578
Langmuir 2005, 21, 578-585
Deposition and Aggregation of Aspirin Molecules on a
Phospholipid Bilayer Pattern
Guangzhao Mao,* Dongzhong Chen,† and Hitesh Handa
Department of Chemical Engineering and Materials Science, Wayne State University,
5050 Anthony Wayne Drive, Detroit, Michigan 48202
Wenfei Dong, Dirk G. Kurth,‡ and Helmuth Möhwald
Max Planck Institute of Colloids and Interfaces, Research Campus Golm,
14424 Potsdam, Germany
Received September 2, 2004. In Final Form: November 2, 2004
Aspirin and 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine (DMPE) are deposited from their alcoholic
mixed solution onto highly oriented pyrolytic graphite (HOPG) by spin coating. The film structure and
morphology are characterized by atomic force microscopy (AFM). The barely soluble DMPE forms a highly
oriented stripe phase as a result of its one-dimensional epitaxy with the HOPG lattice. The bilayer stripe
pattern exposes the cross section of the lipid bilayer lamellae and enables the direct visualization of the
molecular interactions of drug or biological molecules with either the hydrophobic or the hydrophilic part
of the phospholipid bilayer. The bilayer pattern affects the aspirin molecular deposition and aggregation.
AFM shows that the aspirin molecules prefer to deposit and aggregate along the aliphatic interior part
of the bilayer pattern, giving rise to parallel dimer rods in registry with the underlying pattern. The
nonpolar interactions between aspirin and the phospholipid bilayer are consistent with the lipophilic
nature of aspirin. The bilayer pattern not only stabilizes the rodlike aggregate structure of aspirin at low
aspirin concentration but also inhibits crystallization of aspirin at high aspirin concentration. Molecular
models show that the width of the DMPE aliphatic chain interior can accommodate no more than two
aspirin dimers. The bilayer confinement may prevent aspirin from reaching its critical nucleus size. This
study illustrates a general method to induce a metastable or amorphous form of an active pharmaceutical
ingredient (API) by chemical confinement under high undercooling conditions. Metastable and amorphous
solids often display better solubility and bioavailability than the stable crystalline form of the API.
Introduction
The ability to manipulate and characterize materials
at the nanoscale has led to explosive research activities
in the molecular thin films, crystals, and devices. In
supramolecular pharmaceutics, the same active pharmaceutical ingredient (API) molecules are manipulated
by various noncovalent interactions (hydrogen bonding,
van der Waals, π-π stacking, and electrostatic interactions) into different solid-state forms ranging from amorphous to crystalline states.1 The solid-state form of the
API affects its compressibility, solubility, dissolution rate,
chemical stability, and bioavailability. Some new drug
discovery methods involve screening different crystallization conditions in small volume2 and on engineered
surfaces.3 Micropatterns of self-assembled monolayers
have been used to control crystallization by relying on
guest molecules to recognize different surface functional
groups and by microconfinement.4 The integration of
bottom-up to top-down approaches in the developing
technologies, such as the high-throughput screening of
* Corresponding author. E-mail: [email protected].
† Current address: Key Laboratory of Mesoscopic Chemistry of
MOE and Department of Polymer Science & Engineering, College
of Chemistry and Chemical Engineering, Nanjing University,
Nanjing 210093, People’s Republic of China.
‡ National Institute for Materials Science (NIMS), 1-1 Namiki,
Tsukuba, Ibaraki 305-0044, Japan.
(1) Rodrı́guez-Spong, B.; Price, C. P.; Jayasankar, A.; Matzger, A. J.;
Rodrı́guez-Hornedo, N. Adv. Drug Delivery Rev. 2004, 56, 241.
(2) Perepezko, J. H. Mater. Sci. Eng., A 1994, 178, 105.
(3) Ward, M. D. Chem. Rev. 2001, 101, 1697.
(4) Aizenberg, J.; Black, A. J.; Whitesides, G. M. Nature 1999, 398,
495.
drugs, requires the knowledge of placing molecules on
ever diminishing patterns.
