Keywords

1 Introduction

The way in which reactions occur in organic solids has long been of interest. Significant progress in terms of how crystal packing influences reactivity, however, required the development of structure determination using single crystal X-ray diffraction in a suitably routine way so that systematic studies of related solids could be performed. The most cited example is that of the cinnamic acids which was developed in the 1960s. The main conclusion was that reaction could occur in the (perfect) crystal if molecules were suitably arranged for diffusionless reaction. The phrase “topochemical control” was coined to describe such processes. Shortly afterwards, however, reactions which apparently contravened this interpretation were identified and the hypothesis developed that perhaps structural imperfections within the crystal might be the sites at which “non-topochemical” reaction took place. Less well developed has been an assessment of how reactions at the surface of a crystal might differ from that of the bulk (perfect or imperfect). In a similar manner to the need for routine surface studies required the development of a technique with better resolution than that provided by optical microscopy. The development of surface probe microscopy and in particular atomic force microscopy has allowed progress in this area.

2 Topochemical Control – Perfect Packing Dominates

In a series of papers, Schmidt, Cohen and colleagues [1, 2] established that the photochemical reactivity of crystals of cinnamic acid and various cinnamic acid derivatives was a function of the molecular registry as revealed by single-crystal diffraction studies. The crystal structures of numerous substituted cinnamic acids can be categorised into three general types – α, β and γ. Some derivatives can adopt multiple different crystal structures – for example, o-ethoxy-trans-cinnamic acid is polymorphic and crystallises in all three structural types. In each of the structure types, the molecules pack in one dimensional stacks and form pairwise hydrogen bonding interactions across centres of symmetry. Within the stacks the molecules lie parallel with a distance of the order of 3.5 Å between molecular planes. These three types differ, however, in the angle between this normal and the stack axis, in the repeat distance along this axis and in the extent and type of overlap between adjacent molecules in the stack. The β-type resulted in the mirror dimer; the α-type the centrosymmetric dimer; whilst the γ-type was light stable – see Fig. 9.1.

Fig. 9.1
figure 1

Dimerisation of cinnamic acids

As an extension of this lattice controlled dimerization of cinnamic acids the technique of solid-state polymerisation was developed. Hirschfeld and Schmidt [3] suggested that the necessary condition for matrix controlled polymerisation is that no significant disruption of the molecular positions occurred on polymerisation. They referred specifically to crystals of molecules with two potentially reactive centres which were so oriented that they could rotate in place to link up with their neighbours without any linear displacement of the molecular centres – see Fig. 9.2. Two groups of materials that explored this concept were the diacetylenes and divinyl monomers. In the case of divinyl monomers, an important example was that of 2,5-distyrylpyrazine, which readily polymerised in the solid state [4] – with the final crystalline photo-product having a molecular weight several times greater than the solution grown polymer. This could be rationalised on the basis of packing within the perfect lattice. The work of Wegner and co-workers [5] demonstrated that solid state polymerisation of diacetylenes (thermal and photolytic) could also be rationalised on the basis of crystal packing in the perfect lattice, the polymerisation proceeding to high molecular weight products with little change in lattice parameters.

Fig. 9.2
figure 2

Polymerisation reaction

3 Non-topochemical Reactions – Types of Defects in Organic Crystals and Their Possible Role

A recognised challenge, however, was found in the abnormal solid-state photochemical reactions of anthracene and 9-cyanoanthracene (amongst others) [6]. In these systems, analysis of the perfect lattice would suggest stability as molecules are inappropriately arranged to react (either by orientation or separation). Despite this, reactions did in fact take place. One explanation was that those reactions which gave the “non-topochemical” product occurred preferentially at defect sites in the crystal – see Figs. 9.3 and 9.4 [7,8,9,10].

Fig. 9.3
figure 3

Representation of an edge dislocation

Fig. 9.4
figure 4

Defect and product formation

This was supported by the work of Thomas and his group [11, 12] who separately etched and photodimerized matched halves of cleaved melt-grown anthracene crystals (Fig. 9.4). They found that the degree of correspondence between etch pits and dimerization centres was usually greater than 85%. For 9-cyananoanthracene the observed product was further rationalised in terms of stacking faults in the crystals which were generated by the dissociation of [100](221) dislocations (Fig. 9.5 ) [11].

Fig. 9.5
figure 5

Stacking fault in 9-cyanoanthracene leading to appropriate orientation for reaction [11]

Recent micro-Raman studies of the photodimerisation of 9-cynanoanthracene has demonstrated that there is a delay between the onset of chemical reaction and the appearance of the product crystals – the initial reaction can occur within the crystal but there is no diffraction evidence until a reconstructive phase transition occurs [13].

