Robust PEDOT films by covalent bonding to substrates using in tandem sol-gel, surface initiated free-radical and redox polymerization

Article abstract of DOI:10.1039/c1jm12563j

Poly(3,4-ethylenedioxythiophene), PEDOT, films are used as antistatic coatings on electrically insulating substrates such as plastic and glass. A novel method for the synthesis of conducting PEDOT films on insulators relies on sol-gel chemistry to attach a di-Si(OEt)₃ functionalized free radical initiator (AIBN) on oxidized surfaces, followed by: (a) attachment of 3,4-(vinylenedioxy)thiophene (VDOT: an analogue to EDOT susceptible to radical addition through its vinylenedioxy group); and, (b) oxidative (with FeCl ₃) co-polymerization of surface-confined VDOT with 3,4-ethylenedioxythiophene (EDOT). In conjunction with classical photolithography, the method yields thin (～150 nm) yet dense, pinhole-free (confirmed electrochemically), hard (>6H), extremely adhesive (5B), patterned, highly conducting (52 mho cm^-1) films. The process is applied mainly on glass but it works equally well on oxidized metal surfaces (aluminum, steel, Pt). Control studies related to "grafting from" with surface-confined AIBN were conducted by growing inexpensive poly(styrene) and poly(methylmethacrylate) films.

Full text of DOI:10.1039/c1jm12563j

View Article Online / Journal Homepage / Table of Contents for this issue
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 100
www.rsc.org/materials
PAPER
Robust PEDOT films by covalent bonding to substrates using in tandem
sol–gel, surface initiated free-radical and redox polymerization†
*
Anand G. Sadekar, Dhairyashil Mohite, Sudhir Mulik, Naveen Chandrasekaran, Chariklia Sotiriou-Leventis
*
and Nicholas Leventis
Received 6th June 2011, Accepted 6th September 2011
DOI: 10.1039/c1jm12563j
Poly(3,4-ethylenedioxythiophene), PEDOT, films are used as antistatic coatings on electrically
insulating substrates such as plastic and glass. A novel method for the synthesis of conducting PEDOT
films on insulators relies on sol–gel chemistry to attach a di-Si(OEt)3 functionalized free radical initiator
(AIBN) on oxidized surfaces, followed by: (a) attachment of 3,4-(vinylenedioxy)thiophene (VDOT: an
analogue to EDOT susceptible to radical addition through its vinylenedioxy group); and, (b) oxidative
(with FeCl3) co-polymerization of surface-confined VDOT with 3,4-ethylenedioxythiophene (EDOT).
In conjunction with classical photolithography, the method yields thin (ꢀ150 nm) yet dense, pinhole-
free (confirmed electrochemically), hard (>6H), extremely adhesive (5B), patterned, highly conducting
(52 mho cmꢁ1) films. The process is applied mainly on glass but it works equally well on oxidized metal
surfaces (aluminum, steel, Pt). Control studies related to ‘‘grafting from’’ with surface-confined AIBN
were conducted by growing inexpensive poly(styrene) and poly(methylmethacrylate) films.
binding to oxidized surfaces (glass, metal, SiO2) through silane
coupling agents of the type (RO)3-Si-R0, whereas RO- is an
alkoxy group and R0 is either a monomer for a conducting
polymer (typically N-substituted pyrrole), or an organofunc-
tional group able to link the coupling agent to the monomer.
