Formula For Rate Of Dissolution

Getting the Rate Right in Practice

The Formula For Rate Of Dissolution comes down to the Noyes-Whitney equation, and most people mess it up by treating it like pure theory rather than a practical tool you actually need to adapt to real formulations. Here is how it works when you are trying to figure out why your tablet just sits there in the flask instead of dissolving. The equation itself is straightforward: dC/dt = (D x A x (Cs - C)) / h. The rate of dissolution equals the diffusion coefficient multiplied by the surface area of the solid, divided by the thickness of the diffusion layer, all times the concentration gradient between saturation and bulk concentration. That is the formula. Now let us talk about what actually happens when you try to use it.

What the Formula For Rate Of Dissolution Actually Looks Like in the Lab

I spent three weeks tracking a BCS Class II compound that refused to dissolve above 4 mg/mL in standard phosphate buffer. The diffusion coefficient was around 3.2 x 10^-6 cm^2/s at 37 degrees Celsius. Particle size D90 was sitting at about 45 microns. The diffusion layer thickness, under USP apparatus 2 at 100 rpm, was roughly 0.03 cm using the typical estimate for paddle geometry. When I plugged those numbers in, the theoretical rate came out to about 0.12 mg/cm^2/min. The measured rate from the basket method was closer to 0.04 mg/cm^2/min. Three times slower than the model predicted. The gap was not experimental error. The compound was forming a gelatinous layer on the particle surface that acted as an additional diffusion barrier, and that is exactly the kind of thing the basic Noyes-Whitney equation does not account for at all. The formula assumes a clean, planar surface with a static diffusion layer. Real particles do not behave that way. What I ended up doing was switching to a surfactant-containing medium, which collapsed the gel layer and brought the measured rate within 15 percent of the theoretical prediction. That small change in the medium made the difference between a failed formulation and a working one.

Breaking Down Each Variable and Where Beginners Go Wrong

The diffusion coefficient D is usually the easiest variable to get wrong because people pull it from literature values that were measured under completely different conditions. Temperature matters a lot here. A 10 degree Celsius increase typically raises D by about 30 to 40 percent, and most dissolution tests run at 37 degrees but the literature values you find online are often from 25 degrees. If you do not adjust D, your calculated rate will be off by a significant margin. The Stokes-Einstein equation is the standard way to make that adjustment: D = kT / (6 pi eta r), where eta is the viscosity of the medium and r is the hydrodynamic radius of the diffusing species. Viscosity changes with temperature too, so you cannot ignore that. Surface area A is the variable that causes the most trouble in real formulations. The equation assumes a constant surface area, but that assumption falls apart almost immediately once dissolution starts. As particles shrink, A decreases proportionally. For spherical particles, A at any point is proportional to the two-thirds power of the remaining mass. If you are modeling a full dissolution profile, you need to account for this change over time rather than plugging in an initial surface area and hoping for the best. I have seen people use the initial surface area for the entire duration and then wonder why the predicted curve overshoots the early data points and undershoots the later ones. That is exactly what happens when A is treated as constant. The saturation solubility Cs is the other variable that gets overlooked. It is not a fixed number. It changes with pH, ionic strength, temperature, and the presence of cosolvents or surfactants. For weak acids and bases in particular, the apparent solubility can shift dramatically depending on the medium pH relative to the compound pKa. A compound with a pKa of 4.5 will have a very different Cs in pH 1.2 buffer than in pH 6.8 buffer, and that difference propagates directly into the dissolution rate. I once had a case where the initial Cs value used in the calculation was off by a factor of three simply because the measurement was done in a different buffer system than the one used in the actual dissolution test. The resulting prediction was useless.

The diffusion layer thickness h is tricky because it is not directly measurable in most standard dissolution setups. It depends on agitation speed, vessel geometry, and the physical properties of the medium. The common approximation for USP paddle apparatus at 100 rpm is about 0.02 to 0.05 cm, but if you change the paddle speed to 50 rpm, h increases by roughly 40 percent because the turbulence at the solid-liquid interface decreases. If you change to a smaller vessel or a different geometry like a flow-through cell, the h value is completely different. Using a literature h value from a different apparatus without adjusting for your actual setup is one of the most common sources of error I see.

