What industries does newtsim serve?

newtsim serves oil and gas (flow assurance, drilling fluids, subsea pipeline), pharmaceutical and drug discovery, energy storage (battery materials, electrolytes), semiconductor process simulation, green hydrogen (electrolyser and catalyst design), nuclear decommissioning, aerospace and composites, corrosion and materials, carbon capture, agrochemicals, mining and geotechnical, and fermentation and bioprocessing.

How is newtsim different from standard CFD or simulation software?

Standard simulation tools use empirical models at a single scale. newtsim runs the full physics chain from quantum mechanics and molecular dynamics through to plant-scale CFD and structural mechanics, GPU-native, with direct data flow between every scale and no third-party solver dependencies. Constitutive behaviour is predicted from first principles rather than empirically fitted.

What does a fixed-price simulation study include?

A newtsim fixed-price study has a defined scope, clear deliverables, fixed cost agreed upfront, and is delivered in weeks rather than quarters. Studies typically include a scoping call, physics-grounded simulation using the appropriate newtsim solvers, results analysis, and a written report with actionable recommendations.

What is physics-grounded simulation?

Physics-grounded simulation derives material properties and constitutive behaviour from lower-scale physics (molecular dynamics and quantum mechanics) rather than from empirical fitting. This gives accurate predictions for novel materials and extreme conditions outside the range of experimental data, and enables root-cause analysis at the mechanistic level.

How does newtsim approach oil and gas flow assurance simulation?

newtsim runs a continuous physics chain from molecular wax crystallisation to full pipeline transient multiphase flow, with no empirical handoffs between scales. Wax appearance temperatures, gel strength, and deposition kinetics are predicted from molecular dynamics (newtsim Bond), passed directly to mesoscale morphology (newtsim Pack), and then propagated into multiphase CFD (newtsim Stream), giving flow assurance predictions that hold outside the range of laboratory data.

Can newtsim simulate battery electrolyte behaviour and degradation?

Yes. The Bond solver runs molecular dynamics for Li-ion transport and solid-electrolyte interphase (SEI) formation; Pack handles electrode morphology at the mesoscale; these results couple upward to cell-scale electrochemical models. This gives physics-grounded predictions of electrolyte degradation, capacity fade, and thermal behaviour without relying solely on empirical ageing fits.

How does physics-grounded drug discovery simulation differ from docking software?

Molecular docking software treats the protein and ligand as rigid bodies and scores binding poses by geometry. Physics-grounded simulation uses molecular dynamics to capture conformational flexibility and explicit solvation effects that rigid-body docking misses, predicting binding free energy, ADMET properties, and metabolic activation from first principles. This is particularly valuable for targets where induced-fit binding or solvent-mediated interactions drive selectivity.

What makes GPU-native simulation faster than CPU-based CFD?

GPU parallelism maps naturally to the computational kernels underlying molecular dynamics and CFD: large numbers of independent force or flux calculations that can execute simultaneously. newtsim runs all simulation scales GPU-native, avoiding the CPU-to-GPU data transfers that bottleneck traditional solvers which run physics on CPU and only offload selected operations to the GPU. This removes the transfer overhead and allows the full multiscale chain to run without the latency penalties of heterogeneous pipelines.

Does newtsim replace existing simulation tools like ANSYS or COMSOL?

newtsim fills the multiscale gap, connecting molecular-scale physics to the engineering-scale models that tools like ANSYS operate at. It is not a replacement for FEA-only or single-scale workflows where those tools are the right fit. Where newtsim adds value is in problems where constitutive behaviour, material properties, or interfacial physics need to be derived from lower-scale simulation rather than taken from a handbook or empirical fit.

How long does a fixed-price simulation study take?

Most studies complete in 4 to 8 weeks depending on scope. The standard timeline is: scoping call and problem definition in week 1; simulation execution across the relevant physics scales in weeks 2 to 4; results analysis, validation, and report delivery in weeks 5 to 6. More complex multiscale studies with broader operating envelopes may extend to 8 weeks. Scope and timeline are agreed and fixed before work begins.

What deliverables does a simulation study produce?

