A versatile method for tuning the chemistry and size of nanoscopic features by living free radical polymerization

Article abstract of DOI:10.1021/ja028866n

A novel approach is presented for manipulating the size and chemistry of nanoscopic features using a combination of contact molding and living free radical polymerization. In this approach a highly cross-linked photopolymer, based on a methacrylate/acryla

Full text of DOI:10.1021/ja028866n

Published on Web 03/07/2003
A Versatile Method for Tuning the Chemistry and Size of
Nanoscopic Features by Living Free Radical Polymerization
Timothy A. von Werne,† David S. Germack,† Erik C. Hagberg,‡ Valerie V. Sheares,‡
Craig J. Hawker,*,† and Kenneth R. Carter*,†
Contribution from the IBM Almaden Research Center, Center for Polymeric Interfaces and
Macromolecular Assemblies, 650 Harry Road, San Jose, California 95120-6099, and
Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011
Received October 8, 2002; E-mail: hawker@almaden.ibm.com; kcarter@almaden.ibm.com
Abstract: A novel approach is presented for manipulating the size and chemistry of nanoscopic features
using a combination of contact molding and living free radical polymerization. In this approach a highly
cross-linked photopolymer, based on a methacrylate/acrylate mixture, was patterned into submicrometer-
sized features on a silicon wafer using a contact-molding technique. A critical component of the monomer
mixture was the incorporation of an initiator containing monomer into the network structure, which provides
sites for functional group amplification. Features ranging in size from 5 µm to <60 nm were accurately
replicated by this process and living free radical polymerizations, both atom transfer radical and nitroxide-
mediated polymerization (NMP), could be conducted from these initiating sites to yield polymer brushes
which represent a grafted layer of linear chains attached to the original network polymer. Grafts consisting
of polystyrene, poly(methyl methacrylate), and poly(2-hydroxyethyl)methacrylate were grown with controlled
thicknesses ranging from 10 to 143 nm and graft molecular weights of between 18 000 to 290 000 amu.
As a result of this secondary graft process, feature sizes could be tuned from the original 100 nm down to
20 nm, and the surface chemistry varied from hydrophilic to hydrophobic starting from the same initial
master pattern. The thin films and patterned features were characterized by contact angle, ellipsometry,
optical, and atomic force microscopies.
Introduction
structures have their own unique advantages and drawbacks.
For example, top-down approaches are well understood, repro-
The fabrication of nanometer-sized structures and devices is
an area of increasing importance due to its commercial
importance and application to the emerging field of nanoscale
science. Advances in the fields of molecular-scale electronics,
magnetic storage, optoelectronics, and biotechnology all depend
increasingly on the ability to fashion materials with precise
control of feature size and functionality.1 The fabrication
techniques have largely been split into two broad categories,
namely “top-down” approaches and “bottom-up” approaches.
Classical top-down processes include well-known microfabri-
cation techniques such as photolithography2-4 which is currently
used to make essentially all modern microelectronic devices.
Bottom-up approaches are more dependent on the chemistry
and functionality of the materials employed and includes the
use of self-organizing materials that can undergo pattern
formation at nanoscale levels to give technologically interesting
structures.5,6 Both strategies for the fabrication of nanoscale
ducible and scaleable, although advances become more difficult
as device features drop below the 100-nm level. In contrast,
bottom-up approaches show promise in the ability to make sub-
100 nm features, but thus far the ability to utilize bottom-up
approaches to fashion useful devices of known architecture has
proven elusive due to the inherent difficulty of directing the
exact positioning of self-assembling systems over large areas.
In an attempt to fabricate sub-100 nm nanostructures,
especially in the microelectronics arena, researchers have
concentrated on the patterning of thin organic films, which in
turn are utilized as masks or resist materials for other secondary
fabrication processes. Reported methods for the synthesis of
nanoscopically patterned thin organic films include state-of-
the-art photolithographic techniques, advanced lithographic
techniques such as e-beam patterning, and other nonoptical
contact patterning techniques. These contact lithographic tech-
niques include nanoimprint lithography,7 “step and flash”
† IBM Almaden Research Center.
(5) (a) Ma, Q. G.; Remsen, E. E.; Kowalewski, T.; Wooley, K. L. J. Am. Chem.
Soc. 2001, 123, 4627. (b) Thurn-Albrecht, T.; Schotter, J.; Kastle, C. A.;
Emley, N.; Shibauchi, T.; Krusin-Elbaum, L.; Guarini, K.; Black, C. T.;
Tuominen, M. T.; Russell, T. P. Science 2000, 290, 2126.
(6) Quite often discussions concerning bottom-up approaches have been
dominated by self-assembly of molecules and nano-objects through
noncovalent interactions. The polymeriztion of monomers directed to
prepatterned areas can also be considered a bottom-up approach, albeit
different from materials that self-organize at surfaces.
(7) Chou, S. Y.; Krauss, P. R.; Renstrom, P. J. J. Vac. Sci. Technol., B 1996,
14, 4129.
‡ Iowa State University.
(1) (a) Mirkin, C. A.; Rogers, J. A. MRS Bull. 2001, 26, 506. (b) Russell, T.
P. Science 2002, 297, 964.
(2) Klopp, J. M.; Pasini, D.; Byers, J. D.; Willson, C. G.; Frechet, J. M. J.
Chem. Mater. 2001, 13, 4147.
(3) Tully, D. C.; Trimble, A. R.; Frechet, J. M. J. AdV. Mater. 2000, 12, 1118.
