Thicker is better? Synthesis and evaluation of well-defined polymer brushes with controllable catalytic loadings

Article abstract of DOI:10.1002/chem.201202531

Polymer brushes (PBs) have been used as supports for the immobilization of palladium complexes on silicon surfaces. The polymers were grown by surface-initiated atom-transfer radical polymerization (SI-ATRP) and postdecorated with dipyridylamine (dpa) ligands. The pendant dpa units were in turn complexed with [Pd(OAc)₂] to afford hybrid catalytic surfaces. A series of catalytic samples of various thicknesses (ca. 20-160 nm) and associated palladium loadings (ca. 10-45 nmol cm^-2) were obtained by adjusting the SI-ATRP reaction time and characterized by ellipsometry, X-ray reflectivity, X-ray photoelectron spectroscopy, and inductively coupled plasma mass spectrometry (ICP-MS). ICP-MS revealed a near-linear relationship between thickness of the polymer brush and palladium content, which confirmed the robustness of the preparation and postmodification sequence presented herein, rendering possible the creation of functional architectures with predefined catalytic potential. The activities of the catalytic PBs were determined by systematically exploring a full range of substrate-to-catalyst ratios in a model palladium(0)-catalyzed reaction. Quantitative transformations were observed for loadings down to 0.03 mol % and a maximum turnover number (TON) of around 3500 was established for the system. Comparison of the catalytic performances evidenced a singular influence of the thickness on conversions and TONs. The limited recyclability of the hairy catalysts has been attributed to palladium leaching.

Full text of DOI:10.1002/chem.201202531

DOI: 10.1002/chem.201202531
Thicker is Better? Synthesis and Evaluation of Well-Defined Polymer
Brushes with Controllable Catalytic Loadings
Antony E. Fernandes,[a] Ali Dirani,[a] Cꢀcile dꢁHaese,[a] Gladys Deumer,[b]
Weiming Guo,[a] Peter Hensenne,[a] Fady Nahra,[a] Xavier Laloyaux,[a]
Vincent Haufroid,[b] Bernard Nysten,[a] Olivier Riant,*[a] and Alain M. Jonas*[a]
Abstract: Polymer brushes (PBs) have
been used as supports for the immobili-
zation of palladium complexes on sili-
con surfaces. The polymers were grown
by surface-initiated atom-transfer radi-
cal polymerization (SI-ATRP) and
postdecorated with dipyridylamine
(dpa) ligands. The pendant dpa units
flectivity, X-ray photoelectron spectro-
scopy, and inductively coupled plasma
mass spectrometry (ICP-MS). ICP-MS
revealed a near-linear relationship be-
tween thickness of the polymer brush
and palladium content, which con-
firmed the robustness of the prepara-
tion and postmodification sequence
presented herein, rendering possible
the creation of functional architectures
with predefined catalytic potential. The
activities of the catalytic PBs were de-
termined by systematically exploring a
full range of substrate-to-catalyst ratios
in a model palladium(0)-catalyzed re-
action. Quantitative transformations
were observed for loadings down to
0.03 mol% and a maximum turnover
number (TON) of around 3500 was es-
tablished for the system. Comparison
of the catalytic performances evi-
were in turn complexed
with
[Pd(OAc)2] to afford hybrid catalytic
ACHTUNGTRENNUNG
surfaces. A series of catalytic samples
of various thicknesses (ca. 20–160 nm)
and associated palladium loadings (ca.
10–45 nmolcmÀ2) were obtained by ad-
justing the SI-ATRP reaction time and
characterized by ellipsometry, X-ray re-
denced a singular influence of the
thickness on conversions and TONs.
The limited recyclability of the hairy
catalysts has been attributed to palladi-
um leaching.
Keywords: palladium · polymers ·
silicon · supported catalysts · sur-
face chemistry
Introduction
ticularly high local concentrations engendered with such
frameworks offer unique opportunities in supported cataly-
sis.[7] In some cases the PB architecture can significantly en-
hance the outcome of cooperative catalytic processes that
require the close proximity of synergetic active sites.[7c] It is
also possible to incorporate catalytic PB coatings inside the
walls of microreactors for application in continuous flow
processes.[7d,g,h]
In general, PBs are prepared by using the “grafting from”
strategy, which encompasses the preliminary preparation of
an initiator monolayer followed by surface-initiated “living”
polymerization.[8] Modifying the polymerization conditions
(e.g., solvent, concentration, reaction time) allows additional
control over the thickness and architecture of the brushes,
which can be particularly attractive for the configuration of
well-defined structures with determined functional loa-
dings.[7d,g,h] However, despite the remarkable potential of
PBs, there are still very few applications in heterogeneous
catalysis.[7]
Polymer brushes (PBs)[1] represent versatile molecular con-
structions for shaping surface properties for applications in
domains such as actuation,[2] electronics,[3] and biology.[4]
Their interest lies in the possibility of finely adjusting the
polymeric core composition of the brush together with the
possibility of postfunctionalizing the pendant lateral chains
towards their desired use. The covalent attachment of func-
tional molecules (e.g., biomolecules,[5] switches,[6] and cata-
lysts[7]) therefore provides access to a new class of hybrid de-
vices. The dynamic nature of the PBs together with the par-
[a] Dr. A. E. Fernandes, Dr. A. Dirani, C. dꢀHaese, Dr. W. Guo,
Dr. P. Hensenne, F. Nahra, Dr. X. Laloyaux, Prof. B. Nysten,
Prof. O. Riant, Prof. A. M. Jonas
Institute of Condensed Matter and Nanosciences
Universitꢁ catholique de Louvain
Croix du Sud 1, 1348 Louvain-la-Neuve (Belgium)
Fax : (+32)10-45-15-93
Herein we report the robust, stepwise, construction and
detailed examination of catalytic PBs prepared from single-
crystal silicon surfaces and the impact of thickness on activi-
ty.
Planar surfaces hold unique, yet not well understood, po-
tential in catalysis as a result of impressive performances ob-
served upon decoration with monolayers of transition-metal
complexes.[9] The possibility of assembling multifunctional
E-mail: olivier.riant@uclouvain.be
alain.jonas@uclouvain.be
[b] G. Deumer, Prof. V. Haufroid
Louvain Centre for Toxicology and Applied Pharmacology
Universitꢁ catholique de Louvain
Avenue Mounier 52, 1200 Woluwe-Saint-Lambert (Belgium)
Fax : (+32)27-64-53-38
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202531.
16226
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16226 – 16233

