- Email: [email protected]
Investigation into the Effect of Humidity on Drug–Drug Interactions Using the Atomic Force Microscope
Investigation into the Effect of Humidity on Drug–Drug Interactions Using the Atomic Force Microscope PAUL M. YOUNG,1 ROBERT PRICE,1 MICHAEL J. TOBYN,1 MARK BUTTRUM,2 FIONA DEY2 1
Pharmaceutical Technology Research Group, Department of Pharmacy and Pharmacology, University of Bath, BA2 7AY, UK
2
Respiratory Technology, Aventis Pharma, Holmes Chapel, CW4 8BE, UK
Received 25 February 2002; revised 24 June 2002; accepted 24 June 2002
ABSTRACT: The atomic force microscope (AFM) has been used to characterize the cohesive nature of a micronized pharmaceutical powder used for inhalation therapy. Salbutamol sulfate (also referred to as albuterol sulfate), a therapeutic drug commonly delivered from dry powder inhalers (DPI), was chosen as a model system because the cohesion and subsequent de-agglomeration during inhalation are critical aspects to the efficacy of such a delivery system. Salbutamol sulfate drug particulates were mounted on V-shaped AFM cantilevers using a novel micromanipulation technique. Force–distance curves obtained from the measurements between cantilever drug probes and model compacts of salbutamol sulfate were integrated to determine separation energies. The effect of humidity (15–75% RH) on the energy required to separate a drug particle from model drug surface was determined using a custom-built perfusion apparatus attached to the AFM. Separation energy measurements over 10 10-mm areas of the compact surface (n ¼ 4096) exhibited log normal distributions (apparent linear regression, R2 0.97). Significant increases in the median separation energies ( p < 0.05) between the salbutamol sulfate drug probes and salbutamol sulfate model surfaces were observed as humidity was increased. This result is most likely attributed to capillary interactions becoming more dominant at higher humidities. This investigation has shown the AFM to be a powerful technique for quantification of the separation energies between micronized drug particulates, highlighting the potential of the AFM as a rapid preformulation tool. ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 92:815–822, 2003
Keywords: volume
dry powder aerosolization; humidity; atomic force microscope; force
INTRODUCTION The delivery of dry powder particulates to the deep lung is becoming an increasingly important route for the therapeutic treatment of asthma and other bronchial-related diseases. In general terms, the efficacy of these systems is dependent on particle size. To achieve a therapeutic effect, the powder should typically have a mass median Correspondence to: Paul M. Young (Telephone: 44 (0)1225 826826, ext 4831; Fax: 44 (0)1225 826114; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 92, 815–822 (2003) ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association
aerodynamic particle diameter of 5 mm or less to avoid impaction and/or sedimentation in the upper respiratory tract.1 Particulates within this sub-5-mm size are often formed by size reduction of larger crystals by high energy milling processes (micronization). The process of micronization produces high surface area materials with potentially large interparticulate forces, which may have a profound effect on particle–particle interactions in a dry powder delivery system. Three primary interparticulate forces are of importance in dry powder delivery systems; van der Waals dispersive forces, electrostatic forces, and capillary interactions. The contribution of
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
815
816
YOUNG ET AL.
each of these forces to the total interaction is dependent on the material properties and environmental conditions, such as temperature and relative humidity (RH). However, quantitative information about such interactions is often difficult to obtain. Such fundamental measurements of these interactions have previously been limited to bulk techniques, such as entrainment and centrifugal studies.2–4 Again, these techniques often provide empirical data of limited fundamental value. Here, the atomic force microscope (AFM) has been identified as a tool capable of measuring the interactions between individual particulates and substrates at specific environmental conditions. The AFM was developed in 1986, and is commonly used for the high-resolution topographical imaging of surfaces.5 Another application of the AFM is the measurement of forces between a tip and surface under specific media. Specific details of this technique are well documented and can be found in the literature.5–7 In simple terms, the AFM consists of a microfabricated cantilever whose deflection is recorded (usually by laser reflection) as it interacts with a moving substrate. Direct measurement of force can be determined by ramping the tip vertically (z axis), towards and away from a sample. The relative deflection of the cantilever as the tip approaches, makes contact with, and is removed from a surface produces a force–distance curve. The change in hysteresis in cantilever deflection during approach and retraction can be related to the magnitude and type of forces acting between the two surfaces.8 A schematic representation of an ideal force– distance curve is shown in Figure 1. Integration of the area within the adhesion portion of a force– distance curve will be equal to the energy of separation. The AFM force measurement approach can be further modified by incorporating a colloidal material onto the cantilever tip, allowing direct measurement of interactive forces between a particulate and substrate. Previous investigations using cantilevermounted colloidal probes have mainly focused on the physical interactions between inorganic polymeric spheres and atomically flat substrates.9,10 However, the technique has rapidly diversified into areas such as biotechnology8 and xerography.11 Recently, investigators have reported the AFM colloid probe technique as a useful tool in determining variations in adhesion between pharmaceutical systems.12–14 The use of the AFM to JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
determine interactions between lactose, a common pharmaceutical excipient, and gelatin capsules has shown capsule processing can effect particle– capsule interaction.12 Additionally, disparities in the adhesive forces between a silica sphere colloid probe and batches of lactose tablets have been reported.13 Furthermore, recent investigations have investigated lactose–lactose interactions using the colloid probe technique.14 The attachment of a drug particulate to the apex of a tipless cantilever may be a route for the measurement of drug–drug interactions. As part of a program to identify the interactions between such particles, an AFM has been used to study the interactions in a model drug system. Salbutamol sulfate, a therapeutic drug used for the immediate treatment of asthma, is commonly delivered in the form of a dry powder to the respiratory tract. Consequently, the degree of cohesiveness between salbutamol sulfate particles will directly influence the aerosolization performance of the drug when administered to a patient via a dry powder inhaler (DPI). Recent in vitro aerosolization studies have reported humidity to cause a significant decrease in aerosolization efficiency of micronized salbutamol sulfate from a model DPI.15,16 This decrease in aerosolization performance is possibly related to an increase of particulate cohesion due to capillary interactions. Such increases in the cohesion/ adhesion between micronized pharmaceutical
Figure 1. Schematic representation of an ideal force curve between an AFM cantilever drug probe and substrate: (a) approach of probe with no cantilever deflection, (b) jump to contact, (c) point of no applied force, (d) applied cantilever deflection equals applied deflection, (e) negative deflection due to probe adhesion, and (f) retraction of probe with no cantilever deflection.
