An educational Python simulation of a weakly ionized gas whose conductivity and optical emission are converted into threshold-based digital logic.

The project combines simplified plasma kinetics, signal processing, adaptive ODE integration, stochastic uncertainty analysis, and reaction-diffusion modeling to explore a focused question:

Can the physical response of a plasma-like system be digitized into stable AND, OR, XOR, NOT, NAND, NOR, half-adder, full-adder, and stateful logic behavior?

This is a conceptual scientific-computing project. It is not a nuclear transmutation model, laboratory control system, high-voltage design guide, or validated plasma chemistry package.

The logic layer is a calibrated response classifier, not a schematic of series- or parallel-connected discharge tubes.

• A, B, and CARRY_IN change the aggregate physical drive level.
• The plasma ODEs produce conductivity and optical-emission responses.
• Calibrated thresholds and response windows digitize those analog levels.
• AND and OR therefore describe the resulting truth table, not Kirchhoff topologies made from two independently switched tubes.

There is no if V_A >= V_breakdown tube switch in the implementation. A series/parallel gas-tube circuit would require a separate nonlinear circuit layer that solves node voltages and branch currents together with each tube's internal plasma state. Claims about equal voltage division, common ballast resistors, or per-tube maintenance voltage do not apply to the current model.

The current model does include temporal memory: electron, ion, excitation, and metastable states decay over finite rate constants after a pulse. Ionization activation is logistic rather than an instantaneous current step. The Schmitt-trigger example is an additional sensor-level hysteresis stage; it is not presented as the physical discharge's maintenance-voltage curve.

• Zero-dimensional mean-field plasma kinetics
• Ionization, recombination, excitation, and de-excitation
• Electron, positive-ion, negative-ion, and metastable populations
• Dynamic mean electron energy as a two-temperature proxy
• Conductivity and optical-emission sensor models
• Noisy threshold detection and stability metrics
• AND, OR, NOT, NAND, NOR, XOR, half-adder, and full-adder emulation
• Schmitt-trigger hysteresis and conceptual SR-latch memory
• Adaptive BDF, Radau, RK45, and LSODA integration
• Monte Carlo uncertainty propagation with confidence intervals
• CSV-ready coefficient calibration using nonlinear least squares
• One-dimensional reaction-diffusion simulation
• Reproducible seeds, CSV exports, plots, and automated tests

The extended model tracks electron attachment, negative-ion accumulation, metastable populations, and time-varying mean electron energy.

The neutral number density is estimated from the ideal-gas relation:

N = p / (k_B T)

The electric-field input and the voltage-derived field are blended before the reduced field is evaluated:

E_0 = (1 - b) E_input + b (V / d)
(E/N)_Td = E_0 / (N x 10^-21)

Ionization activation is represented by a smooth threshold:

A(E/N) = 1 / (1 + exp(-((E/N) - E_threshold) / width))

The basic normalized charged-particle balance is:

dx_e/dt = R_ion - R_rec - k_e,loss x_e
dx_i/dt = R_ion - R_rec - k_i,loss x_i

The advanced model adds:

electron attachment and detachment
positive-ion / negative-ion neutralization
metastable-assisted stepwise ionization
mean electron-energy relaxation
species-dependent diffusion

Conductivity and optical emission are converted to saturating sensor channels, filtered by a finite detector time constant, contaminated with reproducible Gaussian noise, and digitized using either a single threshold or hysteresis.

The complete assumptions, equations, coefficients, and interpretations are in the Turkish technical report.

The simulation does not draw conventional electronic gates. Instead, zero, one, two, or three active physical inputs create calibrated analog sensor levels. Thresholds or response windows convert those levels into logic states.

The full adder classifies four calibrated physical-response levels and returns the parity bit as SUM and the two-or-more classification as CARRY_OUT.

