COMMUNICATION

The Confusion of Indexing Aspirin Crystals

CLARE AUBREY-MEDENDORP,1 SEAN PARKIN,2 TONGLEI LI1
1
Department of Pharmaceutical Sciences, University of Kentucky, Lexington, Kentucky 40536-0082
2
Department of Chemistry, University of Kentucky, Lexington, Kentucky 40536-0082

Received 5 March 2007; revised 25 April 2007; accepted 1 May 2007

Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21055

ABSTRACT: Much of the existing literature dealing with crystalline aspirin is vague or
ambiguous with regard to indexing of the crystal faces. The inconsistency with which the
indices of the dominant faces have been assigned leads to confusion in analysis of surface
properties. To clarify this, we have conducted crystal growth experiments on aspirin,
and indexed the crystal faces with X-ray diffraction (XRD), paying special attention to
the placement of symmetry elements. The space group was confirmed as P21/c, and the
dominant face was (100). Contact angle measurements made on the two major faces of
aspirin indicate the (100) face to be more hydrophobic than the (001) face, likely due to
the acetyloxy moiety, not the carboxyl, exposed on the (100). ! 2007 Wiley-Liss, Inc. and the
American Pharmacists Association J Pharm Sci 97:1361–1367, 2008
Keywords: wettability; contact angle; crystal packing; indexing; aspirin; X-ray
diffraction

INTRODUCTION properties.4 In order to determine and characterize

The majority of pharmaceutical materials are in
the solid crystalline form. Understanding the
arrangement or packing of molecules in a crystal
is critical because it helps determine the physi-
cochemical and mechanical properties, including
bioavailability.1–3 Depending on the conditions of
preparation, pharmaceutical crystals may exhibit
variations in habit with the consequence that
interfacial properties can be modulated by the
various morphologically important faces. Differ-
ent habits of the same drug compound can have
a huge influence on the dissolution kinetics and
consequent bioavailability. A molecular level knowl-
edge of the crystal surface structure is therefore
crucial to understanding the physical and chemical
Correspondence to: Tonglei Li (Telephone: 859-257-1472;
Fax: 859-257-7585; E-mail: tonglei@uky.edu)
Journal of Pharmaceutical Sciences, Vol. 97, 1361–1367 (2008)
! 2007 Wiley-Liss, Inc. and the American Pharmacists Association
the surface structure of particular crystal faces, the
molecular packing and the face indices must be
assigned properly.
Crystalline aspirin has been studied routinely,
yet despite the persistent investigations since its
discovery in late 1800s,5 confusion remains over
classification of the crystal morphology. In the
1930s, the first attempt was made at defining
the lattice parameters by X-ray diffraction (XRD)
using oscillation and Weissenberg photographs.
From these results it was deduced that the space
group must be P21/a (a nonstandard setting of
space group number 14).6,7 Using the parameters
from this report the predominant face was
identified as (001).8,9 However, the first full
structure determination by single crystal XRD
was not until 1964.10 This report and others that
followed assigned a different setting of the same
space group, that is, P21/c.11–13 In this setting,
the largest face is (100).14,15 A likely cause of the
confusion over face indexing was then the use of

