David J. W. Grant

MODIFICATION OF ACETAMINOPHEN CRYSTALS: INFLUENCE OF GROWTH IN AQUEOUS SOLUTIONS CONTAINING P-ACETOXYACETANILIDE ON CRYSTAL PROPERTIES

Abstract Acetaminophen (paracetamol, P) was crystallized under defined conditions from water containing various concentrations of p -acetoxyacetanilide (A). The uptake of A by growing crystals, determined by HPLC, is approximately proportional to the initial concentration of A in the crystallization medium, 0–1000 mg· dm −3 causing 0–3 × 10 −3 mole fraction to be incorporated into the crystals at 30°C. A at ⩾ 1250 mg · dm −3 causes incorporation to level off at ca. 4.5 × 10 −3 mole fraction, corresponding to the solid solubility limit of A in P at 30°C. Washing the crystals with iso-octane indicates that negligible quantities of A are adsorbed onto the crystal surfaces. Increasing uptake of A (0–0.0015 mole fraction) displaces water from the crystals (from 0.041 to 0.013 mole fraction) and increases the enthalpy of fusion, ΔH f , by 1–9%, the melting point, T m , by 0.1–0.6%, and the entropy of fusion, ΔS f by 1–8%, suggesting increases in lattice strength and order. Above 0.0015 mole fraction of A, water uptake rises (0.013–0.024 mole fraction), while ΔH f , T m , and ΔS f decrease, suggesting weakening and disruption of the crystal lattice. ΔS f is a Linear function of the caLcuLated ideal entropy of mixing of P + A + W with a slope of −6.5. This suggests that a much greater disordering of the crystal lattice is created by the impurity defects as compared with simple random mixing of the molecules.

Thermodynamics, molecular mobility and crystallization kinetics of amorphous griseofulvin.

Griseofulvin is a small rigid molecule that shows relatively high molecular mobility and small configurational entropy in the amorphous phase and tends to readily crystallize from both rubbery and glassy states. This work examines the crystallization kinetics and mechanism of amorphous griseofulvin and the quantitative correlation between the rate of crystallization and molecular mobility above and below Tg. Amorphous griseofulvin was prepared by rapidly quenching the melt in liquid N2. The thermodynamics and dynamics of amorphous phase were then characterized using a combination of thermal analysis techniques. After characterization of the amorphous phase, crystallization kinetics above Tg were monitored by isothermal differential scanning calorimetry (DSC). Transformation curves for crystallization fit a second-order John-Mehl-Avrami (JMA) model. Crystallization kinetics below Tg were monitored by powder X-ray diffraction and fit to the second-order JMA model. Activation energies for crystallization were markedly different above and below Tg suggesting a change in mechanism. In both cases molecular mobility appeared to be partially involved in the rate-limiting step for crystallization, but the extent of correlation between the rate of crystallization and molecular mobility was different above and below Tg. A lower extent of correlation below Tg was observed which does not appear to be explained by the molecular mobility alone and the diminishing activation energy for crystallization suggests a change in the mechanism of crystallization.

Modification of adipic acid crystals: influence of growth in the presence of fatty acid additives on crystal properties

When adipic acid is crystallized from water in the presence of a low concentration of an n-alkanoic acid, the additive is incorporated into the crystals and the crystal habit is changed. The following properties of harvested crystals grown in the presence of various concentrations of hexanoic (C6), octanoic (C8) or undecanoic acid (C11) were studied: dissolution rate (DR) and enthalpy of solution (ΔHs) in water, enthalpy of fusion (ΔHf), melting point (Tm), density (d), specific surface area (A) and scanning electron microscopy (SEM). Increasing concentrations of C6 (0–5380 μmol·dm-3) or C11 (0–67.1 μmol·dm-3) caused increasing uptake of additive and pitting of the crystal surface under SEM, while DR and A of a defined sieve fraction increased and then decreased, whereas ΔHs and ΔHf decreased to minima and then increased. In general, the changes in Tm and entropy of fusion (ΔSf) paralleled the changes in ΔHf. C6 caused negligible changes in d, whereas C11 significantly reduced d suggesting lattice expansion (0–0.8%). Various sieve fractions covering the range < 75 μ m to 850 μm showed negligible differences in ΔHf, ΔHs, Tm and d. C8 (0–139 μ mol · dm−3) elicited behaviour similar to that produced by C6 and C11, except that it reduced DR and increased d (< 0.8%). X-Ray powder diffraction of all crystalline samples failed to detect any changes in lattice dimensions (< 0.5%). Overall, DR did not exactly parallel A or SEM on the one hand nor ΔHf, ΔHs, Tm or d on the other hand, implying that DR depends on both the surface and bulk properties of the modified crystals. The results suggest that growth of drug crystals in the presence of low concentrations of non-toxic additives can be used to control dissolution rate and crystal energy.