The presence of crystalline imperfections in a solid may lead both to local changes in topology at the imperfection and to changes in energy owing to any strain set up or relieved by the imperfection [9]. The chemical potential of a molecule associated with an imperfection will be different to that of molecules in the perfect lattice. In order to anticipate the role that imperfections will play in the properties of an organic solid, it will be necessary to establish completely their character. Some salient properties of imperfections in general, and in particular point defects, linear (dislocations) and planar defects (e.g. stacking faults) are now given. We also note that the concept of a volume defect might apply where solvent occlusion is also present.

Structural imperfections can be classified into three general types based on geometry: point, line and planar.

  1. (i)

    Point defects: In a pure material there are several types of point defects. Vacancies exist where atoms or molecules are missing from their normal positions in the crystal lattice. Molecules may also occupy non-lattice (interstitial) sites (Fig. 9.6). At temperatures above 0 K there is always present an equilibrium number of vacancies in a crystal with the number increasing as the temperature increases. During crystal growth, a quantity of these defects in excess of the equilibrium number may be introduced either through rapidly quenching from high temperatures, through plastic deformation or through radiation damage.

Fig. 9.6
figure 6

Various types of point defects in an idealised atomic-array

Line Imperfections: Line imperfections, or dislocations, are linear regions of elastic distortion caused by atoms that are slightly displaced from their equilibrium lattice positions. The two basic types of dislocations are the edge dislocation and the screw dislocation [9]. An example of an edge dislocation was given in Fig. 9.3. If the dislocation line AB were imagined to bend in the slip plane, then it will be possible for the dislocation line to run parallel to the direction of the Burgers vector of the dislocation. This condition describes the screw dislocation (Fig. 9.7). The Burgers vector of a screw dislocation is parallel to the line of the dislocation. The structural error that arises in the screw dislocation is a spiral ramp of displaced atoms that advances one Burgers vector per circuit of the dislocation.

Fig. 9.7
figure 7

Schematic of a screw dislocation

  1. (ii)

    Planar imperfections: Planar imperfections are any surfaces across which the atom positions in a perfect crystal are not preserved. Two common planar faults found in crystals are twin planes and stacking faults. Another common defect will be grain boundaries – the region between randomly oriented domains within the crystal. Twin planes separate the two parts of a twinned crystal, the parent and its twin – see Fig. 9.8.

Fig. 9.8
figure 8

Microtwins in an organic crystal and associated diffraction pattern

Transmission electron microscopy (TEM) is particularly powerful for imaging dislocations in organic crystals [14]. As an example see Fig. 9.9.

Fig. 9.9
figure 9

Large number of dislocations in a sample of theophylline as seen by TEM [15]

4 Atomic Force Microscopy and Comparison of Surface Reactivity Compared to Bulk

AFM is a member of the family of scanning tunnelling microscopy (STM) in that it relies on a physical probe to explore the surface of interest. [16, 17] The AFM “feels” atomic forces, such as Van der Waals or electrostatic interactive forces between the tip and the surface, by the bending of the cantilever on an AFM probe. See Figs. 9.10 and 9.11.

Fig. 9.10
figure 10

Laser, photodetector and cantilever [17]

Fig. 9.11
figure 11

Various stages of tip-surface contact

This surface sensitivity allows molecular height effects to be observed [17].

Besides imaging individual surfaces, it is also crucial to understand the surface changes of molecular crystals during chemical reactions or in response to environmental alterations. One example of ex situ imaging is offered by the study of (caffeine)⋅(oxalic) acid and (caffeine)⋅(malonic) acid cocrystals, which were demonstrated by Trask et al. to be stable in 98% RH and 75% RH respectively by PXRD [18]. Cassidy et al. monitored the surface response of these cocrystals by ex situ AFM after exposure to these humidities (Fig. 9.12) [19]. Caffeine and oxalic acid were found to exhibit a high degree of mobility on cocrystal surfaces that facilitates recrystallisation events on the surface. This study demonstrated that the information about surface changes that are not apparent from bulk measurements (such as powder X-ray diffraction) can be detected by AFM, and may be important for predicting the stability of active pharmaceutical ingredients (API), as well as at excipient solid-state interfaces.

Fig. 9.12
figure 12

Changes in surface structure of a crystal as a result of exposure to high humidity [19]