Omnisoot: Soot Process Design Package

• Omnisoot is an object-oriented, reduced-order process design tool that couples Cantera-based gas kinetics with soot and carbon black particle dynamics.
• It integrates multiple reactor models and inception mechanisms to ensure full mass and energy closure in gas–particle systems for accurate process design.
• The framework enables parametric studies of reactor conditions, highlighting how different inception pathways and surface kinetics impact soot morphology and yield.

Searching arXiv for the Omnisoot paper and related metadata. Omnisoot is an object-oriented, reduced-order process design package for gas-phase synthesis of carbonaceous nanoparticles that couples detailed gas-phase chemical kinetics, via Cantera, with soot and Carbon Black particle dynamics to predict mass loading, particle size distribution, morphology, and composition under pyrolysis and combustion conditions (Adib et al., 15 Jul 2025). It integrates constant volume, constant pressure, perfectly stirred, and plug flow reactor models with four inception models from the literature and two population balance frameworks, and it is designed to close the mass and energy balances of the gas–particle system while consistently coupling soot inception, surface growth, oxidation, and coagulation with reactor transport models. In contrast to monolithic soot–gas hybrid mechanisms such as BIN approaches embedded in CRECK, Omnisoot emphasizes portable coupling to any Cantera-compatible mechanism and process-design-oriented parametric studies over temperature, pressure, residence time, and feed composition.

1. Integrated scope and process-design role

Omnisoot combines four reactor models—CVR, CPR, PSR, and PFR—with four inception models and two population balance models: a monodisperse model and a sectional model (Adib et al., 15 Jul 2025). Its central purpose is to provide an integrated process design tool that predicts soot mass, morphology, and composition under varying process conditions, while accounting for soot inception, surface growth, and oxidation together with detailed gas-phase chemistry.

The package is framed around a modeling gap: mass and energy balances of the gas–particle system must be closed while soot and gas-phase chemistry remain linked to particle dynamics models that consider the evolving fractal-like structure of soot agglomerates. Omnisoot addresses that gap by making soot formation, gas-phase chemistry, and reactor transport part of a single coupled formulation rather than separate post-processing stages.

This design also makes Omnisoot a Carbon Black process-analysis tool. The reported workflow enables parametric studies over reactor temperature profiles, residence time, pressure, and feed composition, with feedback from soot dynamics to both gas-phase chemistry and energy. A plausible implication is that the package is intended not only for mechanism comparison but also for reactor and operating-window selection when morphology targets matter as much as yield.

2. Reactor formulations and closure of the gas–particle balances

All reactors treat the control volume as the coupled gas–particle mixture, with species scrubbing terms from soot processes denoted by $\dot{s}_k$, and soot variables $\psi$ tracked by dedicated transport equations (Adib et al., 15 Jul 2025). Omnisoot uses soot volume fraction based on gas and reactor volumes, denoted $f_v$ and $\varphi$, to adjust the mass and energy fluxes of the coupled system.

The implemented reactors are a Constant Volume Reactor, Constant Pressure Reactor, Perfectly Stirred Reactor, and Plug Flow Reactor. In the PSR, the nominal residence time is written as

$\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$

Across these reactors, the governing equations incorporate soot source terms directly into the gas-phase balances, rather than solving soot dynamics independently of the reactor state.

A defining feature is full mass and energy closure. Omnisoot augments Cantera’s gas species production rates $\dot{\omega}_k$ with soot scrubbing terms $\dot{s}_k$ from inception, adsorption, HACA growth, and oxidation. Soot sensible and formation energies appear explicitly in the reactor energy equations for CVR, CPR, PSR, and PFR, and gas mass decreases as carbon and hydrogen are transferred to soot. The paper reports that neglecting soot sensible energy significantly biases temperature and soot growth, with overprediction by approximately $150$ K and up to $3\times$ mobility diameter in a $30\%\ \mathrm{CH_4}$ pyrolysis CVR test, which motivates the package’s insistence on full energy closure.

This closure strategy is significant because soot kinetics are highly sensitive to temperature and precursor availability. In Omnisoot, the reactor solution, gas chemistry, and particle dynamics are mutually constrained, so errors in one submodel propagate consistently rather than being hidden by decoupled approximations.