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008 1361

1362 AUBREY-MEDENDORP, PARKIN, AND LI

Table 1. Unit Cell Parameters of Form I of Aspirin
a (Å) b (Å) c (Å) b (8)
11.242 (7) 6.539 (4) 11.245 (9) 95.90
11.446 (13) 6.596 (6) 11.388 (9) 95.33
11.430 (1) 6.591 (1) 11.395 (2) 95.68
11.233 (3) 6.544 (10) 11.231 (3) 95.89
a contact angle measurements.
This work.
b
Reference 10.
c
Reference 11.
d
References 12, 13.
alternative space group settings, exacerbated by
the almost identical metrics of the a and c axes
(Tab. 1), and consequently the functional groups
assigned to the indexed faces. Thus, if the face
indices of aspirin single crystals indexed as P21/a
are used for a crystal structure indexed as P21/c,
any conclusions drawn with regard to surface
properties of a particular face will be in error,
particularly from the viewpoint of structure-
property relationships. Indeed, this apparent
confusion has occurred in several reports on
aspirin crystals, including one of ours.16
Recent studies have probed the chemical nature
of functional groups exposed on the two major
faces of aspirin crystals. As a result of incorrect
face indexing, these studies concluded that the
(100) face was more hydrophilic than the (001)
face due to the presence of the carboxyl group
on the (100). We made a similar mistake when
differentiating the (100) and (001) face in a dis-
cussion of surface wettability.16 The conclusion
that the smaller contact angle observed with
water was on the supposed (100) face, was justi-
fied by the presence of the carboxyl group.
Upon close inspection of the face indexing of
aspirin single crystals with XRD, we believe the
common assignment of the (100) face as more
hydrophilic (due to the exposure of carboxyl
group) is not correct. The confusion in the lit-
erature seems to stem from the original assign-
ment of the space group and indices of the crystal
faces.7–9,17 This communication presents experi-
mental evidence to clarify the difference between
the two major faces of aspirin crystal, and to revise
our conclusion of the contact angle data for further
interpretation of surface wettability.
MATERIALS AND METHODS
Aspirin (acetylsalicylic acid) (>99.0%, Sigma-
Aldrich, St. Louis, MO), ethanol (HPLC grade,
Sigma-Aldrich), acetone (HPLC grade, Sigma-
Aldrich), and deionized distilled water (Barnstead
Millipore Deionized Water System, Dubuque,
(3)a Iowa, 0.2 mm filtered) were used for preparation
(2)b of aspirin crystals. Diiodomethane (99%, Sigma-
(1)c Aldrich), glycerol (98%, Mallinckrodt, Paris, KY),
(2)d and deionized distilled water were used for the
Crystal Growth
Aspirin single crystals were grown by slow eva-
poration from ethanol and acetone. Aspirin was
weighed and dissolved in the liquids, close to its
solubility (0.2 gm/mL in ethanol and 0.18 gm/mL
in acetone), by stirring. The beakers were sealed
with Parafilm1, which had small holes made
using a needle, and stored in ambient conditions.
In addition, aspirin crystals were grown in water
with a supersaturation ratio of 225% based on its
water solubility of 3.3 mg/mL. The beakers were
sealed with Parafilm1 and stored in ambient
conditions. Crystals were harvested after 1 week.
The morphology of aspirin was manually plotted
by the Cerius2 program (Accelrys, Inc., San Diego,
CA) using the single crystal X-ray data that we
determined to resemble the observed habits of the
crystals grown from the different solvents.
Single Crystal X-Ray Diffraction (XRD)
Structure determination and face indexing of
aspirin single crystal was done using a Nonius
KappaCCD diffractometer (Madison, WI) with Mo
Ka X-rays (l ¼ 0.71073 Å). The crystals were
indexed using the COLLECT program.18 Data for
space group assignment and structure determina-
tion were collected in a series of v-scans at 90 K.
These data were integrated, scaled and merged
using the Denzo-SMN package.19 Space group
assignments were made using the maXus and
Shelxtl packages.20,21 Since the a and c axes are
essentially identical in length, and because of the
previous inconsistencies, these assignments were
checked manually. Structures were solved and
refined using Shelxtl. Crystal face indices were
assigned relative to the P21/c cell setting with the
aid of a video capture utility within the COLLECT
program. Video captures of the crystal were taken
along the directions of interest. Miller indices of
well-defined faces were deduced by inspection of
the crystal viewed along specific real and recipro-
cal space vectors (for instance, the (100) face was

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008 DOI 10.1002/jps
 THE CONFUSION OF INDEXING ASPIRIN CRYSTALS 1363

Figure 1. Video images used to index aspirin faces. a, b, and c are unit cell axes, while
a", b", and c" are the corresponding reciprocal axes of the P21/c setting.

found with vectors b and c in this plane). This is

shown in Figure 1.

Powder X-Ray Diffraction (PXRD)

The Miller indices of two major faces of aspirin
single crystals were further verified using a
powder X-ray diffractometer (MultiFlex, Rigaku
Co., The Woodlands, TX) with Cu Ka X-rays
(l ¼ 1.54178 Å). The two major faces of a single
crystal were placed face up on a sample holder,
with the faces horizontal with the surface of the
sample holder. The scan rate was set to 58 per
minute and the scan was taken from 28 to 408 of 2u.
Atomic coordinates from the single crystal X-ray
study was used in Materials Studio 3.5 (Accelrys,
Inc., San Diego, CA) to produce simulated powder
patterns for comparison with the experimental
data.

