Controlling molecular crystal polymorphism with self-assembled monolayer templates

Article abstract of DOI:10.1021/ja0565119

The control of crystal polymorphism is a long-standing issue in solid-state chemistry, which has many practical implications for a variety of commercial applications. At least four different crystalline forms of 1,3-bis(m- nitrophenyl) urea (MNPU), a clas

Full text of DOI:10.1021/ja0565119

Published on Web 12/06/2005
Controlling Molecular Crystal Polymorphism with
Self-Assembled Monolayer Templates
Rupa Hiremath, Joseph A. Basile, Stephen W. Varney, and Jennifer A. Swift*
Contribution from the Department of Chemistry, Georgetown UniVersity,
37th and “O” Streets NW, Washington, D.C. 20057-1227
Received September 27, 2005; E-mail: jas2@georgetown.edu
Abstract: The control of crystal polymorphism is a long-standing issue in solid-state chemistry, which has
many practical implications for a variety of commercial applications. At least four different crystalline forms
of 1,3-bis(m-nitrophenyl) urea (MNPU), a classic molecular crystal system, are known to crystallize from
solution in various concomitant combinations. Herein we demonstrate that the introduction of gold-thiol
self-assembled monolayers (SAMs) of substituted 4′-X-mercaptobiphenyls (X ) H, I, and Br) into the
crystallization solution can serve as an effective means to selectively template the nucleation and growth
of R-, â-, and γ-MNPU phases, respectively. Polymorph control in the presence of SAM surfaces persists
under a variety of solution conditions and consistently results in crystalline materials with high phase purity.
The observed selectivity is rationalized on the basis of long-range two-dimensional geometric lattice matching
and local complementary chemical interactions at the SAM/crystal interfaces.
crystallization in confined spaces, such as capillaries6,7 and
micropores,8 or in the presence of molecular9,10 or polymeric11-13
Introduction
One of the long-standing challenges in organic solid-state
chemistry is the ability to predict and control the occurrence of
polymorphism, the ability of a molecule to crystallize in more
than one packing arrangement.1 This has broad practical
implications for a number of industries, ranging from pharma-
ceuticals (drugs) to textiles (dyes and pigments) to defense
(energetic materials). While in principle it may be possible to
experimentally define the occurrence domain(s)2 under which
a particular form of a molecular solid crystallizes, establishing
such conditions in practice can be far more complicated than
simply specifying a narrow range of solvents, temperatures, and
cooling/evaporation rates. There are also a number of examples
in which two or more polymorphs form concomitantly under
essentially the same crystallization conditions.3 In part, this may
be a consequence of the fact that crystallization begins with
the nucleation stage, usually considered to be a heterogeneous
process, and the total contents of any given solution can never
be completely known.
additives. Polymorph control has also been achieved by epitaxial
nucleation and growth on single-crystal substrates.14,15
Other types of ordered 2D surfaces, such as Langmuir films
and self-assembled monolayers (SAMs), have been used to
selectively nucleate and grow a number of different inorganic
crystals,16-22 though their use as templates for molecular crystals
has been somewhat more limited.23-27 In such studies, the
templating function is generally attributed to strong ionic or
hydrogen-bonding interactions formed across the monolayer/
crystal interface. In our own recent work, we have shown that
(6) Chyall, L. J.; Tower, J. M.; Coates, D. A.; Houston, T. L.; Childs, S. L.
Cryst. Growth Des. 2002, 2, 505-510.
(7) Hilden, J. L.; Reyes, C. E.; Kelm, M. J.; Tan, J. S.; Stowell, J. G.; Morris,
K. R. Cryst. Growth Des. 2003, 3, 921-926.
(8) Ha, J.-M.; Wolf, J. H.; Hillmyer, M. A.; Ward, M. D. J. Am. Chem. Soc.
2004, 126, 3382-3383.
(9) Weissbuch, I.; Popovitz-Biro, R.; Lahav, M.; Leiserowitz, L. Acta Crys-
tallogr. 1995, B51, 115-148.
(10) Davey, R. J.; Blagden, N.; Potts, G. D.; Docherty, R. J. Am. Chem. Soc.
1997, 119, 1767-1772.
(11) Staab, E.; Addadi, L.; Leiserowitz, L.; Lahav, M. AdV. Mater. 1990, 2,
40-43.
While the conventional methods of polymorph screening by
slurry conversion and/or varying solvent, temperature, and
supersaturation have been very much accelerated by the use of
high throughput screening methods,4,5 a variety of complemen-
tary approaches to controlling and/or discovering new poly-
morphs have also been pursued in recent years. These include
(12) Lang, M.; Grzesiak, A. L.; Matzger, A. J. J. Am. Chem. Soc. 2002, 124,
14834-14835.
(13) Price, C. P.; Grzesiak, A. L.; Matzger, A. J. J. Am. Chem. Soc. 2005, 127,
5512-5517.
(14) Bonafede, S. J.; Ward, M. D. J. Am. Chem. Soc. 1995, 117, 7853-7861.
(15) Mitchell, C. A.; Yu, L.; Ward, M. D. J. Am. Chem. Soc. 2001, 123, 10830-
10839.
(16) Heywood, B. R.; Mann, S. AdV. Mater. 1994, 16, 9-20.
