Ostwald's rule of stages, solubility & DSC ranking, and competitive slurry bridging for pharmaceutical polymorph selection
In 1897, Wilhelm Ostwald observed that a system relaxing from a metastable state does not jump directly to the most stable state — it passes through a sequence of intermediate states, each only marginally more stable than the last. Applied to pharmaceutical crystallization, this "rule of stages" predicts that a rapidly cooled or rapidly antisolvent-precipitated supersaturated solution will nucleate the least stable, most soluble polymorph first, even though a thermodynamically more stable (less soluble) form exists and is available to the system.
Classical nucleation theory frames polymorph selection as a competition between bulk free-energy gain and interfacial energy cost:
ΔG* = 16πγ³Vm² / (3(RT ln S)²)
Where γ = solid–liquid interfacial energy, Vm = molar volume, S = supersaturation ratio.
The Ostwald–Volmer rule notes that structurally disordered, high-energy metastable polymorphs typically present a LOWER interfacial energy γ toward the mother liquor than the tightly-packed stable form — because their looser lattice more closely resembles the disordered solution structure. A lower γ means a lower nucleation barrier ΔG*, so nuclei of the metastable form appear first even though they are thermodynamically less favorable once formed.
Practical signature in this system: • Form III (a transient hydrate-like phase) nucleates within 2–5 minutes of quench to 4°C supersaturation ratio S=2.3 • Form I (needles) appears at 15–20 minutes, growing at the expense of Form III • Form II (stable plates) is not observed directly from solution — it only appears after 6–18 hours via solid-state or solution-mediated conversion of Form I
This staged appearance is diagnostic: whenever a crystallization screen shows an early transient phase that vanishes on standing, Ostwald's rule — not experimental error — is usually the explanation.
Time-resolved detection of sequential polymorph appearance requires methods fast enough to catch short-lived intermediates:
• In-situ Raman spectroscopy (PAT probe in the crystallizer): 30 second acquisition, tracks characteristic lattice phonon bands for each form without sampling • Synchrotron time-resolved PXRD: sub-second exposures at a beamline capture Form III → Form I → Form II transitions in a single cooling run • Turbidity + focused-beam reflectance measurement (FBRM): chord-length distribution shifts flag nucleation and the subsequent Ostwald ripening step • Optical microscopy with hot stage: visually confirms needle habit (Form I) replacing granular Form III aggregates
Supersaturation ratio S = C/C* (actual concentration over equilibrium solubility of the referenced form) sets how many forms are seen: at low S (1.1–1.3) only the stable form nucleates (slow, form-selective); at high S (>2) all three forms compete kinetically and Ostwald's full staged sequence is observed.
Ostwald's rule is a kinetic statement, not a thermodynamic one — it never overrides the eventual, sometimes very slow, march toward the lowest free-energy form. In this API, Form III is undetectable after 24 hours at 25°C; Form I itself will keep slowly converting to Form II over weeks unless kinetically trapped by processing choices.
Once multiple polymorphs are isolated (by seeding, solvent, or temperature control), the CMC team must rank their relative stability quantitatively. This combines equilibrium solubility measurements in several solvents, van't Hoff extrapolation across temperature, and differential scanning calorimetry (DSC) of melting point and enthalpy of fusion — then applies the Burger–Ramberger heuristic rules to classify the pair as enantiotropic or monotropic.
Equilibrium solubility of each isolated, phase-pure form is measured by the shake-flask method: excess solid equilibrated with solvent for 72 h at controlled temperature (with PXRD confirmation the solid did not convert during equilibration), then HPLC-assayed supernatant gives C*(T).
van't Hoff plot: ln(x) = −ΔH_soln/R · (1/T) + ΔS_soln/R, where x is mole-fraction solubility.
Measured across 15–45°C in isopropanol: • Form I: ln(x) = −4210/T + 9.8 (R²=0.994) • Form II: ln(x) = −4850/T + 10.6 (R²=0.991)
At 25°C (298 K): x_I / x_II = 1.8 — Form I is 1.8-fold more soluble than Form II, consistent with Form I being the higher free-energy (metastable) polymorph.