A disruption index for quantifying the solid state disorder induced by additives or impurities. II. Evaluation from heat of solution

Abstract In a previous report a dimensionless disruption index (d.i.) was proposed for quantifying the disruptive influence of an additive or impurity (the guest substance) when present in solid solution in the crystal lattice of a host substance at mole fractions, x 2 , less than 0.05. The d.i. value was defined as the rate of change of the difference between the entropy of the solid, S solid , and that of the liquid, S liquid , with respect to the ideal entropy of mixing of the components of the solid, ΔS ideal m , i . e . d . i . = − δ ( S liquid − S solid )/ δ ( ΔS ideal m ). The determination of (S liquid − S solid ) from the heat of fusion and the melting point using differential scanning calorimetry (DSC) or differential thermal analysis (DTA) could itself change the entropy and the concentration of point defects and dislocations by an annealing process. To overcome these problems δ(S solulion − S solid ) = δ (ΔS s ), which is shown to approximate closely to δ(S liquid − S solid ), is determined isothermally (e.g. at 25 or 37° C) using solution calorimetry and measurements of J, the dissolution rate per unit surface area. ΔS s is derived from the heat of solution, ΔH s , and from the Gibbs free energy of solution which is changed by RT · δ(In J) on doping, where R is the gas constant and T is the absolute temperature. The possibilities that ΔS s can be calculated from ΔH s directly assuming enthalpy-entropy compensation, or simply by ignoring the term containing δ(In J), are also considered. The above possibilities are examined using the limited data available for adipic acid doped with hexanoic, octanoic, undecanoic or oleic acid. The d.i. values from solution calorimetry are of the order 10 3 , indicating an enormous potential for lattice disruption, while d.i. values from DSC are generally smaller and decrease with decreasing chain-length of the guest molecule. This suggests that heating in DSC promotes rearrangement of the guest molecules and annihilation of crystal defects. Ignoring δ(In J) changes d.i. by up to 15%. Much larger influences of δ(In J) are given by oleic acid as the guest which tends to concentrate on the surface. In general, the pseudo-disruption index p.d.i., calculated as −δ(ΔH s /T)/δ(ΔS ideal m ), may approximate sufficiently closely to d.i. to be useful. Thus, the p.d.i. value for cephaloridine monohydrate doped with cephaloridine anhydrate, corresponding to slight moisture loss from the lattice of the former, is of the order 1, which suggests very little lattice disruption.

A disruption index for quantifying the solid state disorder induced by additives or impurities. I. Definition and evaluation from heat of fusion