(17) Popovitz-Biro, R.; Lahav, M.; Leiserowitz, L. J. Am. Chem. Soc. 1991,
113, 8943-8944.
(1) Bernstein, J. Polymorphism in Molecular Crystals; Oxford University
Press: New York, 2002.
(2) Sato, K.; Boistelle, R. J. Cryst. Growth 1984, 66, 441-450.
(3) Bernstein, J.; Davey, R. J.; Henck, J.-O. Angew. Chem., Int. Ed. 1999, 38,
3440-3461.
(4) Peterson, M. L.; Morissette, S. L.; McNulty, C.; Goldsweig, A.; Shaw, P.;
LeQuesne, M.; Monagle, J.; Encina, N.; Marchionna, J.; Johnson, A.; Cima,
M. J.; Almarsson, O. J. Am. Chem. Soc. 2002, 124, 10858-10959.
(5) Morissette, S. L.; Soukasene, S.; Levinson, D. A.; Cima, M., J.; Almarsson,
O. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 2180-2184.
(18) Aizenberg, J.; Black, A. J.; Whitesides, G. M. J. Am. Chem. Soc. 1999,
121, 4500-4509.
(19) Kuther, J.; Seshadri, R.; Knoll, W.; Tremel, W. J. Mater. Chem. 1998, 8,
641-650.
(20) Travaille, A. M.; Kaptjin, L.; Verwer, P.; Hulsken, B.; Elemans, J. A. A.
W.; Nolte, R. J. M.; van Kempen, H. J. Am. Chem. Soc. 2003, 125, 11571-
11577.
(21) Bandyopadhyay, K.; Vijayamohanan, K. Langmuir 1998, 14, 6924-6929.
(22) Meldrum, F. C.; Flath, J.; Knoll, W. Langmuir 1997, 13, 2033-2049.
9
10.1021/ja0565119 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 18321-18327
18321

A R T I C L E S
Hiremath et al.
structure of the γ monohydrate phase and, serendipitously,
discovered a previously unknown anhydrous phase, δ, (P21/c:
a ) 4.686 Å, b ) 18.427 Å, c ) 17.72 Å, â ) 90.5°) which
crystallizes as white needles. Depending on the solution
conditions, they observed that the known crystal phases grew
concomitantly in various combinations, but despite attempts to
identify the individual occurrence domains for each of the four
phases, only the â form could be consistently obtained with
high phase purity using conventional crystallization methods.36
Figure 1. Yellow prisms of the R form and white needles of the â form
grow concomitantly from supersaturated ethanol solutions. Scale bar ) 1
mm.
In the present work, we demonstrate that three of the MNPU
phasessR, â and γscan each be selectively nucleated and
reproducibly grown on appropriately functionalized mercapto-
biphenyl SAM templates. The control over nucleation is
rationalized on the basis of favorable intermolecular and
geometric interactions at the SAM/crystal interface.
SAM templates can be used to control the absolute growth
direction of polar crystals of 4-iodo-4′-nitrobiphenyl28 and to
selectively nucleate a less stable form 2-iodo-4-nitroaniline.29
These studies also demonstrate that weaker forces at the SAM/
crystal interface, such as I‚‚‚NO2 and van der Waals interactions,
can be used to direct the nucleation process.
Experimental Section
In the present study, we apply SAM template methods to the
problem of controlling polymorphism in a classic molecular
crystal system, 1,3-bis(m-nitrophenyl) urea (MNPU). First
studied in the late 19th century,30-32 MNPU was identified by
Groth as having at least three different phasessR, â, and γ.
The R (yellow prisms), â (white needles), and γ (yellow
plates) phases are visually distinguishable, but often crystallize
concomitantly from aqueous ethanol solutions (Figure 1). Nearly
a century later, Etter et al.33-35 reported the first crystal structures
for the R (P21/c: a ) 11.495 Å, b ) 13.816 Å, c ) 8.307 Å, â
) 91.92°) and â (C2: a ) 20.95 Å, b ) 7.412 Å, c ) 6.715 Å,
â ) 104.96°) forms as part of a larger study on the hydrogen-
bonding patterns and cocrystallization behavior of diaryl ureas.
At the time our SAM template study was initiated, the structure
of the γ form was unknown. In the course of our work, we
were able to determine that the γ phase is actually a monohy-
drate (C2: a ) 24.876 Å, b ) 7.386 Å, c ) 3.757 Å, â )
96.85°). Bernstein et al.36 also recently obtained the crystal
Synthesis. 1,3-Bis(m-nitrophenyl) Urea (MNPU) was syn-
thesized according to the literature34 by stirring 1:1 mixtures of
3-nitrophenyl isocyanate (Aldrich, 97%) and 3-nitroaniline
(Aldrich, 98%) in benzene for 24 h at room temperature under
dry nitrogen. The product readily precipitates and is easily
isolated in pure form (mp ) 256-258 °C, in agreement with
previous reports).