(4) Tran, H. V.; Hung, R. J.; Chiba, T.; Yamada, S.; Mrozek, T.; Hsieh, Y. T.;
Chambers, C. R.; Osborn, B. P.; Trinque, B. C.; Pinnow, M. J.; MacDonald,
S. A.; Willson, C. G.; Sanders, D. P.; Connor, E. F.; Grubbs, R. H.; Conley,
W. Macromolecules 2002, 35, 6539.
9
10.1021/ja028866n CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 3831-3838
3831

A R T I C L E S
von Werne et al.
imprint lithography,8 nanocontact molding,,9 and derivations of
the above-mentioned microcontact printing of self-assembled
monolayers (SAMs). A recent report demonstrated the construc-
tion of <30-nm structures by a “molecular ruler” approach.10
In that work, metal lines were patterned by e-beam lithography
techniques followed by alteration of the pattern size by adsorbing
organic monolayers on the metal surfaces. Then a second metal
layer was deposited over the size-adjusted features, followed
by lift-off of the organic layer. Fabrication of metallic features
smaller than 30 nm was demonstrated by using this novel
process. While other research groups working in imprint
lithography have mainly focused on improving the physics of
the imprint process, we have concentrated on novel polymeric
materials used in the process, specifically to improve control
of the size and functionality of nanometer-sized features.
In concert with these advances, one bottom-up approach,
which has received considerable attention in recent years, is
the controlled growth of polymer brushes from surface-initiated
sites. This affords polymer brushes of well-defined nanoscopic
structure, molecular weight, and controllable molecular archi-
tecture, which have found application in a variety of different
areas. A number of surface-bound “living” polymerization
initiatorsincludingcationic,11 anionic,12 ring-opening,13 ATRP,14-18
TEMPO,19 and RAFT20,21 have been successfully employed to
give surface-grafted polymers under controlled growth condi-
tions. Several researchers have improved upon this technique
by patterning the initiator on the surface by microcontact printing
techniques followed by controlled polymerization to yield
surfaces decorated with patterned surface-tethered polymers.22,23
These patterning techniques have potential applications in
creation of microstructures, but there are several limitations
inherent with PDMS-based microcontact printing, namely poor
monolayer coverage, limitations in resolution, and lack of
robustness of adsorbed initiator monolayers. Polymers demon-
strated as “inking” stamps for use in microprinting techniques
are generally based on PDMS, which does not easily replicate
nanometer-sized features due to its inherently low modulus and
the tendency for small structures to collapse under the forces
of self-adhesion. Additionally, PDMS readily deforms under
applied pressure, making alignment and controlled contact over
large areas a continued challenge.
Figure 1. Graphical representation of surface tuning for nanoscopic pattern
replication.
Herein we report a new tandem, top-down/bottom-up ap-
proach for controlling the chemistry and size of nanoscopic
features at the sub-100 nm level, which is both simple and
versatile (Figure 1). In this strategy, a top-down nanocontact-
molding process is employed followed by the controlled growth
of polymer brushes from these patterned features (bottom-up).
This is achieved by molding a network photopolymer, containing
an imbedded inimer (having both an initiator and monomer
fragment24), into patterned features. In a secondary reaction step,
polymer brushes of defined chemistry and structure are grown
from the surface of the patterned network, specifically from the
exposed imbedded inimers, thus changing the size of the features
and the surface chemistry of the features. As a consequence of
this strategy, a variety of nanoscopic structures with different
feature sizes and functional groups can be grown from the same,
original molded template.
Experimental Section
Materials. Unless specified, all reagents were used without further
purification. Styrene, methyl methacrylate (MMA), and 2-hydroxyethyl
methacrylate (HEMA) from Aldrich were dried over CaH2 and distilled
under reduced pressure. Ethoxylated bisphenol A dimethacrylate
(Sartomer), N-vinyl pyrrolidone (Aldrich) and 2-ethyl-2-(hydroxy-
methyl)-1,3-propanediol trimethacrylate (TMPA) (Aldrich) were used
as received. 2,2-Dimethoxy-2-phenylacetophenone (DMPA) was re-
crystallized prior to use. 3-Methacryloxypropyl trichlorosilane was used
as received from UCT. All reactions were run under dry N2 unless
otherwise noted. Patterned silicon masters were supplied by SEMAT-
ECH or made by e-beam lithography. We used electronic grade
propylene glycol methyl ether acetate (PGMEA) to make solutions for
the thin film photopolymer resins.
Instrumentation. Gel permeation chromatography (GPC) was
performed on a Waters chromatograph (four Waters Styragel HR
columns HR1, HR2, HR4, and HR5E in series) connected to a Waters
410 differential refractometer with THF as the eluent. Molecular weight
standards were narrow polydispersity polystyrenes. Ellipsometry was
carried out on a Rudolph Auto-EI ellipsometer. Contact angle measure-
ments were recorded on a VCA 2500 video contact angle system with
a drop size of 1.00 µL. Electron micrographs were recorded on a Hitachi
S-4700 scanning electron microscope (SEM). AFM micrographs were
(8) Bailey, T.; Choi, B. J.; Colburn, M.; Meissl, M.; Shaya, S.; Ekerdt, J. G.;
Sreenivasan, S. V.; Willson, C. G. J. Vac. Sci. Technol., B 2000, 18, 3572.
(9) McClelland, G. M.; Hart, M. W.; Rettner, C. T.; Best, M. E.; Carter, K.
R.; Terris, B. D. J. Appl. Phys. Lett. 2002, 81, 1483.
(10) Hatzor, A.; Weiss, P. S. Science 2001, 291, 1019.
(11) Jordan, R.; Ulman, A. J. Am. Chem. Soc. 1998, 120, 243.
(12) Jordan, R.; Ulman, A.; Kang, J. F.; Rafailovich, M. H. J. Am. Chem. Soc.
1999, 121, 1016.
(13) Weck, M.; Jackiw, J. J.; Rossi, R. R.; Weiss, P. S.; Grubbs, R. H. J. Am.
Chem. Soc. 1999, 121, 4088.
(14) Boyes, S. G.; Brittain, W. J.; Weng, X.; Cheng, S. Z. D. Macromolecules
2002, 35, 4960.
(15) von Werne, T. A.; Patten, T. E. J. Am. Chem. Soc. 2001, 123, 7497.
(16) Boerner, H. G.; Duran, D.; Matyjaszewski, K.; da Silva, M.; Sheiko, S. S.
Macromolecules 2002, 35, 3387.
(17) Ejaz, M.; Yamamoto, S.; Ohne, K.; Tsujii, Y.; Fukuda, T. Macromolecules
1998, 31, 5934.
(18) Huang, X.; Wirth, M. J. Macromolecules 1999, 32, 1694.
