PlasmaCCP electric-field magnitude |E| field map — the axisymmetric solution unfolded to a full-diameter section (r −230..230 mm, z 0–70 mm), the field concentrating in the sheath under the powered electrode and at the electrode edge — |E| field map · 300 mm PECVD baseline (3 Torr · 13.56 MHz · 500 W)

Trend simulator for RF CCP chambers

PlasmaCCPSee the plasma before you open the chamber

Sketch a chamber, hit run, and read the field, density, sheath and gas flow a few seconds later — cheap enough to test an idea before you book chamber time.

No account, no card. A guest run is solved in your browser, so the chamber geometry never leaves it.

Try it — no account Where we're right, and where we're not See how it works

Full solve time	~2 s	CI benchmarks	M1–M8
Literature anchors	V1–V10	Presets shipped	24

§1Three questions worth answering first

The ones a screening run can actually settle.

question	what a screening run can actually settle
Where does the field pile up?	Electrode corners, dielectric joints, the sheath edge — see which ones move when you change the geometry.
How does it change across the wafer?	Ion flux and ion energy along the wafer radius, so you can tell edge-heavy from center-heavy.
How much of it should you trust?	The validated range, the warnings and the model's limits sit next to the numbers, not in a footnote.

§2Start from a chamber that looks like yours

24 process presets ship with the simulator — these six cover the common families, and every card is an actual solve of that preset.

DF-CCP 60+2 MHz HAR Etch

60 MHz source + 2 MHz bias, CF₄/Ar high-aspect dielectric etch.

Open preset →Plasma-ALD O2 Step

Close-coupled 6 mm gap at 8 Torr — the conductance-limited regime.

Open preset →GEC Ref Cell — Ar

The standard research geometry (0.1 Torr Ar) — compare against published data.

Open preset →

§3Your geometry never reaches our servers

If you've dropped this kind of screening because the chamber cross-section can't leave the fab, that isn't a problem here.

property	behavior
Guest runs	The solver is compiled to WebAssembly and runs on your machine. Nothing is solved on ours.
Geometry	Not sent. A guest's only request carries 24 scalars for the judgment step — no shapes, no mesh, no snapshot.
History	Nothing is kept. A guest run leaves no history entry — close the tab and it's gone.
Signed in	Run history and saved geometries need the snapshot on your account. If you'd rather not, stay a guest — everything still runs.

§4Three steps, that's it

Draw → Run → See.

01	Draw	Sketch the chamber, or start from one of the presets.
02	Run	One click. The reduced-fluid solver converges in a few seconds.
03	See	Read the maps side by side and decide what to change next.

§5Setup, results and validation in one place

The inputs you ran, the maps they produced, and how far they've been checked — no second tool to open.

300 mm baseline · process overview

Fresh resultSolver: ConvergedValidity: Within rangeModel: Trend-levelMesh: 200 × 176 FVResidual: 3.70e-9RF phase: Not computedWarnings: 0 blocking / 1 notes

parameter	value	unit	condition
Electron temperature (0D model)	1.83	eV	Wafer-facing plasma volume
Mean sheath thickness	1.7	mm	Powered-electrode RF mean
Absorbed RF power	500	W	100.0% of 500 W applied

The results view itself, not a screenshot. The numbers come from a real solve of the baseline preset.

2D field maps	Render \|E\|, electron density, and temperature straight onto the chamber cross-section.
RF phase animationPro	Replay how the sheath and fields breathe over a single RF cycle.
Gas flow & current streamlines	Trace neutral-gas flow and current paths as streamlines.
Parameter sweepsPro	Sweep power, pressure, gap, or frequency at once to map sensitivity curves.
A/B comparePro	Place two runs side by side and isolate the effect of a single change.
Geometry library	Version your saved chambers and pick up any run where you left off.

§6We publish where we're wrong

Grounded in the literature — checked against an independent code.

literature anchors (V1–V10)	10	CI benchmarks (M1–M8)	8
automated tests	3,200+	full solve	~2 s

NOTE: We ran an independent code (Zapdos/MOOSE) against ours and published the result either way — our sheath came out 2.8× too thin, and we left the number in. See Olthoff & Greenberg 1995 (GEC reference cell), Lieberman & Lichtenberg (global model), Turner 2013 (PIC benchmark).

See what we don't claim →

§7One run is never the answer

An engineering call is always A versus B. Wider electrode, tighter gap, different focus ring — these tools exist to make that comparison directly.

tool	what it does
A/B comparePro	Two chambers, same recipe, side by side
Split testPro	Change one thing at a time and isolate its effect
Parameter sweepPro	Walk pressure, power or frequency into a trend line
Calibration	Fit the model to a measured point
Drift diagnosis	Work backwards from a chamber that changed

All of these need an account; the three marked Pro open on the paid plan. Single runs stay free with no account.

§8Pricing

The Free plan is fully usable today. Pro adds the sweep, A/B compare, and split-test workflows plus a higher mesh ceiling — paid checkout is not open yet.

plan	price	includes
Free	$0	Solver runs + every static field map Every process preset and baseline Save 20 private geometries · 30 runs of history Standard mesh up to 48k cells Get started free
Pro	Soon (Coming soon)	Compare, Split Test & sweeps Turn-on transient + GIF & RF phase animation High-res mesh & run history with pinning See details

Compare full pricing →

Run your first chamber now

Run a baseline Read the guide

About this model

PlasmaCCP is a reduced fluid model built for fast trend screening. Results are directional signals — confirm findings with experiments or high-fidelity simulation before sign-off. Read validation and references.