Alkanes and alkane derivatives are known to self-assemble into a long-range ordered stripe phase on highly
oriented pyrolytic graphite (HOPG) as a result of the onedimensional (1-D) epitaxial match between the 1,3-methylene group distance ()0.251 nm) and the distance of the
next nearest neighbor of the HOPG lattice ()0.246 nm).5,6
Recently, this molecular pattern has been used to align
small organic molecules7,8 as well as macromolecules.9-12
The amphiphilic pattern at the solid and liquid interface
has been reproduced in thin solid films by spin coating
and has been used to synthesize sulfide molecular rod
arrays on the copper arachidate template.13 Both the
arachidate pattern and the subsequent inorganic rod
arrays have been characterized by atomic force microscopy
(AFM). AFM is capable of resolving the location and shape
of guest molecular aggregates in relation to a specific type
(5) McGonigal, G. C.; Bernhardt, R. H.; Thomson, D. J. Appl. Phys.
Lett. 1990, 57, 28.
(6) Rabe, J. P.; Buchholz, F. Science 1991, 253, 424.
(7) Wei, Y.; Kannappan, K.; Flynn, G. W.; Zimmt, M. B. J. Am. Chem.
Soc. 2004, 126, 5318.
(8) Hoeppener, S.; Chi, L.; Fuchs, H. ChemPhysChem 2003, 4, 494.
(9) Mena-Osteritz, E.; Meyer, A.; Langeveld-Voss, B. M. W.; Janssen,
R. A. J.; Meijer, E. W.; Bäuerle, P. Angew. Chem., Int. Ed. 2000, 39,
2679.
(10) Kurth, D. G.; Severin, N.; Rabe, J. P. Angew. Chem., Int. Ed.
2002, 114, 3681.
(11) Loi, S.; Butt, H.-J.; Hampel, C.; Bauer, R.; Wiesler, U.-M.; Müllen,
K. Langmuir 2002, 18, 2398.
(12) Severin, N.; Rabe, J. P.; Kurth, D. G. J. Am. Chem. Soc. 2004,
126, 3696.
(13) Mao, G.; Dong, W.; Kurth, D. G.; Möhwald, H. Nano Lett. 2004,
4, 249.
10.1021/la047802i CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/22/2004
Aspirin Deposition on Bilayer Pattern
of functional group on the host pattern. The amphiphilic
pattern of phospholipids serves as a model for the study
of the molecular interactions of drug or biological compounds with biomembranes and cells. The phospholipid
stripe phase is structurally similar to the plane perpendicular to the bilayer lamellae. The bilayer pattern exposes
the hydrophobic and hydrophilic parts simultaneously for
guest molecular recognition and also for AFM imaging.
This enables detailed structural analysis of drug or
biological molecular interactions with specific functional
groups of the phospholipid bilayer.
This paper reports the effect of the phospholipid bilayer
pattern on the deposition and aggregation behavior of
aspirin. Aspirin, also called acetylsalicylic acid or 2-(acetyloxy)-benzoic acid, was first synthesized by Bayer in 1897.
Aspirin is a nonsteroidal anti-inflammatory drug (NSAID),
widely used to treat human inflammatory disorders, such
as blood coagulation, thrombosis, and atherosclerosis, by
inhibiting platelet aggregation.14 Recently, aspirin has
also been found to reduce the risk of heart attack and to
be effective against colorectal cancer. The primary mechanism of aspirin drug action is its interference with the
biosynthesis of inflammatory prostaglandins.15 Another
known effect of NSAIDs is their capability to perturb the
phospholipid ordering in the biomembranes, thus, affecting the normal functions of the membrane proteins.16
Aspirin is a weak acid and lipophilic.17 Aspirin is found
to increase the fluidity of liposomes by inserting itself in
between the hydrocarbon chains of the phospholipid
molecules. Aspirin has only one crystal form,18 which
makes its structural analysis less complicated. The aspirin
crystal consists of hydrogen-bonded dimers. The hydrogenbonded dimers are the most common supramolecular
synthon for monocarboxylic acid crystals.19 The supramolecular synthon is the smallest molecular building block,
whose symmetry and connectivity predispose the symmetry and packing in the final crystal structure. This paper
presents a first study on the adsorption of supramolecular
aggregates in a predefined way on the graphite-based
templates. The study of the aggregation behavior of the
aspirin dimers may shed light on nanoscale confinement
means to engineer crystals with self-replicating building
blocks.