(19) Husemann, M.; Malmstrom, E.; McNamara, M.; Mate, M.; Mecerreyes,
D.; Benoit, D.; Hedrick, J.; Mansky, P.; Huang, E.; Russell, T.; Hawker,
C. J. Macromolecules 1999, 32, 1424.
(20) De Boer, B.; Simon, H. K.; Werts, M. P. L.; van der Vegte, E. W.;
Hadziioannou, G. Macromolecules 2000, 33, 349.
(21) Luo, N.; Hutchinson, J. B.; Anseth, K. S.; Bowman, C. N. J. Polym. Sci.,
Part A: Polym. Chem. 2002, 40, 1885.
(22) Husemann, M.; Mecerreyes, D.; Hawker, C. J.; Hedrick, J. L.; Shah, R.;
Abbott, N. L. Angew. Chem., Int. Ed. 1999, 38, 647.
(23) Jones, D. M.; Huck, W. S. AdV. Mater. 2001, 16, 1256.
(24) Mueller, A. H. E.; Yan, D.; Wulkow, M. Macromolecules 1997, 30, 7015.
9
3832 J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003

Tuning Nanoscopic Features
A R T I C L E S
recorded in tapping mode on a Nanoscope III from Digital Instruments.
FTIR spectra were recorded in transmission mode on a Thermo
NicoletNexus 670 FTIR spectrometer using double polished wedged
silicon wafers.
1H NMR both diastereomers (250 MHz, CDCl3) δ 7.10-7.40 (m, 18H),
4.96 (m, 2H), 4.74 (m, 4H), 3.40 (d, 1H, J ) 10.8 Hz), 3.31 (d, 1H, J
) 10.8 Hz), 2.42 (m, 1H), 1.59 (d, 3H J ) 6.3 Hz), 1.50 (d, 3H J )
6.3 Hz), 1.41 (m, 1H), 1.32 (d, 3H, J ) 6.3 Hz), 1.05 (s, 9H), 0.93 (d,
3H J ) 6.3 Hz), 0.78 (s, 9H), 0.61 (d, 3H, J ) 6.6 Hz), 0.19 (d, 3H J
) 6.6 Hz); 13C NMR (APT) (63 MHz, CDCl3, both diastereomers) δ
146.11 (s), 145.24 (s), 142.55 (s), 135.67(d), 131.12 (d), 128.43 (d),
127.33 (d), 127.16 (d), 127.03 (d), 126.42 (d), 126.21 (d), 83.19 (d),
82.34 (d), 72.10 (d), 63.05 (t), 60.68 (s), 60.52 (s), 32.17 (d), 31.70
(d), 28.50 (q), 28.24 (q), 24.73 (q), 23.09 (q), 22.10 (q), 21.19 (q).
Anal. Calcd for C23H33NO2: C, 77.7; H, 9.35; N, 3.94. Found: C, 77.6;
H, 9.13; N, 3.68.
Synthesis of 2,2,5-Trimethyl-3-(1′-(4′′-methacryoyloxymethyl)-
phenylethoxy)-4-phenyl-3-azahexane, IN-2. In 20 mL of dry THF in
a three-neck flask equipped with a reflux condenser, rubber septa, and
N2 bubbler was dissolved 2.01 g (5.66 mmol) of the hydroxy derivative,
3. Dry triethylamine, 1.51 mL (10.8 mmol) was added with stirring
followed by the dropwise addition of methacryloyl chloride, 1.4 mL
(14.5 mmol). The reaction was allowed to stir at room temperature for
2.5 h, and then 50 mL of methylene chloride was added; the mixture
was washed three times with 100 mL of water. The organic layers were
combined and reduced in volume in vacuo to about 4 mL of crude
product. The product was washed again with two aliquots of 10% NaOH
and 10% HCl and then run through a silica gel flash column with 95%
hexane, 5% ethyl acetate to yield 1.53 g (65.9% yield). 1H NMR (250
MHz, CDCl3) δ 7.3-6.8 (m, 9H), 5.9 (1H, s), 5.4 (1H, s), 5.1 (1H, s),
5-4.9 (2H, d), 4.8-4.6 (1H, m), 3.3-3 (1H, m), 1.8 (3H, s), 1.5-1.2
(4H, m), 1.1 (2H, d), 0.8 (4H, s), 0.7 (2H, d), 0.6 (4H, s), 0.3 (1H, d),
0.1 (1H, d).13C {1H} NMR (62.5 MHz, CDCl3) δ 166.4, 146.3, 145.6,
142.7 (d) 135.3, 134.6, 131.4, 128.7, 127.7 (t), 126.7 (t), 83.7, 82.8,
72.6, 66.6, 60.9, 32.5, 32.1, 28.8, 25.2, 23.6, 23.1, 22.5, 21.6, 14.6.
Anal. Calcd for C27H37NO3: C, 76.6; H, 8.80; N, 3.31. Found: C, 76.4;
H, 8.92; N, 3.45.
Silicon Wafer Preparation. Polished silicon wafers were soaked
in a concentrated sulfuric acid solution containing No-Chromix for 5
min and rinsed extensively with deionized water. The wafers were then
placed in a 2-propanol vapor bath for 5 min and dried in an oven. An
adhesion promoter, 3-methacryloxypropyl trichlorosilane, was then
vapor-deposited on the silicon wafer under a saturated stream of dry
nitrogen.
Mold Fabrication. Patterned silicon masters with the desired features
to be replicated were cleaned as described above. A thin (150 mm)
glass plate was cleaned using the same procedure used above and
exposed to 3-methacryloxypropyl trichlorosilane vapor for 5 min to
form an adhesion layer. A small amount of PP1 resin (0.1 mL) was
added to the surface of the silicon master, and the glass plate was
brought into contact. The viscous liquid was allowed to spread between
the two plates and then exposed at 365-nm light (14 mW/cm2) for 1
min. The glass plate was carefully separated from the silicon master,
yielding a patterned PP1 network mold on a thin glass backing plate.
This mold was then sputter-coated with a 2-4-nm thick layer of Au/
Pd alloy and stored under nitrogen.
