Virtual Dissolution Lab
Run a six-vessel USP apparatus in your browser. Change the paddle speed, medium, temperature and dosage form, introduce the faults that happen on a real bench, and watch the profile and the stage verdict move.
- Apparatus 1 and 2
- 8 fault scenarios
- USP <711> verdict
- Printable report
Dissolution simulation report
Dissolution profile
Analysed results
Verdict at the Q time point
This is a teaching simulation with illustrative release kinetics. It is not a model of any real product and must not be used for batch disposition, method development or any regulatory purpose. Confirm the current acceptance criteria in USP <711> or your own monograph under your own licence.
Eight ways a dissolution run goes wrong
Pick one from the Introduce a fault list and run the test again with everything else unchanged. The point is to learn what each failure looks like on a profile, because that is how you recognise it on a real bench, where nobody tells you a fault is present.
- Speed too high. The profile rises early and the batch looks better than it is. Compare a 50 rpm run with a 75 rpm run.
- Bath too cold. Everything drops together. Near a tight Q this alone can move you from S1 to S2.
- Coning. Release stalls on a plateau below Q while variability climbs — the single most common cause of a false dissolution failure.
- Air bubbles. Results scatter with no pattern and run slightly low.
- Floating unit. Slow, erratic and vessel-specific. The reason sinkers are specified.
- Late sampling. Everything reads high. It flatters a slow batch, which is why actual pull times get recorded.
- Wrong volume in the calculation. A uniform low bias with no visible scatter — invisible in the data and only caught by checking the arithmetic.
- Vibration. The mean looks fine while the %RSD fails. Qualify the bench, not just the instrument.
Run each one, print the report, and keep them side by side. That set of eight profiles is a better training aid for a new analyst than any written SOP.
The model behind the vessels
Each vessel follows a Weibull release curve, the function most commonly fitted to real dissolution data:
Td is the time scale, b the curve shape and the lag the delay before release begins. Your settings move those parameters the way they move on a real apparatus: higher rotation speed shortens Td, lower temperature lengthens it, a medium in which the drug is poorly soluble stretches it further, and a capsule without a sinker both slows and destabilises it. Each of the six vessels gets its own small random variation, so the run has realistic spread rather than six identical curves.
The faults act on the same parameters, which is why they produce the distinct shapes a trained analyst learns to recognise: coning caps the plateau, vibration widens the scatter, a wrong volume shifts every result by a constant factor.
Acceptance staging
At the Q time point the six units are judged the way an immediate-release product is judged: S1 is met when every unit reaches at least Q + 5%. If it is not, the test moves to S2, where twelve units must average at least Q with none below Q − 15%, and then to S3. The verdict panel tells you which stage your run reached and why. The full criteria are in USP <711> — read them under your own licence, and use your own monograph limits for modified-release products.
When you have a profile you like, use Copy profile for the calculators and paste it straight into the f2 similarity calculator to compare it against a reference batch.
Training your analysts on dissolution?
We run this as a live seminar, with your own products and your own failures rather than a simulation.