{"id":"a058115b-c3dc-49b4-a564-41d28c838936","arxiv_id":"2411.08812","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A new analytic model combining exciton diffusion and Marcus charge-transfer rates shows that finite interfacial dissociation limits charge generation in low-offset organic solar cells.","lead":"This paper builds a mathematical model of how light-generated excitons move and split into charges in organic solar cell blends, combining diffusion with chemical reaction rates. It explains why low-energy-offset blends still generate charge slowly and offers a way to estimate domain sizes from transient absorption experiments.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 25 nm domain-size estimate rests on subtracting the ~20% ultrafast GSB offset; if that component is not a clean constant, the fitted R and k_diss,CT are biased.","rationale":"The paper's central theoretical contribution is a combined diffusion/rate-equation model, and the analytic steady-state efficiency expression appears correctly derived: I checked Eq. (20) against the spherical Robin boundary condition and the Marcus rate balance, and the algebra is consistent. The MFPT formula in Eq. (23) is also the standard R^2/(15D) result for uniform initial conditions in an absorbing sphere, and the supplementary mean-survival-time analysis is coherent. These parts support the qualitative conclusion that instantaneous interfacial quenching underestimates charge-generation times in low-offset systems. The experimental application, however, is where the strongest quantitative claim (domain diameter ≈ 25 nm) becomes fragile. The fit uses only two free parameters, R and k_diss,CT, while the ~20% ultrafast GSB component is removed by hand. The model cannot generate such a component with the adopted Marcus parameters, so the subtraction is not a minor data-cleaning step; it is essential to making the data compatible with the model. The paper acknowledges the strong correlation between κ_diss,LE and R, but the correlation with the subtracted offset is not quantified. A two-population refit, or a systematic variation of the offset amplitude, would directly test whether the recovered R is robust. The reader's verdict is already CONDITIONAL, and this concern reinforces, rather than overturns, that condition: the qualitative model conclusions stand, but the 25 nm domain-size estimate should be treated as contingent on the offset-subtraction procedure and the single-sphere morphology assumption.","tokens_in":40795,"tokens_out":7380,"duration_ms":65422,"concrete_test":"Refit the Figure 4 kinetics without subtracting the fast component: add a second disordered-Y6 population with free amplitude, dissociation rate, and decay constant, while keeping R and k_diss,CT as free parameters. If the recovered R leaves the 9–16 nm interval (or shifts by more than the quoted ±0.71 nm uncertainty), the domain-size estimate depends on the offset-subtraction assumption. A simpler robustness check is to repeat the fit with the subtracted offset fixed at 0%, 10%, and 30% of the maximum GSB; if R changes by more than ~20%, the 25 nm claim is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Sec. 5 the model is fitted to PM6:Y6 TAS data with only R and k_diss,CT free. The model predicts GSB = CT + CS starts at zero, but the experimental GSB shows a ~20% instantaneous component that the Marcus parameters (kdiss,LE ≈ 0.2 ps^-1) cannot reproduce. The authors subtract this offset as a constant and attribute it to disordered Y6 domains, then normalize the remaining signal. This subtraction is load-bearing: it directly sets the amplitude and shape of the rising GSB that constrains R and k_diss,CT. If the disordered-phase contribution is not constant on the fitted timescale, or if it overlaps with aggregated-domain kinetics, the recovered R = 12.47 ± 0.71 nm (diameter ≈ 25 nm) is biased. The single-sphere, uniform-generation assumption further conflates a distribution of domain sizes and shapes with one effective radius. The paper itself notes the strong correlation between κ_diss,LE and R, so any error in the subtracted component propagates directly into the headline domain size. The qualitative conclusion that finite interfacial quenching matters for low-offset systems is less affected, since it follows from the model structure and the MFPT analysis in Sec. 4, but the quantitative TAS-based domain-size claim is contingent on this data treatment.