Experimental Section
Materials. Aspirin (Aldrich, +99.5%), 1,2-dimyristoyl-snglycero-3-phosphoethanolamine (DMPE; Sigma, 99%), ethanol
(Pharmco, 100%), and methanol (Mallinckrodt Chemicals, 100%)
are used as received. ZYB grade HOPG (Mikromasch) is handcleaved with an adhesive tape just before film preparation until
a smooth surface is obtained.
Film Preparation. Aspirin and DMPE are dissolved in
methanol or ethanol in different molar ratios. The solubility of
aspirin in alcohol is 1.11 M,17 while DMPE dissolves sparingly
and slowly in alcohol. Approximately 10-5 M DMPE is dissolved
after shaking the solution for 10 min. The DMPE concentration
is maintained at 10-5 M, while the molar ratio of aspirin to DMPE
is varied from 1 to 5000. A total of 100 µL of a freshly prepared
solution, filtered with a 0.22-µm poly(tetrafluoroethylene) filter
(Millipore), is placed on a HOPG substrate rotating at 3000 rpm
(PM101DT-R485 photo resist spinner, Headway Research) at
room temperature for 1 min.
(14) Lasslo, A. Blood Platelet Function and Medicinal Chemistry;
Elsevier: New York, 1984.
(15) Vane, J. R. Nat. New Biol. 1971, 231, 232.
(16) Das, M. M., Dutta, S. K., Eds. Modern Concepts on Pharmacology
and Therapeutics, 24th ed.; Hilton & Co.: Calcutta, 1989; p 26.
(17) Budavari, S., O’Neil, M. J., Smith, A., Heckelman, P. E.,
Kinneary, J. F., Eds. The Merck Index: an Encyclopedia of Chemicals,
Drugs, and Biologicals; Merck & Co., Inc.: Rathway, NJ, 1996; p 144.
(18) Wheatley, P. J. J. Chem. Soc. 1964 (Suppl.), 6039.
(19) Frankenbach, G. M.; Etter, M. C. Chem. Mater. 1992, 4, 272.
Langmuir, Vol. 21, No. 2, 2005 579
Figure 1. Optical micrograph of aspirin recrystallized from
ethanol by slow solvent evaporation.
Characterization. The optical images are captured by an
Olympus BX60 microscope with a SONY DXC-970MD camera
in transmitted light. AFM images are obtained with either an
E scanner (Nanoscope III, VEECO) or a Dimension Scan Head
(Dimension 3100, VEECO). Height, amplitude, and phase images
are obtained in tapping mode in ambient air with silicon tips
(TESP, VEECO). Only height images are shown here unless
specified. The scan rate is 1 Hz. Integral and proportional gains
are approximately 0.4 and 0.7, respectively. The crystal structures
and structural analysis are made with the Materials Studio
software programs from Accelrys.
Results and Discussion
Aspirin Crystals Recrystallized in Alcohol. Aspirin
recrystallized from methanol and ethanol by slow solvent
evaporation forms rectangular plates as shown in Figure
1. The largest aspirin crystal face from alcohol is the (001)
face, followed by two other low-energy faces (100) and
(011).20 Aspirin recrystallized from hexane forms needles.
Aspirin crystals with the needle habit dissolve faster in
water than the crystals with the plate habit.21,22 Therefore,
commercial aspirin tablets contain a significant amount
of needle crystals.
