Organic glasses can grow crystals much faster on a free surface than in the interior, a result of the high mobility of surface molecules. A puzzling property of this process is that it is active in the glassy state, but disrupted as the glass is heated above the glass transition temperature Tg to become a fluid, despite the large increase of mobility. To understand this phenomenon, high-resolution microscopy was used to observe surface crystal growth in amorphous indomethacin (IMC) in two polymorphs (α and γ). α-IMC represents the general case of strong disruption of surface growth by fluidity, while γ-IMC represents the opposite situation where the effect is weak. We observed that heating above Tg causes liquid flow toward the crystals of both polymorphs. The essential difference between the polymorphs is that γ-IMC grows as compact domains and its uniformly advancing growth front is unperturbed by the onset of liquid flow. In contrast, α-IMC grows segregated needles and liquid flow strongly alters the crystal/liquid interface. This effect arises because the slow-growing flanks of needle crystals are wetted and embedded by the liquid. This explanation has support from independent observations that polystyrene spheres (a model for slow-growing crystals) are embedded on the time scale of crystal growth.

Crystal growth kinetics and liquid dynamics of 1,2-diphenylcyclopentene (DPCP) and 1,2-diphenylcyclohexene (DPCH) were characterized by optical microscopy and dielectric spectroscopy. These two molecules are structurally homologous and dynamically similar to the well-studied glassformer ortho-terphenyl (OTP). In the supercooled liquid states of DPCP and DPCH, the kinetic component of crystal growth ukin has a power law relationship with the primary structural relaxation time τα, ukin ∝ τα–ξ (ξ ≈ 0.7), similar to OTP and other fragile liquids. Near the glass transition temperature (Tg), both DPCP and DPCH develop much faster crystal growth via the so-called GC (glass to crystal) mode, again similar to the behavior of OTP. We find that the α-relaxation process apparently controls the onset of GC growth, with GC growth possible only at sufficiently low fluidity. These results support the view that GC crystal growth can only occur in systems where the liquid and crystal exhibit similar local packing arrangements.

Organic glasses can grow crystals much faster on the free surface than in the interior, a phenomenon important for fabricating stable amorphous materials. This surface process differs from and is faster than the glass-to-crystal (GC) growth mode existing in the bulk of molecular glasses. We report that similar to GC growth, surface crystal growth terminates if glasses are heated to gain fluidity. In their steady growth below the glass transition temperature Tg, surface crystals rise above the amorphous surface while spreading laterally and are surrounded by depressed grooves. Above Tg, the growth becomes slower, sometimes unstable. This damage is stronger on segregated needles (α indomethacin, nifedipine, and o-terphenyl) than on crystals growing in compact domains (γ indomethacin). This effect arises because the onset of liquid flow causes the wetting and embedding of upward-growing surface crystals. Segregated needles are at greater risk because their slow-growing flanks appear stationary relative to liquid flow at a low temperature. The disruption of surface crystal growth by fluidity supports the view that the process occurs by surface diffusion, not viscous flow. Compared to the bulk GC mode, surface crystal growth is disrupted less abruptly by fluidity. Nevertheless, to the extent that fluidity damages them, both processes are solid-state phenomena terminated in the liquid state.

Fast crystal growth can abruptly emerge as molecular liquids are cooled to become glasses, a phenomenon presently unknown for other materials. This glass-to-crystal (GC) mode can cause crystallization rates orders of magnitude faster than those predicted by standard models. While GC growth is known for 12 systems, its features vary greatly with growth rates spanning a factor of 104. We report that the general condition for GC growth to exist is that liquid diffusion be slow relative to crystal growth according to D/u < 7 pm. This condition holds for all liquids exhibiting GC growth and suggests that the phenomenon is a solid-state process terminated by fluidity. GC growth must solidify several molecular layers before rearrangement by diffusion. We propose that GC growth propagates by a nonequilibrium crystal/liquid interface 3 nm wide that solidifies by local mobility. These results explain the prevalence of GC growth among organic liquids and guide its discovery in other materials.Keywords: amorphous; crystallization; diffusion; fluidity; glass; indomethacin; o-therphenyl; solid-state crystal growth;

We report a remarkable system of cocrystals containing nicotinamide (NIC) and (R)-mandelic acid (RMA) in numerous stoichiometric ratios (4:1, 1:1 in two polymorphs, and 1:2) with anomalous formation properties. The formation of these cocrystals decreases energy but expands volume, in contrast to most physical processes, but similar to water freezing. The decrease of energy upon cocrystallization agrees with the exothermic mixing of NIC and RMA liquids (a base and an acid). Volume expansion is general for the formation of all NIC cocrystals for which data exist (n = 40): +3.9 Å3/molecule or +17 cm3/kg on average, corresponding to a 2% expansion. This volume expansion correlates with the shortening and strengthening of hydrogen bonds upon cocrystallization, analogous to water freezing. The NIC-RMA binary phase diagram was constructed that contains the congruent and incongruent melting of six crystalline phases. These results are relevant for understanding the nature of cocrystallization and why some molecules are prolific cocrystal formers.