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper develops a deterministic continuum model that couples exciton diffusion in a donor or acceptor domain to interfacial rate equations describing CT state formation, reformation, geminate recombination, and dissociation to free charges. The authors derive an analytical steady-state charge generation efficiency (Eqs. 17-20), analyze characteristic timescales from a dynamic formulation, and fit the model to transient absorption (TAS) data of a PM6:Y6 blend with the Y6 domain radius R and the CT dissociation rate k_diss,CT as the only free parameters. The fit yields R = 12.47 ± 0.71 nm (diameter ≈ 25 nm) and k_diss,CT = 1.12 ± 0.44 ps^-1. The central qualitative conclusion is that the common assumption of instantaneous exciton quenching at the donor-acceptor interface is only valid for high driving forces, and that the intrinsic exciton lifetime is a key parameter enabling efficient generation at low offsets.","tokens_in":41112,"tokens_out":5727,"duration_ms":50639,"significance":"The model is a useful conceptual advance: it combines a tractable diffusion problem with Marcus-theory rates at the interface, and the steady-state expression (Eq. 20) is analytic and internally consistent. The qualitative finding that finite interfacial quenching, not pure diffusion, controls the charge-generation timescale in low-offset systems is well supported by the mean-survival-time analysis in Sec. 4 and by the illustrative simulations. If the quantitative TAS-based domain-size estimate can be made robust, the paper would also provide an experimentally applicable method for estimating domain sizes from charge-generation dynamics. The manuscript is likely to be of interest to the organic photovoltaics community.","major_comments":[{"comment":"The constant 20% offset subtraction is internally inconsistent with the authors' own interpretation of the late-time GSB decay. The model predicts a GSB that starts at zero, while the measured GSB exhibits an instantaneous component that the authors subtract as a time-independent offset and attribute to disordered Y6 domains. However, the same paragraph states that the observed decline in the GSB after ~200 ps is caused by geminate recombination of CT states in these disordered Y6 phases. If those CT states decay, their GSB contribution is time-dependent, so a constant offset subtraction cannot remove the instantaneous component while leaving the decaying component from the same phase. Since the fitted values of R and k_diss,CT depend directly on the amplitude and shape of the normalized rising signal, this data treatment is load-bearing for the quantitative claim. The authors should model the disordered-phase contribution with explicit time-dependent kinetics, or at minimum provide a sensitivity analysis showing that the fitted R and k_diss,CT are stable under alternative, physically motivated treatments of the offset.","section":"Sec. 5, Fig. 4b"},{"comment":"The exclusion of EA data before 0.3 ps is not justified. The EA signal is proportional to the CS population, and its early rise provides direct constraints on k_diss,CT; truncating the early data can bias both the fitted value and the reported uncertainty. The authors should state the reason for the cutoff (e.g., coherent artifact or pump scatter) and test the sensitivity of R and k_diss,CT to the choice of truncation threshold.","section":"Sec. 5, Fig. 4c"},{"comment":"The reported uncertainties on R and k_diss,CT are only the statistical uncertainties of the least-squares fit. The authors acknowledge in the text that kappa_diss,LE and R are strongly correlated when fitted simultaneously, and kappa_diss,LE itself depends on hand-set parameters (d = 1 nm), adopted parameters (|H_DA|, lambda), and the assumed Delta_E_LE-CT. The claim of a 25 nm domain diameter therefore rests on a chain of assumptions whose uncertainty is not propagated. A sensitivity analysis varying d over a reasonable range and using the literature spread of lambda and |H_DA| should be provided before this quantitative result can be considered established.","section":"Sec. 5, Eq. (28)"}],"minor_comments":[{"comment":"The main text states that the excitation wavelength is around 800 nm to selectively excite Y6, while the Materials and Methods section describes pump pulses at 720 nm. These statements should be reconciled.","section":"Sec. 5 vs. Materials and Methods"},{"comment":"The equation for dn_CT/dt contains a plus sign before the decay terms (kref + kf,CT + kdiss,CT)nCT, whereas the correct equation in the main text, Eq. (11), has a minus sign. This typo should be corrected.","section":"Supplementary Note SN2, Eq. (S11)"},{"comment":"The factor '(ratio + 1)' in the expression for n_LE,0 is not defined explicitly; the relationship between the PM6:Y6 mixing ratio and the volume fraction of Y6 should be clarified so that the initial density calculation is reproducible.","section":"Sec. 5, Eq. (27)"},{"comment":"The sentence 'The