Home
Capabilities

What we are actually good at.

Four capabilities, deliberately narrow. We do not do market studies, strategy decks or organisational design. We do technical work on materials, and we turn it into evidence you can act on.

Failure & root-cause analysis

Find the mechanism. Then prove it.

A failure investigation is only finished when someone else can reproduce the failure on demand. Until then you have a theory.

We are usually the second or third team to look at a problem. By the time we arrive the easy explanations have been tried, the obvious process variables have been checked, and there is a customer waiting. Our first job is to widen the hypothesis set back out — including the possibilities your team has already dismissed — and our second is to design measurements that can kill most of them quickly.

Typical problems

Field returns with no reproducible bench failure
Yield excursions that began after an unrelated change
Delamination, cracking, corrosion and interfacial failure
Contamination of unknown origin
Batch-to-batch variation nobody can trace to a variable

Methods we draw on

SEM–EDS, FIB cross-sectioning, TEM, XPS and ToF-SIMS for surfaces and interfaces; XRD for phase and texture; FTIR, Raman, GC–MS and LC–MS for chemistry; DSC, TGA and DMA for thermal and mechanical transitions; CT and acoustic microscopy for internal defects. We commission work through accredited laboratories and supervise the sample preparation ourselves, because that is where most investigations are quietly ruined.

You receive

A written mechanism with the supporting and contradicting evidence set out separately
A reproduction protocol so your team can trigger the failure deliberately
Corrective actions ranked by effect size, cost and time to implement
All raw data, micrographs and laboratory reports, unredacted

Formulation & materials selection

Fewer experiments. Better ones.

Most formulation programmes fail not from lack of effort but from running one-factor-at-a-time trials against a system with strong interactions.

We design experimental programmes that actually resolve the response surface: screening designs to find which of your twelve candidate variables matter, then response-surface or mixture designs to optimise the three or four that do. Where a regulatory change or a supplier exit has forced your hand, we work the substitution problem systematically rather than by trial and error.

Typical problems

A property target that will not move despite months of trials
REACH, RoHS, PFAS or biocide restrictions forcing reformulation
A sole-source raw material being discontinued
Cost-down without a measurable performance loss
Bio-based or recycled feedstock substitution

How we work

Statistically designed experiments, executed either in your laboratory with your staff or through partner facilities. We build the model, we show you the confidence intervals, and we are explicit about where the design cannot tell you anything — which is usually the most useful part of the report.

You receive

An executable experimental design with run order and sample sizes
A fitted model with the operating window and its uncertainty stated
Qualified alternative raw materials with a comparison matrix
Formulation records in a form your own team can maintain

Process & scale-up engineering

The chemistry did not change. Everything else did.

Between a 200 mL flask and a 2 m³ reactor, mixing time, heat removal, shear history and surface-to-volume ratio all move by orders of magnitude. The reaction notices.

We work the transport problem explicitly rather than hoping geometric similarity will carry the process across. That means characterising the rate-limiting step at bench scale, identifying which dimensionless group actually needs to be held constant, and designing the pilot trial to test that specific assumption before capital is committed.

Typical problems

Yield or selectivity collapsing on transfer to pilot or plant
Exotherms that were invisible at bench scale
Coating, drying or curing defects appearing only at line speed
Particle size, morphology or dispersion drifting with batch size
A technology transfer package a receiving site cannot execute

How we work

Reaction calorimetry and kinetic modelling; mixing and residence-time analysis; heat and mass transfer calculations, with CFD where the geometry genuinely warrants it. We attend the pilot trials. Scale-up problems are not solved from a desk.

You receive

A scale-up basis stating which parameters are held constant and why
Pilot trial protocols with defined success and abort criteria
A thermal safety assessment where the chemistry requires one
A technology transfer package written for the receiving site's equipment

Reliability & qualification

A test that predicts something real.

An accelerated test is only meaningful if it drives the same failure mechanism as the field. A great many of them do not.

The common failure is an acceleration factor calculated from an activation energy that was never measured for this material, in this configuration. We start from the mechanism, establish the life–stress relationship experimentally, and only then design the accelerated test. Where the mechanism changes at the stress level you were planning to use, we say so — and the test gets redesigned rather than quietly producing a reassuring number.

Typical problems

A warranty period far longer than any test that supports it
Qualification evidence rejected in a customer audit
Accelerated results that do not correlate with field returns
A new material entering an existing qualified assembly
Standards-mandated testing that misses your dominant failure mode

How we work

Physics-of-failure modelling, Weibull and Arrhenius–Eyring life analysis, HALT and HASS design, damp-heat and thermal-cycling protocols, and design of experiments for qualification matrices. Aligned to IEC, ISO, ASTM, JEDEC or AEC-Q as your market demands.

You receive

A life model with the acceleration factor and its derivation shown
Test protocols traceable to the relevant standard clause by clause
An audit-ready qualification dossier
An explicit statement of the residual risk the programme did not close

Engagement models

Three ways to work with us.

Technical opinion

A short, sharp read on a specific question by a senior engineer. Two to five days. Written, not a phone call you cannot forward.

Second opinion on an internal finding
Feasibility screen before committing budget
Supplier claim assessment

Scoped investigation

The core offer. A defined technical question, a fixed scope, a fixed fee and a named team. Six to fourteen weeks.

Full failure investigations
Formulation or scale-up programmes
Qualification campaigns

Embedded technical support

Named engineers attached to your programme for an agreed number of days per month. For teams carrying a long technical risk.

New product introduction support
Technical due diligence
Capability building for your own staff

Practicalities

The questions we get asked first.

Do you have your own laboratory?

No, and that is deliberate. A consultancy that owns instruments has an incentive to recommend the measurements those instruments can make. We specify the analysis and commission it from accredited laboratories chosen for the problem, then supervise sample preparation and interpret the results.

Where you have capable instruments in-house, we will use yours — it is faster and cheaper, and your team learns the method.

How do you handle confidentiality?

A mutual NDA is signed before the first technical call, not before the proposal. We will not discuss the substance of your problem until it is in place. We do not name clients in marketing material without written permission, which is why the references on this site are anonymised.

We also screen for conflicts before accepting work and will decline an engagement rather than take on two competing clients in the same narrow technical area.

Who owns the intellectual property?

You do. Foreground IP arising from a client engagement is assigned to the client as standard. We retain only our pre-existing methods and general know-how, which we could not un-learn in any case. This is in our standard terms and we do not negotiate it downward.

What if you cannot find the root cause?

It happens. Some failures are genuinely under-determined by the available evidence — the parts have been reworked, the process records are incomplete, or the failed population is too small to support a conclusion.

In that case you get a report saying exactly that, with the hypotheses we eliminated, the ones still standing, and what evidence would be needed to separate them. That is a real result, and it stops you spending another six months on the wrong theory.

Will you give evidence in a dispute?

Yes, in a technical expert capacity, provided the work was conducted to that standard from the outset — which requires chain of custody on samples and a documented method from day one. Tell us at intake if litigation is possible; retro-fitting evidential rigour is not possible.

Not sure which of these you need?

Describe the symptom. Working out which capability applies is our job, and we will tell you honestly if the answer is none of them.

Describe your problem Browse by industry