Surface self-diffusion coefficients have been determined for the organic glass Nifedipine using the method of surface grating decay. The flattening of 1000 nm surface gratings occurs by viscous flow at 12 K or more above the glass transition temperature and by surface diffusion at lower temperatures. Surface diffusion is at least 107 times faster than bulk diffusion, indicating a highly mobile surface. Nifedipine glasses have faster surface diffusion than the previously studied Indomethacin glasses, despite their similar bulk relaxation times. Both glasses exhibit fast surface crystal growth, and its rate scales with surface diffusivity. The observed rate of surface diffusion implies substantial surface rearrangement during the preparation of low-energy glasses by vapor deposition. The Random First Order Transition Theory and the Coupling Model successfully predict the large surface-enhancement of mobility and its increase on cooling, but disagree with the experimental observation of the faster surface diffusion of Nifedipine.

We report the first structural determination of the metastable β polymorph of nifedipine (NIF) by single-crystal X-ray diffraction. Stable, high-quality crystals were grown from the melt in the presence of a polymer dopant. Our β NIF structure is characterized by a unit cell similar to that of the structure recently proposed from powder diffraction, but significantly different molecular conformations. Unlike the stable α polymorph, β NIF undergoes a reversible solid-state transformation near 60 °C. The now available β NIF structure clarifies some confusion concerning NIF polymorphs and enables inquiries into the structural basis for the selective crystallization of β NIF from glasses. We report that another polymorph crystallizes concomitantly with β NIF from the supercooled melt and transforms to β NIF at room temperature; this polymorph also undergoes reversible solid-state transformation.

Co-crystals provide an opportunity to improve the properties of pharmaceuticals and other materials. We report a method for determining the formation enthalpies of co-crystals in which enthalpies of melting are measured for a co-crystal and the physical mixture of its component crystals. Because the two melting processes arrive at the same liquid, the difference of their enthalpy changes is the co-crystal’s formation enthalpy. For the system of nicotinamide (NIC) and R-mandelic acid (RMA), the formation enthalpy at 30 °C is −23 (3) J/g for polymorph 1 and −18 (3) J/g for polymorph 2. These values are comparable with the enthalpy of mixing for NIC and RMA liquids [−49 (4) J/g at 160 °C], indicating the need for correcting for nonideal mixing in calculating formation enthalpies of co-crystals via thermodynamic cycles. This correction is made automatically in our method by performing measurements with physical mixtures of component crystals, as opposed to pure component crystals. One of the NIC-RMA polymorphs (2) was discovered in this work, and its structure and thermodynamic relation to polymorph 1 are reported.

We review recent progress toward understanding and enhancing the stability of amorphous pharmaceutical solids against crystallization. As organic liquids are cooled to become glasses, fast modes of crystal growth can emerge. One such growth mode, the glass-to-crystal or GC mode, occurs in the bulk, and another exists at the free surface, both leading to crystal growth much faster than predicted by theories that assume diffusion defines the kinetic barrier of crystallization. These phenomena have received different explanations, and we propose that GC growth is a solid-state transformation enabled by local mobility in glasses and that fast surface crystal growth is facilitated by surface molecular mobility. In the second part, we review recent findings concerning the effect of polymer additives on crystallization in organic glasses. Low-concentration polymer additives can strongly inhibit crystal growth in the bulk of organic glasses, while having weaker effect on surface crystal growth. Ultra-thin polymer coatings can inhibit surface crystallization. Recent work has shown the importance of molecular weight for crystallization inhibitors of organic glasses, besides “direct intermolecular interactions” such as hydrogen bonding. Relative to polyvinylpyrrolidone, the VP dimer is far less effective in inhibiting crystal growth in amorphous nifedipine. Further work is suggested for better understanding of crystallization of amorphous organic solids and the prediction of their stability.

