Скринінг солей за профілем розчинності залежно від pH — pKa determination, counterion screening, pH-solubility profiling and disproportionation risk for CMC salt selection
Every rational salt-selection campaign begins with a precise measurement of the ionizable group's pKa and the intrinsic solubility of the un-ionized species. These two numbers parameterize every solubility model used downstream — without them, pHmax cannot be calculated or even sanity-checked, and salt screening becomes empirical guesswork.
The pKa of the ionizable nitrogen is measured by potentiometric acid-base titration on a Sirius T3 (or legacy GLpKa) analyzer:
• A known mass of compound is dissolved in a low-ionic-strength aqueous medium (0.15 M KCl) and titrated with standardized 0.5 M HCl / KOH while pH is continuously logged • The Yasuda-Shedlovsky extrapolation is used when the compound is too insoluble to titrate in pure water — titrations are run in increasing methanol:water cosolvent ratios (typically 30/50/70% w/w) and the apparent pKa values are extrapolated back to 0% cosolvent • For very poorly soluble bases (S0 <10 µg/mL), UV-metric titration is preferred: molar absorptivity changes with protonation state are tracked spectrophotometrically as pH is ramped, avoiding the need for the compound to stay in true solution at the titration endpoint • Typical precision: ±0.03 pKa units with 3 replicate titrations
For the model API used in this workflow — a weakly basic tertiary-amine-containing candidate — potentiometric titration returns pKa(BH⁺) = 7.4 ± 0.05, consistent with a moderately basic aliphatic amine environment (compare: piperidine pKa ~11.1, aniline pKa ~4.6 — ring substitution and electron-withdrawing groups shift this considerably).
A Bjerrum (distribution) plot shows the fraction of each ionization state as a continuous function of pH, calculated from the Henderson-Hasselbalch relationship:
Fraction ionized (BH⁺) = 1 / (1 + 10^(pH − pKa)) Fraction neutral (B) = 1 − Fraction ionized
At pH = pKa (7.4), the species is exactly 50:50 BH⁺:B. Two pH units below pKa (pH 5.4), the compound is >99% protonated (BH⁺, the water-soluble salt-forming species); two pH units above pKa (pH 9.4), it is >99% neutral free base — the poorly soluble form that will precipitate once solution pH rises past this window.
This speciation curve is the mechanistic explanation for why every basic-drug salt shows a solubility plateau at low pH (fully ionized, common-ion-limited by the counterion) and a sharp solubility collapse at higher pH (progressively neutral, intrinsic-solubility-limited) — the transition point between these two regimes is pHmax, developed fully in Stage 3.
Shake-flask solubility of the isolated free base is measured in pH 9–10 borate buffer (fully un-ionized region) with 24–72 h equilibration at 25°C and 37°C, residual solid confirmed by XRPD to remain the starting crystal form (no form conversion during the experiment), and supernatant assayed by HPLC-UV against a calibration curve.
Measured intrinsic solubility: S0 = 0.008 mg/mL (8 µg/mL) at 25°C — this places the free base in the BCS Class II/IV "poorly soluble" regime (USP/BCS boundary is commonly cited near 0.1–1 mg/mL depending on dose, using the dose:solubility ratio criterion). This low S0 is precisely the motivation for salt-form development: an appropriately chosen counterion can raise apparent solubility 1,000- to 8,000-fold within the GI-relevant pH window, as shown in later stages.
S0 and pKa are the two fixed physicochemical constants that anchor every pH-solubility curve in this workflow — a counterion only shifts where the curve bends (pHmax); it cannot change the plateau value the curve ultimately falls back to at high pH.
With pKa and S0 in hand, a systematic salt screen challenges the free base against a panel of pharmaceutically acceptable acids spanning a range of pKa and molecular geometry. The goal is not simply "does a salt form" but identifying candidates that crystallize cleanly, reproducibly, and without problematic hydrate/solvate behavior.