A dimensionless disruption index (d.i.) is proposed for quantifying the solid state disorder induced by an additive or impurity (the guest substance), when present in solid solution in the crystal lattice of a host substance at mole fractions, x2, less than 0.05. The d.i. value is defined as the rate of change of the difference between the entropy of the solid and that of the liquid, with respect to the ideal entropy of mixing of the components of the solid, ΔSmideal. From fundamental thermodynamic considerations d.i. is closely approximated by the slope of the plot of the entropy of fusion of the solid, ΔSf, against ΔSmideal for x2 < 0.05. ΔSf is given by the heat of fusion divided by the absolute melting point, while ΔSmideal = − RΣxj ln xj is calculated from the analytical data of the crystals, where xj is the mole fraction of a given component. The linear relationship was tested using the limited literature data available for 7 systems and was found to be obeyed for x2 < 0.05. Values of d.i. range from zero for ideal solutions through about 10−1 for doping of the intermetallic compound InCd3 with either of its components, (somewhat higher, 0.423, for cadmium doped with InCd3), to about 10 for the doping of a stable, ordered organic crystal with an organic additive. The d.i. values for phenacetin doped with benzamide, griseofulvin + lecithin, acetaminophen + water + p-acetoxyacetanilide, and pp-DDT + op-DDT, are 7.94, 5.09, 6.53 and 15.1, respectively. The d.i. values are discussed in relation to the properties of the host and guest. The method of determining d.i. from ΔSf is critically assessed. The d.i. values may be useful in predicting the sensitivity of the crystal lattice of a drug or excipient to the presence of traces of a given impurity in solid solution. If the presence of impurities gives rise to batch-to-batch variations, d.i. values may also be useful for quantifying the observed differences in properties between batches of materials.

True density and thermal expansivity of pharmaceutical solids: comparison of methods and assessment of crystallinity

Abstract Three methods for determining the density of solids, namely displacement of a liquid, displacement of a gas and flotation in a liquid, have been compared using adipic acid doped with various concentrations of oleic acid. The liquid displacement method is tedious and tends to underestimate the true density. Gas pycnometry is rapid, non-destructive and relatively simple to use, but requires expensive instrumentation andalarge sample size, while its accuracy is limited by the relatively low precision of the volume measurements. The flotation method is much simpler in operation, is inexpensive and demonstrates the probabilistic nature of density distribution within a powder sample. Although it is more time consuming than gas pycnometry, the final density values can be accurate to four significant figures using the simplest apparatus. More elaborate set-ups increase the accuracy. The use of two or more suspending liquids in the flotation method enables the density of a given crystal to be determined at two or more temperatures and therefore provides the thermal expansivity. The crystal density at ambient temperature may be interpolated from a plot of density against temperature. The presence of traces of oleic acid in adipic acid crystals marginally lowers the density. Application of the flotation method to acetaminophen crystals shows that the presence of traces of p-acetoxyacetanilide increases the mean density by shifting the density distribution closer to its upper limit. Trace additives in the crystals of both adipic acid and acetaminophen change the thermal expansivity of the most dense crystals in each sample. The use of a single liquid in the flotation method, together with the assumption that the densities of crystals are independent of temperature, may hide significant effects. The results suggest that the thermal expansivity is a more reliable indicator of crystaliinity than density measurements at a fixed temperature.

Entropy of processing: a new quantity for comparing the solid state disorder of pharmaceutical materials

During the pharmaceutical processing of a solid (e.g. crystallization, drying, gain or loss of impurities or additives, milling. compression, heating or irradiation), defects and other imperfections develop, wander or disappear in the crystal lattice. Crystal imperfections contribute to the disorder of the crystal lattice, and entropy is here proposed as a practical and fundamental measure of this effect. The difference between the entropy of a given sample and that of the same amount of the reference material is designated the “entropy of processing (or imperfection)”, ΔSp, of the sample. Thermodynamic cycles are described for evaluating ΔSpof a sample from differential scanning calorimetry or from solution calorimetry with solubility studies. ΔSp may be evaluated for polymorphs, solvates (e.g. hydrates), amorphous forms, glasses, impure, or variously processed samples of a given substance. ΔSp calculated from literature data for processed pharmaceutical solids is found to be positive with respect to a highly pure. stable. crystalline reference material and to range from 0 to 200 J. K−1 · mol−1. Small values of ΔSp (0–10 J · K−1 · mol−1) are given by crystals which have been “doped” with additives or impurities in solid solution. Literature data for milled calcium gluceptate suggest that enthalpy-entropy compensation occurs and may be exploitable. ΔSpof ground samples of chloramphenicol palmitate A or B parallels their dissolution rate and bioavailability. Data for β-lactam antibiotics and calcium gluceptate indicate that ΔSp increases with decreasing X-ray crystallinity, purity and stability and with increasing processing stress. ΔSp may decrease during annealing and/or storage and appears to be useful for quantifying, understanding and predicting batch-to-batch differences in pharmaceutical solids.