3. Inception chemistry, surface kinetics, and oxidation

Omnisoot implements four soot formation submodels: Irreversible Dimerization (Frenklach 1991), Reactive Dimerization (Kholghy et al. 2018), Dimer Coalescence (Blanquart et al. 2009), and E-Bridge Modified (Frenklach 2020) (Adib et al., 15 Jul 2025). These models differ primarily in reversibility and temperature dependence.

Irreversible Dimerization treats PAH dimerization and adsorption as irreversible pathways. Reactive Dimerization separates reversible physical clustering from irreversible carbonization. Dimer Coalescence uses an irreversible multi-step sequence in which dimers form and then coalesce, with adsorption of dimers onto soot. E-Bridge Modified introduces reversible monomer dehydrogenation and irreversible radical recombination to dimers, with adsorption of monomer radicals. The package also allows configurable precursor lists; the default list includes A2, A3, A4, A2R5, A3R5, and A4R5.

Surface growth and oxidation are handled through a HACA mechanism under a “chemical similarity” framework, using PAH-like edge site chemistry on the soot surface. The implemented elementary steps include H-abstraction, acetylene addition, and oxidation by $\psi$0 and $\psi$1. Surface growth depends on active surface sites and local $\psi$2 concentration, while oxidation contributes source terms for soot carbon loss and gas-phase species production or consumption.

The coexistence of reversible and irreversible inception pathways is one of the package’s central comparative features. The reported analyses show that reversibility suppresses effective particle flux at lower temperature, whereas irreversibility can sustain inception in cooler regions. This difference becomes decisive when reactor thermal histories create hot formation zones followed by cooler downstream sections.

4. Population balances, morphology, and particle-dynamics representation

Omnisoot provides two population balance frameworks: the Monodisperse Population Balance Model (MPBM) and the Sectional Population Balance Model (SPBM) (Adib et al., 15 Jul 2025). MPBM is intended to provide fast, accurate mean-property evolution under conditions that rapidly reach SPSD, whereas SPBM tracks PSD and morphology in sections and is used when PSD shape matters.

The SPBM is a fixed-pivot sectional model with explicit sectional source terms for agglomerate number, primary-particle number, and hydrogen content, including coagulation birth and death with mass-conserving redistribution. For context, the paper states the general continuous population balance equation as

$\psi$3

while emphasizing that Omnisoot uses the discrete SPBM with explicit source terms.

Morphology is computed via DEM-derived power laws without explicitly tracking a fractal dimension parameter. The key relations include

$\psi$4

and

$\psi$5

The total surface area per unit gas mass is

$\psi$6

These descriptors feed directly into coagulation and surface kinetics, since HACA rates depend on $\psi$7, while collision diameters and diffusion coefficients depend on $\psi$8, $\psi$9, and $f_v$0.

The package adopts a constant soot density of $f_v$1, approximates soot thermophysical properties by graphite, assumes thermal equilibrium between soot and gas, and imposes an incipient primary particle minimum diameter of $f_v$2 nm. The corresponding minimum carbon count is reported as $f_v$3. These assumptions make the framework computationally tractable while preserving physically interpretable morphology variables for process design.

5. Gas-phase chemistry integration and reactor-scale findings

Omnisoot uses Cantera to evaluate thermochemistry and reaction rates for any mechanism supplied by the user, and the paper demonstrates ABF, Caltech, KAUST, CRECK, ITV, and FFCM2 in methane pyrolysis comparisons (Adib et al., 15 Jul 2025). The gas-phase ODEs are solved alongside soot variables in the selected reactor model, with source and sink coupling through $f_v$4 for PAH consumption, radical or $f_v$5 or CO release, and heat terms through soot sensible and formation energy.

Three case studies define the paper’s main quantitative findings. In methane pyrolysis in a shock tube, using $f_v$6–Ar in a CPR with $f_v$7–$f_v$8 K, $f_v$9–$\varphi$0 bar, residence time approximately $\varphi$1 ms, and the Caltech mechanism, multiple combinations of inception and adsorption adjustment factors $\varphi$2 and $\varphi$3 minimized carbon yield error but produced markedly different morphologies. Carbon yield showed a bell-shaped temperature dependence peaking near approximately $\varphi$4, and primary particle diameter $\varphi$5 increased with $\varphi$6, reaching approximately $\varphi$7 nm near $\varphi$8 K before returning to the $\varphi$9 nm minimum at higher temperature. The E-Bridge Modified model shifted the carbon-yield peak to higher $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$0, reflecting its stronger Arrhenius dependence.