4′-Bromo-4-mercaptobiphenyl (3) was prepared according
to the literature.37 Tetrakis(triphenylphosphine) palladium cata-
lyst (0.2 mmol, Aldrich, 99%) was added (under dry N2) to a
solution of 1-bromo-4-iodobenzene (20 mmol, Alfa Aesar,
97%+) in 10 mL of anhydrous THF and brought to gentle reflux
with stirring. A Grignard solution was prepared from mixing
4-bromothioanisole (20 mmol, Aldrich, 97%) and magnesium
turnings (20 mmol, Aldrich, 98%) in 25 mL of anhydrous
tetrahydrofuran (Aldrich, 99.9%) over 70 min. The Grignard
solution was transferred dropwise to the boiling 1-bromo-4-
iodobenzene solution. After addition was completed, the mixture
was held under reflux for an additional 30 min. After cooling
to room temperature, the mixture was poured into an ice-cold
solution of 5% HCl. The resulting solid intermediate, 4′-bromo-
4-methylthiobiphenyl, was vacuum filtered, washed with dis-
tilled water, dried on the aspirator, and recrystallized from 1:1
(23) Frostman, L. M.; Bader, M. M.; Ward, M. D. Langmuir 1994, 10, 576-
582.
(24) Kang, J. F.; Zaccaro, J.; Ulman, A.; Myerson, A. Langmuir 2000, 16, 3791-
3796.
(25) Lee, A. Y.; Ulman, A.; Myerson, A. S. Langmuir 2002, 18, 5886-5898.
(26) Banno, N.; Nakanishi, T.; Matsunaga, M.; Asahi, T.; Osaka, T. J. Am. Chem.
Soc. 2004, 126, 428-429.
(27) Briseno, A. L.; Aizenberg, J.; Han, Y.-J.; Penkala, R. A.; Moon, H.;
Lovinger, A. J.; Kloc, C.; Bao, Z. J. Am. Chem. Soc. 2005, 127, 12164-
12165.
(28) Hiremath, R.; Varney, S. W.; Swift, J. A. Chem. Mater. 2004, 16, 4948-
4954.
(29) Hiremath, R.; Varney, S. W.; Swift, J. A. Chem. Commun. 2004, 2676-
2677.
(30) Offret, A.; Vittenet, H. Bull. Soc. Chim. Fr. 1899, 21, 151-153.
(31) Offret, A.; Vittenet, H. Bull. Soc. Chim. Fr. 1899, 22, 788-797.
(32) Groth, P. An Introduction to Chemical Crystallography; Wiley: New York,
1906.
(33) Etter, M.; Panuto, T. J. Am. Chem. Soc. 1988, 110, 5896-5897.
(34) Etter, M. C.; Urbanczyk-Lipkowska, Z.; Zia-Ebrahimi, M.; Pannuto, T.
W. J. Am. Chem. Soc. 1990, 112, 8415-8426.
(35) Huang, K.-S.; Britton, D.; Etter, M. C.; Byrn, S. R. J. Mater. Chem. 1995,
5, 379-383.
(36) Rafilovich, M.; Bernstein, J.; Harris, R. K.; Apperley, D. C.; Karamertzanis,
P. G.; Price, S. L. Cryst. Growth Des. 2005, 5(6), 2197-2209.
(37) Ulman, A. Acc. Chem. Res. 2001, 34, 855-863.
9
18322 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Controlling Crystal Polymorphism with SAM Templates
A R T I C L E S
Table 1. Summary of Experimentally Determined Monolayer
Thicknesses Determined by Contact Angle and Ellipsometry
Measurements of SAMs of 1-7
heptane:2-propanol: mp 150-151 °C; 1H NMR (δ ppm, CDCl3)
2.52 (3H), 7.30-7.34 (4H), 7.41-7.56 (4H).
4′-Bromo-4-methylthiobiphenyl (4 mmol) was subsequently
reduced by dissolving it in 12 mL of anhydrous N,N-dimeth-
ylformamide (Aldrich, 99.8%) under dry nitrogen and adding
4 mmol of sodium ethanethiolate (Aldrich, 80%). It was stirred
and refluxed gently for 6 h. After cooling to room temperature,
the mixture was poured into a solution of ice-cold 5% HCl.
The solid was vacuum filtered, washed with distilled water, and
dried on the aspirator. The solid was recrystallized in 1:1
heptane:2-propanol and purified by silica gel column chroma-
SAM
1
2
3
4
5
6
7
contact angle ( 0.8° (lit)a 71° (73) 78° (79) 79° (81) 61° (64) 83° 78° 82°
ellipsometry ( 1 Å (lit)a 13 (14) 15 (15) 14 (14) 14 (14) 11 11 12
a The literature values are taken from Ulman, A. Acc. Chem. Res. 2001,
34, 855-863.
SAMs on the Au surface was estimated using a Gaertner
ellipsometer (Skokie, IL) (He-Ne laser, λ ) 632.8 nm, angle
of incidence 70°). Three separate points on two independent
samples of each type of monolayer were measured and averaged,
using an assumed refractive index of 1.46. The variation in the
measured SAMs thickness was (1 Å.
1
tography (1:1 CH2Cl2:hexane): mp 152-154 °C; H NMR (δ
ppm, CDCl3) 2.52 (1H), 7.30-7.39 (4H), 7.48-7.51 (4H).