Contact Angle Measurements

Sessile drop contact angles were measured on the
two major faces of aspirin. Crystals were cleaved
to expose fresh faces and secured on a microscope
slide using adhesive tape. The contact angles were
obtained using a video-based contact angle system
(OCA, Future Digital Scientific Co., Bethpage,
NY). Drops (5 mL) of water, diiodomethane, and
glycerol were dispensed and placed on the two
faces using a motor driven syringe, respectively.
Water contact angles were measured in a satu-
rated water vapor environmental chamber to
prevent evaporation during spreading. Each drop-
let was recorded for 240 s. The contact-angle
values recorded for droplets on a surface were
obtained automatically via curve-fitting software.

RESULTS AND DISCUSSION

Figure 2. Aspirin crystals grown from (a) ethanol, (b)
Aspirin crystallized in ethanol, water, and acetone acetone, and (c) water. Scale bars denote 1 mm. [Color
results in plate morphology with two major faces figure can be seen in the online version of this article,
and a few different minor faces, shown in Figure 2. available on the website, www.interscience.wiley.com.]

DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008
1364 AUBREY-MEDENDORP, PARKIN, AND LI
The effects of solvent on crystallization of aspirin
demonstrate some morphological differences but
no polymorphism. This was validated in both
the XRD and powder X-ray diffraction (PXRD)
studies.
Aspirin crystals evaluated by single XRD
indicated the crystals were of the monoclinic
space group P21/c with a ¼ 11.242(7), b ¼ 6.539(4),
c ¼ 11.245(9) Å, and b ¼ 95.9(3)8. These values are
in agreement with the literature that previously
investigated the unit cell parameters of form I
using XRD,10,11 neutron diffraction,12,13 and
oscillation and Weissenberg X-ray photographs.7
These values are listed in Table 1. More recently
an elusive second form (form II) of aspirin has
been identified and characterized by single crystal
XRD. Though some controversy exists,22,23 unit
cell parameters (a ¼ 12.095(7), b ¼ 6.491(4),
c ¼ 11.323(6) Å, and b ¼ 111.509(9)8) are super-
ficially quite different from those of form I,
although they are likely related. Nonetheless,
both forms exhibit the same space group, P21/c.24
Comparison of unit cell parameters and atomic
coordinates of the two forms demonstrates a
subtle difference in the molecular packing
between forms I and II. In our crystallization
studies only form I was produced, though the
presence of a very small amount of form II cannot
be discounted.22,23
Figure 3. Growth morphology and crystal structure of aspirin. As indicated, there are
two possible truncations of the (100) face, one with carboxyl groups and another with
acetyloxy groups.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008
Our results demonstrate that for crystals
indexed as P21/c, the largest growth face is
(100) and the second largest is (001), as shown
in Figure 3. Based on these assignments and an
intimate view of the crystal packing, it is apparent
that there are two possible truncations for the
(100) face. One truncation goes through the
carboxyl group and another passes through
the acetyloxy groups. It is important to make
this distinction on the (100) face, especially when
probing surface properties in relation to func-
tional groups on that surface. Although the
standard setting for space group number 14 in
modern crystallography is P21/c (thereby placing
the glide plane along c), it is of course possible
to transform to the nonstandard P21/a setting
(via the transformation matrix (001, 0#10, 100). If
this transformation is made, then the assigned
face indices are also, of necessity, changed. If the
indexing of major faces from earlier reports is used
to correlate the crystal surface structure assigned
with a different space group setting, and no
account is made for the particular setting used,
then any results inferred will be in error. As
shown in Figure 3, if a surface property, such as
wettability, is measured on the (100) face but
explained using the functional groups present on
the (001) face, conclusions will be misleading. The
crucial point is to recognize which functional
DOI 10.1002/jps
 THE CONFUSION OF INDEXING ASPIRIN CRYSTALS
groups are exposed on the major faces whether the
space group is assigned as P21/a or P21/c. The axes
could be assigned according to any number of
conventions, but the functional groups assigned to
the physical faces are invariant. The dominant
face will always have the acetyloxy group (not the
carboxyl group as discussed below) and the other