Counterion selection is constrained to acids with regulatory precedent, per the FDA Inactive Ingredient Database (IID) and historical use in approved NDAs (the "GRAS-adjacent" pharmaceutical salt list). The panel used here spans strong mineral acids to weaker organic diacids, deliberately covering a pKa(acid) range so the resulting salts bracket a wide pHmax range:
• Hydrochloric acid (pKa −6.3, strong) — historically the most common salt-forming acid (~40% of all marketed basic-drug salts) • Methanesulfonic acid / mesylate (pKa −1.9) — strong sulfonic acid, generally high solubility, some genotoxic-impurity (alkyl mesylate) manufacturing controls required • Benzenesulfonic acid / besylate (pKa −2.8) • p-Toluenesulfonic acid / tosylate (pKa −2.8) • Fumaric acid (pKa1 3.0, pKa2 4.4) — dicarboxylic, low toxicity, good crystallinity track record • Succinic acid (pKa1 4.2, pKa2 5.6) • Maleic acid (pKa1 1.9, pKa2 6.1) — historically deprioritized due to maleate metabolic/renal toxicity signals in some series • Citric acid (pKa1 3.1, pKa2 4.8, pKa3 6.4) — triprotic, often hygroscopic • Phosphoric acid (pKa1 2.1)
Each counterion is combined with the free base at a stoichiometric (typically 1:1, occasionally 1:2 for diacids) ratio in solution, and salt formation is attempted by four complementary crystallization routes run in parallel 2–4 mL vials:
1. Cooling crystallization: dissolve both components at 50–60°C in a compendial solvent (EtOH, IPA, acetone/water mixtures), cool to 5°C at 0.1°C/min 2. Anti-solvent addition: dissolve in a good solvent (MeOH), add a miscible anti-solvent (MTBE, heptane) dropwise to induce supersaturation 3. Slurry conversion: suspend both components as solids in a solvent with limited solubility, agitate 48–72 h at ambient temperature (thermodynamically favors the most stable crystal form) 4. Slow evaporation: solvent removed at controlled rate over 3–7 days at ambient temperature
Every isolated solid is triaged by: • XRPD (X-ray powder diffraction) — confirms new crystalline phase distinct from free base and from unreacted counterion; unique diffraction pattern = new salt • DSC (differential scanning calorimetry) — single sharp endotherm consistent with a defined melting point supports a stoichiometric, non-solvated crystal form • TGA (thermogravimetric analysis) — mass loss below 150°C flags solvate/hydrate character requiring further characterization • 1H-NMR — confirms correct salt stoichiometry (integration of counterion vs API protons)
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Hydrochloride | 1:1 stoichiometry | Sharp XRPD, mp 214°C (DSC), anhydrous | Very high aqueous solubility |
| Mesylate | 1:1 stoichiometry | Sharp XRPD, mp 198°C, anhydrous | High solubility, good crystallinity |
| Besylate | 1:1 stoichiometry | Broad XRPD halo — poorly crystalline | Deprioritized (amorphous tendency) |
| Tosylate | 1:1 stoichiometry | Crystalline but forms a hemihydrate | Hydrate stoichiometry risk |
| Fumarate | 1:1 stoichiometry | Sharp XRPD, mp 187°C, anhydrous | Balanced solubility/stability |
| Succinate | 2:1 (API:acid) | Sharp XRPD, mp 176°C, anhydrous | Lower toxicity dicarboxylic acid |
| Maleate | 1:1 stoichiometry | Crystalline, mp 191°C | Deprioritized (toxicity precedent) |
| Citrate | 1:1 stoichiometry | Crystalline but markedly hygroscopic | Deprioritized (moisture uptake) |
The single most decision-critical measurement in salt selection is the full pH-solubility profile of each crystalline salt candidate, because oral absorption occurs across a dynamic pH range (gastric pH ~1–3, rising to duodenal/jejunal pH ~5–7.5) that a salt form must survive without precipitating back to the poorly soluble free base.
Excess solid salt (confirmed crystalline by XRPD before and after the experiment) is added to a series of USP/Ph.Eur. compendial buffers spanning pH 1.0 to 9.0 in 0.5–1.0 pH-unit increments (HCl/KCl pH 1–2, acetate pH 3.5–5.5, phosphate pH 6–8, borate pH 9). Suspensions are agitated at 25°C and separately at 37°C (biorelevant temperature) for 24–72 h until equilibrium is confirmed by two consecutive timepoints agreeing within 5%.
Supernatant is separated by centrifugation/filtration (0.45 µm PVDF, verified for non-adsorption), diluted as needed, and quantified by HPLC-UV against an external calibration curve (typical LOQ <1 µg/mL). Residual solid is re-examined by XRPD — critically, this confirms whether the solid phase in equilibrium with the supernatant is still the intended salt, or has converted to free base (disproportionation, addressed in Stage 4) or a hydrate.