A newtsim simulation study delivers: a validated multiscale model calibrated to your system; an operating envelope with physics-based safety margins; a written technical report with findings and actionable recommendations; and optionally CFD visualisations or an ML surrogate model (newtsim Neural) for fast design-space exploration by your engineering team.

Does newtsim work for green hydrogen electrolyser and catalyst design?

Yes. The Stream solver handles proton exchange membrane and alkaline electrolyser CFD: multiphase flow, bubble dynamics, and membrane transport. Bond handles catalyst surface reaction molecular dynamics, giving adsorption energies and reaction barriers from first principles rather than from empirical correlations. Neural then produces fast-to-run surrogate models for design-space exploration and stack optimisation, reducing the iteration cost for new catalyst compositions and electrode architectures.

newtsim is a physics-grounded multiscale simulation consulting firm and software platform. We deliver fixed-price simulation studies and platform licences for industrial clients, covering the full physics chain from quantum mechanics and molecular dynamics through to plant-scale CFD and structural mechanics. All solvers run GPU-native.

Does newtsim use artificial intelligence or machine learning?

Yes. newtsim Neural produces ML-trained surrogate models derived from high-fidelity physics simulations. These surrogates run in milliseconds rather than hours, enabling rapid design-space exploration and real-time control without sacrificing physical accuracy. The ML models are trained on physics simulation data, not empirical measurements, so they inherit the accuracy and extrapolation properties of the underlying physics.

How do I get started with newtsim?

Book a free technical scoping call at newtsim.com/demo. Bring a description of your engineering challenge, the physics you think are important, and any data you have. The call is 45 minutes and produces a problem definition, proposed physics chain, and outline scope that forms the basis of a fixed-price proposal.

Can newtsim simulate non-Newtonian fluids?

Yes. newtsim specialises in non-Newtonian flow: yield-stress fluids, viscoelastic materials, shear-thinning and shear-thickening systems, and complex multiphase flows. The Stream solver handles macroscale CFD for these systems; Bond provides the molecular-scale rheology predictions that replace empirical viscosity fits; and the full chain gives accurate behaviour outside the calibrated range of laboratory data.

Does newtsim work for small companies or startups?

Yes. Fixed-price studies are available to clients of any size. The fixed-price, fixed-scope model means you know the cost before committing, making it accessible for smaller organisations without large simulation teams. Many newtsim clients are startups and scale-ups who need physics-grounded analysis early in product development before investing in laboratory or pilot-scale work.

What is the difference between molecular dynamics and CFD?

Molecular dynamics (MD) simulates the motion of individual atoms and molecules to compute material properties such as viscosity, diffusivity, reaction rates, and phase behaviour, from first principles. Computational fluid dynamics (CFD) models bulk fluid flow at the engineering scale using those material properties as inputs. newtsim connects the two: MD provides accurate constitutive inputs that CFD uses for engineering-scale predictions, rather than relying on empirical correlations.

What is a surrogate model in simulation?

A surrogate model (also called a reduced-order model or emulator) is a fast mathematical approximation of a complex simulation. newtsim Neural builds ML surrogates trained on high-fidelity physics simulation data. Once trained, a surrogate evaluates in milliseconds instead of hours, enabling real-time optimisation and design-space exploration. Because surrogates are trained on physics data, they are more accurate than empirical correlations, especially outside the training range.

What deliverables does a newtsim simulation study produce?

A newtsim simulation study delivers: (1) a validated multiscale model calibrated to your system; (2) an operating envelope with physics-based safety margins; (3) a written technical report with findings and actionable recommendations; and optionally (4) CFD visualisations and an ML surrogate model for fast ongoing design-space exploration by your engineering team.