Photopolymer Solution. The base photopolymer solution (PP1)
consisted of the following formulation: ethoxylated bisphenol A
dimethacrylate (61%), N-vinyl pyrrolidone (18.5%), 2-ethyl-2-(hy-
droxymethyl)-1,3-propanediol trimethacrylate (18.5%), and 2,2-
dimethoxy-2-phenylacetophenone (2%). The monomers and initiator
were mixed and stored refrigerated in the dark prior to use.
Synthesis of 2-Methacryloxyethyl-2′-bromoisobutyrate (IN-1). A
three neck flask containing 200 mL of dry THF, 4.24 g (32.6 mmol)
of 2-hydroxyethyl methacrylate, and 3.30 g (32.6 mmol) of triethyl-
amine was equipped with a stir bar and cooled to 0 °C in an ice bath.
Using an addition funnel, 8.74 g (38.0 mmol) of 2-bromoisobutyryl
bromide was added dropwise. Upon complete addition, the mixture
was brought to room temperature and allowed to stir for 18 h. The
product was washed with 3 × 100 mL of H2O and dried over anhydrous
MgSO4, and the solvent was evaporated. The remaining pale yellow
oil was distilled under reduced pressure (200 mTorr) at 75 °C, and
5.94 g (65.2%) of the product, IN-1, was collected. 1H NMR (250 MHz,
CDCl3) δ (ppm) 6.07 (1H, s), 5.53 (1H, s), 4.13 (4H, s), 1.94 (6H, d),
1.86 (3H, s). 13C {1H} NMR (62.5 MHz, CDCl3) δ (ppm) 170.2, 135.8,
126.1, 63.4, 61.9, 55.3, 30.6, 18.2. Anal. Calcd for C10H15BrO4: C,
43.0; H, 5.42. Found: C, 42.8; H, 5.64.
Synthesis of 2,2,5-Trimethyl-3-(1′-(4′′-acetoxy)phenylethoxy)-4-
phenyl-3-azahexane, 2. In 34.3 mL of hexamethyphosphoramide
(HMPA) was dissolved 8.58 g (23.0 mmol) of 2,2,5-trimethyl-3-(1-
(4-chloromethyl)phenylethoxy)-4-phenyl-3-azahexane, 1.25 To this was
added 6.74 g (68.7 mmol) of potassium acetate, and the mixture was
stirred for 2 days at room temperature. The reaction mixture was then
loaded onto a silica gel flash column with hexane as the solvent. After
500 mL of hexane was run through the column to remove the HMPA,
the product was separated from the starting material by eluting with a
1:1 mixture of hexane and methylene chloride. The solvent was removed
in vacuo to give the acetoxy derivative, 2, as a thick, clear yellow oil,
1
8.51 g (93.0%); H NMR both diastereomers (250 MHz, CDCl3) δ
7.10-7.40 (m, 18H), 5.12 (d, 2H, J ) 9.3 Hz), 4.91 (ds, 4H, J ) 3.2
Hz), 3.45 (d, 1H, J ) 10.8 Hz), 3.31 (d, 1H, J ) 10.8 Hz), 2.44 (m,
1H), 1.60 (d, 3H J ) 6.3 Hz), 1.52 (d, 3H J ) 6.3 Hz), 1.41 (m, 1H),
1.28 (d, 3H, J ) 6.3 Hz), 1.07 (s, 9H), 0.88 (d, 3H J ) 6.3 Hz), 0.81
(s, 9H), 0.60 (d, 3H, J ) 6.6 Hz), 0.21 (d, 3H J ) 6.6 Hz); 13C NMR
(APT) (63 MHz, CDCl3, both diastereomers) δ 172.54 (s), 146.03 (s),
145.27 (s), 142.80 (s), 142.35 (s), 135.78 (d), 131.04 (d), 128.51 (d),
127.36 (d), 127.31 (d), 127.19 (d), 127.05 (d), 126.50 (d), 126.37 (d),
126.23 (d), 83.23 (d), 82.30 (d), 72.12 (d), 72.10 (d), 63.20 (t), 60.55
(s), 60.48 (s), 46.20 (d), 32.07 (d), 31.77 (d), 28.45 (q), 28.23 (q), 25.48
(q), 24.70 (q), 23.08 (q), 23.01 (q), 22.12 (q), 21.30 (q), 21.19 (q).
Anal. Calcd for C25H35NO3: C, 75.5; H, 8.87; N, 3.52. Found: C, 75.3;
H, 8.82; N, 3.70.
Synthesis of 2,2,5-Trimethyl-3-(1′-(4′′-hydroxymethyl)phenyl-
ethoxy)-4-phenyl-3-azahexane, 3. In 100 mL of a 3:1 mixture of water/
ethanol was dissolved 8.5 g (21 mmol) of the acetoxy derivative, 2.
Potassium hydroxide (3.4 g, 61 mmol) was added, and the mixture
was heated to reflux with stirring. Once the reaction was refluxing a
small amount of ethanol (ca. 10 mL) was added to afford a clear, single-
phase reaction mixture. After 2 h at reflux the reaction was allowed to
cool to room temperature and extracted four times with 80 mL of
methylene chloride and then dried over magnesium sulfate. The solvent
was removed in vacuo, and the product was separated from starting
materials on a silica gel flash column, eluting with a 1:1 mixture of
methylene chloride/hexane going to pure methylene chloride to give
the desired hydroxy functionalized alkoxyamine, 3, 4.63 g (61% yield).
Pattern Transfer. The photopolymer used for the replica layer in
contact patterning consisted of PP1 doped with 10-20 wt % of the
desired inimer, IN-1 or IN-2. This doped mixture was then diluted
with propylene glycol methyl ether acetate (PGMEA) to yield a 3 wt
% solution. This solution was filtered onto a cleaned and prepared
silicon wafer and spun at 3000 rpm for 1 min, yielding a 30-nm-thick
film. The Au/Pd coated mold was brought into contact with the wafer,
and a pressure of 60 psi was applied. The wafer was exposed using
365-nm light (14 mW/cm2) for 1 min, and then the mold was peeled
off leaving an inimer-embedded network replica layer on the surface
of the wafer. Additional information on the contact-molding process
used can be found in our earlier reference.9
Styrene Graft ATRP from Inimer-Embedded Network Thin
Film. A silicon wafer coated with a cured thin film network consisting
(25) Benoit, D.; Chaplinski, V.; Braslau, R.; Hawker, C. J. J. Am. Chem. Soc.