Solubility ratio method (used as a fast pre-DSC screen): SR = C*_metastable / C*_stable • SR > 1 at all tested temperatures with no crossover in the measured range → suggests monotropic behavior across that window (need DSC to confirm the deeper thermodynamic picture) • SR that decreases toward 1 and would extrapolate to crossing at a real, sub-Tm temperature → suggests enantiotropy, with the crossover approximating the transition temperature Tt
DSC of each phase-pure form (10°C/min, crimped Al pan, N2 purge):
• Form I: Tm = 142°C, ΔHfus = 28 kJ/mol, sharp single endotherm • Form II: Tm = 156°C, ΔHfus = 34 kJ/mol, sharp single endotherm • Heating a Form I sample slowly (2°C/min) from 60°C reveals a small solid–solid exotherm near 68°C followed by re-melt at 156°C — evidence of an in-situ conversion to Form II before the Form II melting endotherm
Burger–Ramberger heat-of-fusion rule: if the higher-melting form has the LOWER heat of fusion, the pair is enantiotropic; if the higher-melting form has the HIGHER heat of fusion, the pair is monotropic. Here: Form II (Tm=156°C) has ΔHfus=34 kJ/mol > Form I's 28 kJ/mol — by the heat-of-fusion rule alone this borderline case favors monotropy, so the heat-of-transition rule is applied as the definitive test.
Burger–Ramberger heat-of-transition rule: if the I→II solid-state transition is ENDOTHERMIC, the pair is enantiotropic (a genuine transition temperature Tt exists below both melting points); if EXOTHERMIC, the pair is monotropic (Form I is unstable at all temperatures below its own Tm).
Microcalorimetry of the 68°C solid-state event shows a small ENDOTHERMIC signal (+1.9 kJ/mol) — confirming an enantiotropic system with calculated Tt ≈ 68°C, in close agreement with the solubility-ratio extrapolation.
Two independent rules (heat-of-fusion vs heat-of-transition) disagreed at first glance — a common real-world outcome. The heat-of-transition rule is considered mechanistically more direct because it measures the actual I→II conversion enthalpy rather than inferring it from the two melting events, so Tt=68°C enantiotropic is taken forward as the working model.
Solubility and DSC data predict which form should be more stable, but the definitive experimental proof is the competitive (bridging) slurry: a physical mixture of both forms is suspended in a solvent at a fixed temperature near saturation, and the system is left to equilibrate via solution-mediated transformation — the metastable form dissolves, the dissolved material re-precipitates onto the stable form, and only the stable polymorph remains at the end.
Protocol: a 1:1 w/w physical mixture of Form I and Form II (each independently confirmed phase-pure by PXRD) is suspended in isopropanol at a solid-to-solvent ratio that keeps the system just below saturation for the stable form but above saturation for the metastable form — this "bridge" condition is what drives net conversion.
Mechanism (solution-mediated transformation, an Ostwald-ripening-like process at the polymorph level): 1. Form I (higher solubility) partially dissolves, raising local solution concentration above C*_II 2. Dissolved API re-nucleates/grows onto existing Form II crystal surfaces (lower supersaturation barrier for growth on an existing lattice than fresh nucleation) 3. As Form II grows it consumes dissolved material, pulling concentration back below C*_I, which drives further Form I dissolution 4. The cycle self-sustains until Form I is exhausted — net result: complete polymorphic conversion without ever exceeding true equilibrium for the stable form
Tracked by PXRD peak intensity ratio (Form I 2θ=8.9° / Form II 2θ=11.4°) and by Raman C=O stretch shift (1682 cm⁻¹ Form I vs 1671 cm⁻¹ Form II): • Day 0: 50:50 mixture • Day 3: 71:29 (II:I) • Day 9: >98:2 — effectively complete • Day 21: no further change, confirms Form II is the thermodynamic sink at 25°C
To locate Tt experimentally (rather than relying only on the DSC/solubility extrapolation), parallel bridging slurries are run at several fixed temperatures spanning the predicted 68°C crossover:
• 40°C, 55°C: Form I fully converts to Form II within 5–7 days — confirms Form II is stable below Tt • 60°C: conversion to Form II still dominant but slower (14 days to >95%) • 76°C, 90°C: Form II converts to Form I instead — confirms Form I is the stable form above Tt • 68°C: near-stagnant slurry, both forms persist together for the full 21-day window (equal solubility, zero net driving force) — this is the operational definition of Tt
This temperature-bracketing bridging approach is considered the gold-standard experimental determination of Tt because it measures the real solid-state equilibrium directly, sidestepping any error accumulated in van't Hoff extrapolation or DSC baseline assignment.