Semiconductor Manufacturing

Semiconductor Manufacturing · Case Study

High-Aspect-Ratio Contact Etch Profile Prediction for 5nm DRAM Storage Node

Executive Summary

At the 1z-nm DRAM node — equivalent to aspect ratios of 68:1 for storage node contacts at 15 nm top CD and 1,020 nm depth — the fundamental physics of high-aspect-ratio contact (HARC) etch change character. Published IEDM data from SK Hynix (2021) and Micron (2022) document what engineers at leading DRAM manufacturers confronted in production: traditional recipe optimisation through design-of-experiments splits provides diminishing returns above 60:1 aspect ratio because ion angular distribution broadening and fluorocarbon polymer redeposition couple non-linearly in the deep feature. Micron's D1-alpha node quantified this as a 15–22% etch rate reduction at 68:1 AR versus 52:1 AR under otherwise identical conditions. In practice, the consequence is bottom CD collapse: contacts narrowing from a 14.8 nm target to 6.2 nm at depth — a 58% reduction — with contact resistance 8.3x above specification and storage node continuity yield at 61% against a target above 97%. A 6–9 month node launch delay at 40,000 wafers per month represents $300–450M in deferred revenue. Fourteen process-of-record splits without systematic convergence is the signature of a mechanistic understanding gap, not a recipe problem.

The root cause cannot be resolved by DOE alone because two mechanisms are acting simultaneously and cannot be independently varied on hardware: ion scattering at grazing incidence off the sidewall polymer broadens the effective ion angular distribution from 3.2 degrees FWHM at the wafer surface to 7.8 degrees effective FWHM at 600 nm depth, while fluorocarbon polymer buildup reaches 1.1 nm thickness at 800 nm depth and reduces the effective contact radius independently. Together they account for 62% and 38% of the 6.2 nm total bottom CD loss respectively — but empirical splits that change bias voltage also change polymer deposition rate, making the contributions inseparable without a mechanistic model.

Had feature-scale simulation been applied during the node transition from 52:1 to 68:1 AR, the causal pathway would have been resolved before the mask set was committed. Monte Carlo ion trajectory simulation within an evolving 15 nm contact geometry, coupled to 2D axisymmetric plasma bulk transport modelling, quantifies the ion angular distribution function at each depth increment and tracks polymer layer thickness independently. The simulation identifies the tipping point: above 60:1 AR, the surface IEAD FWHM must be narrowed below 2.5 degrees to prevent effective FWHM at depth from exceeding 5 degrees — a constraint that sets the minimum bias voltage. For this process, the correct operating point is 680 V bias with C4F8 reduced from 15% to 10%, recovering bottom CD to 13.3 nm at 1,020 nm depth and dropping etch stop probability from 12% to 0.8%.

The 33-percentage-point recovery from 61% to 94% storage node contact yield represents approximately $260M per year in at-stake revenue — recoverable only once the mechanistic root cause (ion angular distribution and polymer redeposition coupling) is resolved rather than navigated by continued DOE splits. The simulation's key risk parameters — ion FWHM versus depth, polymer thickness versus C4F8 fraction, bottom CD sensitivity to bias voltage — define the in-line metrology targets for newtsim livesim: real-time wafer-level monitoring that tracks CD at depth via scatterometry proxy metrics, flagging polymer accumulation or ion directionality drift before etch stop events propagate across a production lot.

Charts

Left: dense vs. isolated contact etch rate at baseline and optimised conditions. Right: 62%/38% attribution of bottom CD loss to ion scattering vs. FC polymer redeposition.

Scenario Background

(illustrative reference case)

In this worked example, a DRAM manufacturer planning a next-generation storage node cell retained a simulation consultancy to predict etch profile behaviour before committing to a full mask set redesign. The fictional entity, Meridian Memory Technologies, is calibrated to published 1z-nm node characteristics.

The scenario concerns a 1z-nm DRAM process (15nm half-pitch, 6F² cell layout), equivalent to leading-edge DRAM nodes reported by SK Hynix and Micron at IEDM 2021--2023. The etch tool is a Lam Research ALTUS-series capacitively coupled plasma (CCP) etch system for 300mm wafers, operating at 13.56 MHz RF source with 400 kHz bias frequency. The process gas mixture is C4F8/CH2F2/O2/Ar at 20--60 mTorr, representing standard HARC etch chemistry for SiO2. The hard mask stack comprises SiN/SiO2/amorphous carbon hardmask (ACH) at 2.8 um total stack, requiring ACH selectivity greater than 15:1 to SiO2. The etch target is 1,020 nm depth in SiO2 capacitor dielectric at 15 nm target top CD, yielding a 68:1 aspect ratio.