final or the second (suitable if k−1 f,LE≫⟨t⟩f) equation of table ST1 can then be used to estimate R' is difficult to parse; please rephrase to clearly identify which limiting case corresponds to which equation.","section":"Supplementary Note SN6"},{"comment":"The phrase 'The respective solid line shows' should be 'The solid lines show' for consistency with the plural entries in the caption.","section":"Fig. 3 caption"}],"recommendation":"major_revision","confidential_remarks":"The analytic model and its qualitative conclusions are sound and likely to be useful to the community. The main risk is the empirical domain-size estimate: the constant offset subtraction is not consistent with the authors' own explanation of the late GSB decay, and the reported uncertainties do not account for the acknowledged correlation between kappa_diss,LE and R or for the uncertainty in the adopted Marcus parameters. I recommend requiring a sensitivity analysis and a clear treatment of the disordered-phase contribution before publication. The pump-wavelength discrepancy between Section 5 and Materials and Methods should also be addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core contribution here is real: a clean analytic coupling of 3D exciton diffusion to Marcus-theory interfacial rates, producing a closed-form steady-state efficiency expression (eq. 20). The derivation is consistent, the limiting cases behave, and the comparison with the reduction-factor rate model shows why the common 0.1 prefactor is only a rough stand-in for finite quenching. The MFPT analysis in Sec. 4 also supports the qualitative conclusion that low-offset systems are not diffusion-limited in the naive sense. For an OPV audience, that is a useful framework, and the time-domain interpretation of TAS rise times is a step forward over purely phenomenological rate models.\n\nThe soft spots are concentrated in the experimental application. The PM6:Y6 fit subtracts about 20% of the GSB as a constant offset and excludes early EA data; if that disordered-phase component is not constant or overlaps with aggregated-domain kinetics, the recovered R = 12.47 nm and k_diss,CT = 1.12 ps^-1 are biased. The authors themselves note the strong correlation between R and kappa_diss,LE, so the headline 25 nm diameter is contingent on that parameter and on hand-set d, |H_DA|, and lambda from prior work. The single-sphere assumption is a simplification, though probably acceptable as an effective-radius estimate. I do not see a fatal flaw in the model itself, but the abstract's 'demonstrated' is too strong: the fit is a two-parameter model applied to data that already required the offset subtraction to look like the model's prediction.\n\nThis paper deserves a serious referee. The analytic part is solid and the qualitative physics is likely robust; the experimental section needs revision, ideally with code and data, a sensitivity analysis over kappa_diss,LE and d, and a more careful justification of the offset treatment. I would cite the analytic model if I worked on charge generation modeling, and I would bring it to a reading group focused on OPV kinetics. My recommendation: send it to review, but expect major revision on the fitting and uncertainty quantification.","headline":"Worth engaging for the analytic diffusion-plus-Marcus model; the PM6:Y6 domain-size estimate rests on a load-bearing 20% offset subtraction that needs more scrutiny.","tokens_in":41646,"tokens_out":1059,"would_cite":true,"duration_ms":104914,"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 builds a model that couples exciton diffusion to Marcus-theory charge transfer at the interface and shows that, in low-driving-force organic solar cells, assuming instant quenching at the donor-acceptor boundary makes…","keywords":["organic solar cells","exciton diffusion","Marcus theory","charge transfer states","transient absorption spectroscopy","domain size estimation","non-fullerene acceptors","bulk heterojunction"],"falsifier":"Measure the Y6 domain size in the same PM6:Y6 blend with an exciton-dynamics-independent structural probe, such as grazing-incidence small-angle X-ray scattering or resonant soft X-ray scattering; if the scattering-derived domain diameter disagrees with the fitted $25$ nm beyond the stated uncertainty, the TAS-based estimate and its underlying single-spherical-domain, fixed-Marcus-parameter assumptions are falsified.","tokens_in":40550,"feed_emoji":"☀️","tokens_out":8703,"duration_ms":73271,"temperature":0.7,"pith_summary":"This paper tries to establish a quantitative account of free charge generation in organic solar cells that treats exciton diffusion and interfacial charge transfer as one coupled process rather than two