3′-Bromo-4-mercaptobiphenyl (6) was prepared according
to the literature as for 3, substituting 1-bromo-3-iodobenzene
(Aldrich, 98%) for 1-bromo-4-iodobenzene. Data for the 3′-
bromo-4-methylthiobiphenyl intermediate: mp 123-124 °C; 1H
NMR (δ ppm, CDCl3) 2.52 (3H), 7.25-7.33 (4H), 7.48-7.51
(4H). Data for (6): mp 187-188 °C; 1H NMR (δ ppm, CDCl3)
2.52 (1H), 7.29-7.34 (4H), 7.79-7.51 (4H).
4-Mercaptobiphenyl (1), 4′-iodo-4-mercaptobiphenyl (2), 4′-
nitro-4-mercaptobiphenyl (4), 3′-iodo-4-mercaptobiphenyl (5),
and 3′-nitro-4-mercaptobiphenyl (7) were prepared under analog-
ous reaction conditions. Their syntheses and characterization
have been reported elsewhere.28,29
MNPU Crystal Growth from Solution. The R polymorph
(yellow prisms) was obtained from crystallization in ethanol at
60 °C (mp ) 256-258 °C). The â polymorph (white needles)
was obtained from crystallization in ethyl acetate at room
temperature. Upon heating, â crystals undergo a phase trans-
formation at ∼200 °C before converting to R and melting at
256-258 °C. The γ monohydrate form (yellow plates) is
obtained by a solution-mediated dissolution-recrystallization
of either R and â forms after approximately 1 week of sitting
in 95% ethanol solution. Ironically, this phase transformation
does not seem to occur in aqueous solution. Upon heating, γ
undergoes a phase transformation at ∼90 °C before converting
to R and melting at 256-258 °C. We have extensively searched
for but have not been able to grow crystals of the δ form.
MNPU Crystal Growth on SAMs. Saturated solutions of
MNPU in either ethanol or ethyl acetate solution were prepared
in vials. All SAMs were positioned such that they leaned against
the side of the vial (not laying flat on the bottom), and the vials
were covered with Parafilm. Crystals grown on the function-
alized surface were in the range of 0.5-3 mm and typically
appeared within 5-6 days. To perform experiments at 60 °C,
vials were placed in a digital dry bath (Labnet, Edison, NJ).
SAM substrates were first epoxied to larger strips of mica in
order to prevent the substrates from falling due to any vibrations
in the dry bath. Crystal growth on SAMs was attempted under
multiple solvent and temperature conditions. Crystal growth was
also attempted on two types of control surfacessbare gold
substrates and alkanethiol SAMs of 1-pentanethiol (Aldrich,
98+%), 1-octanethiol (Aldrich, 98.5+%), and 1-dodecanethiol
(Aldrich, 98+%).
X-ray Crystallography. A Siemens SMART Platform CCD
diffractometer was used to collect single-crystal X-ray diffraction
data for R-, â-, and γ-MNPU crystals grown in the presence or
absence of SAM templates. Unit cells matching the known
crystal structures were obtained from the collection of 20
diffraction peaks, and Miller indices were subsequently assigned
to all possible crystal faces.
A full data collection and structure solution of γ-MNPU was
conducted at the University of Minnesota X-ray Crystallographic
Laboratory. The data collection was carried out at 173 K using
Mo KR radiation (graphite monochromator). Intensity data were
corrected for absorption and decay. The structure was solved
and refined based on 1506 strong reflections using Bruker
SHELXTL. All non-hydrogen atoms were located with a
combination of direct methods and full-matrix least-squares/
difference Fourier cycles and were refined with anisotropic
All solvents were obtained from Aldrich, EM Science (ACS
grade), and Fischer Scientific (HPLC grade) and used without
further purification. All melting points reported were determined
by Differential Scanning Calorimetry (Thermal Instruments,
1
Richmond, VA) at a heating rate of 5°/min. H NMR spectra
were measured using a Mercury 300 MHz (Varian) NMR
spectrometer in CDCl3 solution at ambient temperature. All
chemical shifts (δ) are referenced to tetramethylsilane (TMS).
Comparison of the experimental and literature values confirms
the identities of the products.
Monolayer Preparation and Characterization. Gold sub-
strates (3000 Å, 5 mm2) were prepared in the Georgetown
Advanced Electronics Laboratory (GAEL) via high-pressure
gold sputtering. Mica squares (5 mm2) were placed on glass
slides using Post-it adhesive and double-sided tape. Scotch tape
was used to remove the top layer, revealing a fresh layer for
exposure to sputtering. A 30 Å layer of chromium was deposited
at a rate of 3 Å/s, followed by 3000 Å of gold. The average
mean roughness (Ra)38 of the gold substrate (1.4 nm) was
determined in air using contact mode AFM performed on a
Digital Instruments Multimode Nanoscope IIIa (Santa Barbara,
CA). Gold substrates were cleaned in piranha solution (70%
H2SO4, 30% H2O2) for 1 min, rinsed with distilled water then
ethanol, and finally dried under nitrogen. The substrates were
then immersed in an ethanolic thiol solution (2 mM) for 24 h
at room temperature. The substrates were removed from
solution, rinsed with ethanol, and dried under a stream of
nitrogen.