Every basic-drug salt pH-solubility curve has the same characteristic two-regime shape:
• Below pHmax ("salt-limited" / plateau regime): the solid phase in equilibrium is the salt (BH⁺A⁻); solubility is approximately constant and set by the salt's own solubility product, largely independent of pH • Above pHmax ("base-limited" regime): the solid phase in equilibrium converts to the free base (B); solubility falls off steeply, approaching the free base's intrinsic solubility S0 asymptotically at high pH
pHmax is derived (Bogardus & Blackwood 1979; formalized for salt selection by Serajuddin, Adv. Drug Deliv. Rev. 2007) from equating the two solubility expressions at the transition point:
pHmax = pKa + log10[ (S0salt − S0) / S0 ]
where S0salt is the intrinsic solubility of the salt form's "plateau" and S0 is the free base intrinsic solubility. In practice, pHmax is read directly off the empirical curve as the inflection between the plateau and the descending limb — the equation above simply explains *why* different counterions and different S0salt values shift pHmax to different positions.
A higher pHmax means the salt maintains its high-solubility plateau over a wider portion of the physiological pH range before reverting toward free-base-limited solubility — directly favoring more consistent oral absorption, especially for compounds absorbed in the higher-pH small intestine.
Measured pHmax values for the four crystalline candidates: HCl salt pHmax ≈ 4.2, succinate ≈ 4.6, fumarate ≈ 5.1, mesylate ≈ 5.8. The mesylate salt maintains its high-solubility plateau furthest into the physiologically relevant duodenal pH range (5–7.5), a strong early signal in its favor.
Paradoxically, the HCl salt — frequently the single most soluble salt form in a simple shake-flask test at low pH — often shows the lowest pHmax of the series. This is because Cl⁻ is both a strong, fully dissociating counterion (driving a very high plateau solubility, here ~38 mg/mL) and the physiologically ubiquitous anion already present at high concentration in gastric and intestinal fluid (Stage 4 covers the resulting common-ion suppression). The practical consequence: the HCl salt's plateau, while numerically the highest of all candidates, collapses at the lowest pH (pHmax 4.2) — meaning it reverts to free-base-limited solubility earliest as the drug transits from stomach to intestine, precisely where absorption of many BCS II compounds is most sensitive to solubility loss.
Two related but distinct risks must be assessed before a salt can be recommended: solution-phase suppression of solubility by a common ion already present in the GI environment or formulation matrix, and solid-state disproportionation — the salt slowly reverting to free base during storage due to local pH/moisture conditions inside a tablet or capsule.
A sparingly soluble salt BH⁺Cl⁻ exists in solution equilibrium with its constituent ions:
BH⁺Cl⁻(s) ⇌ BH⁺(aq) + Cl⁻(aq), Ksp = [BH⁺][Cl⁻]
For the HCl salt candidate, the apparent Ksp in water at 25°C is approximately 1.4×10⁻³ M². In pure water, [BH⁺] = [Cl⁻] = √Ksp ≈ 37 mM, consistent with the ~38 mg/mL plateau solubility measured in Stage 3. But gastric fluid already contains ~100 mM chloride from HCl secretion and dietary intake — by Le Chatelier's principle, this common ion shifts the equilibrium and suppresses further dissolution:
[BH⁺] = Ksp / [Cl⁻]total ≈ 1.4×10⁻³ / 0.1 ≈ 14 mM (≈ 8 mg/mL as free base equivalent)
This is a >4-fold reduction from the "clean buffer" plateau value — a purely solution-chemistry effect, entirely separate from the pHmax phenomenon in Stage 3, and it disproportionately penalizes salts whose counterion (Cl⁻, and to a lesser extent SO4²⁻ from sulfate salts) is already abundant in physiological fluids or common excipients (e.g., magnesium stearate, croscarmellose sodium formulations with chloride-containing processing residues).
Disproportionation is the process by which a crystalline salt, in the solid state (typically within a tablet or capsule blend), reacts with moisture and local excipient-derived pH to convert back to the free base plus a soluble counterion salt (e.g., sodium chloride, if a basic sodium-containing excipient is present):
BH⁺A⁻(s) + excipient-OH(basic microenvironment) → B(s) + H2O + A⁻(excipient⁺)
Because the free base is far less soluble (here 0.008 mg/mL vs a 14–64 mg/mL salt plateau), even partial disproportionation during shelf life can measurably slow dissolution and reduce bioavailability — a failure mode that will not be caught by a standard assay/purity stability program, only by a dissolution or solid-state (XRPD) stability program.