separate steps. The authors derive an analytical steady-state expression for the charge generation efficiency from a diffusion equation joined to Marcus-theory rate equations at the donor-acceptor interface, and show that the intrinsic exciton lifetime, not just the diffusion length, controls whether a vanishing driving force can still yield high efficiency. Their dynamic analysis leads to the paper's central methodological claim: when the energy offset driving exciton dissociation is small, the measured time for charge generation is much longer than the pure diffusion time, so conventional domain-size estimates that assume instantaneous quenching at the interface are only valid for high driving forces above about $0.2$ eV. Applied to transient absorption data on PM6:Y6, the model reproduces the observed slow hole-transfer rise and yields a fitted Y6 domain radius of $12.47\\pm 0.71$ nm, i.e. roughly a $25$ nm domain diameter, together with a fast CT-state dissociation rate of $1.12\\pm 0.44$ ps$^{-1}$.","feed_headline":"Diffusion-only quenching fails for low-offset solar cells","feed_subtitle":"Combined diffusion-Marcus model fits PM6:Y6 data, placing Y6 domains near 25 nm and fast CT splitting.","key_machinery":"The central object is a reaction-diffusion system on an acceptor domain $\\Omega\\subset\\mathbb{R}^3$: a diffusion equation $\\partial_t n_{\\mathrm{LE}}=D\\Delta n_{\\mathrm{LE}}+G_{\\mathrm{LE}}-R_{\\mathrm{LE}}$ for the singlet exciton density, coupled to a surface density of charge-transfer states $n_{\\mathrm{CT}}$ through a Robin-type boundary condition $-D\\nabla n_{\\mathrm{LE}}\\cdot\\nu=\\kappa_{\\mathrm{diss,LE}}n_{\\mathrm{LE}}-k_{\\mathrm{ref}}n_{\\mathrm{CT}}$ on the donor-acceptor interface, with $n_{\\mathrm{CT}}$ evolving by the rate equation $\\partial_t n_{\\mathrm{CT}}=\\kappa_{\\mathrm{diss,LE}}n_{\\mathrm{LE}}-(k_{\\mathrm{ref}}+k_{\\mathrm{f,CT}}+k_{\\mathrm{diss,CT}})n_{\\mathrm{CT}}$. The carrying identity is the closed-form steady-state efficiency, which depends only on the ratios $L_D/R$ and $\\kappa_{\\mathrm{diss,LE}}/(R\\,k_{\\mathrm{f,LE}})$ together with the reformation probability $P_{\\mathrm{ref}}$ and the CT dissociation probability $P_{\\mathrm{diss,CT}}$; this compact reduction is what lets the authors isolate the intrinsic lifetime as the parameter that controls low-offset performance. A second analytic object, the mean survival time of an exciton in a sphere, provides the diffusion time scale used to show that finite interfacial quenching, not pure diffusion, sets the measured rise times.","core_discovery":"The central discovery is that the efficiency and dynamics of free charge generation in phase-separated donor-acceptor blends are governed by the interplay between exciton diffusion to the interface and the finite, Marcus-limited rates of exciton dissociation, CT reformation, and CT dissociation. In the steady state, the charge generation efficiency reduces to an explicit function of only four parameter combinations: the diffusion length over the domain radius, a velocity ratio $\\kappa_{\\mathrm{diss,LE}}/(R\\,k_{\\mathrm{f,LE}})$, the reformation probability, and the CT dissociation probability. The model shows that a long intrinsic exciton lifetime raises efficiency at low offset by permitting multiple dissociation attempts, whereas diffusion length alone cannot. Dynamically, for low offsets the ground-state-bleach rise time is considerably longer than the mean diffusion time because weak quenching prevents a concentration gradient from forming; hence interpreting long experimental rise times of roughly $100$ ps as pure diffusion overestimates the diffusion-limited time and invalidates instant-quench domain sizing unless $\\Delta E_{\\mathrm{LE-CT}}>0.2$ eV. Fitting the model to PM6:Y6 transient absorption data gives $R=12.47\\pm 0.71$ nm and $k_{\\mathrm{diss,CT}}=1.12\\pm 0.44$ ps$^{-1}$.","pith_inferences":["A direct test would be to extract the domain radius from time-resolved photoluminescence survival times using the mean-survival-time expression in the supplement and compare it with the TAS-derived $25$ nm on the same film; agreement would validate the model, while disagreement would reveal which observable is more sensitive to the disordered-phase assumption.","The strong correlation between $\\kappa_{\\mathrm{diss,LE}}$ and $R$ implies that independent morphology measurements or independently determined Marcus parameters are needed to make the domain-size estimate unique, since multiple (rate, radius) pairs can fit the same transient.","If the roughly $20\\%$ ultrafast component is indeed from disordered Y6, then changing the excitation wavelength or probing a sub-ensemble should shift the fitted