The transition from the 1y-nm to the 1z-nm node increases the aspect ratio from 52:1 to 68:1 for the same nominal contact diameter (15 nm). Internal observation showed systematic bottom CD collapse at depth greater than 450 nm, with bottom CD measured at 6.2 nm vs. the 14.8 nm target -- a 58% narrowing -- and yield at the storage node contact layer of 61% vs. a target above 97%.

The 68:1 aspect ratio represents the leading edge of HARC etch capability as of 2022--2023. Published DRAM roadmap data (TechInsights, SEMI) confirm the industry is pursuing aspect ratios of 60:1 to 80:1 in the 1z/1-alpha node range, with aspect ratio increasing approximately 10:1 per node generation as cell height is maintained while diameter shrinks. Micron's D1-alpha node, the most advanced as of 2023, operates in this AR regime.

Challenge

The HARC etch process, while achieving correct top CD and mask selectivity, produced systematic and reproducible profile distortion:

Failure Mode	Measured Value	Specification / Target	Business Impact
Bottom CD at 1,020 nm depth	6.2 nm	14.8 nm (58% narrowing)	Contact resistance 8.3x above spec; capacitor fill impossible
Etch stop events	12% of contacts stopped at 600--700 nm depth	<1% etch stop rate	12% contacts electrically open before Cu fill
Sidewall taper angle	87.3 degrees	89.5 degrees (target near-vertical)	8.6 nm cumulative CD loss across 1,020 nm depth
Micro-loading (isolated vs. dense)	18% etch rate difference	<5% loading differential	Isolated contacts over-etch; dense arrays under-etch simultaneously
SNC contact resistance	8.3x specification	<1.1x specification	All contacts in affected zone fail parametric test
SNC continuity yield	61%	>97%	36 pp shortfall; each 1% improvement worth ~$8M/year at volume

The 14 process of record splits attempted by the internal team covered a broad parameter space without systematic convergence. Bias power was varied from 300 to 600 W, where bottom CD improved marginally above 500 W but etch stop frequency increased. Pressure was swept from 20 to 50 mTorr, with lower pressure slightly improving directionality but increasing micro-loading. C4F8 fraction was explored from 10 to 20%, where lower C4F8 reduced etch stop events but also reduced selectivity to the hardmask. Addition of CH2F2 from 0 to 5% did not systematically improve bottom CD.

The failure to converge across 14 splits indicated a mechanistic understanding gap: the team was navigating a high-dimensional parameter space without a physical model to identify the causal pathway. Two hypotheses were held simultaneously without resolution: (A) polymer plug formation is the primary mechanism, and (B) ion angular distribution degradation is primary. The simulation study was commissioned to resolve this ambiguity.

DRAM pricing and market share depend critically on node advancement timing. A 6--9 month delay in 1z-nm node launch at 40,000 wafers/month production target (at $50M/month revenue projection) represents $300--450M in delayed revenue. Each percentage point of SNC yield contributes approximately $8M/year to operating margin at volume. The 36 pp yield gap represents approximately $288M/year at stake.

Real-World Basis

HARC etch profile distortion is a thoroughly documented challenge at aspect ratios beyond 50:1.

SK Hynix presented storage node contact challenges at aspect ratios above 60:1 at IEDM 2021, describing profile simulation as a necessary tool for process window definition at this AR regime. Traditional recipe optimisation through DOE-based splits provides diminishing returns above 60:1 AR because ion angular distribution, polymer deposition, and feature geometry couple non-linearly.

Micron's 1-alpha generation data identified etch micro-loading and aspect-ratio-dependent etch rate (ARDE) as primary yield limiters at sub-15nm contact diameter, quantifying ARDE as a 15--22% etch rate reduction at 68:1 AR versus 52:1 AR under otherwise identical conditions -- consistent with the 18% micro-loading differential in this scenario.

Monte Carlo simulation of ion scattering at 50:1 AR has shown that ion shadowing is the dominant mechanism for bottom CD reduction: the effective ion angular distribution at the contact bottom broadens from the surface FWHM of approximately 3 degrees to approximately 8--12 degrees at 50:1 AR due to multiple grazing-incidence sidewall collisions. This is consistent with the 68:1 AR predictions in this study (3.2 degrees surface to 7.8 degrees effective at 600 nm depth).