Spin-Coated DMPE Films. Spin coating of the 10-5
M DMPE methanol solution produces the stripe phase on
HOPG with close to a monolayer coverage. Figure 2a-c
displays the spin-coated film structures as captured by
AFM. Similar film structures are obtained also from
ethanol. The stripe phase domain is rectangular in shape
with an average domain size of 200 nm. The domain size
is consistent with the single crystalline domain size of
ZYB grade HOPG. Two sides of the rectangle are
terminated clearly at the stripe edges, while the other
sides are less obvious because of higher boundary energy.
The domain orientation displays the threefold symmetry
of the HOPG lattice, as expected from the 1-D epitaxy
(Figure 2a). The domain height is 0.7 nm by the sectional
height analysis (Figure 2b). At full monolayer coverage,
the film appears to be identical to HOPG at the micrometer
scale, showing the typical folded layer ledges of HOPG.
The AFM tip is capable of sweeping the DMPE molecules
away when scanning at a force ) 100 nN repeatedly in
contact mode. The AFM tip digs a square trench with the
same size of the previous scans as a result. Sectional height
(20) Masaki, N.; Machida, K.; Kado, H.; Yokoyama, K.; Tohda, T.
Ultramicroscopy 1992, 42, 1148.
(21) Kim, Y.; Machida, K.; Taga, T.; Osaki, K. Chem. Pharm. Bull.
1985, 33, 2641.
(22) Yesook, K.; Matsumoto, M.; Machida, K. Chem. Pharm. Bull.
1985, 33, 4125.
580
Langmuir, Vol. 21, No. 2, 2005
Mao et al.
Figure 2. AFM images of DMPE films on HOPG spin coated from 10-5 M methanol solution. (a) A 600-nm scan showing submonolayer
domains. The angle between domains as shown is 59.48°. (b) The cross-sectional height profile along the dotted line in part a. The
domain height is 0.7 nm. (c) A 200-nm scan of the DMPE bilayer stripes. The periodicity is 5.2 nm as determined from the 2-D
Fourier transform pattern in the inset. The arrow points to the direction of the interlinear spacing R.
analysis shows that the unperturbed film portion is 0.7
nm above the square trench. It is found that the tip does
not damage the HOPG surface at the same force level.
The measured thickness is much smaller than the DMPE
chain length, which excludes the normal monolayer
structure with vertically oriented hydrocarbon chains. The
interchain spacing in triclinic alkane crystals is 0.42 and
0.48 nm in the direction perpendicular and parallel to the
zigzag carbon chain plane, respectively.23 The film thickness of 0.7 nm is consistent with the DMPE double chains
lying parallel to the HOPG basal plane but in a tilted
configuration to maximize the chain packing. The centerto-center distance between neighboring stripes is measured to be 5.2 nm by the two-dimensional (2-D) Fourier
(23) Kitaigorodskij, A. I.; Mnjukh, J. B. Bull. Acad. Sci. U.S.S.R.,
Div. Chem. Sci. 1959, 1992.
transform analysis (Figure 2c inset). This periodicity, of
which 3.3 nm belongs to the hydrophobic tails,24 agrees
well with the DMPE bilayer thickness as determined by
X-ray scattering.25 Energetically the lipid bilayer should
be continuous across its hydrophobic interior in a tailto-tail configuration with a small separation existing
between the opposing headgroups from neighboring
bilayers. The same energy consideration also requires the
stripe phase terminating at the hydrophilic headgroups.
Figure 2c shows that the domain edge matches the edge
and not the center of a bright stripe. It is evident that the
center of the bright stripe should correspond to the
(24) The hydrophobic layer thickness, LH, in the DMPE bilayer can
be calculated as follows: LH ) 2 × [12 + (9/8)] × 0.1265 nm ) 3.32 nm.
(25) Helm, C. A.; Tippmann-Krayer, P.; Möhwald, H.; Als-Nielsen,
J.; Kjaer, K. Biophys. J. 1991, 60, 1457.
