Surface Characterization of Aspirin Crystal Planes by Dynamic

Anal. Chem. 2000, 72, 3419-3422

Articles

Surface Characterization of Aspirin Crystal Planes by Dynamic Chemical Force Microscopy Ardeshir Danesh,† Martyn C. Davies,† Steven J. Hinder,† Clive J. Roberts,†,* Saul J. B. Tendler,† Phil M. Williams,† and Michael J. Wilkins§

Laboratory of Biophysics and Surface Analysis, School of Pharmaceutical Sciences, University of Nottingham, University Park, Nottingham NG7 2RD, U.K., and SmithKline Beecham Pharmaceuticals, New Frontiers Science Park, Harlow, Essex, U.K.

Tapping mode (TM) atomic force microscopy (AFM) has been applied in a novel fashion to characterize and distinguish the (001) and (100) surfaces of individual aspirin crystals. The surface characterization was achieved by amplitude-phase, distance (a-p,d) measurements employing gold-coated AFM probes functionalized with selfassembled monolayers (SAM). Experiments using model probes coated with sCH3 and sCOOH terminated SAMs have been performed on the two aspirin crystal planes (001) and (100). Results indicate that the hydrophobic sCH3 terminated AFM probes had a greater degree of interaction with the crystal plane (001), whereas the s COOH terminated AFM probes had a larger interaction with the crystal plane (100). Interpretation of these data, based upon the chemistries of the probes, correlates with current understanding of the crystal surface chemistry derived from X-ray diffraction data and dissolution rate studies. The recrystallization environment of a crystalline material can often significantly influence the crystallization process and can be used to control the crystal structure (polymorphs), particle size, shape, and properties.1 Aspirin2 (acetyl salicylic acid) is one example where both crystal shape and the dissolution behavior are dependent upon the type of solvent used in its recrystallization.3 It is well documented in the literature that the dissolution velocity of aspirin crystals in water recrystallized from a nonpolar solvent (e.g., hexane) is approximately 50% faster than that when recrystallized from a polar solvent (e.g., ethanol).4 This difference was originally explained by the possible existence of two different crystal structures (polymorphs) as suggested by Tawashi.5 However, such findings were not substantiated because of insufficient * Corresponding author: (Tel.) (44) 115 9515048, (fax) (44) 115 9515110, (e-mail) [email protected]. † University of Nottingham. § SmithKline Beecham Pharmaceuticals. (1) Halebian, J. J. Pharm. Sci. 1975, 64, 1269-1287. (2) Florey, K. Anal. Profiles Drug Subst. 1979, 8, 1-5. (3) Watanabe, A.; Yamaoka, Y.; Takada, K. Chem. Pharm. Bull. 1982, 30, 29582963. (4) Yesook, K.; Matsumoto, M.; Machida, K. Chem. Pharm. Bull. 1985, 33, 4125-4131. (5) Tawashi, R. Science (Washington, D.C.) 1968, 8, 76. 10.1021/ac991498u CCC: $19.00 Published on Web 06/27/2000