major face will always have the methyl and phenyl
groups. This is illustrated in Figure 3.
Results from the PXRD work confirmed the
XRD studies and verified that the dominant face is
(100) for the P21/c cell setting. As illustrated in
Figure 4, experimental peaks of the single crystal
(100) and (001) faces correspond well to the
equivalent peaks in the simulated patterns. Due
to the systematic absences for P21/c, odd num-
bered reflections of the (00l) family, such as (001)
and (003), are systematic absences when generat-
ing XRD peaks; these hypothetical peaks are
colored green in the simulated pattern. The single
peak observed in the experimental pattern of the
(001) surface at 15.688 represents the (002) face,
and has an equivalent peak in the simulated
pattern. There should also be an observed peak,
similar to the simulated pattern at 328, but the
peak was too small to detect and was not observed
Figure 4. Experimental and simulated PXRD patterns of the (100) and (001) faces of
aspirin. Note the simulated (001) and (003) are hypothetical due to the systematic
absence.
DOI 10.1002/jps
1365
in the experimental pattern. On the other hand,
all four peaks from experimental data for the (100)
face agree with equivalent peaks from the
simulated data: (100) at 7.848 versus simulated
7.808, (200) at 15.768 versus simulated 15.848,
(300) at 23.608 versus simulated 23.808, and (400)
at 31.888 versus simulated 32.08. These results
verify the findings from the XRD study, and
confirm application of the P21/c space group as
opposed to using P21/a. The key feature to
distinguish the two faces is the (100) peak at
7.808. If there were no systematic absences
present, the (001) peak would be observed close
to the (100) peak at 7.908.
Contact angles were measured with various
solvents on the two major faces of aspirin, and
the results are summarized in Table 2. All of
the contact angles measured on the (100) face are
larger than those on the (001) face. Due to the
polar nature of the solvents used, the measure-
ments indicate the (100) face is less hydrophilic
than the (001), emphasizing the importance of
knowing the surface crystal structure. Although
the (100) face may expose the carboxyl acid group,
the contact angle measurements indicate other-
wise. It appears to be the acetyloxy truncation
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008
1366 AUBREY-MEDENDORP, PARKIN, AND LI
Table 2. Contact Angles on the (100) and (001) Faces
of Aspirin Measured in Selected Solvents
Water Diiodomethane Glycerol
(100) 66.9 $ 2.9 41.2 $ 1.7 65.1 $ 1.1
(001) 56.4 $ 2.4 36.2 $ 1.9 56.5 $ 0.9
that is exposed, which leads to the (100) face being
less hydrophilic than the (001) face. The hydrogen
bonding between carboxyl groups in the crystal is
strong; cleavage through this bonding to expose
the carboxyl group is unlikely. Once again, the
relationship between the interfacial properties
and surface structure could be misconstrued when
using the improper labeling and thereby identify-
ing the wrong surface functional groups. If the
more hydrophilic face is mistaken as the (100),
carboxyl groups would be thought as the dominant
surface group and that it would be involved in
solvent interactions. Our results show that the
carboxyl group is not exposed and therefore a
lower contact angle by a polar liquid is not
expected.
CONCLUSIONS
Analysis of the crystal structure of aspirin by
single X-ray and powder XRD has clarified the
confusion over the face indices of aspirin crystals.
The dominant face of aspirin crystals grown in
ethanol, water, and acetone is (100) when the
space group P21/c is assigned. The elucidation of
the indices also allowed proper assignment of
functional groups to the major faces and raised
awareness of two possible truncations on the (100)
face, one revealing the carboxyl group and the
other the acetyloxy moiety. Our contact angle
results further show that the (100) is less
hydrophilic than the (001) face when a polar
liquid is used in the measurement, suggesting
that the carboxyl group truncation is not mea-
sured. The common thought of the (100) face being
more hydrophilic in previous studies, including
one of ours,16 is incorrect. The mistake probably
stems from the inconsistent indexing caused by
the use of a nonstandard space group setting. The
face indexing of aspirin crystal is not incorrect,8,9
because their assignment is relative to their
choice of P21/a. However, if these face indices
are used to examine surface moieties based on
a P21/c description without the necessary index