Aspirin Deposition on Bilayer Pattern
Langmuir, Vol. 21, No. 2, 2005 581
Figure 3. AFM images of aspirin films on HOPG. (a) A 500-nm scan showing disordered molecular aggregates spin coated from
10-3 M methanol solution. (b) The cross-sectional height profile along the line in part a. The aggregate height is 0.5 nm. (c) A 5-µm
scan showing crystalline layers spin coated from 10-2 M ethanol solution. The arrow points to a top crystalline layer with a step
edge angle close to 90°. Some edges display a 60° angle as marked. (d) The cross-sectional height profile along the dotted line in
part c. The minimum step height of the aspirin crystal is 1.1 nm.
hydrophobic center of the lipid bilayer, while the dark
gap should correspond to the hydrophilic headgroup
region.
Spin-Coated Aspirin Films. No aspirin molecules are
observed on HOPG when the aspirin concentration is below
10-3 M. At 10-3 M, the HOPG surface is sparsely covered
by the molecular aggregates presumably belonging to
aspirin. The aggregates are irregular in shape and do not
bind strongly to HOPG (Figure 3a). A minimum force in
tapping mode can still sweep the aggregates away. The
average height of the aggregates is 0.5 nm (Figure 3b).
Above 2 × 10-3 M, localized crystalline layers of aspirin
of at least 10 nm in thickness appear on HOPG. Figure
3c shows the aspirin film spin coated at 10-2 M in ethanol.
The layers have orthogonal boundaries with a minimum
differential thickness of 1.1 nm (Figure 3d). This is
consistent with the texture orientation of the aspirin
crystal (001) face. Some 60° angles also exist, which
indicates likely azimuthal correlation between the aspirin
(001) face and the HOPG basal plane.
Spin-Coated Aspirin and DMPE Films. The molar
ratio of aspirin to DMPE is varied from 1 to 5000 with the
DMPE concentration fixed at 10-5 M in methanol or
ethanol. When the aspirin-to-DMPE molar ratio is less
than 5, only the DMPE stripe phase is observed on HOPG.
On the other hand, when the aspirin-to-DMPE molar ratio
is greater than 50, the film morphology is largely that of
the aspirin crystalline layers. When the molar ratio is
between 5 and 50, the film shows both the molecular
aggregates of aspirin on top of the DMPE stripe domains
and the platelike aspirin crystals on bare HOPG areas
between the DMPE domains, as illustrated by the scheme
in Figure 4a. The spin-coated films from binary solutions
generally show phase separation. If the two components
are similar to each other in solubility and surface affinity,
lateral phase separation occurs. A significant difference
in solubility or surface affinity between the two components results in vertical phase separation and the formation of a wetting layer.26-28 This AFM study reveals a
largely vertical phase separation between aspirin and
DMPE, with DMPE forming the wetting layer. This is
due to the large difference in alcohol solubility between
(26) Jones, R. A. L.; Norton, L. J.; Kramer, E. J.; Bates, F. S.; Wiltzius,
P. Phys. Rev. Lett. 1991, 66, 1326.
(27) Steiner, U.; Klein, J.; Eiser, E.; Budkowski, A.; Fetters, L. J.
Science 1992, 258, 1126.
(28) Walheim, S.; Böltau, M.; Mlynek, J.; Krausch, G.; Steiner, U.
Macromolecules 1997, 30, 4995.
582
Langmuir, Vol. 21, No. 2, 2005
Mao et al.
Figure 4. AFM images of aspirin and DMPE mixed films on HOPG with an aspirin-to-DMPE molar ratio ) 10. (a) Schematic
graph of topographical features observed in the mixed film of aspirin and DMPE. (b) An 800-nm scan. (c) A 300-nm scan. (d) A
150-nm scan. The arrows point out that the aspirin molecular aggregates sit directly on top of the DMPE stripes, not in between.