1999, 121, 3904.
9
J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003 3833

A R T I C L E S
von Werne et al.
of PP1 doped with 10 wt % ATRP inimer (IN-1), was added to a
Schlenk flask containing 10.2 mg (7.1 × 10-5 mol) of CuBr and a stir
bar. Styrene (2.50 mL, 21.8 mmol), pentamethyldiethylenetriamine
(PMDETA) (13.0 µL, 7.0 × 10-5 mol), and ethyl-2-bromoisobutyrate
(4.0 µL, 2.2 × 10-5 mol) were added via syringe, and 2.5 mL of
p-xylene was added as a solvent. The flask was purged of oxygen via
three cycles of freeze, pump, thaw, and backfilling with N2. The reaction
was sealed and heated at 90 °C for 4 h and then cooled. The wafer
was removed from the flask and rinsed liberally with acetone and THF.
Any physisorbed polymer was removed by exhaustive extraction with
THF in a Soxhlet apparatus.
MMA Graft ATRP from Inimer-Embedded Network Thin Film.
MMA polymerizations were performed as reported for styrene with
the exception that 2.50 mL (22.2 mmol) of MMA was substituted for
styrene and the polymerization was conducted at 60 °C for 4 h.
HEMA Graft ATRP from Inimer-Embedded Network Thin
Film.26 A silicon wafer coated with a cured thin film network consisting
of PP1 doped with 10 wt % ATRP inimer (IN-1), was added to a
Schlenk flask containing 10.2 mg (7.1 × 10-5 mol) of CuBr and a stir
bar, and the flask was deoxygenated by purging with N2. In a 10-mL
round-bottom flask, CuBr (9.0 mg, 6.3 × 10-5 mol), dipyridyl (14.0
mg, 1.2 × 10-4 mol), and HEMA (1.5 mL, 10.8 mmol) were added
followed by 1.0 mL of H2O and 2.5 mL of methanol as solvents, and
the reaction mixture purged with N2. The contents of the round-bottom
flask were added to a Schlenk flask via a purged syringe, and the
polymerization was left to stir at room temperature for 4 h, after which
time the wafer was removed and rinsed liberally with acetone and
methanol.
Figure 2. Representation of contact-molding process.
Scheme 1. Photopolymer Resin Composition
Alkoxyamine Initiated Graft Polymerization from Inimer-
Embedded Network Thin Film. A silicon wafer coated with a cured
thin film network consisting of PP1 doped with 10 wt % alkoxyamine-
based inimer (IN-2) was added to a Schlenk flask containing styrene
(3.0 mL, 26.2 mmol), free alkoxyamine initiator (0.80 mg, 0.0023
mmol), and chlorobenzene as a solvent (2.0 mL). A stir bar was added,
and the flask was deoxygenated via three cycles of freeze, pump, thaw,
and backfilling with N2. The flask was sealed and heated at 125 °C for
16 h, after which time the wafer was removed and rinsed liberally with
THF and acetone.
cross-linking density, and the N-vinyl pyrrolidone serves to
increase chain propagation rates during cure when using 2,2-
dimethoxy-2-phenylacetophenone (DMPA) as a photointiator
at 365 nm.
Results and Discussion
The primary patterning technique used in this approach is a
top-down contact-molding process which involves the use of a
patterned polymeric mold to template a secondary liquid
photopolymer resin layer that is subsequently UV-polymerized
while in contact with the mold to give pattern transfer (Figure
2).8 The patterned polymeric mold is formed by casting a
photopolymer resin (PP1) on a hard master, in our case either
SiO2 or Si, and photopolymerizing the resin to give a polymeric
network mold that has negative features of the original master.
Many molds can be made from a single master, and subse-
quently many imprints can be made from a single mold. As a
consequence, this process allows low throughput, high-cost
lithographic methods, such as electron beam and X-ray lithog-
raphy, to be used to fabricate a single master, which can then
be reproduced to fabricate many replicas. The composition of
the PP1 resin formulation was optimized for a number of criteria
including network hardness after cure, shrinkage, cure rate, and
UV absorption.9 The optimal composition is shown in Scheme
1.
The contact-molding process required the availability high-
quality patterned masters. For our initial experiments the master
consisted of etched silicon wafers with test features ranging from
5.0 to 0.35 µm, with a feature depth of 1.0 µm. To make the
reusable mold for the replication process, the PP1 resin mixture
was applied to the surface of the master, and a thin glass backing
plate was brought into contact with the resin, forming a
sandwich structure. The PP1 was cured under UV light, and
after curing, the mold on the backing plate was carefully
removed from the silicon master. The glass backing plate had
been treated with an adhesion promoter to ensure the cured
network polymer layer would have good adhesion to the glass
versus the Au/Pd coated master. The polymer network mold
represented a negative relief image of the silicon master. This
mold was then coated with a thin (2-4 nm) layer of Au/Pd to
prevent subsequent adhesion of the mold to any photopolymer
resin layers during pattern replication. The network polymer
mold on the glass backing plate was then used to replicate the
desired pattern into a second photopolymer layer by contact
molding. The photopolymer resin PP1 was diluted to 3 wt %
in PGMEA and spin-coated onto a silicon wafer that had been
cleaned and treated with adhesion promoter. The mold was
brought into contact with the thin film photopolymer resin on
The ethoxylated bisphenol A dimethacrylate imparts hardness,
etch resistance, and rigidity. The trimethacrylate provides a high
(26) Fuji, Y.; Watanabe, K.; Baek, K.-Y.; Ando, T.; Kamigaito, M.; Sawamoto,
M. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 2055-2065.