Representative results:

Requirements:

• Python 3.10 or newer
• NumPy
• Pandas
• Matplotlib
• SciPy

git clone https://github.com/AybarsBarut/Ionization-Based-Logic-System-Simulation.git
cd Ionization-Based-Logic-System-Simulation
python -m venv .venv

Activate the virtual environment:

# Windows PowerShell
.\.venv\Scripts\Activate.ps1

# Linux or macOS
source .venv/bin/activate

Install the project:

python -m pip install --upgrade pip
python -m pip install -e .

Run every scenario, logic table, parameter sweep, uncertainty analysis, calibration example, and spatial simulation:

python run_experiments.py --output-dir outputs

After editable installation, the equivalent command is:

plasma-logic-sim --output-dir outputs

Run the automated tests:

python -m unittest -v

from plasma_model import PlasmaConfig, simulate_plasma

config = PlasmaConfig(
    gas="argon",
    pressure_pa=4_000.0,
    temperature_k=300.0,
    electric_field_v_m=175_000.0,
    applied_voltage_v=875.0,
    pulse_frequency_hz=2_000.0,
    pulse_width_s=2.0e-4,
    measurement_noise_std=0.025,
    random_seed=2026,
)

result = simulate_plasma(config)

print(result.summary)
print(result.data.head())

Advanced adaptive integration:

from advanced_models import AdvancedModelOptions, simulate_advanced_plasma

result = simulate_advanced_plasma(
    config,
    options=AdvancedModelOptions(solver_method="BDF"),
)

print(result.data[
    [
        "electron_density_m3",
        "negative_ion_density_m3",
        "metastable_fraction",
        "electron_energy_ev",
    ]
].tail())

The experiment runner writes reproducible CSV, PNG, and text artifacts, including:

• Time-series plasma states
• Scenario summaries
• Logic truth tables
• Full-adder calibration levels
• Threshold and field-pressure sweeps
• Parameter sensitivities
• Adaptive-solver comparisons
• Monte Carlo samples and confidence summaries
• Coefficient-calibration diagnostics
• Schmitt-trigger and latch sequences
• One-dimensional spatial snapshots

Generated artifacts are intentionally excluded from Git because they can be recreated with one command. Selected figures are versioned in docs/images.

.
|-- plasma_model.py          # Base kinetics, metrics, and logic emulation
|-- advanced_models.py       # Adaptive ODE, uncertainty, fitting, and 1D model
|-- run_experiments.py       # Reproducible experiment and plotting pipeline
|-- test_plasma_model.py     # Automated regression and physics sanity tests
|-- MODEL_RAPORU.md          # Detailed Turkish equations and interpretation
|-- docs/images/             # Selected versioned result figures
|-- requirements.txt
`-- pyproject.toml

• The default random seed is fixed.
• Input parameters are stored in dataclasses.
• All generated tables use explicit column names and SI units.
• Solver comparisons are exported to CSV.
• Sixteen automated tests cover state bounds, logic behavior, adaptive solvers, calibration, uncertainty propagation, and spatial diffusion.

• Coefficients are effective educational parameters, not validated reaction cross sections.
• The model is not suitable for laboratory safety decisions or device control.
• It does not solve Kirchhoff equations for a network of discrete gas tubes, ballast resistors, or shared electrical nodes.
• It does not implement a Paschen-law breakdown-voltage curve or an explicit per-tube breakdown/maintenance-voltage pair.
• The one-dimensional model does not solve Poisson's equation or electrode sheath dynamics.
• Air chemistry is represented by aggregate effective species.
• Optical emission is relative rather than radiometrically calibrated.
• The included coefficient-fitting example uses synthetic observations.

• Detailed model report in Turkish
• Contributing guide
• Security and safe-use policy
• Citation metadata

If this educational simulator supports a report, course, or prototype, cite it using the metadata in CITATION.cff.

Issues and pull requests are welcome. Please keep proposed physics extensions clearly labeled as conceptual, empirical, or experimentally calibrated. See CONTRIBUTING.md before submitting changes.

Released under the MIT License.

Plasma simulation, ionization modeling, threshold logic, plasma logic gates, scientific Python, reaction-diffusion, adaptive ODE solvers, signal processing, Monte Carlo uncertainty, conductivity modeling, optical emission, and educational computational physics.