Contact angle measurements of the SAMs on gold were
determined by the half-angle measuring method, using a CAM-
PLUS MICRO Contact Angle Meter by Tantec, Inc. (Schaum-
burg, IL). All contact angles reported are the averages of 10
readings made on three independently prepared SAMs. The
accuracy of the measurements is (0.8°. The thickness of the
(38) Priest, C. I.; Jacobs, K.; Ralston, J. Langmuir 2002, 18, 2438-2440.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18323

A R T I C L E S
Hiremath et al.
atoms. The distances are smallest in the γ monohydrate form
(9.060 Å) and increase in the anhydrous â (9.937 Å), δ (11.322
Å), and R (11.629 Å) polymorphs. Readers are referred to ref
36 for additional details regarding molecular and lattice energy
calculations, graph set analyses, and thermal characterizations
of the various MNPU crystal forms.
For the purposes of understanding template-directed growth,
our interest in these materials is focused not on their bulk
structures, but rather on the topographical and chemical
characteristics of the different crystal growth faces observed in
the various MNPU forms. Unit cells obtained from X-ray
diffraction measurements and goniometry were used to assign
Miller indices to all faces of solution-grown R-, â-, and
γ-MNPUs. Crystals of R and â crystals can be grown from either
ethanol or ethyl acetate, and the choice of growth solvent does
not appear to alter their morphologies in any significant way.
R-MNPU grows as yellow prisms bounded by (010), (1h11),
(111h), (100), (110), (001), (011), and symmetry related faces
(see Figure 2, top). â-MNPU grows as white needles elongated
along the crystallographic a axis and is bounded by (100), (1h00),
(010), (01h0), (001), and (001h) faces (see Figure 2, middle).
γ-MNPU grows in a yellow plate-like morphology as a result
of a solution-mediated transformation of either the R or â form
in ethanol. Crystals of the γ form tend to grow slightly elongated
in the crystallographic c direction and exhibit (100), (1h00), (010),
(01h0), (001), and (001h) faces (see Figure 2, bottom). Previous
morphological characterization of the γ form by Groth32
suggested the presence of (110) faces, although we have not
observed this. We do commonly observe higher order (h0l) faces
in the γ form, but the exact Miller indices of these faces are
hard to specify with certainty since they often appear curved.
Notably, though both â and γ are polar phases, neither exhibits
the polarity morphologically. All efforts to reproduce the
solution growth of the δ-MNPU phase discovered by Bernstein
et al. have so far been unsuccessful, although this is not the
first time a particular polymorph can be grown in one laboratory
and not another.40
Growth of MNPU on SAM Templates. In our de novo
design of SAM templates that might be used to selectively
nucleate individual polymorphs with high phase purity, we
assumed that 2D lattice matching and/or complementary chemi-
cal matching across the SAM/crystal interface would be two
critical factors. Any polymorph selectivity observed in the
presence of a SAM would presumably originate from favorable
interactions between the two contacting surfaces. The chemical
functionality of gold-thiol monolayers is relatively easy to
control with organic synthesis. For predictable control over the
lattice dimensions, we regarded arenethiols to be potentially
better suited to molecular crystals than the more widely used
alkanethiols. Depending on the alkyl chain length, headgroup,
and solvent, alkanethiols can adopt very different tilt angles with
respect to the surface normal, a factor that also changes the 2D
lattice periodicity on the surface. With arenethiols, close-packing
considerations necessitate a more perpendicular molecular
alignment at the surface. Previous studies of 4′-X-4-mercapto-
biphenyl SAMs reveal that small tilt angles and herringbone
packing arrangements are observed for a number of these
monolayers.37,41-43 The slightly larger size of the 2D lattice
Figure 2. Crystal packing diagrams and observed morphologies of R-
(top),35 â- (middle),35 and γ- (bottom) MNPU constructed from fractional
coordinates. Morphology figures were generated using WinXMorph, version
1.1.39
displacement parameters. All hydrogen atoms were placed in
ideal positions and refined as riding atoms with relative isotropic
displacement parameters. The final full-matrix least-squares
refinement converged to R1 ) 0.0341 and wR2 ) 0.0903 (F2,
all data).
Results and Discussion
Molecular Conformation, Morphology, and Surface Struc-
tures of MNPU Forms. Packing diagrams of the three known
forms appear in Figure 2. One of the most striking differences
among the crystal structures is the conformation of the MNPU
molecule. In solution, MNPU can adopt many different con-
formations due to the low rotational barrier about the central
C-N amide bonds; however, in the solid state, the molecular
conformation is fixed. In the R phase, both of the nitro groups
are oriented anti with respect to the central carbonyl bond. In
â and γ forms, both of the nitro groups are oriented syn with
respect to the carbonyl group. In the δ form (not shown), one
nitro group is syn to the carbonyl while the other is anti. These
very different molecular conformations also change the in-
tramolecular O2N‚‚‚NO2 distances between the two nitrogen
(39) Kaminsky, W. J. Appl. Crystallogr. 2005, 38, 566-567.
(40) Dunitz, J. D.; Bernstein, J. Acc. Chem. Res. 1995, 28, 193-200.