9
3834 J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003

Tuning Nanoscopic Features
A R T I C L E S
Scheme 2. Structures of Inimers IN-1 and IN-2
Scheme 3. Synthesis of Inimer IN-2
the wafer, pressure (60 psi) was applied, and the mold/
photopolymer/wafer sandwich was exposed to UV light. After
the photocure, the mold was released to yield a positive image
of the original pattern in the cured thin photopolymer on the
wafer. If necessary, any residual polymer remaining in the flat
areas between the molded features was removed using a
descumming process, such as dry plasma etching.
To allow controlled graft polymerization from the surface-
molded polymer networks, functionalized methacrylate mono-
mers containing a dormant living free radical initiator, termed
inimers (initiators-/-monomers), were added to the PP1 pho-
topolymer resin mixture. This resulted in the embedded inimers
being covalently bound throughout the cross-linked polymer
network, with a certain fraction of those groups being advan-
tageously located at the surface of the network and available
for subsequent reactions. The inimers utilized in this study, IN-1
and IN-2, are shown in Scheme 2 and are based on either atom
transfer radical polymerization (ATRP) or NMP.
ATRP has been shown to be a facile route to well-defined
polymers for a range of monomers at moderate temperatures.27
Accordingly, the synthesis of 2-methacryloxyethyl-2′-bro-
moisobutyrate (IN-1) was accomplished through the coupling
of 2-hydroxyethyl methacrylate (HEMA) with 2-bromoisobu-
tyryl bromide which gives IN-1 in 65% isolated yield after
distillation. As a complement to the ATRP-based system, NMP-
based procedures were also examined since it does not require
the addition of a catalyst that can leave trace metals in the
polymer.28 The alkoxyamine-based inimer 2,2,5-trimethyl-3-(1′-
(4′′-methacryloylmethyl)phenylethoxy)-4-phenyl-3-azahex-
ane, IN-2, was synthesized in a three-step process as shown in
Scheme 3.
The successful use of the embedded inimer in this contact-
molding approach relies on two critical features: (1) the ability
to form stable cross-linked networks in the presence of the
inimers and (2) retention of activity for the inimers after the
photopolymerization process. To investigate the effect of the
inimer groups on network formation, the methacrylate deriva-
tives, 10-20 wt % of either inimer (IN-1 or IN-2) were added
to PP1 resin formulations. These embedded inimer formulations
were subsequently polymerized by exposure to UV radiation,
and the cured cross-linked networks were observed to have no
noticeable decrease in toughness or hardness, indicating that
the addition of up to 20 wt % inimer does not have any
detrimental effect on mold formation or mechanical properties.
Table 1. Contact Angle Measurements of Photopolymer Films
water contact angle (deg)
inimer
inimer wt %
monomer
initial
after polymerization
N/A
N/A
20
20
20
20
styrene
HEMA
MMA
styrene
styrene
69.7
68.6
68.9
69.2
69.3
69.2
44.5
70.1
87.6
96.1
ATRP
ATRP
ATRP
NMP
It is however presumed that the inimer molecules are well
dispersed throughout the entire cured resin with a fraction of
them advantageously located at or near the surface being
accessible for subsequent “living” graft polymerizations.
The stability of the inimers to the photopolymerization process
was then probed by studying the growth of polymer brushes
from inimer-embedded networks. In this case thin films with
no patterns (i.e., flat) were initially used since this allows for
the investigation of surface characteristics and polymer growth
by changes in film thickness and contact angle measurements.
These flat surfaces were prepared by molding thin films of the
photopolymer resin while in contact with a mold that consisted
of a flat, polished silicon wafer. In addition to inimer-embedded
thin films, we routinely prepared undoped networks of cured
PP1 to be used as control specimens. All of the films were
characterized using water contact angle measurements to
investigate the surface properties of the films and ellipsometry
to determine the thickness of the thin films. Table 1 shows the
measured contact angle for both the films containing the
embedded inimers and the control sample before and after
grafting reactions. The incorporation of 20% of the ATRP-based
inimer into the cross-linked photopolymer network did not have
an appreciable effect on the surface energy of the thin film as
no change in the contact angle was observed, all samples having
a contact angle of ∼69°. The corresponding film thickness of
the doped and the undoped films are presented in Table 2.
(27) (a) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. ReV. 2001, 101, 3689-
3746. (b) Matyjaszewski, K.; Xia, J. Chem. ReV. 2001, 101, 2921-2990.
(28) (a) Hawker, C. J.; Bosman, A. W.; Harth, E. Chem. ReV. 2001, 101, 3661-
3688. (b) Hawker, C. J. Acc. Chem. Res., 1997, 30, 373.
9
J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003 3835

A R T I C L E S
von Werne et al.
Table 2. Ellipsometry Film Thickness Data
important to note that the control sample containing no
embedded inimer showed no appreciable change in surface
energy after ATRP polymerization. These results support the
assertion that the differences in contact angle are due to actual
modification of the surface and not due to free polymer/
reactants/solvent or other contaminants that are merely adsorbed
onto the network surface during or after the polymerization.
film thickness (nm)
solution
solution
inimer inimer wt, % monomer initial after polymerization polymer, M polymer, M /M
n
w
n
N/A
N/A
20
20
20
40
styrene 26.3
HEMA 26.1
MMA 35.8
styrene 27.9
MMA 25.2
styrene 17.2
26.1
39.5
64.5
48.2
54.4
N/A
N/A
N/A
N/A
1.22
1.12
1.14
1.48
ATRP
ATRP
ATRP
ATRP
NMP
3.27 × 104
3.10 × 104
3.30 × 104
2.90 × 105
20
160
To further confirm the presence of covalently attached
polymer brushes, FTIR spectroscopy was used to characterize
the composition of the polymer films before and after grafting.
The FTIR spectra of the initial film collected in transmittance
mode displayed peaks characteristic of a methacrylate network
polymer, with the dominant peaks assigned to the carbonyl
stretch (1725 cm-1) and the C-H stretch (2950 cm-1). After
the network layer was subjected to surface polymerization
conditions, new peaks were observed which corresponded to
those expected for the corresponding polymer brush; for
example, in the case of styrene new peaks were observed at
1602 and 3026 cm-1 which correspond to the CdC and aryl
C-H stretching bands for polystyrene.