Under reaction conditions akin to sol–gel chemistry, the RO-
groups of the coupling agent condense with hydroxyl groups on
the surface of oxidized substrates and modify them with R0.12–14
In that context, in 1982 Wrighton used N-linked trimethoxy-
silane functionalized pyrrole,15 and that paradigm has been
pursued actively by others.16–19
1. Introduction
Improved adhesion of polymeric films on surfaces increases their
useful lifetime and can be brought about by non-covalent
interactions (e.g., electrostatic, dipolar, van der Waals), but most
effectively by covalent bonding between the polymer and the
substrate.1 The latter approach may be particularly beneficial to
conducting polymers (e.g., polypyrroles, polythiophenes, poly-
aniline), which are used commercially mainly as antistatic coat-
ings, but also as hole injection layers (HIL) in organic light
emitting diodes (OLED) and solar cells, as electrodes in elec-
trolytic capacitors, in thick-film electroluminescent panels, in
organic thin film transistors and in printed boards in lieu of direct
metallization.2–4
Here, drawing from our previous work on coating con-
formally three-dimensional assemblies of silica nanoparticles
with polymers using surface initiated polymerization (SIP),20 we
use R0 ¼ 2,20-azobisisobutyronitrile (AIBN) and introduce an
alternative method for coating glass and oxidized metals with
polymers. For this, a bidentate free radical initiator, Si-AIBN, is
attached covalently to a surface, and upon heating induces
‘‘grafting from’’ polymerization of monomers in contact with
the surface.21 The method is predicated on the fact that poly-
mers are end-capped with an initiator fragment. Experimental
parameters are worked out with poly(methylmethacrylate)
(PMMA) and poly(styrene) (PS). Subsequently, the findings are
used to improve the adhesion of poly(3,4-ethylenedioxy-
thiophene), PEDOT, arguably the most successful commercially
conducting polymer manufactured and marketed by Bayer
under the trade name Baytronꢀ P. Baytronꢀ P is an aqueous
solution of the poly(styrenesulfonate), PSS, salt of the oxidized-
conducting form of PEDOT and is used in antielectrostatic
coatings.22
In that regard, Jaehne used phosphonic acid functionalized
pyrrole for binding to oxidized substrates (Ti/TiO2, Ta/Ta2O5
and Al/Al2O3),5,6 while Labaye demonstrated chemisorption of
acrylate-functionalized pyrrole on electrode surfaces by electro-
chemical methods.7 In terms of covalent bonding, Fabre,8–10 and
others11 have used organolithium reagents [e.g., 5-(N-pyrrolyl)
pentyllithium, and 2-thienyllithium] to bind pyrrole and thio-
phene on silicon via reaction with hydrogen-terminated silicon
surfaces. However, the most commonly used approach involves
Department of Chemistry, Missouri University of Science and Technology,
Rolla, MO, 65409, USA. E-mail: leventis@mst.edu; cslevent@mst.edu
† Electronic supplementary information (ESI) available: Profilometry of
a typical Si-AIBN film spin-coated on a glass slide and of various PMMA
films grown by various methods. GPC of PS extracted by scraping
material off glass slides. See DOI: 10.1039/c1jm12563j
100 | J. Mater. Chem., 2012, 22, 100–108
This journal is ª The Royal Society of Chemistry 2012

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
350 nm (by Dektak profilometry, see Figure S.1 in Electronic
Supporting Information – ESI†). For its own merit, but also in
order to work out experimental details, the method was first
applied successfully on conventional polymers such as poly-
styrene (PS) and polymethylmethacrylate (PMMA), and only
then onto high value added conducting PEDOT films. Glass
slides 200 ꢂ 200 ꢂ 1/1600 have been used as the typical substrate,
even though films have been also grown with ease onto oxidized
metal surfaces.
Since EDOT is not polymerizable by a free radical process, we
have synthesized and used EDOT bifunctional analogues,
3,4-(vinylenedioxy)thiophene (VDOT), or 3,4-bis(vinyloxy)thio-
phene (BVOT), which have free radical reactive sites and a thio-
phene moiety that can be co-polymerized oxidatively with
EDOT.23
2.1 Control case: PMMA and PS films on Si-AIBN modified
glass
For this, two methods were followed. The first involves dipping
and incubating initiator-coated glass slides in solutions of the
monomer, while the second one involves spin-coating of
a viscous prepolymer solution on top of the initiator layer fol-
lowed by heating. For control, glass slides with no initiator were
spin-coated with the prepolymer solution and heated. In all
cases, at the end slides were cleaned by sonication in toluene to
remove unbound polymer. No polymer layer was visible on the
control slides. Initiator bearing slides appear macroscopically
uniform, and profilometry shows that the prepolymer route
yields thicker films (up to 4000 nm) than simple incubation in
monomer solutions (150 nm, see Figure S.1 in ESI†). Since the
prepolymer route is not a viable approach for the conducting
polymer films described below, most further characterization,
unless indicated otherwise, is based on films produced by dipping
and incubating in monomer solutions.