Published fluorocarbon polymer deposition models for HARC SiO2 etch measured CF2 sticking coefficient on SiO2 at 0.06--0.10 (this study uses 0.08) and predicted polymer layer thickness at 800 nm depth at 0.9--1.3 nm for standard C4F8/O2 chemistry -- consistent with the simulation prediction of 1.1 nm.

Lam Research demonstrated that tailored voltage waveforms (TVW) for pulsed plasma bias at >60:1 AR narrow the ion energy distribution compared to sinusoidal bias, recovering 2--4 nm of bottom CD at equivalent etch selectivity. Published 3D Monte Carlo simulation studies for Ar/C4F8/O2 HARC etch at 37:1 AR validated feature-scale MC methodology with mean absolute error of approximately 0.7 nm versus STEM cross-section measurements across 10 depth positions, establishing methodological credibility for the approach deployed in this study.

Simulation Approach

The HARC etch simulation methodology couples plasma bulk transport (2D reactor scale) to feature-scale Monte Carlo particle tracing (sub-micron geometry) through a hierarchical handoff protocol:

Phase 1 -- Plasma bulk transport model (weeks 1--2):

The first phase constructs a 2D axisymmetric fluid model of the CCP reactor at operating conditions (20--60 mTorr, C4F8/O2/Ar with CH2F2 additive). The model solves coupled transport equations for electrons and 12 ion species to extract the quantities that matter at the wafer surface: ion energy and angular distributions (IEAD) for each etch-relevant species at 6 bias voltage conditions (300--800 V), neutral radical flux maps at 4 pressure conditions (20--60 mTorr), and time-averaged sheath voltage and thickness.

The IEAD at the wafer surface shows FWHM of 3.2 degrees at 450 V bias, narrowing to 2.1 degrees at 680 V -- demonstrating the key result that higher bias voltage improves ion directionality by accelerating ions through a thicker, more stable sheath.

Phase 2 -- Feature-scale Monte Carlo etch simulation (weeks 2--5):

The second phase performs 3D Monte Carlo particle tracing for ion and neutral transport within an evolving 15 nm-diameter, 1,020 nm-deep contact hole.

Ion scattering is modelled as a mix of specular and diffuse reflection at the sidewall polymer layer, with scattering angle distributions calibrated to molecular dynamics data for F+ impact on fluorocarbon polymer. The etch rate model combines physical sputtering with chemical etch, applying the ion-assisted synergy where ion-enhanced chemical etch rate is 3--5x larger than the sum of physical and thermal components alone. Fluorocarbon polymer redeposition is tracked using a CF2 sticking coefficient of 0.08 on SiO2; polymer layer thickness is tracked at each depth increment and removed by ion sputtering. The profile evolves self-consistently, with CD recorded at every 20 nm depth for comparison to STEM data. The parameter sweep covers 6 bias voltages by 4 pressures by 4 C4F8 fractions, for 96 simulation conditions total.

Key simulation finding -- mechanistic attribution of bottom CD loss:

Mechanism	Contribution to Bottom CD Loss	Physical Cause
Ion angular distribution broadening	62% (3.9 nm of 6.2 nm total)	Multiple grazing-angle sidewall collisions deflect ion trajectories away from vertical; ions arrive at contact bottom with effective FWHM 7.8 degrees vs. 3.2 degrees at surface
Fluorocarbon polymer redeposition	38% (2.4 nm of 6.2 nm total)	CF2 radicals deposit on sidewall polymer; at 800 nm depth, simulated polymer thickness 1.1 nm reduces effective contact radius by 2x polymer thickness = 2.2 nm
Combined (non-linear)	100% (6.2 nm measured)	Ion scattering off polymer-coated sidewall creates secondary scattering amplifying both effects

Phase 3 -- Micro-loading analysis (weeks 5--6):

A neutral radical depletion model captures pattern-density-dependent etch rate. In dense contact arrays (3% open area fraction), F and O radical fluxes are depleted by consumption at the array surface, reducing effective etch species flux by 22% vs. an isolated contact (40% open area). The micro-loading model implements an etch rate correction factor as a function of local open area fraction (OAF): rate_correction = 1 - 0.65 x (1 - OAF/OAF_ref). Calibrated against a 14-point density split matrix, the model yields R-squared = 0.91.