ATOMIC FORCE MICROSCOPY CHARACTERIZATION
materials at elevated humidities were also demonstrated using the centrifugal technique,3,17 where the adhesion of micronized drug and/or excipient particles were removed from drug or excipient compact surfaces as a function of humidity. Previous studies to investigate capillary interactions have been conducted using the AFM.18,19 However, these studies have generally concentrated on either tip–substrate interactions18 (tip radius of curvature, 20–40 nm) or larger pharmaceutical drug probe–substrate interactions19 (probe diameter >100 mm). To quantify the behavior of salbutamol sulfate particulates in a DPI system, a study was undertaken to generate a cohesion profile for the micronized (5 mm) drug at varied humidity using the AFM colloid probe technique. Multiple force– distance curves between AFM cantilever-mounted drug particulates and model drug compacts (across prespecified areas) were conducted using a force volume technique. This technique provides a large matrix (n ¼ 4096) of separation energy values at each humidity, allowing comparison and statistical analysis.
MATERIALS AND METHODS Materials Micronized salbutamol sulfate (Aventis Pharma, Holmes Chapel, UK) was used as supplied. Physical Characterization The mass median aerodynamic particle diameter was determined using laser time of flight measurements (Aerosizer Mach 2, Amherst Process Instruments, Tewkesbury, UK). The Aerosizer was equipped with an API Aerodisperser device. Samples were analyzed over a 120-s period, at a high deagglomeration setting equivalent to a pressure drop across the orifice of 5 psi. Particle morphology was investigated by scanning electron microscopy (SEM; Jeol 6310, Jeol, Japan). Samples were deposited on a carbon sticky tab and gold coated (Edwards Sputter Coater, UK) prior to imaging at 10 KeV. Detailed topographical information of the salbutamol sulfate model compact was investigated using an AFM (Multimode SPM with Nanoscope III controller, Digital Instruments, UK). The model compacts were mounted on a carbon sticky tab and Imaged using AFM Tapping ModeTM with
817
a high aspect ratio silicon probe (OTESP, Digital Instruments, UK), at a scan rate of 0.7 Hz. Specific surface areas for the micronized salbutamol sulfate were measured by 5-point Brunauer–Emmett–Teller (BET) nitrogen adsorption (Gemini 2360, Micromeritics, Dunstable, UK) at 77 K in triplicate. Samples were dried for 24 h prior to analysis at 408C under a dry nitrogen stream (FlowPrep 060 Micromeritics, Dunstable, UK). Moisture sorption profiles of the salbutamol sulfate were determined by dynamic vapor sorption (DVS; DVS-1, Surface Measurement Systems Ltd., London, UK). Approximately 10 mg of sample was weighed into the sample cell and subjected to a 0–90% RH cycle; mass equilibration at each humidity was determined with a dm/dt of 0.002. Preparation of Drug Compacts The use of model compacts for the determination of interparticulate forces is well documented3,17 and was chosen to circumvent issues that would arise when measuring the adhesion between individual micronized particles. Model surfaces of salbutamol sulfate were prepared by direct compression (TA HDi Texture analyser, Stable Micro Systems, Surrey, UK). Approximately 250 mg of salbutamol sulfate was weighed into a 10-mm die and compacted at a compression rate of 0.5 mm s1 (500 kg; dwell time, 180 s). Prepared model surfaces were stored in tightly sealed containers for 24 h prior to use. Utmost care was taken when handling the model surfaces to avoid contamination. Preparation of Drug Probes A multistage micromanipulation method for the attachment of micrometer-sized drug particulates to tipless cantilevers was adapted from the method described by Preuss and Butt.20 Essentially, particle mounting was achieved by manipulating a small quantity of epoxy resin followed by a representative 5-mm drug particle onto a nominal 0.58 N/m spring constant, tipless cantilever (DNP-020 Digital Instruments, Cambridge, UK) using the custom-built microscope assembly show diagrammatically in Figure 2. In addition, a charge-couple device (CCD) video, reflection microscope (Digital Instruments, Cambridge, UK) with a 500 long working distance lens was used throughout the micromanipulation process to evaluate tip quality, quantity of glue, and tip JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
818
YOUNG ET AL.
Figure 2. Schematic diagram of the micromanipulation PTFE slide holder assembly.
particle integrity prior to and post curing. Prepared drug probe tips were stored in tightly sealed containers for 24 h prior to use. Possible variations in spring constant were minimized by obtaining a wafer of tipless cantilevers (>500 tips), so batch-to-batch tip thickness concerns were eliminated. Randomly chosen tips (n ¼ 5) from across the wafer indicated <14% variance in spring constant using the thermal method.21
conducted over a 10 10-mm area with the following settings: approach–retraction cycle, 2 mm; cycle rate, 8.14 Hz; and constant compliance region, 60 nm. Experimental validation of these settings indicated neither cycle time (0.1–10 Hz) or constant compliance loading (1–100 nm) affected force curve measurements. Such variables have directly influenced colloid adhesion in previous studies,9 but this result was not found here and was not particularly expected. Salbutamol sulfate is a relatively brittle, crystalline material and would not appreciably deform under low stress. The area under each curve was integrated using custom-built batch conversion software and exported to produce a 64 64 block of separation energy values corresponding to the X and Y positions of the force distance curves. In addition, offline analysis of individual force curves was conducted for qualitative purposes because curve shape can indicate many interactive physical properties. A series of separation energy measurements between three salbutamol sulfate drug probes and three salbutamol sulfate model surfaces were conducted at 15, 30, 45, 60, and 75% RH.