9
18324 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Controlling Crystal Polymorphism with SAM Templates
A R T I C L E S
Table 2. Crystal Growth Conditions of MNPU Relative to the Type
of SAM Used, Initial Crystal Phase(s) of MNPU Used, Solvent,
and Temperature
SAM
initial solute phase
solvent
temp (
°C)
growth
1
R, â, or γ or any
ethanol or
25 or 60
25 or 60
25 or 60
R
combination of phases ethyl acetate
R, â, or γ or any ethanol or
combination of phases ethyl acetate
R, â, γ or any ethanol or
combination of phases ethyl acetate
R, â, γ or any ethanol or
combination of phases ethyl acetate
2
3
â
γ
4, 5, 6, 7,
bare gold,
or alkanethiol
controls
25 or 60 none
dimensions afforded by aromatic SAMs (compared to alkanethi-
ols) would also make them potentially more compatible with
the typical unit cell dimensions (5-20 Å) of most molecular
crystals. SAMs of 3′-X-4-mercaptobiphenyl SAMs29 present the
same types of functional groups on a 2D surface, albeit with a
different orientation, and were additionally explored. Ellipsom-
etry measurements and contact angle characterization suggest
small tilt angles with respect to the normal to the gold surface,
though we do not know the extent to which the general shape
of these molecules might affect the long-range 2D order on the
monolayer.
SAMs of 1-7 were introduced into saturated ethanol or ethyl
acetate solutions of MNPU prepared in conventional glass vials
and positioned so as to expose the gold surface to the bulk
solution. To minimize the possibility of unintentional seeding
due to incompletely dissolved MNPU, saturated solutions were
prepared from pure R, â, or γ phases, or polymorphic mixtures.
Solutions were covered with Parafilm and maintained at either
room temperature (24 ( 1 °C) or at 60 °C in a digital dry bath.
After sitting undisturbed for a period of 5-6 days, visible
millimeter-sized crystals were consistently observed on SAMs
1, 2, and 3. No MNPU crystals were ever observed on SAMs
4-7. Additionally, no crystals were ever found to grow on bare
gold substrates or alkanethiol monolayer (1-pentanethiol, 1-oc-
tanethiol, 1-dodecanethiol) controls.
Figure 3. Selective nucleation and growth of R-MNPU on a 4-mercapto-
biphenyl SAM template, SAM 1. Top: Micrograph of an R-MNPU prism
attached to SAM 1. View is down the b axis. Scale bar ) 0.1 mm. Bottom:
Schematic of the R-(010)/SAM 1 interface. Red arrows indicate the direction
of the molecular dipole moment.
SAMs of 1, 2, and 3 were each found to serve as selective
nucleating templates for a single but unique MNPU phase
regardless of the initial solute phase. Crystals of R-MNPU
formed on monolayers of 1, â-MNPU grew selectively on 2,
and γ-MNPU grew on 3. Multiple crystallization experiments
prepared under identical conditions confirmed the reproducibility
of this effect (Table 2). The resultant crystal form obtained was
completely independent of the solvent choice and/or temperature
conditions explored. In all cases where growth was observed,
crystals adopted preferred orientations with respect to the SAM
templates, thereby illustrating the importance of the interface
in the nucleation process. The three interfaces observed were
SAM 1/R-(010), SAM 2/â-(01h0), and SAM 3/γ-(01h0). We note
that because the bulk and surface structures of R, â, and γ phases
are different, it is sheer coincidence that the observed faces in
contact with the SAMs share the same Miller indices.
Figure 4. Selective nucleation and growth of â-MNPU on a 4′-iodo-4-
mercaptobiphenyl SAM template, SAM 2. Top: Micrograph of a â-MNPU
needle attached to SAM 2. View is down the b axis. Scale bar ) 0.3 mm.
Bottom: Schematic of the â-(01h0)/SAM 2 interface. The red arrow indicates
the direction of the molecular dipole moment.
(41) Sabatani, E.; Cohen-Boulakia, J.; Bruening, M.; Rubinstein, I. Langmuir
1993, 9, 2974-2981.
(42) Lee, S.; Puck, A.; Graupe, M.; Colorado, R.; Shon, Y.-S.; Lee, T. R.; Perry,
S. S. Langmuir 2001, 17, 7364-7370.
Rationalizing Polymorph Selectivity. To discern the origin
of the selectivity of each SAM template for a particular phase,
(43) Shaporenko, A.; Heister, K.; Ulman, A.; Grunze, M.; Zharnikov, M. J.
Phys. Chem. B 2005, 109, 4096-4103.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18325

A R T I C L E S
Hiremath et al.
R-(1h11), R-(111h), â-(010), and γ-(001). Epitaxy screening
against the 4′-chloro-4-mercaptobiphenyl parameters yielded five
potential coincident matches: R-(010), R-(100), â-(010), γ-(001),
and γ-(010). We note that epitaxy calculations do not distinguish
between the same family of crystal faces, some of which in â
and γ phases are chemically distinct. Notice that all three of
the experimentally observed contacting surfaces (R-(010),
â-(01h0), and γ-(01h0)) appear in the list, though it is not possible
to predict the single best experimental match based on geometric
considerations alone.
Analysis of the functional group densities on the different
surfaces provides additional insight into the nature of the SAMs’
templating effect. Although the strength of a single nitro‚‚‚
halogen interaction is small (∼5.7 kJ mol-1),47,48 short NO2‚‚‚
halogen contacts are frequently observed in aromatics bearing
these substituents. Such interactions at SAM/crystal interfaces
have also been exploited to control the orientation and polar
growth direction of molecular crystals.28 SAMs of 2 and 3 were
observed to be in contact with â-(01h0) and γ-(01h0), respectively.