CRYSTAL ENGINEERING STUDIES WITH AN EXCIPIENT MATERIAL (ADIPIC ACID)

In a previous paper (Fairbrother and Grant, 1978) the effect of traces of nalkanoic acid additives on the habit of growing adipic acid crystals was described. At the threshold concentration (ranging from 500 pg/ml for valeric acid to 2 ug/ml for undecanoic acid) the additives promote the development of a number of additional crystal faces on the regular hexagonal plates. The most important of these truncate the corners formed by the (100) and (110) faces. At higher additive concentrations the adipic acid crystallises in the form of cigar-shaped spars with smoothly rounded edges. Further increase in additive concentration leads to the formation of fused pairs Of spherules. An attempt has been made to study the mechanism by which the n-alkanoic acids modify the crystal habit.

Modification of adipic acid crystals. II. Influence of growth in the presence of oleic acid on crystal properties

Abstract Adipic acid (AA) was crystallized under defined conditions from water containing 0–110 μmol · dm −3 oleic acid (OA). With increasing concentration of OA, the water content of the crystals was constant (0.047 ± 0.005 mole fraction), while the uptake of OA increased linearly, and the dissolution rate (DR) and specific surface area (SSA) of a defined sieve fraction decreased. The decrease in DR exceeded the decrease in SSA. Washing the crystals with chloroform removed an appreciable proportion of the poorly water-soluble OA from the surface of the crystals and produced a marked increase in DR. OA at 15 μmol · dm −3 caused a doubling of DR, whereas higher concentrations reduced DR of the washed crystals to the original value. Increasing incorporation of OA into the AA crystals reduced their density, indicating a decrease in crystallinity. Increasing incorporation of OA also reduced the melting point, T m , the enthalpy of fusion, ΔH f , and the enthalpy of solution, ΔH s , of the crystals, indicating an increase in energy of the crystalline bulk, corresponding to an increase of lattice strain. Higher concentrations of OA reversed the energetic effects. The reduction in the entropy of fusion by small amounts of OA was greater than the increase in the sum of the ideal partial molar entropies of the components of the crystals by a factor approaching 1000. This indicates that OA greatly disrupts the order of the crystal lattice of AA.

Modification of acetaminophen crystals. III. Influence of initial supersaturation during solution-phase growth on crystal properties in the presence and absence of p-acetoxyacetanilide

Abstract Acetaminophen (P) crystals have been grown from pure aqueous solutions containing 0 and 500 mg · dm −3 p -acetoxyacetanilide (A) at various stirring speeds (200–400 rpm) and the physical properties examined. When A is omitted from the solutions, an increase of the stirring rate has no significant effects on the crystal habit (polyhedral prisms), but markedly raises the water content (by a factor of 10), lowers the enthalpy of fusion, ΔH f , melting point, ( igT m ) and entropy of fusion, ΔS f , (by 8–18%, 0 respectively) and decreases and then increases the intrinsic dissolution rate ( IDR ) of the P crystals. However, in the presence of A, the crystals assume an acicular morphology, independent of the agitation rate of the solutions. Whilst eliciting no further changes in the habit of these crystals, application of more vigorous stirring to the solutions leads to a moderate decrease in the uptake of A, a slight increase in water content, a slight decrease in the ΔH f , T m and ΔS f as well as an initial reduction and a su IDR . Compared with the polyhedral prisms crystallized in the absence of A, these crystals invariably demonstrate a considerably higher IDR , consistent with their needle-like habit and much lower water content. These observations demonstrate the pharmaceutical significance of agitation during crystallization in modulating the physical properties and aqueous dissolution rate of the P crystals.