Mass-Conserving Dissolution, Precipitation, and Redissolution

Why Track Each Drug Pool?

A concentration-time curve alone does not say where drug went. A falling dissolved concentration could reflect precipitation, absorption, or another transfer out of the measured fluid. A later rise could reflect redissolution, continued dissolution of the administered solid, or a change in fluid volume. A useful dissolution–precipitation model therefore tracks the drug amount in each pool and specifies the transfers between them.

The central requirement is accounting: dissolution moves drug from an undissolved dose into solution; precipitation moves dissolved drug into a solid precipitate; and redissolution moves precipitated drug back into solution. None of those transfers creates or destroys drug. Absorption moves drug out of the local fluid compartment, but it can be represented as a transfer into an absorbed pool so that the overall balance remains explicit.

The model below assumes a single homogeneous fluid compartment, a fixed volume, and a specified solubility over the time interval considered. Concentration refers to dissolved drug available to precipitate or redissolve. The rate laws must use compatible units: for example, if amounts are in mg, volume in L, and time in min, rates are in mg/min and concentrations are in mg/L. The treatment does not require a particular nucleation-and-growth law. Let \(P\) denote the precipitation rate supplied by the chosen precipitation model; it is a nonnegative amount per time.

Coupled Transfer Equations

Let \(D\) be the dissolution rate from the administered solid, \(R\) the redissolution rate from precipitated solid, and \(J_{\mathrm{abs}}\) the rate of transfer from the fluid compartment into the absorbed pool. Use nonnegative rates. A basic direction rule is that dissolution and redissolution require undersaturation, whereas precipitation requires supersaturation:

Here \(k_d\) and \(k_r\) are mass-transfer coefficients, while \(A_u\) and \(A_p\) are the available surface areas of the administered solid and precipitate. With concentration in mg/L and area in cm², each coefficient has units L/(min·cm²) when rates are in mg/min. Surface area can depend on the remaining solid amount and particle-size representation. The earlier material on particle-size distributions and dissolving surface area provides ways to specify that dependence. The rate is zero when its solid pool is empty. For precipitation, \(P=0\) when \(C\leq S\); when \(C>S\), its magnitude can depend on supersaturation and on the nucleation and growth state.

These equations state where every local change comes from. In particular, precipitation is subtracted from the dissolved pool and added to the precipitated pool using the same rate \(P\). Redissolution is subtracted from the precipitate and added to the dissolved pool using the same rate \(R\). Using a different rate in each equation for the same transfer breaks mass conservation.

For a simple absorption representation, take \(J_{\mathrm{abs}}=k_aM_d\), where \(k_a\) has units min\(^{-1}\). Define \(M_{\mathrm{abs}}\) as cumulative absorbed amount, with \(dM_{\mathrm{abs}}/dt=J_{\mathrm{abs}}\). Absorption is then a loss from the local fluid compartment but not a loss from the tracked total. Other exits, such as fluid transit out of the compartment, can be treated in the same way by adding an outflow pool or explicitly defining a local balance that excludes that pool.

Why the Equations Conserve Mass

The key justification is to add the equations for all pools through which the drug is tracked. For a closed fluid compartment with no absorption, the total local amount is \(M_u+M_d+M_p\). With absorption included, add \(M_{\mathrm{abs}}\) as well.

Each internal transfer cancels because it leaves one pool at the same rate it enters another. Thus, provided no untracked source or sink is present, the total equals its initial value at every time. In a closed system without absorption, the same argument gives \(d(M_u+M_d+M_p)/dt=0\). This is a structural property of the equations, not a consequence of choosing particular numerical rate constants.

Supersaturation and Redissolution

The sign of \(C-S\) determines which direction is thermodynamically favored in this basic model. At \(C>S\), precipitation can remove dissolved drug; at \(C<S\), existing solids can dissolve. At \(C=S\), the concentration-driven terms above are zero. A precipitation model may also include induction time or state variables describing nuclei and crystal growth, but any precipitation amount removed from solution must still be added to \(M_p\). The earlier material on nucleation and growth supplies possible descriptions for \(P\); the mass balance is independent of that choice.

Absorption can occur at the same time as precipitation. In that case, dissolved drug may decrease both because it enters the precipitate and because it leaves the local fluid. Separating these flows is important: only the first is an internal transfer among local pools, while the second is a local exit and an input to the absorbed pool.

Implementation Checks and Common Failure Modes

A model can have plausible-looking concentration curves while violating its accounting. Check the amount balance as well as the concentration profile. For a simulation with absorption, compute the residual \(E(t)=M_u(t)+M_d(t)+M_p(t)+M_{\mathrm{abs}}(t)-M_{\mathrm{total}}(0)\). In a closed model, this should be zero analytically and close to numerical precision in computation. With a numerical solver, inspect the residual across the full run and repeat with tighter tolerances or smaller time steps if it grows.

• Negative solid or dissolved amounts: An explicit numerical step can remove more mass than a pool contains. Reduce the time step, use a positivity-preserving method, or limit the transfer over the step to the available amount. Do not silently set a negative amount to zero without accounting for the mass changed by that correction.
• Precipitation during undersaturation: Applying a positive precipitation rate when \(C<S\) contradicts the direction rule of this model. Check the concentration units, solubility units, and supersaturation sign.
• Precipitate disappears but keeps redissolving: Set \(R=0\) when \(M_p=0\). Otherwise the differential equation can drive \(M_p\) negative.
• Unexplained loss from solution: A concentration decrease is not automatically precipitation. Check whether absorption or another exit was included and whether its amount is tracked consistently.
• Mass residual grows with step size: This often signals an implementation error or a transfer step too large for the numerical method. Equal-and-opposite rates should cancel in the equations even when their values vary over time.

Choosing a Model Structure

Explicit amount tracking is not the only possible representation, but alternatives answer different questions and have different limitations.

An equilibrium clamp can be useful as an idealized limiting assumption when equilibration is treated as instantaneous. It should not be confused with a kinetic model: it removes the time course of precipitation and may mask induction or delayed redissolution. A net-loss term can fit a dissolved profile in some settings, but without \(M_p\) it cannot represent solid accumulation and subsequent return to solution. Explicit pools are therefore appropriate when the timing and amount of redissolution matter.

The basic formulation also has boundaries. A fixed \(V\) is an assumption; if volume changes, compute \(C=M_d/V(t)\) and account for dilution or concentration through the changing volume rather than treating it as a drug-mass transfer. Likewise, solubility may depend on conditions such as pH or temperature, and the surface areas and rate laws may be uncertain. Those refinements change the rates or concentration calculation, not the requirement that every transfer have matching source and destination terms.

Common Mistake: Treating Precipitation as Destruction

A tempting but incorrect equation is \(dM_d/dt=-P\), with no corresponding equation for a precipitate pool. It may be described as “drug lost by precipitation,” but that wording incorrectly treats precipitation as drug destruction. If the model later needs to represent redissolution, it has no record of how much solid is available to return to solution.

The correction is to write \(dM_d/dt=\cdots-P\) and \(dM_p/dt=P-R\), then add both equations when checking the total. Precipitation is a transfer, not disappearance. Absorption may be a local exit, but if total administered drug is being balanced, its amount must likewise be counted in an absorbed pool.