After initial surface characterization, the thin films were
subjected to ATRP polymerization conditions. The polymer-
coated wafers were heated in a Schlenk flask with monomer,
solvent, and catalyst (when necessary) under a nitrogen atmo-
sphere. “Sacrificial” free initiator was added to the polymeri-
zation system to ensure that a living polymerization was
achieved and to control film thickness.19 After polymer brush
formation, the wafers were extensively rinsed and carefully dried
to ensure that any physioadsorbed polymer or solvent was
removed from the films. This is a important step due to the
presence of sacrificial initiator during the polymerization
reaction, which leads to the production of nonsurface attached
polymer in solution. In each case, a control wafer with a thin
film of undoped PP1 was subjected to identical ATRP condi-
tions to determine if there was any change in the surface
properties or the film thickness after being subjected to
polymerization with different monomers and at different tem-
peratures. For styrene, HEMA, and MMA under ATRP condi-
tions and for styrene under NMP polymerization conditions at
125 °C, no change in the undoped films was observed after
polymerization. Characterization data of the films before and
after surface polymerization is presented in Tables 1 and 2.
For the networks that did contain embedded inimers, contact
angle measurements gave a qualitative indication of significant
chemical changes to the surface of the thin film and the
associated growth of polymer brushes (Table 1). In the case of
the wafer exposed to MMA/ATRP conditions, a contact angle
of 70° was observed after surface polymerization. This repre-
sented little, if any change in the surface energy and may be
expected since a MMA brush was being grafted to a methacryl-
ate-based network. However, in the case of both the styrene
and HEMA polymerizations, significant surface energy changes
were noted. The film grafted with the hydrophilic HEMA brush
would result in a high surface energy film and as expected the
water contact angle of this sample decreased from an initial
value of 69° to 45°. Conversely, grafting the more hydrophobic
polystyrene brush using both ATRP and NMP techniques caused
the surface energy to drop as evidenced by the increase in the
water contact angle to 88° and 96° respectively. These values
are in good agreement with reported contact angle measurements
of 90°-96° of pure polystyrene films.29 The polymer brush
coverage appears to be more dense in the thicker sample as
evidenced by the higher contact angle measured (48 nm and
88° for the ATRP polymerization vs 160 nm and 96° for the
NMP polymerization. Repeated measurements on a number of
samples over various areas of the substrate showed that uniform
modification of the inimer embedded surface was achieved
which demonstrates the utility of this bottom-up approach. It is
The growth characteristics of the polymer brushes were
examined by ellipsometry and the thickness changes of the
polymer films related to the molecular weight of the grafted
linear polymer chains (Table 2). In all cases the film thickness
data supported the findings of the contact angle and FTIR
measurements and the control sample, PP1 containing no
embedded inimer, showed no significant change in film thick-
ness with a film thickness of 26 nm before and after polymer-
ization. An increase in film thickness was observed for all of
the inimer-doped films subjected to the polymerization condi-
tions. The HEMA-grafted film showed the smallest increase of
13 nm after reaction, while the nitroxide-mediated polymeri-
zation had the largest change of 143 nm. Previous studies
involving well-defined polymers grafted from initiators im-
mobilized on solid surfaces in the presence of free initiator in
the solvent phase have shown that the molecular weight of the
polymer isolated from solution is a good approximation of the
polymer grafted from the embedded inimers.18 In the case of
the MMA and styrene ATRP polymerizations, the free polymer
grown concurrently in solution was isolated and analyzed using
GPC to determine the molecular weight and the molecular
weight distribution. The low polydispersity (Mw/Mn) of the
samples indicated that control of the polymerization, i.e., its
“living” nature was preserved. To examine the nature of the
relationship between observed film thickness and polymer
molecular weight, a series of wafers with a thin film of PP1
doped with 10 wt % inimer, IN-1, were reacted with styrene
under ATRP in the presence of free initiator. The reactions were
stopped at different times and the reaction solution, which
contained nonsurface-attached or “free” polymer, was analyzed
by GPC. The clean and dried wafer film thicknesses were
measured and compared with the molecular weight of the free
polymer formed in solution. As can be seen in Figure 3, a linear
correlation between observe film thickness and polymer mo-
lecular weight with films ranging from 15 to 38 nm correspond-
ing to molecular weights of 18-37 K, respectively.
The ability to employ embedded inimers and for those inimers
to remain active after the contact-molding process now permitted
growth from patterned features to be examined. The same
contact-molding technique as described above was employed
(29) Jarvis, N. L.; Fox, R. B.; Zisman, W. A. In Contact Angle, Wettability,
and Adhesion; ACS Advances in Chemistry Series 43; Fowkes, F. M., Ed.;
American Chemical Society: Washington, DC, 1964.
9
3836 J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003

Tuning Nanoscopic Features
A R T I C L E S
Figure 3. Molecular weight (Mn) vs thickness of surface ATRP polystyrene.
with a variety of patterned molds to replicate three-dimensional
structures of inimer-embedded photopolymer. The resulting
patterned polymer surfaces, containing embedded initiating sites,
could then be used to initiate the grafting of well-defined
polymer brushes.
Figure 4. AFM of contact-molded pillars. (A) and (B) surface plots of the
array of 60-nm pillars before polymer brush growth; (C) and (D) surface
plots of the array of 75-nm pillars after brush growth.
These patterned films were subjected to the same surface
polymerization conditions as described above for the flat surface
studies with free initiator present in solution. After polymeri-
zation, the solution polymer was isolated and analyzed, and the
molecular weights and molecular weight distribution were found
to be similar to those observed for the flat surface samples,
indicating that films of comparable thickness were grown on
the surface of the patterned features when compared to flat
surfaces. During the contact-molding process, it was not always
possible to eliminate all traces of the photopolymer in the flat
areas between the features where, ideally, full contact between
mold and substrate is desired. To remove this thin, 1-3 nm
layer of unwanted residue, a dry-etch process for its removal
was examined, and again it was necessary to ensure that
exposure of the inimer-embedded nanostructures to a harsh O2
plasma etch would not destroy the initiating sites at the surface
which are needed for subsequent polymerization. Test samples
were prepared consisting of IN-1-doped PP1 photopolymer and
the films exposed to an O2 plasma; both etch rates and chemical
compatibility were then investigated. For IN-1-doped film
having an initial thickness of 52.3 nm, etching for 30 s resulted
in a decrease in the measured film thickness to 22.1 nm (ca. 10
Å/s). Significantly, reaction of the etched film under ATRP
conditions with MMA gave behavior similar to that observed
for nonetched films. In this case, the grafted PMMA layer was
found to have a thickness of approximately 20 nm, and the Mn
of the polymer grown in solution was 24 000 (Mw/Mn ) 1.27).
