geometry. It claims a ~105-fold SNR improvement over un-balanced fiber-laser SRS and ~12-fold over autobalanced SRS on DMSO; demonstrates 240 nm lateral resolution; shows spectral fidelity under air-condenser collection; images live cells and urea-cleared rat brain to >200 μm depth; and proposes 8 M urea as a thermal enhancement medium. The text includes thermal-lensing simulations, hyperspectral imaging with LASSO unmixing, and low-NA objective demonstrations.","tokens_in":16999,"tokens_out":3427,"duration_ms":32771,"significance":"If the quantitative claims are substantiated, this is a notable practical advance: replacing bulky OPO sources with a fiber laser and relaxing collection NA would expand coherent Raman chemical imaging to multi-well plates, thick cleared tissues, and contact-free clinical sampling. The paper includes self-contained measurements (SRS modulation depth as simulation input, qualitative validation of NA optimization), and the core imaging demonstrations appear internally consistent. Strengths include a clear motivation, direct comparison to prior autobalanced SRS, and explicit discussion of limitations such as probe coherence, laser power constraints, and background absorption. However, the headline SNR ratios and the urea thermal-enhancement mechanism require stronger evidence before the central claims can be fully accepted.","major_comments":[{"comment":"The ~105-fold and ~12-fold SNR improvements are the quantitative basis for the abstract's 'two-order-of-magnitude' claim, but the comparison is under-specified. The text and Fig. S3 list laser powers and duty cycle but not modulation frequencies, lock-in time constant/filter slope, detection bandwidth, number of replicates, or error bars. Since Fig. 3b shows SRP SNR strongly depends on modulation frequency (peaking at 600 kHz), while the SRS modulation-depth measurement in Fig. S2 was made at 125 kHz, a comparison at different frequencies could conflate modality advantage with operating-point choice. Please provide a fully specified protocol with identical settings for all three modalities and include replicate statistics; also clarify whether the autobalanced SRS data were acquired with the same system or taken from reference [21].","section":"Section 2.3, Fig. S3"},{"comment":"The claim that urea provides 'thermal enhancement' (abstract and Figs. 6-7) is not supported by the data. Table S1 lists the thermo-optic coefficient (dn/dT) for 8 M urea as '-' and no direct water-versus-urea SNR or contrast comparison is shown. Lower heat capacity (1420 vs 4184 J/kg/K) alone does not guarantee higher SRP signal if urea's dn/dT is smaller than water's; the relative signal intensity column is not given for urea. Please either measure dn/dT or provide an experimental water/urea SNR comparison. Otherwise, restrict the claim to tissue clearing and state that thermal enhancement is a hypothesis.","section":"Section 2.6/2.7, Table S1"}],"minor_comments":[{"comment":"The Discussion states '1~2 orders SNR improvement' while the abstract claims 'two-order-of-magnitude improvement'; please harmonize these statements.","section":"Discussion"},{"comment":"The text refers to a '0.12 MHz high pass filter'; please verify the units, since 0.12 MHz (120 kHz) seems plausible but is written unusually.","section":"Section 4.1"},{"comment":"The phrase 'axions and dendrites' should read 'axons and dendrites'.","section":"Section 2.7"},{"comment":"The protocol titled 'Preparation of rat brain slice with tissue clearance' actually describes ovarian tumor cryosectioning; this appears to be a labeling error and should be corrected.","section":"Supplementary Method"},{"comment":"Panels c and d are described as 'spectral fidelity' and 'intensity-concentration relation,' but the text discusses limit of detection (LOD); please define LOD in the caption and report the linear-fit parameters if available.","section":"Fig. 3 caption"},{"comment":"The LASSO regularization parameter β is stated to be 'empirically optimized' without giving its value or selection criterion; please report it to improve reproducibility.","section":"Section 4.7"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Xiaowei Ge and colleagues have put together a genuinely useful systems paper. The new bits are: a tunable fiber laser driving SRP, a low-NA long-working-distance air condenser that keeps spectral fidelity where autobalanced SRS blurs, and urea used as both a clearing agent and a thermal-enhancement medium. The inverted geometry with a 28 mm air gap is a real practical step toward multi-well and thick-tissue imaging. The demonstrations are internally consistent: 240 nm lateral resolution, live-cell lipid trajectories, and 200 um depth in urea-cleared brain all support the platform claim.\n\nThe strongest part of the paper is the spectral-fidelity comparison in Fig. 4. It directly shows SRP preserving the PMMA Raman spectrum under 0.55 NA collection while SRS does not. That is a clean, important result. The thermal-lensing simulation, which takes the measured SRS modulation depth as input and is matched qualitatively by experiment, is also reasonable.