RESULTS AND DISCUSSION Control of Relative Humidity The effect of RH on particle substrate separation energy was investigated using a custom-built perfusion apparatus connected to the AFM. Filtered nitrogen gas was split at source and passed independently through a dessicator, containing silica gel, and a bubble chamber with moisture trap. Both the dry and 100% humidified nitrogen were passed through separate needle-valve flow controllers and fed into a sealed AFM scanning head containing a temperature and humidity monitor. Total flow into the AFM head was 1 cm3 min1 at 25 0.28C. Temperature of the humidified nitrogen was 258C to eliminate the possibility of condensation. Separation Energy Measurements Multiple force curves for the interactions between a salbutamol sulfate drug probe and salbutamol sulfate model surface were conducted by using the AFM in force volume mode (Multimode SPM with Nanoscope III controller, Digital Instruments, UK). Individual force curves (n ¼ 4096) were JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Physical Characterization To quantify the effect of humidity on particle– particle interactions, the powders were first characterized for particle size, morphology ,and moisture sorption. The Aerosizer mass median aerodynamic diameter of the salbutamol sulfate was 3.76 0.09 mm (n ¼ 3), suggesting the micronized drug was of suitable size for inhalation purposes.1 A representative scanning electron micrograph of the ‘‘as supplied’’ drug material and colloid drug probe are shown in Figures 3A and 3B, respectively. It can be seen from Figure 3A that the drug exhibits a micrometer-size (1–10 mm), columnar crystal shape. A photomicrograph of a typical AFM colloid drug probe (drug probe 1), imaged after the AFM study, is shown in Figure 3B. All of mounted drug particulates used in the AFM study exhibited approximate dimensions of 5 2 mm and clearly stood proud of the epoxy resin on the cantilever. A representative AFM topographical image of for the salbutamol sulfate model compact surface, shown in Figure 4, suggests an irregular surface
ATOMIC FORCE MICROSCOPY CHARACTERIZATION
819
Figure 4. Representative AFM topographical image of the salbutamol sulfate model compact surface (X, Y, Z ¼ 10 10 10 mm).
Figure 3. Representative SEM image of (A) salbutamol sulfate powder (5000) and (B) salbutamol sulfate drug probe immobilized on the tip of a pyramidal AFM cantilever (5000).
morphology with a low root mean square roughness (10 10 mm areas) of 101 16 nm (n ¼ 5). This result was of a similar order to the compact roughness reported in previous AFM studies.14,19 However, it should be noted that such values are dependent on both the technique used and methods employed. In addition, AFM images of the salbutamol sulfate (not shown here) at two extremes of humidity (15 and 75% RH) indicated no change in particle shape. Although previous investigations17 have demonstrated humidity to alter the general surface morphology of some pharmaceutical materials, this was not expected for salbutamol sulfate because DVS analysis indicated the surface to be relatively hydrophobic across the range 15–75% RH. DVS profiles, calculated as percentage of dry mass, for salbutamol sulfate are shown in Figure 5. The water sorption isotherms follow a sigmoidal
class L3 curve,22 suggesting multilayer water sorption onto the crystal surfaces. Monolayer coverage of water can be calculated from the slope and intercept of the linear BET single-point equation from data obtained around the first inflection. Values (R2 ¼ 1) of 0.06% w/w water for monolayer coverage and 2.24 m2 g1 for surface area were determined from the DVS data using a projected water molecule surface area of 11.13 1020 m2.23 These results correlated well with specific surface area measurements of 3.76 0.028 m2 g1 (n ¼ 3) obtained by nitrogen adsorption. Regression analysis indicates monolayer coverage (0.06% w/w) to occur at 23% RH, suggesting that separation energy measurements
Figure 5. Dynamic vapour sorption (DVS) isotherm of salbutamol sulfate. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
820
YOUNG ET AL.
Figure 6 as cumulative frequency separation energy values for salbutamol sulfate drug probe 1 (15–75% RH). The assumption of log normal distribution was further justified using apparent linear regression analysis of log–probability plots (5–95%), and full 4-parameter sigmoidal regression analysis, where R2 values of 0.97 were obtained using both methods. As a result, a median value, or 50% cumulative undersize (e0.5), was selected as the most appropriate descriptor for separation energy measurements because the large positive skew in the data may adversely affect the mean values. The effect of humidity on the e0.5 is shown graphically in Figure 7. In all cases, change of humidity had a significant effect on the e0.5 (CI (confidence interval) ¼ 0.95). Furthermore, a 6.4 1.1 times increase in mean e0.5 (n ¼ 3) occurred between 15 and 75% RH, indicating capillary forces become a dominating factor at higher humidities. Such capillary forces may be directly induced by the multilayer adsorption of water onto the salbutamol sulfate surfaces as humidity was increased (as indicated by DVS in Figure 5). Essentially, these observations suggest that the increase in particle cohesion at higher humidities lead to an increase in energy required for deagglomeration. For dry powder inhaler formulations, this may result in a decrease in fine particle fraction, as previously reported.15,16 It is important to note, however, that such measurements are undertaken between individual drug probes and model drug surfaces and therefore can not be directly related to the complex
conducted by AFM at 30, 45, 60, and 75% RH represented multilayer water coverage. Separation Energy Measurements Separation energy measurements between three salbutamol sulfate drug probes and salbutamol sulfate model compacts were investigated with the AFM at 15, 30, 45, 60, and 75% RH and are summarized in Table 1. Integration of the force distance curves measured over a 10 10-mm area at each humidity (n ¼ 4096) produced a relatively wide distribution of separation energy values. Furthermore, separation energy histogram analysis showed the data spread across more than one order of magnitude in an asymmetrical, positively skewed distribution. Analysis of the separation energy data block (64 64 force curves) suggested no relationship between force curve number and separation energy value. In addition, offline analysis of individual force curves conducted at the beginning, middle, and end of each run indicated no rank change in separation energy value or curve shape. Such observations would suggest that the salbutamol sulfate drug probe did not deform or appreciably tribocharge under such conditions. Furthermore, comparison of the separation energy data block values with the individual force curves, in offline analysis, indicated no surface ‘hotspots’ that may have been observed if contamination was present. Cumulative frequency–log distributions of the separation energy values indicated a log normal distribution. This distribution is demonstrated in
Table 1.