Phase 4 -- Optimised process recommendation (weeks 7--8):

A response surface model (RSM, quadratic polynomial fit) is fitted to CD(z) profiles across the 96-condition parameter sweep. Gradient-based optimisation minimises bottom CD loss subject to three constraints: etch stop probability below 1% (polymer plug formation constraint), mask selectivity greater than 8:1 (ACH hardmask erosion constraint), and micro-loading below 5% (from pressure optimisation). The optimal condition identified is bias 680 V, pressure 55 mTorr, C4F8 10%.

Simulation Caveats

The HARC feature-scale Monte Carlo methodology is well-validated at aspect ratios up to 50:1 in the peer-reviewed literature. At 68:1 AR, additional considerations apply.

At 68:1 AR with 15 nm diameter, the solid angle subtended by the contact bottom from the top opening is approximately 0.05 steradians. Monte Carlo convergence requires more than 10⁶ ion trajectories per simulation to achieve less than 5% statistical error in the ion flux at the contact bottom; this was verified by convergence testing.

The fluorocarbon polymer composition (CF_x stoichiometry, cross-link density) evolves during etching and affects ion scattering probability. The model uses a fixed CF2 composition; real-time polymer evolution is not captured, introducing +/-15% uncertainty in the polymer contribution to bottom CD loss.

At very high AR (>65:1), twisting and tilting of contact features (distortion perpendicular to the contact axis) have been reported in the literature. This study's 2D (cylindrically symmetric) model does not capture twisting modes; a separate 3D simulation would be required. The 400 kHz bias generates time-varying sheath conditions that are captured in the time-averaged IEAD but not in full time-resolved detail; pulsed bias analysis would require a separate time-resolved plasma model.

Key Predictions / Results

Etch Rate and Ion Angular Distribution vs. Aspect Ratio

Primary simulation vs. measurement comparison:

Metric	Baseline Process (450 V, 20 mTorr, C4F8 15%)	Optimised Prediction (680 V, 55 mTorr, C4F8 10%)	Target / Specification
Bottom CD at 1,020 nm depth	6.2 nm	13.3 nm	14.8 nm (delta = 1.5 nm from target)
Bottom CD at 600 nm depth	10.4 nm	14.1 nm	14.8 nm
Profile taper angle	87.3 degrees	89.1 degrees	>=89.5 degrees
Etch stop probability (polymer plug)	12%	0.8%	<1%
Micro-loading (isolated vs. dense rate delta)	18%	4% (at 55 mTorr pressure)	<5%
SNC contact resistance	8.3x specification	1.1x specification	<=1.2x specification
Projected SNC continuity yield	61%	94%	>97%
Ion FWHM at wafer surface	3.2 degrees	2.1 degrees (at 680 V bias)	--
Ion effective FWHM at 600 nm depth	7.8 degrees	4.3 degrees (narrower due to higher bias directionality)	--
FC polymer thickness at 800 nm depth	1.1 nm	0.4 nm (lower C4F8 fraction reduces deposition)	<0.5 nm (target)

CD(z) profile by depth -- baseline vs. optimised:

Depth (nm)	Baseline CD (nm)	Optimised Prediction CD (nm)	Target CD (nm)
0 (top)	14.9	15.0	15.0
100	14.5	14.8	14.8
200	13.8	14.6	14.8
300	13.0	14.4	14.8
400	11.8	14.2	14.8
500	10.2	13.9	14.8
600	8.6 (12% of contacts stop here)	13.5	14.8
700	7.4	13.4	14.8
800	6.8	13.4	14.8
900	6.5	13.3	14.8
1,020 (bottom)	6.2	13.3	14.8

HARC Etch CD Profile: Baseline vs. Optimised

The root cause is quantified as follows. The ion angular distribution at the wafer surface (FWHM 3.2 degrees at 450 V) broadens effectively to 7.8 degrees at 600 nm depth due to multiple grazing-incidence sidewall collisions. Combined with fluorocarbon polymer buildup (simulated thickness 1.1 nm at 800 nm depth), the effective ion flux at the contact bottom is reduced by 74% relative to the open wafer surface. Increasing bias voltage from 450 V to 680 V narrows the surface IEAD FWHM from 3.2 to 2.1 degrees, reducing effective FWHM at 600 nm depth from 7.8 to 4.3 degrees. Simultaneous reduction of C4F8 fraction from 15% to 10% reduces CF2 radical flux and polymer deposition rate, bringing polymer thickness at 800 nm depth from 1.1 nm to 0.4 nm.