transformation, the (100) face will be mistaken as
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008
(001), and vice versa. This is the cause of apparent
confusion in earlier studies, including one of our
own.16
This report also demonstrates the importance of
X-ray crystallography in solving both the crystal
structure and the relationship between alterna-
tive cell settings and indices assigned to crystal
faces. Without consistent indexing, erroneous
conclusions are inevitable. For the case of aspirin,
the confusion almost certainly arose from the
similarity in length of the a and c axes.
ACKNOWLEDGMENTS
The research was supported by the PhRMA Foun-
dation. The authors would like to thank Ms. Ellen
Savelli for her work with contact angle measure-
ments.
REFERENCES
1. Byrn SR, Pfeiffer RR, Stowell JG. 1999. Solid-State
Chemistry of Drugs. 2nd edition. West Lafayette,
IN: SSCI, Inc.
2. Vippagunta SR, Brittain HG, Grant DJW. 2001.
Crystalline solids. Adv Drug Deliv Rev 48:3–26.
3. Suryanarayanan R, Byrn SR. 2001. Characteriza-
tion of the solid state. Adv Drug Deliv Rev 48:1–136.
4. Kiang YH, Shi HG, Mathre DJ, Xu W, Zhang D,
Panmai S. 2004. Crystal structure and surface
properties of an investigational drug—A case
study. Int J Pharm 280:17–26.
5. Hoffmann F. Acetyl Salicylic Acid. US Patent
644,077, February 27, 1900.
6. Kozu S, Takane K. 1935. Cell-dimensions and
space-group of acetylsalicylic acid. Proc Imper Acad
11:381–382.
7. Nitta I, Watanabe T. 1937. Unit cell and space
group of acetylsalicylic acid. Sci Pap Inst Phys
Chem Res 31:125–128.
8. Watanabe A, Yamaoka Y, Kuroda K. 1980. Study of
crystalline drugs by means of polarizing micro-
scope. III. Key refractive indices of some crystalline
drugs and their measurement using an improved
immersion method. Chem Pharm Bull 28:372–378.
9. Watanabe A, Yamaoka Y, Takada K. 1982. Crystal
habits and dissolution behavior of aspirin. Chem
Pharm Bull 30:2958–2963.
10. Wheatley PJ. 1964. The crystal and molecular
structure of aspirin. J Chem Soc Suppl 6036–6048.
11. Kim Y, Machida K, Taga T, Osaki K. 1985. Struc-
ture redetermination and packing analysis of
aspirin crystal. Chem Pharm Bull 33:2641–2647.
DOI 10.1002/jps
 THE CONFUSION OF INDEXING ASPIRIN CRYSTALS
12. Harrison A, Ibberson R, Robb G, Whittaker G,
Wilson C, Youngson D. 2002. In situ neutron
diffraction studies of single crystals and powders
during microwave irradiation. Faraday Discuss
122:363–379.
13. Wilson CC. 2002. Interesting proton behaviour in
molecular structures. Variable temperature neu-
tron diffraction and ab initio study of acetylsalicylic
acid: Characterising librational motions and com-
paring protons in different hydrogen bonding
potentials. New J Chem 26:1733–1739.
14. Meenan P. 1997. Crystal morphology predictive
techniques to characterize crystal habit: Applica-
tion to aspirin (C9H8O4). ACS Symp Ser 667:2–
17.
15. Hammond RB, Pencheva K, Roberts KJ. 2006. A
structural-kinetic approach to model face-specific
solution/crystal surface energy associated with
the crystallization of acetyl salicylic acid from
supersaturated aqueous/ethanol solution. Crystal
Growth Des 6: 1324–1334.
16. Li T, Liu S, Feng S, Aubrey CE. 2005. Face-
integrated Fukui function: Understanding wett-
ability anisotropy of molecular crystals from den-
sity functional theory. J Am Chem Soc 127:1364–
1365.
DOI 10.1002/jps
1367
17. Niini R. 1931. The crystal form and optical con-
stants of aspirin. Z Kristallogr 79:532–536.
18. Nonius BV. 1997. KappaCCD Server Software.
Windows 3.11 edition. The Netherlands: Delft.
19. Otwinowski Z, Minor W. 1998. Denzo-SMN pro-
gram package. Processing of X-ray diffraction data
collected in oscillation mode. Methods Enzymol
276:302.
20. Mackay S, Dong W, Edwards C, Henderson A,
Gilmore C, Stewart N, Shankland K, Donald A.
1999. MaXus, Comprehensive crystallography soft-
ware. 4.0 edition. Glasgow: University of Glasgow.
21. Sheldrick GM. 1990. SHELXTL PC, An integrated
system for solving, refining, and displaying crystal
structure from diffraction data. 4.1 edition. Madi-
son: Siemens Analytical X-ray Instruments.
22. Bond AD, Boese R, Desiraju GR. 2007. On the
polymorphism of aspirin. Angew Chem Int Ed
46:615–617.
23. Bond AD, Boese R, Desiraju GR. 2007. On the
polymorphism of aspirin: Crystalline aspirin as
intergrowths of two ‘‘polymorphic’’ domains. Angew
Chem Int Ed 46:618–622.
24. Vishwesher P, McMahon JA, Oliveira M, Peterson
ML, Zaworotko MJ. 2005. The predictably elusive
form II of aspirin. J Am Chem Soc 127:16802–16803.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 97, NO. 4, APRIL 2008