Overall, the process of applying VDOT-EDOT films involves
simple sol–gel chemistry and chemical oxidation, and can be
integrated easily with conventional microfabication methods as
demonstrated herewith with photolithographic patterning. Being
covalently bonded to the substrate, the resulting VDOT-EDOT
films are mechanically robust, hard, extremely adhering to the
underlying surface, while they show similar electrical conduc-
tivity to PEDOT/PSS films obtained from Baytronꢀ P.
The chemical identify of the films was confirmed with IR
spectroscopy (Fig. 1) by scraping material off glass slides. The IR
spectra match those of commercial samples, although in the case
of PS we see additional broad absorptions in the 3440 cmꢁ1 and
1100 cmꢁ1 range, probably owing to scraped off initiator layer
that contains silica. The spatial arrangement of the polymer in
the films (brushes versus globules) was investigated through
molecular weight determinations and comparison of the film
thickness with the polymer radius of gyration.24 Thus, washing
PS films in boiling THF gives only small amounts of low
2. Results and discussion
Covalent attachment of polymers to surfaces can be carried out
either by ‘‘grafting to’’ or by ‘‘grafting from’’ methods. In both
cases, prior to coating, the surface needs to be modified with
appropriate functionality. In the ‘‘grafting to’’ approach bulk
polymerization is carried out in the solution in contact with the
surface; then, periodically, randomly growing polymeric chains
engage the surface functionality and get attached to the
substrate. The bulk of the polymer though is formed in the
solution, hence the method is not appropriate for custom-
synthesized, high value added monomers. Thus, here we have
opted for a ‘‘grafting from’’ approach using a free radical poly-
merization method, whereas by employing a bidentate free
radical initiator all polymer should be surface-confined, and
precious unreacted monomer could be recovered. Clearly, this
approach is based on the fact that in free radical polymerization
polymeric chains are capped with an initiator fragment. As
a suitable initiator for that purpose we use Si-AIBN, which is
synthesized in one step from 2,20-azobiscyanovaleric acid and
3-aminopropyltriethoxysilane (APTES) as described before.20
Si-AIBN was stored as a THF stock solution, which was typically
applied on glass slides by spin coating (see Experimental). Films
were aged in the dark and their thickness was found at around
Fig. 1 FTIR spectra of polystyerene (PS), PMMA and VDOT-EDOT
scraped off films on glass slides, in comparison with commercial materials
as indicated.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 100–108 | 101

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
molecular weight polymer (up to 1,000 - by Gel Permeation
Chromatography). Treatment with aqueous HF etches the film
off completely, but again consistently with similar previous
attempts with PS-crosslinked silica aerogels made by the
Si-AIBN method,20 failed to give any higher molecular weight
polymer. However, scraping the film off with a razor blade, and
extracting with THF yielded almost monodispersed PS with Mn
ꢀ 110,000 and Mw ꢀ 115,000 (see Figure S.2 and Table S.1 in
ESI†). The radius of gyration, Rg, was calculated via the rela-
tionship given by Terao and Mays for PS in THF at 25 ꢃC (Rg ¼
0.0118 ꢂ (Mw)0.6),24 and it was found equal to 12.8 nm, that is
much smaller than the thickness of the films (ꢀ150 nm) pointing
to a brush-like structure. Conversely, a brush-like arrangement
also implies high surface grafting density, sa, of the polymer
chains. The latter is calculated via sa ¼ 4/ps2, where s is the mean
distance between the polymer chains on the surface; in turn, s is
calculated from the layer (film) thickness, L (150 nm in our case),
Fig. 2 Au-sputtered, polystyrene-coated glass slides after lift-off by
ultrasonication in toluene. Left: The clear zone was not derivatized with
Si-AIBN prior to polystyrene deposition from prepolymer (see Experi-
mental); polystyrene and its Au-coating survive only over areas where
polystyrene is covalently bonded to the substrate via Si-AIBN. Right:
a similar slide, processed in parallel to the one at left, with uniform
initiator coating.
,
via Unsworth’s relationship, s ¼ [N3a5/L3]1/2 25 where a and N are
2.2 Electrically conducting polymer films on glass via SIP
the effective monomer length (0.26 nm, by molecular modeling
using ChemDraw) and the degree of polymerization (1100),
respectively. Thus, in our case s ¼ 0.68 nm and sa ¼ 2.75 chains/
nm2, which, interestingly, is close to the –OH group surface
density on amorphous sol–gel-derived silica (from tetramethy-
lorthosilicate), whereas all Q3-silicon sites are terminated with
hydroxyl groups (4.6 OH/nm2).26 (The difference is attributed to
the length of the AIBN tether in Si-AIBN.) When the mean
distance between polymer chains, s, is much less than twice the
Flory radius (RF ¼ aN3/5), the grafted polymer is in brush form.27
Indeed, s/2RF ¼ 0.020, that is ꢄ1, hence confirming again the
same conclusion reached above via MW and Rg considerations
about a brush-like structure.