Non-Sink Conditions and Why the Standard Equation Breaks Down

The Noyes-Whitney equation assumes sink conditions, meaning C remains close to zero relative to Cs throughout the dissolution process. That assumption holds well when the dissolution medium volume is large enough or the compound is highly soluble. But in biorelevant media or when you are working with poorly soluble compounds, C can approach Cs quickly, and the driving force (Cs - C) collapses. When C reaches even 70 percent of Cs, the dissolution rate drops by more than half. The equation still technically applies, but the predictions become much less useful because the model does not account for the feedback of increasing concentration on the diffusion layer itself. Under non-sink conditions, you need to solve the differential equation numerically rather than relying on the simplified integrated form. The integrated equation dM/dt = (D x A x Cs / h) is only valid when C is negligible. If you are running a test with only 250 mL of medium and the dose is 200 mg of a compound with Cs of 0.5 mg/mL, you are clearly not in sink conditions. The medium becomes saturated well before the dose is fully dissolved, and the dissolution curve will plateau early. This is not a measurement artifact. It is what the equation predicts when you include the C term properly. I learned this the hard way during a solubility optimization study where I kept seeing unexpected plateaus in the dissolution profile and initially blamed it on the apparatus rather than realizing the medium simply could not hold the drug.

Matrix Systems and Irregular Particles

The Noyes-Whitney equation was derived for simple solids dissolving in a clear fluid. It does not handle matrix tablets, lipid systems, or particles with highly irregular shapes very well. In a matrix tablet, the dissolution surface recedes inward in a way that does not follow simple geometric rules, and the surface area may actually increase over time as the polymer network erodes. I worked on a project with a hydrophilic matrix system where the surface area peaked at around 40 percent completion and then declined, which is the opposite of what the standard equation predicts for any simple particle shape. The model failed to capture the initial rate acceleration entirely. For irregular particles, the effective surface area is always larger than the geometric surface area calculated from an equivalent spherical diameter. The difference can be substantial. Sphericity values for pharmaceutical powders typically range from 0.6 to 0.9, and a sphericity of 0.6 can mean the actual surface area is nearly twice what you would calculate from the volume-equivalent sphere. If you are using laser diffraction data to estimate surface area, make sure you know whether the instrument output is based on volume-equivalent diameter or surface-equivalent diameter. They are not the same, and confusing them will throw off your A value significantly.

When the Formula Works and When You Should Move On

The Noyes-Whitney based approach is reliable for freely soluble compounds in simple formulations under sink conditions with reasonably spherical particles and well-controlled hydrodynamics. Under those circumstances, you can predict relative dissolution rates within about 20 percent using literature values for D and a reasonable estimate for h. That level of accuracy is usually sufficient for initial formulation screening. But if you are dealing with a poorly soluble compound in non-sink conditions, a complex matrix system, an irregular particle morphology, or a formulation where the diffusion layer is modified by excipients or precipitation, the basic formula will give you answers that are qualitatively reasonable but quantitatively unreliable. In those cases, the standard workaround is to shift to an empirical or semi-empirical approach. You can use the van't Hoff equation to model the temperature dependence of solubility, apply the Higuchi model for matrix systems, or simply rely on direct experimental measurement with a design of experiments approach rather than trying to predict everything from first principles. The formula is a starting point, not a finish line. I have found that spending an afternoon running a small set of dissolution experiments at different agitation speeds and temperatures usually provides more actionable information than a week of calculated predictions based on estimated parameters. Temperature control is another area where small deviations matter more than people expect. The USP requires 37.0 +/- 0.5 degrees Celsius, but even a deviation of 1 degree can change the dissolution rate by 5 to 10 percent for many compounds. That is because both D and Cs are temperature-dependent, and their combined effect can push the rate in one direction or the other depending on the thermodynamics of dissolution. If your lab thermostat is not calibrated regularly, your data will have unexplained batch-to-batch variation that looks like formulation inconsistency when it is actually just a temperature drift issue.

Media composition is the final variable that deserves attention. The presence of surfactants like SDS or polysorbate 80 increases the apparent solubility of lipophilic compounds and can also reduce the effective diffusion layer thickness by altering the interfacial tension. I have seen formulations where adding just 0.5 percent w/v of a surfactant to the dissolution medium increased the initial dissolution rate by a factor of four, not because the intrinsic solubility changed by four times, but because the surfactant disrupted the stagnant layer and improved wetting simultaneously. Without understanding which mechanism was dominant, it is easy to misinterpret the results and draw the wrong conclusion about the formulation itself.