A slurry that fails to convert after 21 days is ambiguous — it could mean the two forms are genuinely isoenergetic (at or near Tt) or simply that the kinetics are too slow at that temperature (common below ~10°C, where solubility and molecular mobility are both low). Extended slurries (60–90 days) or seeded slurries are used to break this ambiguity before declaring a form "trapped."
All of the solubility, DSC, and slurry-bridging data are consolidated into a single free-energy (G) versus temperature diagram — the standard tool for communicating polymorph risk to formulation, process, and regulatory teams. This diagram makes explicit whether the system is enantiotropic (a true crossover exists below the melting points, so relative stability depends on temperature) or monotropic (one form is more stable at every temperature up to its melting point).
The free-energy curve for each form is not measured directly — it is reconstructed from solubility (via ΔG = −RT ln(x_form/x_reference)) and calorimetric data, then extrapolated using the Gibbs–Helmholtz relation. Each curve is drawn with a slope proportional to −S(T) (entropy) and curvature set by heat capacity differences; melting points appear as the temperature where G_solid(T) = G_liquid(T).
Key features on this system's diagram: • Two G(T) curves for Form I and Form II cross exactly once, at Tt=68°C, well below both melting points (142°C and 156°C) — the single-crossover pattern is the definitional signature of an enantiotropic pair • Below 68°C: G_II(T) < G_I(T) → Form II is thermodynamically stable; Form I is metastable and will eventually convert • Above 68°C (and below Tm,I=142°C): G_I(T) < G_II(T) → Form I becomes the stable form • Because Tt < Tm for both forms, both forms can in principle be observed at their own melting points without an intervening solid-solid transition erasing the distinction — a classic enantiotropic diagnostic (vs. monotropic systems, where the metastable form's G(T) curve never dips below the stable form's curve at any temperature short of an unreachable, hypothetical extrapolated crossover above Tm)
The practical metastable zone for Form I at room temperature (25°C, well below Tt=68°C) is real: Form I is NOT the room-temperature thermodynamic form, but its conversion kinetics are slow enough (weeks, per Stage 3 slurry data) that it survives as a kinetically trapped, shelf-viable metastable polymorph if processing and storage conditions are controlled.
Turning the thermodynamic picture into a practical risk assessment requires overlaying kinetics:
• Solid-state conversion rate (Form I→II) roughly doubles per 8–10°C per an Arrhenius fit of the accelerated stability data (40°C/75%RH, 25°C/60%RH, 5°C/ambient) • At 25°C/60%RH: <1% conversion projected over 24 months (acceptable per ICH Q1A shelf-life criteria) • At 40°C/75%RH (accelerated stress): 15–20% conversion at 6 months — flags that hot/humid excursions during transport or tropical-zone distribution are the dominant real-world risk, not steady controlled storage • Moisture is a strong catalyst here: trace water mediates local dissolution/recrystallization even in a nominally solid dosage form, so packaging with desiccant and a moisture barrier blister is identified as a key risk control
The metastable zone width (MSZW) — the supersaturation range in which Form I can be produced and isolated without spontaneous Form II nucleation — was independently mapped during process development: MSZW ≈ ΔT of 8–12°C in isopropanol at the target concentration, setting the allowable cooling-rate and hold-time windows for reproducible manufacture of the desired form.
