crystal structure, and the surfaceinteraction between the solvent andacetaminophen molecules in thediffusion layer. These results werecompared to computer simulations of the same process to gain a betterunderstanding of how the resultingsurface texture provides informationabout the dissolution mechanisms of acetaminophen. In the cases of

water, acetic anhydride, and pyridine,the etch pits were formed parallel to the a and c crystallographic axes of the crystal. However, the etchingpatterns formed by acetone,dichloroethane, and ethyl acetate werenot clearly understood and did notmatch the structure expected from thecomputer simulations. It is believed thatin these cases, the etching processwas disrupted by the absorption of solvent molecules on the crystalsurface, resulting in unexpectedetching patterns.

This hypothesis was further tested by observing the etching patternsproduced by water with the tailor-made additives, acetanilide and 4-methyl acetanilide.8 Drastically differentpatterns resulted from small amounts ofthese additives. This effect is believedto be caused by the absorption ofadditive molecules that diffuse into thecrystal lattice and disrupt the originalsupramolecular interaction network,changing the etch pits fromparallelograms (in pure water) torectangular, square, or circular forms.

In another example, AFM has also been applied to gain anunderstanding of the differences indissolution behavior of the (001) and (100) crystal surfaces of aspirin.9

Dissolution studies have shown a 50percent larger dissolution rate for the(100) crystal plane with respect to the(001) surface. TappingMode imagesshow a smooth (001) surface with 7.3-angstrom molecular steps (Figure3). The (100) surface was muchrougher with 50- to 100-angstrom

Figure 2. AFM images of the (010) crystal surface of acetaminophen single crystals after partialdissolution by a) water, b) acetone, and c) pyridine. The resulting surface texture provides informationabout the dissolution mechanisms of acetaminophen and how the dissolution is related to its crystalstructure7. 60µm scans. Images courtesy T. Li, K.R. Morris, and K. Park, Purdue University

(temperature, pH), and at anappropriate rate for the drug to beeffectively absorbed. Dissolution rateand behavior have been shown to be affected by surface area, temper-ature, crystal structure, pH and thenature of the solvents. 5 Thus, propercharacterization of these factors has a very important role in a drug’soverall efficacy.

Traditionally, dissolution testing hasbeen carried out at the macroscale byconstant surface area or particulatedissolution methods to examine thebehavior of drug products underdifferent conditions. However, there are several advantages of studying the dissolution processes by AFM.Dissolution is commonly tested at37±0.5˚C, which is made possiblewith AFM by heating the fluidenvironment during experimentation.Since the AFM can operate in air and fluid environments, the operatorhas the choice of how to conduct theexperiment. It can be conducted in-situ,i.e. by imaging the changing surfacestructure in the dissolution mediumwhich provides information about thedynamics of the process. Or it can be conducted ex-situ,by viewing the resulting surfacetopography at different stages of theetching process by removing thesample from solution and conducting

the imaging in air. Since the AFMprovides this information at thenanometer-to-micron scale, it canprovide information about thedissolution mechanisms at the molecularscale. For these reasons, it isincreasingly being utilized bypharmaceutical researchers andmanufacturers.6–9

Dissolution Studies

The influence of different solvents onacetaminophen single crystals wasinvestigated by AFM to gain a betterunderstanding of the dissolutionmechanisms at the molecular level.7

Partial dissolution tests were conductedon the (010) crystal surface with water,dichloroethane, pyridine, acetone, ethyl acetate, and acetic anhydride.Dissolution patterns were produced that are related to the interactionbetween the solvent molecules and thecrystal structure.

As an example, Figure 2 showsparallelogram, square, and hexagonaldissolution patterns formed by water,acetone, and pyridine, respectively. Inmany cases, the thickness of the stepsseen in these images is approximately10 angstroms, which is consistent withthe dissolution of a layer of unit cells.The dissolution process producedetching patterns that are related to the

a

bc

steps. Thus, the (100) surface has agreater surface area, which maycontribute to the faster dissolution rate.

Further investigation of these surfaceswas conducted by the functionalizationof the AFM probes with –CH3terminated groups and –COOHterminated groups to producehydrophobic and hydrophilic sensors,respectively. Using functionalizedprobes to study the chemicalinteractions with the surface iscommonly referred to as ChemicalForce Microscopy (CFM)10. By usingamplitude-phase distance relationshipsbetween the crystal planes and thefunctionalized probes, it was observedthat the hydrophobic –CH3 terminatedprobes had a strong attractiverelationship with the (001) surface,whereas the hydrophilic –COOHterminated probe had a strongattractive relationship with the (100)surface. Thus, the strong hydrophilic(100) surface would lead to easierwetting in an aqueous environmentresulting in a faster dissolution rate.

The dynamics of the dissolutionprocess of aspirin (001) and (100)crystal planes were further tested byconducting real-time AFM dissolutionstudies in a 0.05M HCl solution.6

By conducting the imaging duringdissolution, the dissolution mechanismsand rate were studied. For the (001)plane, dissolution occurred by therecession of molecular step edges,

Figure 3.TappingMode images of aspirin crystal surface. a) (001) crystal surface showing 7.3Å monomolecular steps, b) (100) crystal surface irregularsteps ranging from 50 to 100Å9. 4µm height (left) and Phase (right) images. Images courtesy C.J. Roberts, Univ. of Nottingham, UK.

Figure 4. In situ dissolution of the (001) in 0.05M HCl by receding step edges. The white arrowshows the progression of the same crystal terrace in each images at times of a) 0, b) 66, c) 132,d) 197, e) 263, and f) 329 seconds.6 10µm scans. Images courtesy C. Roberts, University ofNottingham, UK.

resulting in a dissolution rate of 0.45nm/s (Figure 4). For the (100) plane,dissolution occurred by the crystalterraces sinking at a rate of2.93nm/s. Intrinsic dissolution rateswere determined to be 1.35 x 10-7

g s -1 cm-2 and 8.36 x 10 -7 g s-1 cm-2

for the (001) and (100) planes,respectively. Thus, the (100) planedissolved at a rate six times greaterthan the (001) plane.

This difference in the dissolution ratecould be contributed partially to themolecular orientation of the crystalfaces. The (001) crystal face hasbenzene rings at the surface, which

are not easily removed in the 0.05MHCl solution, whereas, the ester groupsare exposed as the terraces on the(001) face, which are more easilyremoved, resulting in dissolution of thereceding step edges. The exposure ofmany ester groups at the surface of the(100) crystal face leaves all points onthe surface susceptible to removalduring dissolution, resulting in a fasterdissolution rate and a differentmechanism for material removal. Also,as discussed in the previous example,the increased surface roughness andthe hydrophilic nature of the (100)surface contribute to a faster dissolutionrate as well.

a b