(e) The cross-sectional height profile along the line in part c. The height difference between the top and the bottom layer is 0.7
nm.
aspirin and DMPE. For example, when the DMPE
concentration is 10-5 M and the aspirin-to-DMPE molar
ratio is 50, the solution is saturated with DMPE but highly
undersaturated with aspirin (aspirin concentration/
solubility ) 0.000 45). DMPE precipitates upon the slight
solvent evaporation. Its epitaxial interaction with HOPG
warrants the formation of the wetting monolayer as well
as its self-organization into the stripe pattern. Aspirin
precipitates only when its solubility limit is reached with
further solvent evaporation, either on the DMPE layer or,
if the surface is not fully covered, on HOPG. AFM images
show that aspirin molecules recognize the amphiphilic
pattern inherent in the DMPE stripe phase. Aspirin
molecules adsorb only on the hydrophobic interior of the
DMPE bilayer stripe while avoiding the hydrophilic
boundaries. This molecular recognition should be due to
the strong lipophilicity of aspirin. The bilayer pattern
should promote the 1-D aggregation of aspirin dimers,
while limiting the 2-D and three-dimensional crystallization. Inhibition of the growth of the (001) face is
desirable because this face is less water-soluble than the
(100) face. Here the nonpolar interactions are sufficient
to stabilize an otherwise highly unstable molecular rod
structure. Meta-stable and amorphous forms of APIs are
often more desirable than the stable crystalline form
because they tend to show higher bioavailability.
Aspirin Deposition on Bilayer Pattern
Figure 4b-d shows the film made from a mixed solution
with the aspirin-to-DMPE molar ratio of 10, as imaged at
800-nm, 300-nm, and 150-nm scan sizes, respectively. The
individual domains display the threefold azimuthal
orientation as defined by the stripe edges. The domain
has a rough appearance, which is different from the
parallel stripe pattern of the DMPE layer. A closer look
(Figure 4c,d) shows an additional layer partially covering
the DMPE stripes. The second layer is 0.7 nm above the
first stripe layer (the stripe layer thickness is still 0.7 nm)
as measured in Figure 4e. Pure DMPE spin coated at 10-5
M, on the other hand, forms a submonolayer structure.
Presumably the bottom layer belongs to the DMPE wetting
layer and the second layer is made of aspirin aggregates.
Figure 4d shows that the height undulation of the rodlike
aspirin aggregates is in phase with that of the underlying
DMPE stripes. This means that the aspirin aggregates
are situated directly on top of the hydrophobic center of
the DMPE bilayer stripe. This is consistent with the
lipophilic nature of aspirin, which interacts through
nonpolar interactions with the hydrophobic interior of
biomembranes and the hydrophobic pockets of the protein
receptors. When the aspirin-to-DMPE molar ratio is
increased to 20, elongated particles with ill-defined shapes
are distributed randomly but evenly on the surface, as
shown in Figure 5a. The orientation of these particles is
analyzed by measuring the orientation angles of those
particles whose long axes are clearly defined. The orientation histogram is presented in Figure 5b. The mutual
orientation angle of the majority of the particles maintains
the threefold symmetry of the HOPG basal plane. Figure
5c,e shows two characteristic features at higher magnifications. Figure 5c shows that aspirin forms a continuous
layer on the DMPE stripe phase. The layer is 0.7 nm above
the DMPE stripes (Figure 5d). The long sides of the aspirin
layer are parallel to the DMPE stripe and, thus, better
defined. The other sides display no clear boundaries
against the DMPE background. Figure 5e shows rectangular aspirin crystals located between the DMPE stripe
domains. The height of the crystals, 1.1 nm, is one unit
length along the [001] axis. It is clear that the DMPE
bilayer pattern inhibits the crystallization of aspirin by
stabilizing the rodlike molecular aggregates.