In ethylene pyrolysis in a flow reactor, with $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$1 in $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$2, reactor length $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$3 m, diameter $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$4 mm, $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$5 K, and flow rates $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$6 L/min, only irreversible inception models—Irreversible Dimerization and Dimer Coalescence—predicted bimodal PSD at $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$7 L/min. Reactive Dimerization and E-Bridge Modified predicted nearly unimodal PSDs because inception became negligible outside the hot zone. The paper further reports that in cooler downstream zones the irreversible models maintained high inception flux, while reversible models decayed by more than $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$8 orders of magnitude.

In ethylene combustion in a PSR followed by a PFR, with PSR volume $\tau_{psr} = \frac{\rho V}{\dot{m}_{in}}.$9 ml, nominal residence time approximately $\dot{\omega}_k$0 ms, downstream PFR length $\dot{\omega}_k$1 m, diameter $\dot{\omega}_k$2 cm, average axial velocity approximately $\dot{\omega}_k$3 m/s, and equivalence ratios $\dot{\omega}_k$4, all inception models captured the measured unimodal PSDs downstream. At $\dot{\omega}_k$5, the measured PSD statistics were $\dot{\omega}_k$6 nm and $\dot{\omega}_k$7, while the model outputs ranged from $\dot{\omega}_k$8–$\dot{\omega}_k$9 nm and $\dot{s}_k$0–$\dot{s}_k$1, depending on inception model. Under these conditions, more than $\dot{s}_k$2 of soot mass was acquired via HACA at the PFR outlet.

Together, these cases support the paper’s main interpretive claim: fits to yield alone are insufficient, because different inception and adsorption pathways can match yield while implying different PSD evolution, primary-particle diameters, and aggregate morphologies.

6. Software architecture, diagnostics, and limitations

The package is organized into modules or classes for the Cantera chemistry interface; reactor models; particle dynamics with MPBM and SPBM; PAH growth models; and surface reactions including HACA growth and oxidation by $\dot{s}_k$3 and $\dot{s}_k$4 (Adib et al., 15 Jul 2025). Users select the gas mechanism, reactor type, inception model, population balance model, and flow, temperature, and pressure histories, and can adjust $\dot{s}_k$5, $\dot{s}_k$6, and precursor lists.

Its outputs include time or axial histories of gas species, temperature, soot mass loading $\dot{s}_k$7, PSD for SPBM, morphology descriptors $\dot{s}_k$8, composition $\dot{s}_k$9, and diagnostics such as elemental and energy balance residuals. The package also produces carbon pathway maps separating inception, HACA, adsorption, and oxidation contributions to soot carbon. The reported validation shows elemental and energy conservation residuals typically below $150$0, and collision-kernel checks indicate that Fuchs interpolation matches DEM across Knudsen regimes while the harmonic mean remains within approximately $150$1 errors in the transition regime.

The practical guidance given in the paper distinguishes MPBM from SPBM and irreversible from reversible inception models. MPBM is recommended for rapid parametric sweeps or when SPSD is reached quickly, whereas SPBM is recommended when PSD shape or bimodality matters, or when residence times are short. For flow reactors with cool downstream regions and desired bimodality, irreversible models are preferred; for hot, short-residence scenarios where inception ceases early, the models tend to agree more closely on PSD shape, and attention shifts to surface area and HACA in matching soot volume fraction.

The paper also identifies several limitations. Fixed-pivot SPBM can lose mass if new particles exceed the last section; soot density is fixed at $150$2; the incipient primary diameter floor is $150$3 nm; necking and sintering are not modeled; oxidation by $150$4, $150$5, and $150$6 is omitted; and surface reactivity $150$7 is represented through empirical relations. In addition, gas chemistry uncertainty remains a dominant source of uncertainty in precursor fluxes, and the paper emphasizes that measured morphology, primary particle size, aggregate structure, and PSD are essential for constraining model parameters in addition to yield.

A plausible implication is that Omnisoot is best understood not as a single soot mechanism, but as a comparative and design framework: it makes mechanism choice, reactor choice, and morphology diagnostics part of one mass- and energy-consistent computational environment.