As organic liquids are cooled to become glasses, crystal growth at the free surface can be substantially faster than in the interior, a phenomenon uncommon for other materials and for which different explanations exist. We have measured the surface and bulk growth rates of three polymorphs in carbamazepine glasses. Crystal density has no controlling effect on the extent to which surface crystal growth is enhanced over bulk crystal growth, in contradiction to models that relate fast surface crystal growth to the release of crystallization-induced tension.

The crystallization of glasses and amorphous solids is studied in many fields to understand the stability of amorphous materials, the fabrication of glass ceramics, and the mechanism of biomineralization. Recent studies have found that crystal growth in organic glasses can be orders of magnitude faster at the free surface than in the interior, a phenomenon potentially important for understanding glass crystallization in general. Current explanations differ for surface-enhanced crystal growth, including released tension and enhanced mobility at glass surfaces. We report here a feature of the phenomenon relevant for elucidating its mechanism: Despite their higher densities, surface crystals rise substantially above the glass surface as they grow laterally, without penetrating deep into the bulk. For indomethacin (IMC), an organic glass able to grow surface crystals in two polymorphs (α and γ), the growth front can be hundreds of nanometers above the glass surface. The process of surface crystal growth, meanwhile, is unperturbed by eliminating bulk material deeper than some threshold depth (ca. 300 nm for α IMC and less than 180 nm for γ IMC). As a growth strategy, the upward-lateral growth of surface crystals increases the system’s surface energy, but can effectively take advantage of surface mobility and circumvent slow growth in the bulk.

To study the influence of polymer additives on bulk and surface crystal growth in organic glasses (amorphous solids), which are being investigated for delivering poorly soluble drugs and in this role must resist crystallization. Recent studies have discovered new modes of crystal growth that emerge as organic liquids are cooled to form glasses: one existing in the bulk (GC growth) and another at the surface, both leading to crystal growth much faster than predicted by standard theories.Bulk and surface crystal growth rates were measured in nifedipine glasses doped with polyvinylpyrrolidone (PVP) of different molecular weights. AFM enabled observation of the microstructure of surface-growing crystals.Polymer additives influence bulk and surface crystal growth differently. For every weight percent of PVP added, surface crystal growth of nifedipine slows by two times at 12°C below Tg, whereas bulk crystal growth slows by 10 times. In contrast to the polymers, the VP dimer had little effect on crystal growth.Polymer additives inhibit crystal growth in nifedipine glasses more strongly in the bulk than at the surface. The effectiveness of crystallization inhibitors depends not only on intermolecular interactions but also on molecular sizes.

Raman microscopy and isotope labeling have been used for the first time to measure water self-diffusion in carbohydrate glasses. Together with pulsed-gradient stimulated-echo NMR, this method yielded the self-diffusion coefficients of water in amorphous maltose over 8 orders of magnitude, from the liquid to the glassy state. There are consistencies and major differences between our data and those obtained by evaporative drying. Water diffusion is remarkably fast in maltose glasses, decouples from maltose diffusion, and is not strongly affected by the glass transition.

Diamond and graphite are polymorphs of each other: they have the same composition but different structures and properties. Many other substances exhibit polymorphism: inorganic and organic, natural and manmade. Polymorphs are encountered in studies of crystallization, phase transition, materials synthesis, and biomineralization and in the manufacture of specialty chemicals. Polymorphs can provide valuable insights into crystal packing and structure−property relationships. 5-Methyl-2-[(2-nitrophenyl)amino]-3-thiophenecarbonitrile, known as ROY for its red, orange, and yellow crystals, has seven polymorphs with solved structures, the largest number in the Cambridge Structural Database. First synthesized by medicinal chemists, ROY has attracted attention from solid-state chemists because it demonstrates the remarkable diversity possible in organic solids. Many structures of ROY polymorphs and their thermodynamic properties are known, making ROY an important model system for testing computational models. Though not the most polymorphic substance on record, ROY is extraordinary in that many of its polymorphs can crystallize simultaneously from the same liquid and are kinetically stable under the same conditions. Studies of ROY polymorphs have revealed a new crystallization mechanism that invalidates the common view that nucleation defines the polymorph of crystallization. A slow-nucleating polymorph can still dominate the product if it grows rapidly and nucleates on another polymorph. Studies of ROY have also helped understand a new, surprisingly fast mode of crystal growth in organic liquids cooled to the glass transition temperature. This growth mode exists only for those polymorphs that have more isotropic, and perhaps more liquid-like, packing. The rich polymorphism of ROY results from a combination of favorable thermodynamics and kinetics. Not only must there be many polymorphs of comparable energies or free energies, many polymorphs must be kinetically stable and crystallize at comparable rates to be observed. This system demonstrates the unique insights that polymorphism provides into solid-state structures and properties, as well as the inadequacy of our current understanding of the phenomenon. Despite many studies of ROY, it is still impossible to predict the next molecule that is equally or more polymorphic. ROY is a lucky gift from medicinal chemists.