Salbutamol Sulfate Separation Energy Values Relative Humidity, %
Tip
Parameter a
e0.5 (nJ) GSDb R2 c e0.5 (nJ)a GSDb R2 c e0.5 (nJ)a GSDb R2 c
1 1 1 2 2 2 3 3 3
15
30
45
60
75
2153 2.3 0.997 1628 2.4 0.994 2091 4.0 0.976
3065 2.2 0.999 2325 2.4 0.973 3451 3.1 0.977
3650 2.2 0.998 5262 2.1 0.977 7067 3.2 0.972
4990 2.1 0.997 5981 2.1 0.997 8294 2.9 0.980
11188 2.1 0.997 12726 2.0 0.994 12779 2.7 0.975
a
e0.5, median separation energy (n ¼ 4096). Geometric standard deviation (e0.841/e0.159)0.5. R2, linear regression of lognormal distribution (e0.05 and e0.95).
b c
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
ATOMIC FORCE MICROSCOPY CHARACTERIZATION
821
CONCLUSION
Figure 6. Separation energy distribution for salbutamol sulfate drug probe on salbutamol sulfate model surface at 15, 30, 45, 60, and 75% RH at 25 < 0.28C (n ¼ 4096).
dynamics that exist during the aerosolization process. Furthermore, the use of compaction to produce a relatively smooth contiguous surface may alter the surface properties of the salbutamol sulfate. Because of these factors, the compaction rate and dwell time were specifically chosen to minimize such damage. It is believed that the possible recrystallization of any amorphous regions under low compaction rates was limited by the use of micronized salbutamol sulfate that had been stored for at least a year at 45% RH before the studies were conducted. Although it was possible that amorphous material existed on the drug surface, DVS traces of the raw micronized material (as shown in Figure 5) showed no such characteristic hysteresis indicating recrystallization.
Figure 7. (e0.5) median separation energy values for salbutamol sulfate drug probes on model drug compacts at 15, 30, 45, 60, and 75% RH (n ¼ 4096).
The relationship between humidity and the cohesion of salbutamol sulfate particles, as determined with the AFM, can be attributed to the relative contribution of the interparticulate forces that exist between the two particle contact areas. As the humidity is increased (15–75% RH), capillary forces become a dominating factor because of the multilayer sorption of water onto the crystal surfaces. Ultimately, this increase in separation energy will influence the aerosolization and subsequent efficacy of such powders and should be considered during formulation. This investigation has shown atomic force microscopy to be a powerful technique for quantification of the separation energies between micronized drug particulates. Additionally, the use of a custom-built perfusion apparatus, connected to the AFM, has allowed the measurement of these interactions in environments relative to pharmaceutical material performance.
REFERENCES 1. Pritchard JN. 2001. The influence of lung deposition on clinical response. J Aerosol Med 14:S19– S26. 2. Staniforth JN. 1995. Performance-modifying influences in dry powder inhalation systems. Aerosol Sci Technol 22:346–353. 3. Podczeck F, Newton JM, James MB. 1997. Influence of relative humidity of storage air on the adhesion and autoadhesion of micronized particles to particulate and compacted powder surfaces. J Colloid Interface Sci 187:484–491. 4. Clarke MJ, Peart J, Cagnani S, Byron PR. 2002. Adhesion of powders for inhalation: An evaluation of drug detachment from surfaces following deposition from aerosol streams. Pharm Res 19:322– 329. 5. Binnig G, Quate CF, Gerber C. 1986. Atomic force microscope. Phys Rev Lett 56:930–933. 6. Ducker WA, Senden TJ, Pashley RM. 1991. Direct measurement of colloidal forces using an atomic force microscope. Nature 353:239–241. 7. Butt H-J. 1992. Measuring electrostatic, van der Waals, and hydration forces in electrolyte solutions with an atomic force microscope. Biophys J 60: 1438–1444. 8. Heinz WF, Hoh JH. 1999. Spatially resolved force spectroscopy of biological surfaces using the atomic force microscope. Trends Biotechnol 17:143–150. 9. Biggs S, Spinks G. 1998. Atomic force investigation of the adhesion between a single polymer sphere JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
822
10.
11.
12.
13.
14.
15.
16.
YOUNG ET AL.
and a flat surface. J Adhesion Sci Technology 12: 461–478. Yalamanchili MR, Veeramasuneni S, Azevedo MAD, Miller JD. 1998. Use of atomic force microscopy in particle science and technology research. Colloids Surf, A 133:77–88. Mizes H, Ott M, Eklund E, Hays D. 2000. Small particle adhesion: Measurement and control. Colloids Surf, A 165:11–23. Ibrahim TH, Burk TR, Etzler FM, Neuman RD. 2000. Direct adhesion measurements of pharmaceutical particles to gelatin capsule surfaces. J Adhesion Sci Technol 14:1225–1242. Louey MD, Mulvaney P, Stewart PJ. 2001. Characterisation of adhesional properties of lactose carriers using atomic force microscopy. J Pharm Anal 25:559–567. Sindel U, Zimmermann I. 2001. Measurement of interaction forces between individual powder particles using an atomic force microscope. Powder Technol 117:247–254. Jashnani RN, Byron PR, Dalby RN. 1995. Testing of dry powder aerosol formulations in different environmental conditions. Int J Pharm 113:123–130. Jashnani RN, Byron PR. 1996. Dry powder aerosol generation in different environments: Performance comparisons of albuterol, albuterol sulfate, albuterol adipate and albuterol salt. Int J Pharm 130: 13–24.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
17. Podczeck F, Michael Newton J, James MB. 1997. Variations in the adhesion force between a drug and a carrier particles as a result of changes in the relative humidity of the air. Int J Pharm 149: 151–160. 18. Sedin DL, Rowden KL. 2000. Adhesion forces measured by atomic force microscopy in humid air. Anal Chem 72:2183–2189. 19. Berard V, Lesniewska E, Andres C, Pertuy D, Laroche C, Pourcelot Y. 2002. Dry powder inhaler: influence of humidity on topology and adhesion studied by AFM. Int J Pharm 232:213–224. 20. Preuss M, Butt H-J. 1998. Direct measurement of particle-bubble interactions in aqueous electrolyte: dependence on surfactant. Langmuir 14:3164– 3174. 21. Hutter JL, Bechhoefer J. 1993. Calibration of atomic-force microscope tips. Rev Sci Instrum 64:1868–1873. 22. Giles CH, MacEwan TH, Nakhwa SN, Smith D. 1960. Studies in adsorption. Part XI.* A system for classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J Chem Soc 3973–3993. 23. Webster CE, Drago RS, Zerner MC. 1998. Molecular dimensions for adsorptives. J Am Chem Soc 120: 5509–5516.