Comparison Methodology

The primary validation compares the higher-fidelity Monte Carlo feature-scale simulation against lower-fidelity metrology (STEM cross-section, XPS, retarding field analyser). Published experimental data at lower aspect ratios provides secondary confirmation.

Validation metric	Simulation prediction	Measured reference	Agreement metric
CD(z) profile -- baseline	6.2 nm at 1,020 nm depth	6.2 nm +/- 0.8 nm (STEM, 40 contacts)	Exact match at tip; mean absolute error 0.9 nm across 15 depth points
Polymer thickness at 800 nm	1.1 nm (simulated)	1.0 +/- 0.2 nm (XPS sidewall composition)	Within measurement uncertainty
FC film F/C ratio	CF1.8 (simulated)	CF1.7 +/- 0.1 (XPS)	Within 6% of F/C ratio
Micro-loading etch rate delta	18% dense vs. isolated (simulated)	18% +/- 3% (14-point density split)	R-squared = 0.91 across all density conditions
IEAD FWHM at surface	3.2 degrees at 450 V (simulated)	3.0 +/- 0.4 degrees (energy-resolved retarding field analyser)	Within 0.2 degrees
Published benchmark (50:1 AR)	Profile at 50:1 AR	Published experimental data	Profile within 8% across 12 depth positions

STEM cross-section samples comprised 40 contacts sectioned across 3 wafer locations (centre, mid-radius, edge) on each of 5 qualification wafers. STEM operated at 200 kV in HAADF mode with spatial resolution of approximately 0.5 nm, and CD was measured at 15 depth positions per contact from the HAADF contrast profile.

Deliverables

Plasma transport model report: IEAD at wafer surface for baseline and 3 optimised conditions (450 V baseline, 550 V intermediate, 680 V optimised); species flux maps (F, CF2, O radicals; F+, CF3+, Ar+ ions); sheath voltage and thickness vs. bias frequency; delivered as PDF report + CSV data tables.
Feature-scale simulation database: CD(z) profiles for full 6x4x4 process parameter matrix (bias voltage x pressure x C4F8 fraction), 96 conditions total; delivered as CSV with 20 nm depth resolution + VTK visualisation files for 3D profile rendering in ParaView.
Root cause analysis report: Mechanistic attribution of bottom CD loss to ion scattering (62%) vs. polymer redeposition (38%), with supporting simulation evidence; CD(z) decomposition showing separate contributions; physical explanation of coupling between bias voltage, ion directionality, and polymer growth balance.
Optimised POR specification: Bias 680 V, pressure 55 mTorr, C4F8 10%; process control limits (3-sigma): bias +/-20 V, pressure +/-2 mTorr, C4F8 fraction +/-0.5%; sensitivity table showing bottom CD response to each control parameter.
Micro-loading correction recipe: Pressure adjustment algorithm for density-dependent etch rate compensation; lookup table for OAF to pressure offset; recommended OAF-aware etch segmentation (split etch into dense-array and isolated zones with different pressure setpoints).
Yield projection report: SNC continuity yield vs. process window position; predicted yield at 27 combinations (3 bias x 3 pressure x 3 C4F8) in the neighbourhood of optimal; recommended SPC control chart limits for on-going production monitoring using inline CD-SEM proxy metrics.

Timeline: 8 weeks from receipt of etch tool characterisation data (IEAD measurements or RF probe data confirming sheath voltage), STEM cross-section images (40 contacts minimum across 3 wafer positions), and 14-split process characterisation data (CD-SEM measurements for RSM calibration).

This case study is an illustrative reference scenario demonstrating newtsim's simulation methodology. All company names, personnel, and specific operational data are fictional. The incident descriptions draw on publicly documented real-world events cited in the frontmatter.