Amorphous pharmaceuticals, a viable approach to enhancing bioavailability, must be stable against crystallization. An amorphous drug can be stabilized by dispersing it in a polymer matrix. To implement this approach, it is desirable to know the drug’s solubility in the chosen polymer, which defines the maximal drug loading without risk of crystallization. Measuring the solubility of a crystalline drug in a polymer is difficult because the high viscosity of polymers makes achieving solubility equilibrium difficult.Differential Scanning Calorimetry (DSC) was used to detect dissolution endpoints of solute/polymer mixtures prepared by cryomilling. This method was validated against other solubility-indicating methods.The solubilities of several small-molecule crystals in polymers were measured for the first time near the glass transition temperature, including d-mannitol (β polymorph) in PVP, indomethacin (γ polymorph) in PVP/VA, and nifedipine (α polymorph) in PVP/VA.A DSC method was developed for measuring the solubility of crystalline drugs in polymers. Cryomilling the components prior to DSC analysis improved the uniformity of the mixtures and facilitated the determination of dissolution endpoints. This method has the potential of providing useful data for designing physically stable formulations of amorphous drugs.

Amorphous solids are generally more soluble and faster dissolving than their crystalline counterparts, a property useful for delivering poorly soluble drugs. Amorphous drugs must be stable against crystallization, for crystallization negates their advantages. Recent studies found that crystal growth in amorphous indomethacin is orders of magnitude faster at the free surface than through the bulk and this surface-enhanced crystallization can be inhibited by an ultrathin coating. Herein, we report a second system that exhibits the same phenomena. Crystal growth at the free surface of amorphous nifedipine (NIF) was at least 1 order of magnitude faster than that through the bulk below the glass transition temperature Tg (42 °C). A thin coating of gold (10 nm) reduced the surface crystal growth rate to the bulk crystal growth rate. Surface-enhanced crystal growth was more pronounced near and below Tg than substantially above Tg, which suggests that this growth mode is more important for the glassy state. Our results support the view that a thin layer of molecules near the surface have higher mobility than the bulk molecules and can enable faster crystal growth. The higher mobility of surface molecules and the resulting fast crystal growth can be suppressed by an ultrathin coating.Keywords: Amorphous solid; coating; glass transition; nifedipine; surface-enhanced crystal growth;

Seeding the liquid of d-mannitol with its polymorphs (α, β, and δ) revealed several cases of cross-nucleation between polymorphs. Only seeds of the α polymorph (the structure of the intermediate thermodynamic stability and the fastest growth rate) yielded the same polymorph in new growth. Seeds of the δ polymorph yielded the α polymorph in new growth. Seeds of the β polymorph yielded the β polymorph in new growth at small undercoolings (a few degrees below its melting point); at lower temperatures, the α polymorph nucleated on β seeds and its amount in the product increased with decreasing temperature. Seeding with single crystals of the β polymorph (rods elongated along c) showed that the orientation of the seed crystal affected how much the seed polymorph could grow before cross-nucleation occurred. Cross-nucleation makes seeding ineffective for achieving polymorphic selectivity but is sometimes avoidable by choosing suitable crystallization conditions.

A dispersion‐corrected density functional theory method has been used to study the formation energies and volumes of cocrystals. For four cocrystals of nicotinamide (NIC) and (R)‐mandelic acid, a broad agreement is found between experimental and computed values. We report that cocrystals containing NIC are anomalous as their formation generally decreases energy but expands volume. In this respect, the formation of NIC cocrystals is in contrast to most physical processes, but similar to water freezing. As in the case of water freezing, the cocrystallization with NIC leads to stronger hydrogen bonds and looser molecular packing, a combination that is likely responsible for the anticorrelation between energy and volume. NIC has two conformers 4 kJ/mol apart in energy and both can form cocrystals, with the resulting structures having comparable formation energies and volumes. These results are relevant for understanding the nature of cocrystallization and why NIC is a prolific cocrystal former. © 2014 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm Sci 103:2896–2903, 2014