Pharmaceutical Technology Research Group, Department of Pharmacy and Pharmacology, University of Bath, BA2 7AY, UK
2
Respiratory Technology, Aventis Pharma, Holmes Chapel, CW4 8BE, UK
Received 25 February 2002; revised 24 June 2002; accepted 24 June 2002
ABSTRACT: The atomic force microscope (AFM) has been used to characterize the cohesive nature of a micronized pharmaceutical powder used for inhalation therapy. Salbutamol sulfate (also referred to as albuterol sulfate), a therapeutic drug commonly delivered from dry powder inhalers (DPI), was chosen as a model system because the cohesion and subsequent de-agglomeration during inhalation are critical aspects to the efficacy of such a delivery system. Salbutamol sulfate drug particulates were mounted on V-shaped AFM cantilevers using a novel micromanipulation technique. Force–distance curves obtained from the measurements between cantilever drug probes and model compacts of salbutamol sulfate were integrated to determine separation energies. The effect of humidity (15–75% RH) on the energy required to separate a drug particle from model drug surface was determined using a custom-built perfusion apparatus attached to the AFM. Separation energy measurements over 10 10-mm areas of the compact surface (n ¼ 4096) exhibited log normal distributions (apparent linear regression, R2 0.97). Significant increases in the median separation energies ( p < 0.05) between the salbutamol sulfate drug probes and salbutamol sulfate model surfaces were observed as humidity was increased. This result is most likely attributed to capillary interactions becoming more dominant at higher humidities. This investigation has shown the AFM to be a powerful technique for quantification of the separation energies between micronized drug particulates, highlighting the potential of the AFM as a rapid preformulation tool. ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 92:815–822, 2003
Keywords: volume
dry powder aerosolization; humidity; atomic force microscope; force
INTRODUCTION The delivery of dry powder particulates to the deep lung is becoming an increasingly important route for the therapeutic treatment of asthma and other bronchial-related diseases. In general terms, the efficacy of these systems is dependent on particle size. To achieve a therapeutic effect, the powder should typically have a mass median Correspondence to: Paul M. Young (Telephone: 44 (0)1225 826826, ext 4831; Fax: 44 (0)1225 826114; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 92, 815–822 (2003) ß 2003 Wiley-Liss, Inc. and the American Pharmaceutical Association
aerodynamic particle diameter of 5 mm or less to avoid impaction and/or sedimentation in the upper respiratory tract.1 Particulates within this sub-5-mm size are often formed by size reduction of larger crystals by high energy milling processes (micronization). The process of micronization produces high surface area materials with potentially large interparticulate forces, which may have a profound effect on particle–particle interactions in a dry powder delivery system. Three primary interparticulate forces are of importance in dry powder delivery systems; van der Waals dispersive forces, electrostatic forces, and capillary interactions. The contribution of
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
815
816
YOUNG ET AL.
each of these forces to the total interaction is dependent on the material properties and environmental conditions, such as temperature and relative humidity (RH). However, quantitative information about such interactions is often difficult to obtain. Such fundamental measurements of these interactions have previously been limited to bulk techniques, such as entrainment and centrifugal studies.2–4 Again, these techniques often provide empirical data of limited fundamental value. Here, the atomic force microscope (AFM) has been identified as a tool capable of measuring the interactions between individual particulates and substrates at specific environmental conditions. The AFM was developed in 1986, and is commonly used for the high-resolution topographical imaging of surfaces.5 Another application of the AFM is the measurement of forces between a tip and surface under specific media. Specific details of this technique are well documented and can be found in the literature.5–7 In simple terms, the AFM consists of a microfabricated cantilever whose deflection is recorded (usually by laser reflection) as it interacts with a moving substrate. Direct measurement of force can be determined by ramping the tip vertically (z axis), towards and away from a sample. The relative deflection of the cantilever as the tip approaches, makes contact with, and is removed from a surface produces a force–distance curve. The change in hysteresis in cantilever deflection during approach and retraction can be related to the magnitude and type of forces acting between the two surfaces.8 A schematic representation of an ideal force– distance curve is shown in Figure 1. Integration of the area within the adhesion portion of a force– distance curve will be equal to the energy of separation. The AFM force measurement approach can be further modified by incorporating a colloidal material onto the cantilever tip, allowing direct measurement of interactive forces between a particulate and substrate. Previous investigations using cantilevermounted colloidal probes have mainly focused on the physical interactions between inorganic polymeric spheres and atomically flat substrates.9,10 However, the technique has rapidly diversified into areas such as biotechnology8 and xerography.11 Recently, investigators have reported the AFM colloid probe technique as a useful tool in determining variations in adhesion between pharmaceutical systems.12–14 The use of the AFM to JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
determine interactions between lactose, a common pharmaceutical excipient, and gelatin capsules has shown capsule processing can effect particle– capsule interaction.12 Additionally, disparities in the adhesive forces between a silica sphere colloid probe and batches of lactose tablets have been reported.13 Furthermore, recent investigations have investigated lactose–lactose interactions using the colloid probe technique.14 The attachment of a drug particulate to the apex of a tipless cantilever may be a route for the measurement of drug–drug interactions. As part of a program to identify the interactions between such particles, an AFM has been used to study the interactions in a model drug system. Salbutamol sulfate, a therapeutic drug used for the immediate treatment of asthma, is commonly delivered in the form of a dry powder to the respiratory tract. Consequently, the degree of cohesiveness between salbutamol sulfate particles will directly influence the aerosolization performance of the drug when administered to a patient via a dry powder inhaler (DPI). Recent in vitro aerosolization studies have reported humidity to cause a significant decrease in aerosolization efficiency of micronized salbutamol sulfate from a model DPI.15,16 This decrease in aerosolization performance is possibly related to an increase of particulate cohesion due to capillary interactions. Such increases in the cohesion/ adhesion between micronized pharmaceutical
Figure 1. Schematic representation of an ideal force curve between an AFM cantilever drug probe and substrate: (a) approach of probe with no cantilever deflection, (b) jump to contact, (c) point of no applied force, (d) applied cantilever deflection equals applied deflection, (e) negative deflection due to probe adhesion, and (f) retraction of probe with no cantilever deflection.