Structural Analysis and Modeling. It is possible to
understand the AFM images further by studying the
molecular packing structures of aspirin and DMPE in the
solid state. The structural analysis provides a likely model
for the aspirin molecular aggregation on the DMPE bilayer
pattern. The DMPE crystal structure is unavailable, and,
therefore, the crystal structure of 1,2-dilauroyl-sn-glycero3-phosphoethanolamine acetic acid (DLPE) is used instead. DLPE has the same molecular structure as DMPE
with the exception that it is shorter by two carbons in
each aliphatic chain. DLPE forms a monoclinic crystal
structure P21/c with the following unit cell parameters:
a ) 47.700 Å, b ) 7.770 Å, c ) 9.950 Å, and β ) 92.000°.29,30
One reaches a hydrophobic layer thickness of 32.920 Å
from the crystal structure similar to the value by X-ray
scattering25 by adding an ethylene unit ()4.992 Å) to each
of the DLPE chains. The hydrophilic headgroup thickness
of the DMPE bilayer is 14 Å according to the molecular
model. The periodicity in the AFM images is slightly larger
than the expected bilayer thickness of DMPE, which is
(29) Elder, M.; Hitchcock, P.; Mason, R.; Shipley, G. G. Proc. R. Soc.
London, Ser. A 1977, 354, 157.
(30) CCDC LAPETM10 contains the supplementary crystallographic
data for DLPE. These data can be obtained free of charge via www.ccdc.
cam.ac.uk/data_request/cif, by e-mailing [email protected],
or by contacting The Cambridge Crystallographic Data Centre, 12, Union
Road, Cambridge CB2 1EZ, U.K.; fax, +44 1223 336033.
Langmuir, Vol. 21, No. 2, 2005 583
also its crystal unit cell length along the [100] direction.
The hydrocarbon chains in the DLPE crystal lie along the
[100] axis perpendicular to the bilayer plane. The height
of the DMPE layer measured by AFM matches the unit
cell length along the [010] direction, about one and a half
the hydrocarbon chain diameter. The plane containing
the two hydrocarbon chains of a DMPE molecule should
be tilted against the HOPG basal plane for maximum
packing. The zwitterionic phosphoethanolammonium
headgroup dipoles are parallel to the bilayer surface and
along the [010] axis in the DLPE crystal. The ammonium
groups form short bonds (bond length ) 2.7 Å) with
neighboring phosphate groups with partial hydrogen, and
the rest have ionic bonding character.31 This rigid bonding
structure among headgroups may enhance the epitaxial
ordering of DMPE on HOPG. It can be concluded that
DMPE exhibits an azimuthal orientation on the HOPG
basal plane with the hydrocarbon chains along the
interhexagonal direction of the HOPG lattice.
Aspirin recrystallizes from ethanol into a monoclinic
crystal structure P21/c with the following unit cell
parameters: a ) 11.430 Å, b ) 6.591 Å, c ) 11.395 Å, and
β ) 95.68°.18,32 Aspirin forms the inversion-symmetric,
hydrogen-bonded carboxylic acid dimer as illustrated in
Figure 6a. The (001) and (100) faces are displayed in Figure
6b as viewed along the [010] direction. The (001) face
contains the methyl and phenyl groups, while the (100)
face contains the ester groups. Aspirin dimers most likely
prefer to face the DMPE bilayer with the (001) face to
maximize the nonpolar-nonpolar interactions. A previous
AFM study by Danesh et al. shows that the methylterminated AFM tip interacts more favorably with the
more hydrophobic (001) face, while the carboxyl-terminated AFM tip is attracted more to the (100) face.33 Aspirin
dimers are free to aggregate through π-π stacking along
the [010] axis along the DMPE stripe direction; however,
the aggregation in the direction perpendicular to the stripe
is inhibited by the hydrophilic bilayer boundaries. The
DMPE hydrocarbon region is 3.3 nm wide, which can fit
up to two unit cells along the aspirin [100] axis. Aspirin
aggregates may not be able to reach the required critical
nucleus size for crystallization to occur on the DMPE
pattern. Therefore, aspirin forms only molecular aggregates on DMPE at the same concentration where
crystals are observed on the bare HOPG substrate. The
AFM-determined aggregate height, 0.7 nm, less than the
unit cell length along [001], also indicates that crystallization has not occurred. The height is close to the size
of aspirin, 7.433 Å, as measured lengthwise from the
methyl to the phenyl group from its crystal structure.
