The most common outcome of a first graphene trial is not success or failure. It is ambiguity. The team adds graphene, sees a small change in one property and a small loss in another, and cannot tell whether the result reflects the material, the dispersion, the loading, or batch noise. Six months later the project is quietly shelved.
This is avoidable. A pilot trial is an experiment, and it should be designed like one.
Step 1: Write Down the Decision Before the Experiment
Start with the business question, not the material. “Does graphene improve our product?” is not answerable. “Can we reduce coating thickness by 20% while keeping salt-spray performance, at an added cost below X per square metre?” is.
A good decision statement names three things:
- The property that matters, measured by the test your customers or standards actually use.
- The threshold that would justify a change. A 3% improvement that disappears in normal batch variation is not a result.
- The cost ceiling, so that a technically successful trial that cannot pay for itself is recognised as a failure early.
If you cannot write this down, you are not ready to buy material.
Step 2: Characterise Your Baseline
Run your current formulation several times through the same process and tests you will use in the trial. This gives you the natural variation of your own system.
Teams skip this step more than any other, and it is the reason most results are unreadable. Without a measured baseline spread, you cannot tell whether a 5% change is signal or noise.
Step 3: Choose Materials Deliberately
Pick two or three grades, not one. Different grades — nanoplatelets of different thickness, a functionalised grade, a dispersion — behave differently. Testing one grade and concluding “graphene doesn’t work” generalises from a sample of one.
Get the data sheet and the certificate of analysis for the actual batch. Our guide to reading data sheets explains what to look for. Keep a retained sample of every batch you test.
Start with pre-dispersed forms where possible. A masterbatch or dispersion separates “does graphene help my system?” from “can my plant disperse graphene?” Answer the first question first.
Step 4: Bracket the Loading
Graphene properties rarely improve in a straight line with loading. Many systems peak at a low loading and then decline as agglomeration takes over.
Test at least three or four loadings spanning a wide range — for example 0.05%, 0.2%, 0.5%, and 1% by weight for a polymer composite. A single loading tells you almost nothing about where the optimum is.
Step 5: Use Proper Controls
Every trial needs at least two controls:
- The unmodified baseline, processed identically, including any extra mixing steps. Extra mixing alone can change properties.
- A cheap filler at equal loading or equal cost, such as carbon black, talc, nanoclay, or an incumbent additive. This answers the question that matters commercially: is graphene better than what you could do more cheaply?
If a dispersant or carrier comes with the graphene, include a control with the dispersant alone. Surfactants can change properties on their own.
Step 6: Verify Dispersion
Do not assume the graphene dispersed. Use at least one simple check — optical microscopy of a thin section, a rheology measurement, or a conductivity-versus-loading curve. A result without dispersion data cannot be interpreted, because a negative outcome may simply mean the flakes never separated.
Step 7: Measure More Than the Target Property
Graphene can improve one property while degrading another. Measure the target property plus the properties most likely to suffer: impact strength, elongation, viscosity, cure time, colour, and processability. A coating with better barrier performance that can no longer be sprayed is not an improvement.
Step 8: Replicate
Make each formulation more than once, from separately prepared batches. One good sample is an anecdote. Three consistent ones are evidence.
Step 9: Test on Production Equipment Before Deciding
Laboratory plaques and bench-mixed coatings are useful for screening. They do not reproduce the shear, temperature history, and orientation of real production. Before any go decision, make parts on your actual line, or the closest scale-down available.
Step 10: Decide Against the Criteria You Wrote in Step 1
Compare results to the threshold and cost ceiling you set before the trial. Resist the temptation to redefine success after seeing the data. If the trial missed but showed a clear trend, write down what the next trial would test and why — and set a limit on how many rounds you will fund.
A Realistic Timeline and Budget
A well-run first evaluation for a single application typically involves a few kilograms of material, several weeks of formulation and testing, and a second round to confirm. The material is rarely the largest cost; testing time is. Spending a little more on design up front saves a great deal on inconclusive repeat trials.
What to Ask Suppliers to Contribute
Good suppliers will support a well-designed trial with application data, recommended dispersion methods, and sometimes technical staff time. Ask for that support — and be wary of any supplier who prefers you not to run a cheaper-filler control.
This article is part of our Market & Business coverage. For material pricing, see Graphene Pricing: Why Costs Range from $5/kg to $100,000/m².