© 2000 American Chemical Society

physical data and a lack of any X-ray diffraction evidence.6-9 Other studies have indeed illustrated that the differences observed in the dissolution velocities of aspirin recrystallized from different solvents may be due to the extent of expression of different crystal planes.3,4,10 Dissolution studies of different aspirin crystal planes have revealed a dissolution rate at least 50% larger for crystal plane (100) compared with that for crystal plane (001).4,11 In addition, morphological examination of aspirin crystals reveals plate-shaped crystals (where crystal plane (001) has had optimum growth) and thin needle-shaped crystals (where crystal plane (100) has had optimum growth), as a result of recrystallization from ethanol and hexane, respectively.3,10 X-ray diffraction analysis of aspirin (recrystallized from ethanol) has identified a monoclinic space group P21/c, with the following unit cell parameters: a ) 11.430 Å, b ) 6.591 Å, c ) 11.395 Å, β ) 95.68°.12 In this work, we use tapping mode atomic force microscopy (TM-AFM) for the rapid in situ qualitative characterization of aspirin crystal planes and discuss the implications of the results for interpreting their dissolution behavior. AFM13 has become an important tool for imaging14,15 and for the direct measurement of discrete intermolecular forces.16-21 The development of TM(6) Pfeiffer, R. J. Pharm. Pharmacol. 1971, 23, 75-79. (7) Mitchell, A. G.; Milaire, B. L.; Saville, D. J.; Griffiths, R. V. J. Pharm. Pharmacol. Lett. 1971, 23, 534-537. (8) Schwartzman, G. J. Pharm. Pharmacol. 1972, 24, 169-173. (9) Payne, R. S.; Rowe, R. C.; Roberts, R. J.; Charlton, M. H.; Docherty, R. J. Comput. Chem. 1999, 20, 262-273. (10) Meenan, P. ACS Symposium Series 667; American Chemical Society: Washington, DC, 1997; Chapter 1. (11) Masaki, N.; Machida, K.; Kado, H.; Yokoyama, K.; Tohoda, T. Ultramicroscopy 1992, 42, 1148-1154. (12) Yesook, K.; Machida, K.; Taga, T.; Osaki, K. Chem. Pharm. Bull. 1985, 33, 2641-2647. (13) Binnig, G.; Quate, C. F.; Gerber, C. Phys. Rev. Lett. 1986, 56, 930-933. (14) Butt, H. J.; Jaschke, M.; Ducker, W. Bioelectrochem. Bioenerg. 1995, 38, 191-201. (15) Hansma, H. G.; Hoh, J. H. Annu. Rev. Biophys. Biomol. Struct. 1994, 23, 115-139. (16) Dammer, U.; Popescu, O.; Wagner, P.; Anselmetti, D.; Guntherodt, H. J.; Misevic, G. Science (Washington, D.C.) 1995, 267, 1173-1175. (17) Dammer, U.; Hegner, M.; Anselmetti, D.; Wagner, P.; Dreier, M.; Huber, W.; Guntherodt, H. J. Biophys. J. 1996, 70, 2437-2441. (18) Florin, E. L.; Moy, V. T.; Gaub, H. E. Science (Washington, D.C.) 1994, 264, 415-417. (19) Allen, S.; Chen, X.; Davies, J.; Davies, M. C.; Dawkes, A. C.; Edwards, J. C.; Roberts, C. J.; Sefton, J.; Tendler, S. J. B.; Williams, P. M. Biochemistry 1997, 36, 7457-7463.

Analytical Chemistry, Vol. 72, No. 15, August 1, 2000 3419

AFM22-24 has generally overcome the problems of sample damage, which can occur when imaging soft or weakly adsorbed samples in contact mode AFM. In addition, the availability in TM-AFM of simultaneously acquired phase data provides further surface related physicochemical information. By mapping the phase shift of cantilever oscillation compared with the driving signal during a tapping mode scan, phase imaging facilitates detection of variations in composition, adhesion, friction, and viscoelasticity.25 Further surface-related information may also be obtained from selected points on a sample surface by measuring the amplitude and/or phase lag of cantilever oscillation against tip-sample distance. We refer to such data as amplitude-phase, distance (ap,d) measurements.25 It is worth noting that, since this work was carried out in air under ambient conditions, the alternative more quantitative AFM approach of studying probe-sample interactions by force-distance measurements is not possible due to the greater susceptibility of this method to the masking effect of capillary forces. On a practical level, the relatively stiff AFM cantilevers employed for TM-AFM imaging in air do not provide highresolution force-distance data. We have previously demonstrated how phase imaging and a-p,d measurements obtained in TM-AFM may be used to discriminate between two polymorphs of the drug cimetidine.25 In this paper, we illustrate how a-p,d measurements using chemically modified probes can be applied to characterize the surface chemistry of individual aspirin crystal planes. The deliberate chemical modifications of the AFM probes, chemical force microscopy (CFM),26,27 has proven to be a useful tool for characterizing surface chemistry by measurements of intermolecular interactions at nanometer scale.28-30 Although CFM is typically used in contact mode AFM, Leiber et al. have employed CFM in TM-AFM.31 Finot and McDermott have employed phase imaging in TM-AFM to map surface chemistry.32 Utilizing selfassembled monolayer (SAM) films of alkanethiols,33,34 we exploit a-p,d measurements in a dynamic mode of CFM. The function(20) Cappella, B.; Baschieri, P.; Frediani, C.; Miccoli, P.; Ascoli, C. IEEE Eng. Med. Biol. 1997, 58-65. (21) Allen, S.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M. Trends Biotechnol. 1997, 15, 101-105. (22) Zhong, Q.; Inniss, D.; Kjoller, K.; Elings. V. B. Surf. Sci. 1993, 290, L688L692. (23) Magonov, S. N.; Elings, V.; Whangbo, M. H. Surf. Sci. 1997, 375, L385L391. (24) Akari, S. O.; Van der vegte, E. W.; Grim, P. C. M.; Belder, G. F.; Koutsos, V.; Ten Brinke, G.; Hadziioannous, G. Appl. Phys. Lett. 1994, 65, 19151917. (25) Danesh, A.; Chen, X.; Davies, M. C.; Roberts, C. J.; Sanders, G. W. H.; Tendler, S. J. B.; Williams, P. M.; Wilkins, M. J. Langmuir 2000, 16, 866870. (26) Frisbie, C. D.; Rozsnynai, L. F.; Noy, A.; Wrighton, M. S.; Lieber, C. M. Science (Washington, D.C.) 1994, 265, 2071-2078. (27) Noy, A.; Frisbie, C. D.; Rozsnynai, L. F.; Wrighton, M. S.; Lieber, C. M. J. Am. Chem. Soc. 1995, 117, 7943-7951. (28) Chen, X.; Patel, N.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M.; Davies, J.; Dawkes, A. C.; Edwards, J. C. Appl. Phys. A 1998, 66, S631-S634. (29) Clear, S. C.; Nealey, P. F. Journal Colloid Interface Sci. 1999, 213, 238250. (30) Beake, B. D.; Leggett, G. J. Phys. Chem. Chem. Phys. 1999, 1, 3345-3350. (31) Noy, A.; Sanders, C. H.; Vezenov, D. V.; Wong, S. S.; Lieber, C. M. Langmuir 1998, 14, 1508-1511. (32) Finot, M. O.; McDermott, M. T. J. Am Chem. Soc. 1997, 119, 8564-8565. (33) Bain, C. D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-335. (34) Poirer, G. E.; Pylant, E. D. Science (Washington, D.C.) 1996, 272, 11451148.

