Solar Cell Simulation Data

An example multilayer-device input and spreadsheet simulation exports for a Spiro / MAPICl / TiO2 stack. The repository stores tabular artifacts and explanatory slides. It does not contain an LLM, a training pipeline, or the simulation code that generated the workbooks.

Contents

All data are under SpiroMAPIClTiO2/.

File	Contents
layer_unit_test.csv	One device configuration: 7 records and 34 fields
Current flux density.xlsx	16 current/flux worksheets, each 50 x 480
Electrostatic potential & Electric field.xlsx	Fion, Vion, rho, each 50 x 481
Potential & Energy.xlsx	Ecb, Evb: 50 x 481; EFn, EFp: 10 x 481
Recombination.xlsx	btb, srh, vsr, tot, each 50 x 481
SemiSimu-output.pptx	Five slides describing output quantities and plotting terminology

These dimensions count data, not a header row. The first row and first column contain values. Some numerical entries are stored as Excel text and others as numeric cells.

Device input

The CSV describes two electrodes, two transport layers, two interfaces, and one active layer:

electrode -> Spiro -> interface -> MAPICl -> interface -> TiO2 -> electrode

It is one configuration, not seven independent device samples. Non-electrode layer_points entries are 100, 40, 200, 40, and 100. Their sum is 480, but the actual spatial coordinate vector is not included in the exports.

Field group	CSV columns
Geometry and mesh	layer_type, material, thickness, layer_points, xmesh_coeff
Energy parameters	Phi_EA, Phi_IP, EF0, Et
Densities and ionic ceilings	Nc, Nv, Ncat, Nani, c_max, a_max
Mobilities	mu_n, mu_p, mu_c, mu_a
Dielectric, generation, and recombination parameters	epp, g0, B, taun, taup, sn, sp, vsr_zone_loc
Plot colors	Red, Green, Blue
Global/device options	optical_model, xmesh_type, side, N_ionic_species

The first electrode row specifies Beer-Lambert, erf-linear, left, and one ionic species. Many fields are blank where values are not supplied; do not indiscriminately replace those blanks with zero.

Raw thickness values are 2.00E-05, 2.00E-07, 3.40E-05, 2.00E-07, and 1.00E-05. The CSV has no explicit units row or schema. Do not interpret these values directly as nanometers or assume a conversion without the generating model's conventions.

Output quantities

The slides reference dfana.calcr(sol, mesh_option) and dfplot operations. Those functions are not included here. The slides provide context, not an executable interface or a complete machine-readable schema.

Workbook	Worksheet names
Current/flux	Jn, Jp, Jc, Ja, Jdisp, Jtot, Jndrift, Jndiffu, Jpdrift, Jpdiffu, Jadrift, Jadiffu, Jcdrift, Jcdiffu, Jddisplay, Jddtot
Electrostatics	Fion, Vion, rho
Energy levels	Ecb, Evb, EFn, EFp
Recombination	btb, srh, vsr, tot

The presentation associates current terms with electrons, holes, cations, anions, and displacement current. btb, srh, and vsr denote band-to-band, Shockley-Read-Hall, and interfacial surface recombination. EFn and EFp denote electron and hole quasi-Fermi levels.

Preserve the literal worksheet name Jddisplay when reading the file. The slides use Jddisp, so assuming the slide label is the workbook key will fail.

The slides describe spatial profiles across rows and temporal series down columns. No explicit time, bias, or spatial-coordinate arrays are shipped with the workbooks. The 480/481-column and 10/50-row distinctions must be resolved before aligning, combining, or interpolating quantities.

Read the data in Python

git clone https://github.com/ShaneLogic/solar-cell-LLM-database.git
cd solar-cell-LLM-database
python -m pip install pandas openpyxl

from pathlib import Path
import pandas as pd

root = Path("SpiroMAPIClTiO2")
layers = pd.read_csv(root / "layer_unit_test.csv")
print(layers[["layer_type", "material", "thickness", "layer_points"]])

book = pd.ExcelFile(root / "Current flux density.xlsx", engine="openpyxl")
print(book.sheet_names)

# No header row: convert numeric strings and reject unexpected text.
jtot = pd.read_excel(book, sheet_name="Jtot", header=None)
jtot = jtot.apply(pd.to_numeric, errors="raise")
assert jtot.shape == (50, 480)

efn = pd.read_excel(root / "Potential & Energy.xlsx", sheet_name="EFn", header=None)
efn = efn.apply(pd.to_numeric, errors="raise")
assert efn.shape == (10, 481)

The default header=0 would silently consume the first data row. Explicit conversion handles the mixed Excel cell types without hiding unexpected strings as missing data.

Interpretation and reuse

• These are simulation artifacts for one example configuration, not documented experimental measurements or a curated multi-device training benchmark.
• A complete generating-solver revision, run protocol, coordinate vectors, unit schema, and parameter-to-output provenance manifest are not supplied.
• Do not assume that the two total-current representations or matrices with different shapes are interchangeable.
• For ML or LLM-assisted analysis, add verified units, axes, sample/state definitions, provenance, and split rules before assigning scientific targets or making performance claims.
• A dataset license is not declared by a standalone license file in this snapshot.