ATOMIC FORCE MICROSCOPY CHARACTERIZATION
materials at elevated humidities were also demonstrated using the centrifugal technique,3,17 where the adhesion of micronized drug and/or excipient particles were removed from drug or excipient compact surfaces as a function of humidity. Previous studies to investigate capillary interactions have been conducted using the AFM.18,19 However, these studies have generally concentrated on either tip–substrate interactions18 (tip radius of curvature, 20–40 nm) or larger pharmaceutical drug probe–substrate interactions19 (probe diameter >100 mm). To quantify the behavior of salbutamol sulfate particulates in a DPI system, a study was undertaken to generate a cohesion profile for the micronized (5 mm) drug at varied humidity using the AFM colloid probe technique. Multiple force– distance curves between AFM cantilever-mounted drug particulates and model drug compacts (across prespecified areas) were conducted using a force volume technique. This technique provides a large matrix (n ¼ 4096) of separation energy values at each humidity, allowing comparison and statistical analysis.
MATERIALS AND METHODS Materials Micronized salbutamol sulfate (Aventis Pharma, Holmes Chapel, UK) was used as supplied. Physical Characterization The mass median aerodynamic particle diameter was determined using laser time of flight measurements (Aerosizer Mach 2, Amherst Process Instruments, Tewkesbury, UK). The Aerosizer was equipped with an API Aerodisperser device. Samples were analyzed over a 120-s period, at a high deagglomeration setting equivalent to a pressure drop across the orifice of 5 psi. Particle morphology was investigated by scanning electron microscopy (SEM; Jeol 6310, Jeol, Japan). Samples were deposited on a carbon sticky tab and gold coated (Edwards Sputter Coater, UK) prior to imaging at 10 KeV. Detailed topographical information of the salbutamol sulfate model compact was investigated using an AFM (Multimode SPM with Nanoscope III controller, Digital Instruments, UK). The model compacts were mounted on a carbon sticky tab and Imaged using AFM Tapping ModeTM with
817
a high aspect ratio silicon probe (OTESP, Digital Instruments, UK), at a scan rate of 0.7 Hz. Specific surface areas for the micronized salbutamol sulfate were measured by 5-point Brunauer–Emmett–Teller (BET) nitrogen adsorption (Gemini 2360, Micromeritics, Dunstable, UK) at 77 K in triplicate. Samples were dried for 24 h prior to analysis at 408C under a dry nitrogen stream (FlowPrep 060 Micromeritics, Dunstable, UK). Moisture sorption profiles of the salbutamol sulfate were determined by dynamic vapor sorption (DVS; DVS-1, Surface Measurement Systems Ltd., London, UK). Approximately 10 mg of sample was weighed into the sample cell and subjected to a 0–90% RH cycle; mass equilibration at each humidity was determined with a dm/dt of 0.002. Preparation of Drug Compacts The use of model compacts for the determination of interparticulate forces is well documented3,17 and was chosen to circumvent issues that would arise when measuring the adhesion between individual micronized particles. Model surfaces of salbutamol sulfate were prepared by direct compression (TA HDi Texture analyser, Stable Micro Systems, Surrey, UK). Approximately 250 mg of salbutamol sulfate was weighed into a 10-mm die and compacted at a compression rate of 0.5 mm s1 (500 kg; dwell time, 180 s). Prepared model surfaces were stored in tightly sealed containers for 24 h prior to use. Utmost care was taken when handling the model surfaces to avoid contamination. Preparation of Drug Probes A multistage micromanipulation method for the attachment of micrometer-sized drug particulates to tipless cantilevers was adapted from the method described by Preuss and Butt.20 Essentially, particle mounting was achieved by manipulating a small quantity of epoxy resin followed by a representative 5-mm drug particle onto a nominal 0.58 N/m spring constant, tipless cantilever (DNP-020 Digital Instruments, Cambridge, UK) using the custom-built microscope assembly show diagrammatically in Figure 2. In addition, a charge-couple device (CCD) video, reflection microscope (Digital Instruments, Cambridge, UK) with a 500 long working distance lens was used throughout the micromanipulation process to evaluate tip quality, quantity of glue, and tip JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
818
YOUNG ET AL.
Figure 2. Schematic diagram of the micromanipulation PTFE slide holder assembly.
particle integrity prior to and post curing. Prepared drug probe tips were stored in tightly sealed containers for 24 h prior to use. Possible variations in spring constant were minimized by obtaining a wafer of tipless cantilevers (>500 tips), so batch-to-batch tip thickness concerns were eliminated. Randomly chosen tips (n ¼ 5) from across the wafer indicated <14% variance in spring constant using the thermal method.21
conducted over a 10 10-mm area with the following settings: approach–retraction cycle, 2 mm; cycle rate, 8.14 Hz; and constant compliance region, 60 nm. Experimental validation of these settings indicated neither cycle time (0.1–10 Hz) or constant compliance loading (1–100 nm) affected force curve measurements. Such variables have directly influenced colloid adhesion in previous studies,9 but this result was not found here and was not particularly expected. Salbutamol sulfate is a relatively brittle, crystalline material and would not appreciably deform under low stress. The area under each curve was integrated using custom-built batch conversion software and exported to produce a 64 64 block of separation energy values corresponding to the X and Y positions of the force distance curves. In addition, offline analysis of individual force curves was conducted for qualitative purposes because curve shape can indicate many interactive physical properties. A series of separation energy measurements between three salbutamol sulfate drug probes and three salbutamol sulfate model surfaces were conducted at 15, 30, 45, 60, and 75% RH.