Figure 6c displays the possible model of the aspirin
aggregation on the DLPE bilayer, which is consistent with
the AFM study.
Conclusions
This paper describes the molecular aggregation behavior
of aspirin molecules on the bilayer stripe pattern of DMPE.
The pattern is formed spontaneously by spin coating the
DMPE alcoholic solution on HOPG. When aspirin is added
to the DMPE solution, it deposits either on top of the DMPE
stripe pattern or on the uncovered HOPG surface during
spin coating. Aspirin molecules do not deposit randomly
but prefer to aggregate along the hydrophobic center of
(31) Pascher, I.; Lundmark, M.; Nyholm, P.-G.; Sundell, S. Biochim.
Biophys. Acta 1992, 1113, 339.
(32) CCDC ACSALA01 contains the supplementary crystallographic
data for aspirin.
(33) Danesh, A.; Davies, M. C.; Hinder, S. J.; Roberts, C. J.; Tendler,
S. J. B.; Williams, P. M.; Wilkins, M. J. Anal. Chem. 2000, 72, 3419.
584
Langmuir, Vol. 21, No. 2, 2005
Mao et al.
Figure 5. AFM images of aspirin and DMPE mixed films on HOPG with an aspirin-to-DMPE molar ratio ) 20. (a) A 2-µm scan
showing elongated but otherwise ill-defined aspirin aggregates. (b) Particle orientation analysis of the previous image. (c) A 300-nm
scan showing an aspirin layer on top of the DMPE stripes. (d) The cross-sectional height analysis along the line in part c. The aspirin
layer height is 0.7 nm. (e) A 250-nm scan showing rectangular aspirin crystals deposited on HOPG between the DMPE stripe
domains. The phase image is shown here because the corresponding height image does not show the aspirin crystal shape as clearly.
the DMPE bilayer. It shows that the aspirin molecules
can bind to a specific region of the lipid bilayer through
the nonpolar van der Waals interactions. While the bilayer
pattern promotes the aspirin dimer aggregation via π-π
Aspirin Deposition on Bilayer Pattern
Langmuir, Vol. 21, No. 2, 2005 585
Figure 6. Aspirin structural analysis by the Materials Studio software program. (a) The basic building block of the aspirin crystal,
the aspirin dimer. The hydrogen bonds are marked by the dotted lines. (b) The functional groups on the (100) face and (001) face
as viewed along the [010] direction. (c) A structural model of aspirin rodlike molecular aggregates on the DLPE bilayer pattern
on HOPG. The HOPG surface is not drawn to avoid crowding. Only one dimer row is drawn for every bilayer stripe even though
the stripe width can accommodate up to two dimer rows. The DLPE layer is created by cleaving the crystal along its (010) face.
A vacuum slab is created above the DLPE supercell. Nine aspirin dimers along the [010] axis are selected from the aspirin crystal
structure and are docked above the DLPE supercell. A vacuum slab with the DLPE (010) face at the bottom enables the docking
of aspirin dimers within the vacuum above so that the dimer does not see a periodic image of the surface above it. The dimer row
is arranged to parallel the [001] direction of the DLPE layer.
stacking along the length of the stripe, the narrow width
of the stripe inhibits the crystallization of aspirin, probably
by limiting the association among aspirin dimers in the
width direction of the bilayer stripe. The experimental
approach described here may offer a simple strategy to
study the effect of surfactant mesophases and liposomes
on the crystal engineering and encapsulation of the APIs.
It is especially relevant to the solid-state preparation
methods that require a small volume and high undercooling. One may utilize the various noncovalent interac-
tions between the API molecule and the surfactant or
polymer additive molecule to control the degree of
crystallinity and even the solid form to achieve different
aggregate structures from the same API molecule.
Acknowledgment. This work is partially supported
by the National Science Foundation (CTS-0221586 and
CTS-0216109). G.M. acknowledges the financial support
from the German-American Fulbright Commission.
LA047802I