Risk assessment protocol: • Slurry disproportionation experiments: suspend the salt in water or biorelevant media (FaSSGF/FaSSIF) for 3–4 weeks at room temperature with periodic XRPD sampling — appearance of free-base diffraction peaks signals a disproportionation-prone salt • Microenvironmental pH (pHM) measurement: a flat-surface pH electrode pressed against a compacted powder blend or intact tablet measures the local pH the salt actually experiences, which can differ by 2–3 units from the bulk aqueous pH due to excipient buffering (e.g., magnesium stearate raises local pH; the disintegrant/filler blend matters) • Accelerated stability (40°C/75% RH, open dish) with XRPD and dissolution monitoring at 1, 3, and 6 months
A salt is considered lower-risk for disproportionation when its pHmax lies comfortably below the microenvironmental pH generated by the intended excipient blend — typically a margin of ≥1–2 pH units is targeted.
Basic excipients are a recurring, underappreciated driver of disproportionation: magnesium stearate slurries can reach local pH 9–10. Formulators of high-disproportionation-risk salts frequently substitute sodium stearyl fumarate (lower local pH impact) or reduce magnesium stearate level, and avoid alkaline superdisintegrants in direct contact with the salt.
No single measurement — not even pHmax — determines the final salt form on its own. A defensible CMC recommendation integrates solubility/pHmax with manufacturability, hygroscopicity, disproportionation margin, counterion regulatory precedent, and cost, then documents the trade-offs explicitly for the regulatory filing.
Serajuddin's widely cited framework (Adv. Drug Deliv. Rev. 2007, "Salt formation to improve drug solubility") formalizes salt selection as a weighted assessment across:
• Solubility/pHmax — favors mesylate (pHmax 5.8) and fumarate (pHmax 5.1) over HCl (4.2) and succinate (4.6) for maintaining solubility across the GI pH range • Manufacturability — crystallinity, filterability, and flow properties; all four crystalline finalists (HCl, mesylate, fumarate, succinate) passed this gate in Stage 2, while besylate, tosylate, maleate, and citrate were deprioritized • Hygroscopicity — DVS profiling (a dedicated downstream workflow) further ranks the finalists; sulfonate salts like mesylate are sometimes, though not universally, more hygroscopic than carboxylate salts • Disproportionation margin — mesylate's pHmax (5.8) sits furthest below typical excipient microenvironmental pH, giving the largest safety margin before Stage 4 risk manifests • Counterion precedent/toxicity — mesylate has extensive FDA Inactive Ingredient Database precedent (>15 approved NDAs use a mesylate salt), though genotoxic alkyl mesylate impurity control (ICH M7) must be built into the synthetic route; fumarate and succinate carry essentially no toxicity concerns as endogenous metabolic intermediates (Krebs cycle) • Cost and IP — mesylic acid is inexpensive and unpatented as a counterion; a novel salt form itself may still support secondary patent claims
The pHmax concept is not academic — several marketed drugs illustrate its clinical consequences directly:
• Ritonavir: notoriously underwent a late-stage polymorphic transition (Form II) that reduced solubility and forced product reformulation/recall in 1998, an object lesson in why exhaustive solid-form and solubility characterization is performed before NDA filing, not after • Haloperidol: marketed as both free base (poorly soluble, used in depot/oil formulations) and lactate salt (aqueous-soluble, used for oral/parenteral solutions) — a direct illustration of salt form being chosen to match the intended route and pH environment of administration • Ziprasidone: developed as the mesylate salt specifically to achieve adequate aqueous solubility for an intramuscular formulation, after the free base and several other salt candidates failed to reach the required concentration — a close real-world analog to the workflow in this simulation
These cases underscore that pHmax and salt-form solubility are not abstract CMC exercises — they directly determine whether a formulation can achieve its target exposure, and errors discovered late in development are extremely costly to correct.
Integrating all five stages, the mesylate salt is recommended as the lead development salt form:
• Highest pHmax (5.8) of the crystalline candidates, maintaining salt-plateau solubility across nearly the entire physiologically relevant GI pH range (1–7.5) • Confirmed crystalline, anhydrous, single sharp DSC endotherm at 198°C — favorable manufacturability • Largest disproportionation margin relative to typical tablet-blend microenvironmental pH • Strong regulatory precedent via FDA IID; established ICH M7 control strategy exists industry-wide for mesylate ester genotoxic impurities
Fumarate is retained as a backup salt form (pHmax 5.1, excellent toxicological profile as an endogenous metabolite) should hygroscopicity or long-term stability data in later DVS/stability workflows disfavor the mesylate.
The core lesson of pH-solubility profiling: the salt with the highest solubility in a single buffer is not necessarily the best development candidate. pHmax — the pH at which that high solubility is lost — is frequently the more decision-relevant number, because it predicts behavior across the dynamic, multi-compartment pH environment the dosage form actually experiences in vivo.