aggregated-domain radius only within the stated uncertainty; a large shift would indicate that the constant-offset subtraction is not clean."],"forward_implications":["Domain-size estimates that assume instant quenching at the donor-acceptor interface are only reliable when the driving force exceeds about $0.2$ eV; below that, fitted sizes will be distorted because slow dissociation itself extends the measured rise time.","Long intrinsic exciton lifetimes, not just long diffusion lengths, can push charge generation efficiency close to unity at zero or slightly negative offsets, making lifetime a design lever for low-voltage-loss solar cells.","Standard rate-equation models that fold diffusion into a fixed reduction factor reproduce efficiency trends but cannot match the explicit domain-size dependence of the present model.","For PM6:Y6, the model attributes the roughly $100$ ps hole-transfer completion to combined diffusion and dissociation rather than diffusion alone, and places the Y6 domain diameter near $25$ nm with rapid CT dissociation on the order of $1$ ps."],"supporting_citations":[{"why":"Supplies the Marcus-theory rate equations used to set $k_{\\mathrm{diss,LE}}$ and $k_{\\mathrm{ref}}$ as functions of the energy offset.","marker":"[41]"},{"why":"Provides the explicit-exciton-diffusion treatment and lifetime emphasis that this model generalizes to three-dimensional domains with Marcus kinetics.","marker":"[17]"},{"why":"Supplies the reference rate model for comparison and the $|H_{\\mathrm{DA}}|=0.01$ eV electronic coupling adopted in the fits.","marker":"[19]"},{"why":"Gives experimental transient-absorption evidence that hole-transfer rates drop sharply when the offset approaches zero.","marker":"[15]"},{"why":"Demonstrates experimentally that long exciton lifetimes enable charge generation at negligible offsets, supporting the paper's lifetime-centered conclusion.","marker":"[16]"},{"why":"Derives the semi-absorbing Robin boundary condition that couples exciton diffusion to interfacial dissociation in this model.","marker":"[29]"},{"why":"Provides the micro-to-macro derivation in which the exciton dissociation constant is proportional to the CT localization thickness $d$, motivating the $1$ nm value.","marker":"[20]"},{"why":"Supplies the measured Y6 diffusion coefficient $D=0.017$ cm$^2$/s used in the PM6:Y6 fits.","marker":"[53]"},{"why":"Supplies the singlet-singlet annihilation coefficient used to model transient absorption at experimental fluences.","marker":"[54]"},{"why":"Reports X-ray-scattering domain sizes for PM6:Y6 against which the fitted $25$ nm diameter is benchmarked.","marker":"[57]"}],"fun_headline_variants":["Low-offset solar cells: exciton lifetime beats diffusion limits","Diffusion-only quenching fails for low-offset organic cells","Exciton lifetime rescues low-offset solar cells","PM6:Y6 charges form via diffusion and Marcus kinetics","Low-offset organic cells: diffusion alone can't explain charge yield"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fitted $25$ nm Y6 domain size rests on the assumption that the measured PM6:Y6 transient absorption can be cleanly separated into a constant roughly $20\\%$ ultrafast component from disordered Y6 and a remainder governed by one spherical aggregated Y6 domain with a single set of Marcus parameters, including the hand-set $1$ nm CT localization thickness and literature values for $|H_{\\mathrm{DA}}|$ and $\\lambda$.","fun_headline_variants_meta":{"raw":{"variants":["Low-offset solar cells: exciton lifetime beats diffusion limits","Diffusion-only quenching fails for low-offset organic cells","Exciton lifetime rescues low-offset solar cells","PM6:Y6 charges form via diffusion and Marcus kinetics","Low-offset organic cells: diffusion alone can't explain charge yield"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00157,"raw_usage":{"total_tokens":6366,"prompt_tokens":1144,"completion_tokens":5222,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":760,"completion_tokens_details":{"reasoning_tokens":5138}},"tokens_in":760,"tokens_out":5222,"duration_ms":34396,"temperature":1.0,"reasoning_tokens":5138,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T21:18:49.602725+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the Y6 domain size in the same PM6:Y6 blend with an exciton-dynamics-independent structural probe, such as grazing-incidence small-angle X-ray scattering or resonant soft X-ray scattering; if the scattering-derived domain diameter disagrees with the fitted $25$ nm beyond the stated uncertainty, the TAS-based estimate and its underlying single-spherical-domain, fixed-Marcus-parameter assumptions are falsified.","supporting_citations":[],"review_version":1}