\n\nThe soft spots are where you'd expect. The headline 105x/12x SNR improvement over SRS is not fully documented. Fig. S3 lists pump/Stokes/probe powers but not modulation frequency, lock-in bandwidth, time constant, or replicate counts for each mode. Because both SRP and SRS SNR depend on operating point, the ratio could be inflated if the SRS arm was measured at a less favorable frequency. This is a documentation gap, not a red flag in the physics; the independent probe beam is a plausible noise-isolation mechanism. But the abstract's 'two-order-of-magnitude' line should be backed by a controlled, fully specified comparison with error bars.\n\nThe urea story is half-supported. The clearing benefit is demonstrated. The thermal-enhancement claim relies on lower heat capacity, yet Table S1 leaves dn/dT for 8 M urea blank and no water-vs-urea SNR or contrast measurement is shown. The relative signal intensity column doesn't even list a value for urea. It is plausible that urea enhances signal, but it is currently an assumption, not a measurement.\n\nMinor point: data and code are 'available upon request'—fine for a methods paper, but a public repository would make the SNR numbers easier to verify.\n\nThis paper deserves a serious referee. The engineering advance is real, the spectroscopy is careful, and the limitations are at least partially acknowledged in the discussion. My recommendation: send it to peer review, and in revision require the SNR comparison protocol and a direct measurement or literature value for urea's thermo-optic coefficient.","headline":"Solid fiber-laser SRP systems paper; the quantitative SNR headline needs a fully specified comparison protocol before I'd trust it.","tokens_in":17625,"tokens_out":3727,"would_cite":true,"duration_ms":33209,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Switching SRS readout from pump-beam loss to a probe-beam thermal-lens measurement yields ~105-fold higher SNR than unbalanced fiber-laser SRS and ~12-fold higher than autobalanced SRS, while allowing low-NA, long-working-distance optics.","keywords":["stimulated Raman photothermal microscopy","fiber laser","thermal lensing","long working distance","label-free imaging","tissue clearing","urea","SNR improvement"],"falsifier":"Measure the thermo-optic coefficient of 8 M urea near room temperature and compare SRP SNR on a standard sample (e.g., DMSO or 3 µm PMMA beads) in water versus 8 M urea with identical pump/Stokes/probe powers, modulation, and collection NA; if the urea signal is not proportionally higher, the thermal-enhancement claim is refuted.","tokens_in":16545,"feed_emoji":"🔬","tokens_out":10625,"duration_ms":94144,"temperature":0.7,"pith_summary":"Fiber-laser-based stimulated Raman photothermal (SRP) microscopy detects the thermal lens left behind after Raman vibrational relaxation, rather than the small intensity loss of the pump beam. Because this readout uses a separate continuous-wave probe beam, it sidesteps the intensity noise of the compact fiber laser that otherwise limits SRS sensitivity, giving ~105-fold better signal-to-noise than fiber-laser SRS without balanced detection and ~12-fold better than autobalanced SRS on dimethyl sulfoxide. The same geometry removes the need for high-numerical-aperture collection, so a 0.55-NA air condenser with a 28 mm working distance suffices, opening the sample space to multi-well plates, flow cells, and thick tissues. The paper also identifies 8 M urea as a dual-purpose immersion medium that clears tissue and enhances the photothermal signal, enabling >200 µm deep volumetric imaging of cleared rat brain.","feed_headline":"Fiber-laser Raman gains ~105x signal by reading heat, not loss","feed_subtitle":"Thermal-lens readout lets compact lasers image thick cleared tissue with low-NA optics.","key_machinery":"The central object is the thermal lens generated by SRS-driven vibrational relaxation. A CW probe beam, slightly offset from the pump/Stokes focus, experiences a spatially varying refractive-index gradient and its far-field intensity changes; that change is demodulated on a lock-in amplifier. Because the modulation is imposed on the pump/Stokes beams and sensed by the low-noise probe, the fiber laser's intensity noise is effectively bypassed. Simulations in the paper place the optimal collection NA at ~0.32 for this thermal-lens contrast, which is why a 0.55-NA long-working-distance air condenser can replace the oil-immersion condenser required for SRS.","core_discovery":"The paper's central claim is that the photothermal readout converts SRS from a technique shackled to low-noise solid-state lasers and high-NA immersion optics into one that works with a noisy, compact fiber laser and an air-gap condenser. In SRP, the pump and Stokes beams deposit heat through stimulated Raman absorption, and a third continuous-wave probe measures the resulting thermal-lens-induced redistribution of the probe's far field. The probe's noise floor is far below the fiber laser's, so the dominant SRS noise source is removed without balanced detection. On pure DMSO at the C-H vibration, the authors measure a 0.72% SRS modulation depth and a 3.3% SRP modulation depth under identical pump/Stokes conditions, and the SNR gain over fiber-laser SRS is ~105-fold (unbalanced) and ~12-fold (autobalanced). They further show that at NA 0.55 air collection, SRP preserves the Raman spectrum of PMMA beads while autobalanced SRS does not, and that urea-cleared rat brain slices can be imaged past 200 µm depth.","pith_inferences":["The urea 'thermal enhancement' claim would be