3420

Analytical Chemistry, Vol. 72, No. 15, August 1, 2000

alized gold-coated AFM probes were used as hydrophobic or hydrophilic sensors for measuring the degree of interaction with each crystal plane. EXPERIMENTAL SECTION Materials. Absolute ethanol (Hayman, UK), dodecanethiol (Aldrich, Dorset, UK), 11-mercapto-1-undecanoic acid (Aldrich) were used as received. Aspirin crystals were obtained from recrystallization of USP specification aspirin at room temperature from a solution of 95% ethanol.5 The crystals were immobilized on AFM sample stubs using double-sided adhesive tape. Initially the crystal was mounted so that the narrowest plane (crystal plane (100)) was accessible to the AFM probe. The same crystal was then laid flat with the biggest plane (crystal plane (001)) made accessible. AFM measurements were performed using a Nanoscope IIIa MultiMode AFM (Digital Instruments, Santa Barbara, CA). All images were acquired in air using tapping mode with the E-type scanner (10 µm × 10 µm × 2 µm) and silicon TESP probes (Digital Instruments). These are 125-µm-long rectangular-shaped cantilevers with a nominal force constant and resonant frequency of approximately 50 N/m and 300 kHz, respectively. All images were taken at a scan rate between 1 and 2 Hz, with a 512 × 512 pixel resolution. The surface of each crystal plane was initially imaged using standard silicon probes, then a-p,d measurements were taken using SAM functionalized gold-coated probes. Preparation of Monolayer Functionalized Probes. Gold coated AFM probes were immersed into 1 mM ethanolic solutions of dodecanethiol or 11-mercaptoundecanoic acid for 24 h before being rinsed in ethanol, forming sCH3 and sCOOH terminated SAM surfaces, respectively. The influence of humidity on a-p,d measurements was investigated by performing measurements on the same area at normal relative humidity (R.H. ) 29%) and at low humidity (R.H. ) 4.5%). Humidity was reduced by enclosing the apparatus in an airtight enclosure containing dried silica gel and was monitored using a humidity and temperature indicator HMI 31 (Vaisala, Finland). To allow the comparison of obtained results, the drive frequency was kept at the cantilever’s oscillation resonance, which was determined at tip-sample distances below 100 nm. To confirm reproducibility of the data, the experiments were repeated with different tips, but the data used for comparison purposes are obtained using the same tip. RESULTS AND DISCUSSION Typical AFM images of crystal planes (001) and (100) are shown in Figures 1 and 2, respectively. The step edges (indicated by an arrow) in Figure 1a (crystal plane (001)) are uniform in appearance and have a vertical height of 7.3 Å, which is comparable with the size of an aspirin molecule (7.0 Å), suggesting a single molecular layer. However, the step edges in Figure 2a (crystal plane (100)) are irregular in appearance and range in height from 50 to 100 Å. These steps on the surface of crystal plane (100) contribute to a higher surface roughness than the (001) surface. The a-p,d curves obtained from the surface of aspirin crystal face (001) and face (100) are demonstrated in Figure 3. At point a in Figure 3a, the cantilever is vibrated at its resonant frequency (away from the sample surface) and is moving toward the surface