Similar behavior was observed for the alkoxyamine function-
alized networks (IN-2), which indicates that these surface
confined, living free radical initiators are stable to a wide range
of conditions and processing steps.
as described above, and used to mold a thin layer of IN-1-doped
PP1. Figure 4 shows an AFM image of the replicated pillars
and demonstrates the excellent retention of the size and shape
for the 60-nm pillars that can be achieved using this contact-
molding procedure with the modified PP1 formulation. The
patterned substrate was then reacted with MMA for 4 h under
ATRP conditions at 60 °C, and after thorough solvent extraction
and drying, the modified nanostructures were examined by
AFM. Interestingly, the pillars still had a 95-nm period;
however, the size of the pillars was observed to increase from
60 to 75 nm with the interpillar spacing decreasing accordingly
from 35 to 20 nm; this change of 15 nm is attributed to the
growth of the polymer brush from the surface of the nanostruc-
tures.
To ensure that the nanoscopic size of patterned features could
be controllably tuned, a series of ATRP reactions of MMA with
varying molecular weights for the resulting PMMA chains were
performed simultaneously on flat and patterned surfaces. The
flat films were 30-nm thick films of IN-1-doped PP1, and the
patterned surfaces consisted of molded IN-1-doped PP1 with a
repeating series of 100-nm wide lines. AFM and SEM images
of identical areas of the patterned wafers are shown in Figure
5, A and B. The line edge roughness observed in the images
was intentional and an artifact of the original masters made by
e-beam lithography, the roughness being reproduced in suc-
ceeding daughter images. These intentional defects were
employed as a means to determine the smallest size of features
that could be reproduced and to give a range of feature sizes
for subsequent growth from. In fact features as small as 5 nm
have been faithfully reproduced by this patterning method.
The patterned and flat samples were then subjected to polymer
brush formation by ATRP and the images in Figure 5, C and
D, show the samples after grafting of the linear PMMA chains
to the network structures. As can be observed, all of the features
show an increase in size with the size increase being uniform
To visualize the growth process and demonstrate the coupling
of a top-down molding process with a bottom-up polymer brush
growth process, patterned samples consisting of arrays of close
packed pillars were prepared. The master used was manufactured
by e-beam lithography and consists of rows of pillars with a
95-nm period, an average pillar width of 60 nm, and pillar-to-
pillar separation of 35 nm.9 A mold was cast from the master,
9
J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003 3837

A R T I C L E S
von Werne et al.
expected in the case of the lines since the thickness change
results from uniform film growth on both sides of the features.
As a result the size of the nanoscopic lines can be controllably
varied from their initial size of 100 nm to as large as 180 nm
by simply varying the reaction conditions (sacrificial initiator,
reaction time, monomer concentration). Since the period of the
features does not change, a corresponding change in line spacing
from 100 down to 20 nm is obtained. This combination of top-
down and bottom-up techniques now provides a reproducible
method for tuning feature sizes from a single master to suit
lithographic requirements even to sub-50-nm features.
Conclusions
In summary, we have demonstrated the utility of combining
a top-down contact-molding process with a bottom-up surface-
initiating grafting strategy to form three-dimensional patterns
in which the chemistry and size of the nanoscopic features can
be accurately tuned. Pattern replication of features ranging from
5 µm to <60 nm can be achieved by contact molding a
photopolymer mixture containing reactive inimers which lead
to surface active sites for secondary polymerization. Both ATRP-
type and nitroxide-mediated living free radical polymerizations
can be initiated from the inimer-embedded surfaces leading to
the formation of well-defined polymer brushes consisting of
polystyrene, MMA, and HEMA linear chains. This leads to
surfaces with different chemistries and physical properties. The
grafts could be grown from either flat or nanopatterned
substrates with controlled thickness ranging from 10 to 143 nm
and with graft molecular weights estimated to range between
18 000 to 290 000 amu. The ability to control the growth of
polymer brushes from contact-molded features can then be
utilized to tune the size of nanoscopic features down to 20 nm.
This novel combination of top-down and bottom-up strategies
now gives researchers new tools for the reproducible fabrication
of nanoscopic structures. Current work focuses on creating
templates for pattern replication and the development of process
steps for the transfer of these patterned images into a variety of
substrates for patterned magnetic media and molecular electron-
ics applications.
Figure 5. AFM and SEM images of contact-molded lines before (A), (B),
and after ATRP brush growth (C), (D).
Figure 6. Demonstration of molecular weight and nanoscopic feature size
control. Red line shows change in film thickness as a function of brush
molecular weight. Blue line shows the change in molded feature width as
a function of brush molecular weight.
regardless of the original feature size, which corresponds to a
decrease in line-to-line separation.
Acknowledgment. Financial support from the MRSEC
Program of the National Science Foundation under Award
Number DMR-9808677 for the Center for Polymeric Interfaces
and Macromolecular Assemblies, and the IBM Corporation are
gratefully acknowledged. Thanks to Teddie Magbitang for the
GPC results and Justine Shaw for additional substrate prepara-
tion and processing.
The molecular weight of the polymer brush was varied from
43 to 88 K, and the resulting thickness changes of the flat films
and the increase in line width for the patterned feature sizes
are plotted in Figure 6. Again the correlation between molecular
weight of the grafted chains and film/feature thickness change
is excellent with a linear relationship in both cases. Of particular
importance was the observation that the change in the line width
was about double that of the corresponding flat films. This is
JA028866N
9
3838 J. AM. CHEM. SOC. VOL. 125, NO. 13, 2003