Importantly, both of these MNPU phases are polar. The polar
axis in each of these crystals parallels the crystallographic b
axis. Consequently, the (b faces expose different functionalities,
with only the (01h0) faces presenting NO2 groups on the surface.
Of all the potential surfaces listed in Table 3, the â-(01h0) and
γ-(01h0) faces express some of the highest NO2 densities. While
the NO2 density of γ-(100) is marginally higher than that of
â-(01h0), making it a potential theoretical match, the nearly
parallel (4.3°) orientation of nitro groups on this surface
presumably limits its ability to coordinate strongly with the
halogens on the underlying substrate. Similar reasoning might
also be used to discount a hypothetical preference for nucleation
on the â-(100) face, in which nitro groups are high in density
but oriented slightly toward the interior of the crystal and not
outward on the crystal surface.
Figure 5. Selective nucleation and growth of γ-MNPU on a 4′-bromo-4-
mercaptobiphenyl SAM template, SAM 3. Top: Micrograph of a γ-MNPU
needle attached to SAM 3. View is down the b axis. Scale bar ) 0.3 mm.
Bottom: Schematic of the γ-(01h0)/SAM 3 interface. The red arrow indicates
the direction of the molecular dipole moment.
each of the natural surfaces of R, â, and γ polymorphs was
analyzed in terms of its specific 2D lattice dimensions and
functional group orientations and densities. Molecular models
of each crystal surface were constructed using Mercury 1.3
software and the relevant cif file. Such models make the intrinsic
assumption that no gross surface reconstructions occur in
solution and that the surface structures are nearly identical to
that of the bulk plane. Some of the significant structural
parameters for each observed crystal surface are provided in
Table 3.
The geometric lattice matching program EpiCalc44 was used
to determine which of the crystal surfaces of the various
polymorphs are epitaxially matched with the SAM. While the
long-range two-dimensional molecular ordering on a single
crystal surface is likely to be higher than that on a self-assembled
monolayer, SAMs of 4′-X-mercaptobiphenyls are expected to
exhibit domains with a high degree of 2D order.37,43 Other
authors45,46 have previously reported the two-dimensional lattice
parameters for 4′-methyl-4-mercaptobiphenyl and 4′-chloro-4-
mercaptobiphenyl monolayers as 4.8 × 10.0 Å, R ) 90° and
5.5 × 8.0 Å, R ) 90°, respectively. On the basis of the trends
observed in the SAMs characterization data (Table 1), param-
eters for all of the SAM surfaces used in this study are expected
to have dimensions in approximately the same range. Compu-
tational screening for epitaxial relationships between all MNPU
faces and the cited 4′-methyl-4-mercaptobiphenyl SAMs pa-
rameters resulted in five potential coincident matches: R-(010),
While there are a number of general similarities between the
â-(01h0) and γ-(01h0) surfaces, recall that the molecular confor-
mation in these phases is different. The intramolecular O2N‚‚‚
NO2 distance between nitro groups in a given molecule in the
â polymorph conformation is greater than that in γ (9.937 vs
9.060 Å). It is possible that the expanded O2N‚‚‚NO2 distance
in â is slightly better suited to coordinating the larger iodine
atom (r ) 2.15 Å).49 Coordination of the smaller bromine atom
(r ) 1.95 Å) might be preferred when the molecular conforma-
tion is more compact such as in γ. The contact angle measure-
ments of SAMs 2 and 3 are nearly identical, so it is unlikely
that SAM 3 would have a higher affinity for preexisting water
which might encourage growth of the hydrated NMPU phase.
Explaining the preference for the observed R-(010) growth
on the nonpolar surface of SAM 1 is more difficult. We
originally hypothesized that the MNPU phase most likely to
nucleate on this surface would be that which possesses the
lowest NO2 density. If this were the case, SAMs of 1 would
have theoretically preferred â-(100), â-(010), or γ-(001) sur-
faces, which do not present any NO2 groups on the surface. All
but two of the remaining surfaces have a lower NO2 group
(47) Hulliger, J.; Langley, P. J. Chem. Commun. 1998, 2557-2558.
(48) Allen, F. H.; Lommerse, J. P. M.; Hoy, V. J.; Howard, J. A. K.; Desiraju,
G. R. Acta Crystallogr. 1997, B53, 1006-1016.
(49) Pauling, L. The Nature of the Chemical Bond and the Structure of Molecules
and CrystalssAn Introduction to Modern Structural Chemistry, 2nd ed.;
Oxford University Press: London, 1940.
(44) Hillier, A.; Ward, M. D. Phys. ReV. B 1996, 54, 14037-14051.
(45) Azzam, W.; Fuxen, C.; Birkner, A.; Rong, H.-T.; Buck, M.; Woll, C.
Langmuir 2003, 19, 4958-4968.
(46) Kang, J. F.; Ulman, A.; Liao, S.; Jordan, R.; Yang, G.; Liu, G. Langmuir
2001, 17, 95-106.