RESULTS AND DISCUSSION Control of Relative Humidity The effect of RH on particle substrate separation energy was investigated using a custom-built perfusion apparatus connected to the AFM. Filtered nitrogen gas was split at source and passed independently through a dessicator, containing silica gel, and a bubble chamber with moisture trap. Both the dry and 100% humidified nitrogen were passed through separate needle-valve flow controllers and fed into a sealed AFM scanning head containing a temperature and humidity monitor. Total flow into the AFM head was 1 cm3 min1 at 25 0.28C. Temperature of the humidified nitrogen was 258C to eliminate the possibility of condensation. Separation Energy Measurements Multiple force curves for the interactions between a salbutamol sulfate drug probe and salbutamol sulfate model surface were conducted by using the AFM in force volume mode (Multimode SPM with Nanoscope III controller, Digital Instruments, UK). Individual force curves (n ¼ 4096) were JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
Physical Characterization To quantify the effect of humidity on particle– particle interactions, the powders were first characterized for particle size, morphology ,and moisture sorption. The Aerosizer mass median aerodynamic diameter of the salbutamol sulfate was 3.76 0.09 mm (n ¼ 3), suggesting the micronized drug was of suitable size for inhalation purposes.1 A representative scanning electron micrograph of the ‘‘as supplied’’ drug material and colloid drug probe are shown in Figures 3A and 3B, respectively. It can be seen from Figure 3A that the drug exhibits a micrometer-size (1–10 mm), columnar crystal shape. A photomicrograph of a typical AFM colloid drug probe (drug probe 1), imaged after the AFM study, is shown in Figure 3B. All of mounted drug particulates used in the AFM study exhibited approximate dimensions of 5 2 mm and clearly stood proud of the epoxy resin on the cantilever. A representative AFM topographical image of for the salbutamol sulfate model compact surface, shown in Figure 4, suggests an irregular surface
ATOMIC FORCE MICROSCOPY CHARACTERIZATION
819
Figure 4. Representative AFM topographical image of the salbutamol sulfate model compact surface (X, Y, Z ¼ 10 10 10 mm).
Figure 3. Representative SEM image of (A) salbutamol sulfate powder (5000) and (B) salbutamol sulfate drug probe immobilized on the tip of a pyramidal AFM cantilever (5000).
morphology with a low root mean square roughness (10 10 mm areas) of 101 16 nm (n ¼ 5). This result was of a similar order to the compact roughness reported in previous AFM studies.14,19 However, it should be noted that such values are dependent on both the technique used and methods employed. In addition, AFM images of the salbutamol sulfate (not shown here) at two extremes of humidity (15 and 75% RH) indicated no change in particle shape. Although previous investigations17 have demonstrated humidity to alter the general surface morphology of some pharmaceutical materials, this was not expected for salbutamol sulfate because DVS analysis indicated the surface to be relatively hydrophobic across the range 15–75% RH. DVS profiles, calculated as percentage of dry mass, for salbutamol sulfate are shown in Figure 5. The water sorption isotherms follow a sigmoidal
class L3 curve,22 suggesting multilayer water sorption onto the crystal surfaces. Monolayer coverage of water can be calculated from the slope and intercept of the linear BET single-point equation from data obtained around the first inflection. Values (R2 ¼ 1) of 0.06% w/w water for monolayer coverage and 2.24 m2 g1 for surface area were determined from the DVS data using a projected water molecule surface area of 11.13 1020 m2.23 These results correlated well with specific surface area measurements of 3.76 0.028 m2 g1 (n ¼ 3) obtained by nitrogen adsorption. Regression analysis indicates monolayer coverage (0.06% w/w) to occur at 23% RH, suggesting that separation energy measurements
Figure 5. Dynamic vapour sorption (DVS) isotherm of salbutamol sulfate. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
820
YOUNG ET AL.
Figure 6 as cumulative frequency separation energy values for salbutamol sulfate drug probe 1 (15–75% RH). The assumption of log normal distribution was further justified using apparent linear regression analysis of log–probability plots (5–95%), and full 4-parameter sigmoidal regression analysis, where R2 values of 0.97 were obtained using both methods. As a result, a median value, or 50% cumulative undersize (e0.5), was selected as the most appropriate descriptor for separation energy measurements because the large positive skew in the data may adversely affect the mean values. The effect of humidity on the e0.5 is shown graphically in Figure 7. In all cases, change of humidity had a significant effect on the e0.5 (CI (confidence interval) ¼ 0.95). Furthermore, a 6.4 1.1 times increase in mean e0.5 (n ¼ 3) occurred between 15 and 75% RH, indicating capillary forces become a dominating factor at higher humidities. Such capillary forces may be directly induced by the multilayer adsorption of water onto the salbutamol sulfate surfaces as humidity was increased (as indicated by DVS in Figure 5). Essentially, these observations suggest that the increase in particle cohesion at higher humidities lead to an increase in energy required for deagglomeration. For dry powder inhaler formulations, this may result in a decrease in fine particle fraction, as previously reported.15,16 It is important to note, however, that such measurements are undertaken between individual drug probes and model drug surfaces and therefore can not be directly related to the complex
conducted by AFM at 30, 45, 60, and 75% RH represented multilayer water coverage. Separation Energy Measurements Separation energy measurements between three salbutamol sulfate drug probes and salbutamol sulfate model compacts were investigated with the AFM at 15, 30, 45, 60, and 75% RH and are summarized in Table 1. Integration of the force distance curves measured over a 10 10-mm area at each humidity (n ¼ 4096) produced a relatively wide distribution of separation energy values. Furthermore, separation energy histogram analysis showed the data spread across more than one order of magnitude in an asymmetrical, positively skewed distribution. Analysis of the separation energy data block (64 64 force curves) suggested no relationship between force curve number and separation energy value. In addition, offline analysis of individual force curves conducted at the beginning, middle, and end of each run indicated no rank change in separation energy value or curve shape. Such observations would suggest that the salbutamol sulfate drug probe did not deform or appreciably tribocharge under such conditions. Furthermore, comparison of the separation energy data block values with the individual force curves, in offline analysis, indicated no surface ‘hotspots’ that may have been observed if contamination was present. Cumulative frequency–log distributions of the separation energy values indicated a log normal distribution. This distribution is demonstrated in
Table 1.