strengthened by a direct water-versus-urea SNR measurement on the same sample; if urea's thermo-optic coefficient is near water's, the advantage may come mostly from clearing and refractive-index matching rather than heat capacity.","The noise-immunity argument is not specific to fiber lasers: any compact source with high intensity noise but short pulses, such as microchip or supercontinuum lasers, could adopt the same three-beam thermal-lens readout.","The optimal NA of ~0.32 suggests that tuning condenser NA for each solvent, rather than using the maximum 0.55, could improve contrast further; the paper does not explore that parameter.","Using a short-coherence-length probe laser would remove the interference fringes the authors observed at water-cell interfaces, potentially improving low-NA SRP on heterogeneous samples without oversampling."],"forward_implications":["Compact fiber-laser SRP microscopes can deliver chemical contrast without the bulk and cost of solid-state OPO systems.","The long 28 mm working distance and air condenser make contact-free, open-top sample formats practical, including multi-well plates and microfluidic chips, without re-aligning optics between samples.","Urea-based clearing plus thermal enhancement extends volumetric label-free imaging past 200 µm in brain tissue, with depth limited mainly by the objective working distance.","Detection limits around 11 mM (C-H) and 14 mM (C-D) at 20 µs/wavenumber point to practical metabolic imaging with deuterium labeling on a portable source."],"supporting_citations":[{"why":"introduced the SRP detection scheme and demonstrated its sensitivity","marker":"[30]"},{"why":"defined the fiber-laser intensity-noise problem and the balanced-detection baseline","marker":"[19]"},{"why":"supplied the autobalanced fiber-laser SRS configuration used for SNR comparison","marker":"[21]"},{"why":"provided the urea tissue-clearing protocol and clearing-enhanced vibrational imaging concept","marker":"[33]"},{"why":"is the dual-output fiber laser source whose noise and tuning performance the system relies on","marker":"[35]"},{"why":"established the high-NA collection requirement in SRS for suppressing XPM background","marker":"[25]"},{"why":"underpins the photothermal detection principle with single-molecule absorption measurements","marker":"[31]"}],"fun_headline_variants":["Fiber-laser Raman goes photothermal: 105x SNR gain","Photothermal readout lets fiber lasers do deep tissue Raman","Fiber-laser SRP: 105x better SNR without balance detection","Compact laser Raman imaging: thermal lens beats noise","Heat, not loss: fiber-laser Raman gains 100x"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that 8 M urea boosts the heat-lens signal beyond what its tissue-clearing effect alone would provide; this depends on urea's thermo-optic coefficient, which the paper does not measure, so the claimed thermal enhancement could vanish if that coefficient is smaller than water's.","fun_headline_variants_meta":{"raw":{"variants":["Fiber-laser Raman goes photothermal: 105x SNR gain","Photothermal readout lets fiber lasers do deep tissue Raman","Fiber-laser SRP: 105x better SNR without balance detection","Compact laser Raman imaging: thermal lens beats noise","Heat, not loss: fiber-laser Raman gains 100x"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000636,"raw_usage":{"total_tokens":2932,"prompt_tokens":947,"completion_tokens":1985,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":563,"completion_tokens_details":{"reasoning_tokens":1897}},"tokens_in":563,"tokens_out":1985,"duration_ms":14615,"temperature":1.0,"reasoning_tokens":1897,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:42:19.152517+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the thermo-optic coefficient of 8 M urea near room temperature and compare SRP SNR on a standard sample (e.g., DMSO or 3 µm PMMA beads) in water versus 8 M urea with identical pump/Stokes/probe powers, modulation, and collection NA; if the urea signal is not proportionally higher, the thermal-enhancement claim is refuted.","supporting_citations":[{"cited_title":"Zhu et al., Stimulated Raman photothermal microscopy toward ultrasensitive chemical imaging, Science Advances 9, eadi2181 (2023)","cited_arxiv_id":null,"evidence_quote":"introduced the SRP detection scheme and demonstrated its sensitivity"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defined the fiber-laser intensity-noise problem and the balanced-detection baseline"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplied the autobalanced fiber-laser SRS configuration used for SNR comparison"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provided the urea tissue-clearing protocol and clearing-enhanced vibrational imaging concept"},{"cited_title":"Brinkmann et al., Portable all -fiber dual-output widely tunable light source for coherent Raman imaging, Biomed","cited_arxiv_id":null,"evidence_quote":"is the dual-output fiber laser source whose noise and tuning performance the system relies on"},{"cited_title":"Bertoncini, S","cited_arxiv_id":null,"evidence_quote":"established the high-NA collection requirement in SRS for suppressing XPM background"},{"cited_title":"Gaiduk, M","cited_arxiv_id":null,"evidence_quote":"underpins the photothermal detection principle with single-molecule absorption measurements"}],"review_version":1}