A common regulatory pitfall is treating "enantiotropic" as synonymous with "risky" and "monotropic" as synonymous with "safe." In practice an enantiotropic pair with Tt far above any realistic storage/processing temperature (here, 68°C vs. 25°C ambient) can be perfectly manufacturable and stable — the diagram, not the label, is what should drive the control strategy.
With the full stability ranking, transition temperature, and conversion kinetics in hand, the CMC team makes the final call on which polymorph advances into formulation and manufacturing. This is rarely a pure thermodynamics decision — the more soluble metastable form often gives better dissolution and absorption, so the choice balances bioavailability upside against documented conversion risk, and is locked in with analytical control per ICH Q6A.
The Ostwald-rule irony is that the polymorph the system WANTS to make first (Form I, or transiently Form III) is often also the one with the pharmaceutically desirable higher solubility and faster dissolution rate — a 1.8× solubility advantage for Form I can translate into materially higher Cmax and AUC for a poorly-soluble (BCS Class II) compound.
Options considered by the development team: 1. Develop Form II (thermodynamically stable): lower risk of uncontrolled in-process or shelf conversion, simpler long-term stability story, but must confirm dissolution/exposure is still adequate — sometimes compensated with particle size reduction (micronization) or a solubilizing formulation (surfactant, amorphous solid dispersion) rather than relying on a metastable crystal form for bioavailability 2. Develop Form I with tight control: exploit the solubility advantage, but requires demonstrating the metastable form is kinetically stable under ALL intended storage/handling conditions for the product shelf life, including manufacturing unit operations (wet granulation, milling, compression heat) that can seed conversion 3. Amorphous or co-crystal alternative: sidesteps the polymorph ranking question entirely, at the cost of a different (often harder) stabilization problem
In this case Form II was selected: the 1.8× solubility gap was judged formulation-addressable (particle size + surfactant in the dissolution medium closed most of the exposure gap in biorelevant media testing), while the alternative — carrying a metastable Form I with real conversion risk during 24-month ambient shelf life and warm-climate distribution — was judged the larger overall program risk.
Once a form is selected, ICH Q6A (Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and Products) requires a documented control strategy whenever polymorphism could affect quality, safety, or efficacy:
• Drug substance release specification: PXRD identity test confirming Form II pattern; limit test for Form I content (LOD/LOQ validated at 2%/0.5% w/w by spiked-standard PXRD or ssNMR) • Crystallization process control: Form II seed crystals (0.5% w/w) added at a defined supersaturation setpoint within the mapped MSZW, cooling profile validated to avoid entering the Form I nucleation zone • In-process controls: PAT (in-line Raman) during the final crystallization step confirms form identity before isolation, catching any Form I contamination before it reaches drying/milling • Stability protocol: long-term (25°C/60%RH, 36 months) and accelerated (40°C/75%RH, 6 months) stability batches include PXRD/DSC form-identity testing at each pull point, not just assay and degradation products • Drug product considerations: any high-shear or high-heat unit operation (wet granulation, hot-melt extrusion, compression) is separately verified not to induce Form I→(back-conversion) or amorphization, since processing stress can sometimes override the bulk thermodynamic preference locally
This full documentation package — form diagram, Tt determination, slurry bridging data, and the specification/control strategy — is what regulatory submissions (ANDA/NDA drug substance sections) expect to see whenever more than one polymorph of a drug substance is known to exist.
Famotidine, ritonavir, and rotigotine are textbook cases where an unanticipated late-appearing stable polymorph (Ostwald's rule playing out during commercial manufacturing, not R&D) forced product reformulation or, in ritonavir's case, a market withdrawal in 1998. Establishing the full stability ranking and control strategy BEFORE launch, as done here, is standard practice specifically to avoid repeating that failure.