Figure 1. 4 µm × 4 µm TM-AFM image of aspirin crystal plane (001) (a) height image (Z range ) 10 nm), (b) corresponding phase image (Z range ) 5 degrees).

Figure 2. 4 µm × 4 µm TM-AFM image of aspirin crystal plane (100) (a) height image (Z range ) 100 nm), (b) phase image (Z range ) 60 degrees).

at a steady rate (thicker line) with no observable changes in its amplitude or phase. At point b the cantilever is close enough to the surface that it experiences forces of both attractive and repulsive nature, such as hydrophobic, electrostatic, van der Waals, and capillary force interactions.35-39 The region between points b and c is described as the attractive section of the a-p,d curve because the phase shift indicates the dominant forces are attractive.35,36,40,41 In this region, the amplitude dampens and the phase lag increases (positive phase shift) as the tip moves toward the surface. At point c, the phase lag starts to decrease, indicating a change in tip-sample interactions from attractive-dominated to repulsive-dominated.35,36,40,41 The region between points c and d is described as the repulsive section. The principle repulsive force is the elastic response of the sample to an increasing indentation by the tip37 and to a smaller extent of van der Waals forces.39 In this region, both the amplitude and phase lag continue to decrease, (negative phase shift) as the tip continues to move forward, until point d, a predetermined point of maximum amplitude reduction. The tip is then retracted from the sample surface, and the amplitude and phase shift of the tip are depicted by the thinner line. (35) Garcia, R.; San Paulo, A. Ultramicroscopy 2000, 82, 79-83. (36) Anczykowski, B.; Cleveland. J. P.; Kruger, D.; Elings, V.; Fuchs, H. Appl. Phys. A 1998, 66, S885-S889. (37) Kuhle, A.; Sorensen, A. H.; Bohr, J. J. Appl. Phys. 1997, 81, 6562-6569. (38) Behrend, O. P.; Ouleney, F.; Gourdon, D.; Dupas, E.; Kulik, A. J.; Germand, G.; Burnham, N. A. Appl. Phys. A. 1998, 66, S219-S221. (39) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: London, 1991, Chapter 1. (40) Chen, X.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M.; Davies, J.; Dawkes, A. C.; Edwards, J. C. Ultramicroscopy 1998, 75, 171182. (41) Kuhle, A.; Sorensen, A. H.; Zandbergen, J. B.; Bohr, J. Appl. Phys. A 1998, 66, 329-332.