Among common conducting polymers, polyaniline is conducting
in its two-electron oxidized, doubly protonated form (emeraldine
salt) and is sensitive to acids and bases.28 On the other hand,
polypyrroles have limited shelf-life,29 while polythiophenes
appear the most versatile.23 Yet, the long-term stability of poly-
thiophenes is compromised by the fact that the oxidized-con-
ducting form is susceptible to nucleophilic attack in the 3- and
4- positions,30 as for example by water, leading to degradation
and therefore to environmental instability. Preventing nucleo-
philic attack at those positions by substitution is a viable option,
and thus poly(3-methylthiophene) is much more stable than
polythiophene, and in fact is used in devices (e.g., transistors)
submerged in water.31 The ultimate stability, of course, would
result by blocking both the 3- and 4- positions, however, methyl
hydrogens in poly(3,4-dimethylthiophene) get on the way of one
another, forcing adjacent thiophenes out of the planarity
required for extended conjugation and electrical conductivity.
Thus, in the spirit used frequently in similar situations (e.g.,
aromatization of [10]annulene,32 or synthesis of dodecahydro-
3a,9a-diazaperylene – the most easily oxidizable p-phenylenedi-
amine,33) EDOT has been designed to eliminate such steric
restrictions and PEDOT turns out the most successful commer-
cially conducting polymer.
As implied by the difficulty to remove films for the polymer
characterization studies above, the adhesion of the polymer films
on the glass slides is exceptional: a Tape Test (ASTM Designa-
tion: D 3359-97) that calls for drawing with a sharp object 100
squares (1 mm2 each) on the film and counting the number of
squares surviving after the test, showed that all 100 squares
remained intact, while no squares stayed on the surface from
a control film made from a prepolymer solution on a clean,
initiator-free slide. (For that control experiment, slides were not
washed at the end – if washed the film is removed.) Clearly, the
different adhesion characteristics have to be attributed to initi-
ator (Si-AIBN) mediated covalent bonding of the polymer films
to the substrate.
Thus, although in analogy to styrene in Section 2.1, 3-vinyl-
thiophene would probably be sufficient to demonstrate the
A macroscopic visual confirmation of the adhesion of the
polymer, with further implications in terms of integrating this
process with classical microfabrication, was conducted as
follows: a 200 ꢂ 200 glass slide was partially masked with a tape
through its middle before spin-coating with the initiator solu-
tion. Subsequently, the tape was removed and the slide was
processed with the viscous prepolymer as above. At the end,
the slides were dried and sputter coated with Au. Ultra-
sonication in toluene dissolves the non-covalently bonded
polymer film from the initiator-free zone and lifts off the Au
layer above. Over the initiator bearing zones, the polymer
together with its Au coating survived sonication completely.
Fig. 2 shows the results and suggests that films derived by this
method can withstand some of the most-harsh conditions used
in microfabrication.
properties of a conducting polythiophene film covalently
attached to a substrate via SIP, nevertheless the utility of such
films would be limited, hence it was deemed appropriate to work
with PEDOT for the reasons outlined above. As all thiophenes,
EDOT is polymerizable through oxidative coupling in the 2- and
5- positions (not a free radical process). Typically, bulk oxidation
of EDOT is carried out with FeCl3,34 since the polymer is formed
in its oxidized state, the colorless-clear EDOT solution turns
opaque-blue. Since EDOT cannot be attacked by radicals, we
resorted to VDOT (although similar results have been also
obtained with BVOT – see Scheme 1). Clearly, as far as the
oxidizable (conducting) moiety (thiophene) and its substitution
pattern are concerned, EDOT and VDOT (or BVOT) are
equivalent. Indeed, electrochemically (Fig. 3) both EDOT and
VDOT show an identical oxidation onset in acetonitrile at 1.2 V
102 | J. Mater. Chem., 2012, 22, 100–108
This journal is ª The Royal Society of Chemistry 2012

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Scheme 1 Synthesis of VDOT and BVOT.