Salbutamol Sulfate Separation Energy Values Relative Humidity, %
Tip
Parameter a
e0.5 (nJ) GSDb R2 c e0.5 (nJ)a GSDb R2 c e0.5 (nJ)a GSDb R2 c
1 1 1 2 2 2 3 3 3
15
30
45
60
75
2153 2.3 0.997 1628 2.4 0.994 2091 4.0 0.976
3065 2.2 0.999 2325 2.4 0.973 3451 3.1 0.977
3650 2.2 0.998 5262 2.1 0.977 7067 3.2 0.972
4990 2.1 0.997 5981 2.1 0.997 8294 2.9 0.980
11188 2.1 0.997 12726 2.0 0.994 12779 2.7 0.975
a
e0.5, median separation energy (n ¼ 4096). Geometric standard deviation (e0.841/e0.159)0.5. R2, linear regression of lognormal distribution (e0.05 and e0.95).
b c
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
ATOMIC FORCE MICROSCOPY CHARACTERIZATION
821
CONCLUSION
Figure 6. Separation energy distribution for salbutamol sulfate drug probe on salbutamol sulfate model surface at 15, 30, 45, 60, and 75% RH at 25 < 0.28C (n ¼ 4096).
dynamics that exist during the aerosolization process. Furthermore, the use of compaction to produce a relatively smooth contiguous surface may alter the surface properties of the salbutamol sulfate. Because of these factors, the compaction rate and dwell time were specifically chosen to minimize such damage. It is believed that the possible recrystallization of any amorphous regions under low compaction rates was limited by the use of micronized salbutamol sulfate that had been stored for at least a year at 45% RH before the studies were conducted. Although it was possible that amorphous material existed on the drug surface, DVS traces of the raw micronized material (as shown in Figure 5) showed no such characteristic hysteresis indicating recrystallization.
Figure 7. (e0.5) median separation energy values for salbutamol sulfate drug probes on model drug compacts at 15, 30, 45, 60, and 75% RH (n ¼ 4096).
The relationship between humidity and the cohesion of salbutamol sulfate particles, as determined with the AFM, can be attributed to the relative contribution of the interparticulate forces that exist between the two particle contact areas. As the humidity is increased (15–75% RH), capillary forces become a dominating factor because of the multilayer sorption of water onto the crystal surfaces. Ultimately, this increase in separation energy will influence the aerosolization and subsequent efficacy of such powders and should be considered during formulation. This investigation has shown atomic force microscopy to be a powerful technique for quantification of the separation energies between micronized drug particulates. Additionally, the use of a custom-built perfusion apparatus, connected to the AFM, has allowed the measurement of these interactions in environments relative to pharmaceutical material performance.
REFERENCES 1. Pritchard JN. 2001. The influence of lung deposition on clinical response. J Aerosol Med 14:S19– S26. 2. Staniforth JN. 1995. Performance-modifying influences in dry powder inhalation systems. Aerosol Sci Technol 22:346–353. 3. Podczeck F, Newton JM, James MB. 1997. Influence of relative humidity of storage air on the adhesion and autoadhesion of micronized particles to particulate and compacted powder surfaces. J Colloid Interface Sci 187:484–491. 4. Clarke MJ, Peart J, Cagnani S, Byron PR. 2002. Adhesion of powders for inhalation: An evaluation of drug detachment from surfaces following deposition from aerosol streams. Pharm Res 19:322– 329. 5. Binnig G, Quate CF, Gerber C. 1986. Atomic force microscope. Phys Rev Lett 56:930–933. 6. Ducker WA, Senden TJ, Pashley RM. 1991. Direct measurement of colloidal forces using an atomic force microscope. Nature 353:239–241. 7. Butt H-J. 1992. Measuring electrostatic, van der Waals, and hydration forces in electrolyte solutions with an atomic force microscope. Biophys J 60: 1438–1444. 8. Heinz WF, Hoh JH. 1999. Spatially resolved force spectroscopy of biological surfaces using the atomic force microscope. Trends Biotechnol 17:143–150. 9. Biggs S, Spinks G. 1998. Atomic force investigation of the adhesion between a single polymer sphere JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
822
10.
11.
12.
13.
14.
15.
16.
YOUNG ET AL.
and a flat surface. J Adhesion Sci Technology 12: 461–478. Yalamanchili MR, Veeramasuneni S, Azevedo MAD, Miller JD. 1998. Use of atomic force microscopy in particle science and technology research. Colloids Surf, A 133:77–88. Mizes H, Ott M, Eklund E, Hays D. 2000. Small particle adhesion: Measurement and control. Colloids Surf, A 165:11–23. Ibrahim TH, Burk TR, Etzler FM, Neuman RD. 2000. Direct adhesion measurements of pharmaceutical particles to gelatin capsule surfaces. J Adhesion Sci Technol 14:1225–1242. Louey MD, Mulvaney P, Stewart PJ. 2001. Characterisation of adhesional properties of lactose carriers using atomic force microscopy. J Pharm Anal 25:559–567. Sindel U, Zimmermann I. 2001. Measurement of interaction forces between individual powder particles using an atomic force microscope. Powder Technol 117:247–254. Jashnani RN, Byron PR, Dalby RN. 1995. Testing of dry powder aerosol formulations in different environmental conditions. Int J Pharm 113:123–130. Jashnani RN, Byron PR. 1996. Dry powder aerosol generation in different environments: Performance comparisons of albuterol, albuterol sulfate, albuterol adipate and albuterol salt. Int J Pharm 130: 13–24.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 92, NO. 4, APRIL 2003
17. Podczeck F, Michael Newton J, James MB. 1997. Variations in the adhesion force between a drug and a carrier particles as a result of changes in the relative humidity of the air. Int J Pharm 149: 151–160. 18. Sedin DL, Rowden KL. 2000. Adhesion forces measured by atomic force microscopy in humid air. Anal Chem 72:2183–2189. 19. Berard V, Lesniewska E, Andres C, Pertuy D, Laroche C, Pourcelot Y. 2002. Dry powder inhaler: influence of humidity on topology and adhesion studied by AFM. Int J Pharm 232:213–224. 20. Preuss M, Butt H-J. 1998. Direct measurement of particle-bubble interactions in aqueous electrolyte: dependence on surfactant. Langmuir 14:3164– 3174. 21. Hutter JL, Bechhoefer J. 1993. Calibration of atomic-force microscope tips. Rev Sci Instrum 64:1868–1873. 22. Giles CH, MacEwan TH, Nakhwa SN, Smith D. 1960. Studies in adsorption. Part XI.* A system for classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J Chem Soc 3973–3993. 23. Webster CE, Drago RS, Zerner MC. 1998. Molecular dimensions for adsorptives. J Am Chem Soc 120: 5509–5516.