Figure 3 parts a and b are a-p,d curves acquired using dodecanethiol-coated AFM probes from crystal planes (001) and (100), respectively. The a-p,d measurements from crystal plane (001) (Figure 3a) have a much longer attractive region, indicating the greater affinity of the sCH3 terminated SAM for this crystal plane and the presence of stronger attractive-dominated interactions. The interaction between 11-mercapto-1-undecanoic acid SAM coated AFM probes and crystal planes (001) and (100) are shown in parts c and d of Figure 3, respectively. Here, the reverse has occurred and a large attractive region is present for a-p,d measurements obtained from crystal plane (100) and a sCOOH group terminated SAM functionalized AFM probe. It is clear that the a-p,d measurements appear to indicate a difference in the surface chemical structure of the aspirin crystal planes (001) and (100). To study this effect further, previous X-ray crystallographic investigations of aspirin12 were reviewed revealing that aspirin molecules exist as dimers in the solid state, through intermolecular hydrogen bonding between the carboxylic acid groups. The major crystal planes are (001) and (100), with the former being primarily dominated by the phenyl and methyl functional groups, whereas the latter has methyl and oxygen groups exposed (Figure 4) (adapted from Figure 4 Masaki et al.11). These data can be used to explain the differences observed in a-p,d curves, where a larger attractive region due to the hydrophobic interactions is present between the sCH3 terminated SAM coated probes and the phenyl, methyl dominated crystal plane (001). The opposite is true for sCOOH terminated SAM functionalized probes, where hydrogen bonding between the acid group and the oxygen of the ester group exposed at crystal plane (100) is believed to be responsible for the large attractive region. The sCOOH group is expected to remain in nonionized state;42 therefore, the interactions observed are mainly nonionic. Since these measurements were performed in air, the contributions of surface-adsorbed water and capillary force needed to be assessed. The a-p,d measurements obtained at low relative humidity (R.H. ) 4-6%) (data not shown) illustrate that the profiles of these curves remain the same; the only difference observed was a reduction in the length of the attractive region, which was of a similar extent in all cases. Another potential factor is the use of the dodecanethiol or 11-mercaptoundecanoic acid SAMs on the AFM probe (as have been employed by others31,43,44), which raises the possibility of the mechanical properties of the SAM affecting the observed interactions. For example, the degree of crystallinity, or liquidlike behavior, of such films on AFM probes has yet to be assessed, although similar films on gold nanoparticles have been found to be liquid-like.44 Here, through the use of the same methyl- or acid-terminated probe to interrogate the different aspirin crystal planes, effects due to the SAM chain length may be removed. Therefore, the observed differences in the a-p,d measurements are due to distinct variations in the surface chemistry of the crystal planes, rather than an influence of any surface-adsorbed water. The data obtained from such surface chemistry characterization may contribute to an understanding of the noted differences (42) Kane, V.; Mulvaney, P. Langmuir 1998, 14, 3303-3311. (43) He, H. X.; Li, C. Z.; Song, J. Q.; Mu, T.; Wang, L.; Zhang, H. L.; Liu, Z. F. Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A 1997, 294, 99-102. (44) Templeton, A. C.; Hostetler, M. J.; Kraft, C. T.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 1906-1911.

Analytical Chemistry, Vol. 72, No. 15, August 1, 2000

3421

Figure 3. a-p,d curves from the surface of (a) crystal plane (001) using sCH3 terminated SAM, (b) crystal plane (100) using sCH3 terminated SAM, (c) crystal plane (001) using sCOOH terminated SAM, and (d) crystal plane (100) using sCOOH terminated SAM.

parallel with readjustment of the relative areas of the developed faces in such a way as to minimize the total surface energy under the new solvent-crystal interactions.3 Therefore, since the readjustments of the relative areas of a single crystal recrystallized from hexane are larger when it is immersed in water than they are when it is recrystallized from ethanol, this may lead to its higher rate of dissolution.

Figure 4. Crystal structure of aspirin depicted along the b-axis projection. Horizontal and longitudinal dotted lines indicate (001) and (100) planes, respectively. (adapted from Figure 4, N. Masaki et al.11)

in dissolution velocity of the two crystal planes. First are differences in surface roughness, which lead to differences in surface area. Crystal plane (100) exhibits a greater surface roughness than crystal plane (001) and therefore has a larger surface area exposed, which contributes to its faster rate of dissolution. This surface roughness data also correlates with that of Masaki et al.11 who reported a disordered surface on crystal plane (100) compared with that on (001). Second, the differences in the hydrophilic and hydrophobic natures of the two surfaces may lead to easier wetting of the more hydrophilic surface in an aqueous environment and hence the higher rate of dissolution. Third, when a single crystal recrystallized from a nonpolar solvent is immersed in water, the dissolution of the crystal as a whole takes place in 3422 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000

CONCLUSIONS Utilizing AFM probes functionalized with two model surfaces, sCH3 and sCOOH, we have demonstrated that SAM functionalized probes can be employed to characterize different planes of individual aspirin crystals by a-p,d measurements with TM-AFM. The a-p,d measurements showed that crystal plane (001) had a higher degree of interaction for the hydrophobic sCH3 terminated SAM functionalized probe and crystal plane (100) had a stronger interaction with the hydrophilic sCOOH terminated SAM functionalized probe. These observations demonstrate the ability of the technique in distinguishing surface chemistry of the crystal planes and illustrate its contribution to our understanding of crystal structures at a molecular level. In addition to complementing the data obtained using bulk techniques such as X-ray diffraction crystallography, this type of surface-property investigation provides an insight into the dissolution behavior of crystalline materials. ACKNOWLEDGMENT The authors thank SmithKline Beecham and the EPSRC for the provision of a studentship for A.D. We acknowledge the help of Dr. Barrie Kellam with recrystallization experiments. Received for review December 30, 1999. Accepted March 31, 2000. AC991498U