polymer, and indeed the substrate remained electrically insu-
lating. At that point it was reasoned that the random and fairly
rigid stereochemical arrangement of the thiophene moieties on
the glass surface (refer to Scheme 2) was preventing them
from reaching and coupling with one another via their 2- and
5- positions. Since VDOT and EDOT have the same onset for
oxidation (Fig. 3), EDOT could be grafted to surface-confined
VDOT (provided that the latter is indeed on the surface), by
simultaneous FeCl3-oxidation of both moieties. The ability of
EDOT and VDOT to co-polymerize was evaluated by electro-
chemical co-polymerization of a 1 : 1 mol/mol solution of the
monomers, and upon electrochemical characterization in fresh
electrolyte (no monomers present) the resulting films show
a redox wave between those of the two independent films (see
Fig. 3, inset). As expected, of course, those films demonstrate
poor adhesion to the electrode surface. On the other hand, films
obtained by chemical polymerization of EDOT engaging surface
bound VDOT would remain covalently bonded to the substrate
and therefore should be extremely robust.
Fig. 3 Magnified first-sweep cyclic voltammetry at the foot of the anodic
waves (i.e., not the complete voltammograms) showing the common
onset of oxidation of VDOT and EDOT in Ar-degassed acetonitrile
containing 0.1 M TBAP as supporting electrolyte. Other parameters:
sweep rate 0.1 V sꢁ1; Pt electrode (0.0314 cm2); concentrations of VDOT
and EDOT 0.1 M. The higher current in the return sweep (i.e., in the
negative direction) is due to electrodeposited polymer. Inset: Cyclic vol-
tammetry at 0.1 V sꢁ1 of films as shown, grown on Pt disk electrodes from
CH3CN/0.1 M TBAP solutions of EDOT (0.2 M), VDOT (0.2 M) and
EDOT-VDOT (0.1 M in each) by repetitive sweeping of the potential
Scheme 3 summarizes the practical implementation of the
process outlined above, and includes patterning of the resulting
films by classical positive resist photolithography. Thus, dipping
a colorless VDOT modified glass slide in a FeCl3 solution in
acetonitrile, followed by addition of EDOT produced an opa-
que-blue solution that settles into a blue precipitate upon
standing, and left a deep blue coating on the VDOT-modified
glass slide. Brief ultrasonication in acetonitrile removed all
loosely held polymer from the glass slide, leaving behind
a transparent light-blue film that is not affected by further
ultrasonication (even lasting for several hours). The films are
macroscopically uniform (Fig. 4) and microscopically fairly
smooth (inset in Fig. 4, and Fig. 5). By profilometry, the film
thickness was similar to what was measured for PMMA and PS
films obtained by incubation of Si-AIBN modified glass slides in
the monomer solutions (about 150 nm, compare inset in Fig. 4
with Figure S.1 in ESI†). Patterning is carried out by applying
photoresist to protect the areas where the film is intended to
survive, followed by etching off (with HF) the photoresist
unprotected areas (Scheme 3). Remaining photoresist over the
blue film was removed with acetone. A photograph of a typical
patterned film, made using a photomask from previous studies,42
is included in Fig. 4.
from 0 V to 2.5 V vs. aq. Ag/AgCl at 0.1 V sꢁ1
.
vs. Ag/AgCl (first potential sweep). Unlike EDOT, however,
VDOT carries an isolated double bond prone to radical addition.
VDOT was synthesized from thiophene in satisfactory yields
by modification of literature procedures according to Scheme
1.35–39 It is noted that traditional synthesis of EDOT is a five-step
process,40 but more recently a convenient synthesis has been
described through a common intermediate with the VDOT
process, namely compound 4 (3,4-dimethoxythiophene, Scheme
1),41 so that the overall synthesis of VDOT is only two steps
longer than the shortest route to EDOT.
Based on the findings with PS and PMMA, neat VDOT (or
BVOT) was applied and spread evenly with a brush over glass
slides, which had been spin-coated previously with the initiator
(Si-AIBN). However, after heating, aging and FeCl3 treatment,
VDOT-modified glass slides failed to give the characteristic blue
color expected from the oxidized-conducting form of the
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 100–108 | 103