9
18326 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Controlling Crystal Polymorphism with SAM Templates
A R T I C L E S
Table 3. Select Parameters for R-, â-, and γ-MNPU Crystal Surfacesa
orientation of C
−
NO2
no. of unique NO2 groups
projecting from surface
(NO2 density)
cell
bond wrt surface
surface
parameters
(0° ) |, 90° )
)
R-MNPU
(100) or (1h00)
(010) or (01h0)
(001) or (001h)
(110) or (1h1h0)
(011) or (01h1h)
(1h11) or (11h1h)
(111h) or (1h1h1)
8.31 × 13.82, 90°
38.7°
2 (1 per 57.42 Å2)b
2 (1 per 47.75 Å2)
2 (1 per 79.46 Å2)
3 (1 per 40.72 Å2)
3 (1 per 38.27 Å2)
3 (1 per 71.67 Å2)
3 (1 per 71.67 Å2)
8.31 × 11.50, 91.92°
11.50 × 13.82, 90°
20.8° (#1); 70.7° (#2)
41.6° (#1); 18.4° (#2)
45.2°, 23.2° (#1); 54.3°, -14.9° (#2)
70.5°, -19.3° (#1); 39.8°, 22.7° (#2)
42.1°, 44.0° (#1); 80.6°, -8.0° (#2)
44.1°, 42.2° (#1); 80.6°, -7.8° (#2)
11.50 × 13.82, 129.76°
8.31 × 16.12, 121.01°
13.95 × 16.12, 107.01°
13.95 × 16.12, 107.01°
â-MNPU
γ-MNPU
(100) or (1h00)
(01h0)
4.71 × 6.72, 90°
-11.9 °
43.3°
46.5°
0
6.72 × 20.95, 104.96°
4.71 × 20.95, 90°
4 (1 per 34.00 Å2)c
2 (1 per 49.34 Å2)
(001) or (001h)
(100) or (1h00)
(01h0)
(001) or (001h)
3.77 × 7.39, 90°
4.3°
66.1°
23.7°
1 (1 per 27.86 Å2)
4 (1 per 23.28 Å2)d
2 (1 per 91.93 Å2)
3.77 × 24.88, 96.85°
7.39 × 24.88, 90°
a Two-dimensional lattice dimensions are derived from fractional coordinates. The orientation of the C-N(NO2) bond with respect to the surface is
reported for each independent type of nitro group. NO2 densities are based on the number of NO2 groups presented on the outermost surface per unit cell
area. b The (200) plane of R-MNPU exposes no NO2 groups. c Only the (01h0) face of â-MNPU presents NO2 groups, none are projected on (010). d Only
the (01h0) face of γ-MNPU presents NO2 groups, none are projected on (010).
density, so based on this criteria, the observed contact with the
R-(010) surface seems an unlikely choice. With no obvious
reason for the nucleation preference, we offer two potential
hypotheses. MNPU molecules in the R form adopt an anti-
anti conformation. The dipole moment of any one molecule is
oriented nearly parallel to the CdO bond direction, but the
crystal lacks a net dipole due to centrosymmetry. The molecular
dipoles in R-MNPU crystals lie nearly parallel to the (010) plane.
This may be the least unfavorable orientation for a highly
polarized molecule to align against a relatively nonpolar surface.
Given that the R-(010) surface is also the most highly cor-
rugated, it may be more energetically favorable to undergo a
surface reconstruction to a structure that is more compact.
However, with the experimental methods currently available for
characterizing 2D interfaces in solution, it is difficult to probe
this further with sufficient molecular-level detail.
grown on these different SAMs adopt specific crystal orienta-
tions, which reflect their preferred geometry at the nucleation
stage. The R-MNPU form grows on SAM 1, the â-MNPU form
on SAM 2, and the γ-MNPU form on SAM 3. The observed
orientations have been rationalized on the basis of two-
dimensional lattice matching (epitaxy), complementary func-
tional group interactions, and/or dipole moments across the
SAM/crystal interface. We believe this template-based strategy
used to control heterogeneous nucleation events is sufficiently
general and may be a useful means to direct the course of
crystallization in other polymorphic systems.
Acknowledgment. J.A.S. is grateful for the financial support
provided by the National Science Foundation, the Camille and
Henry Dreyfus Foundation, and the Henry Luce Foundation.
S.W.V. and J.A.B. thank Georgetown University for under-
graduate research support through GUROP. We thank Victor
G. Young, Jr. at the X-ray Crystallographic Laboratory at the
University of Minnesota for solving the structure of γ-MNPU,
and Michal Rafilovich and Joel Bernstein for advanced informa-
tion on the δ crystal phase.
Conclusions
This study establishes that by exploiting chemical and
geometric interactions at 2D self-assembled monolayer interfaces
bearing different functionalities it is possible to direct the
nucleation and growth of a single crystalline phase in what is
otherwise a concomitant polymorphic system. Three different
crystal phases of MNPU, which grow concomitantly in various
combinations from solution, could each be selectively resolved
with high phase purity on gold-thiol SAM templates composed
of different 4′-X-4-mercaptobiphenyls. All MNPU polymorphs
Supporting Information Available: The cif file for the
γ-MNPU structure has been deposited in the Cambridge
Crystallographic Database. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA0565119
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18327