FULL PAPER
catalytic surfaces[10] by means of lithographic techniques is a
supplementary benefit of using planar supports. In our en-
deavor to explore the operation of catalytically active silicon
surfaces, we envisioned the use of PBs as supports for ho-
mogeneous organometallic complexes. The systems present-
ed herein were prepared by the complexation of [PdACHTNUTRGNEUNG(OAc)2]
to pendant dipyridylamine (dpa) ligands incorporated on to
PBs of various thicknesses and evaluated in a model reac-
tion based on the palladium(0)-mediated deprotection of
Alloc-coumarin 1 (Scheme 1).[9g]
Scheme 1. Catalytic polymer brush on a planar silicon surface. A sche-
matic of the palladium-catalyzed deprotection of Alloc-coumarin 1 is
shown.
Results and Discussion
Scheme 2. Stepwise synthesis of [Si-PHEMA-dpa-Pd] catalytic brushes.
In this report, poly(2-hydroxyethyl methacrylate) (PHEMA)
PBs were grown on single-crystal silicon surfaces and post-
decorated with 2,2’-dipyridylamine (dpa) derivative 3 for
the complexation of palladium catalysts (Scheme 2).[11]
(Figure 1). The samples were then postfunctionalized as de-
scribed above to obtain [Si-PHEMA-dpa-Pd] surfaces. The
evolution of the dry thickness of the layer gives a primary
and rapid indication of the successful stepwise surface modi-
fication. The validity of the ellipsometric determination of
the thickness was controlled by measuring the thicker brush-
es by XRR, which provides a model-independent thickness
from interference fringes (see the Supporting Information).
PHEMA brushes were selected as they allow the simple
introduction of suitable molecules by classical coupling with
the hydroxy substituents of the lateral chains.[8a,12] An atom-
transfer radical polymerization (ATRP) initiator silane was
first self-assembled on to freshly cleaned silicon substrates
by gas-phase silanization at 808C. The [Si-Initiator] samples
were subsequently used for the surface-initiated atom-trans-
fer radical polymerization (SI-ATRP)[8] of HEMA mono-
mers. The pendant hydroxy groups of [Si-PHEMA] were
then activated with N,N’-disuccinimidyl carbonate (DSC)[12]
by using a standard protocol and the resulting samples di-
rectly reacted with dpa derivative 3 in DMF in the presence
of Et3N to afford [Si-PHEMA-dpa]. Finally, the substrates
were immersed in a solution of [PdACHTNURGTNE(UNG OAc)2] in CH2Cl2 to
give the catalytic surfaces [Si-PHEMA-dpa-Pd] after exten-
sive washing in a Soxhlet apparatus.
The growth and modification of the PHEMA PBs were
confirmed by ellipsometry, X-ray reflectivity (XRR), X-ray
photoelectron spectroscopy (XPS), and inductively coupled
Figure 1. Evolution of the dry thickness of PBs following time-controlled
SI-ATRP and subsequent postmodification. Diamonds: [Si-PHEMA];
squares: [Si-PHEMA-dpa]; triangles: [Si-PHEMA-dpa-Pd]. Closed sym-
bols: ellipsometry; open symbols: XRR, displaced laterally by 7 min for
clarity. The lines are fits to power laws (atb).
plasma mass spectrometry (ICP-MS).
[Si-PHEMA] samples with dry thicknesses in the range of
10–80 nm were prepared by time-controlled SI-ATRP
A
series of
ACHTUNGTRENNUNG
Chem. Eur. J. 2012, 18, 16226 – 16233
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16227

O. Riant, A. M. Jonas et al.
The increase in the thickness upon functionalization arises
from the stretching of the individual chains to reduce steric
congestion following the incorporation of the dpa ligand
and palladium complexation. Accordingly, the thicknesses of
the samples reported in this paper refer to the thickness of
the parent [Si-PHEMA] brush, for example, 80[Si-PHEMA-
dpa-Pd] refers to the postmodified 80 nm [Si-PHEMA] sur-
face, but its actual thickness is around 150 nm. Based on the
reported density of PHEMA and the molar volume of the
repeat unit (102 cm3 molÀ1),[13] the thickness of the pristine
brushes can be converted into the number of PHEMA
repeat units per surface area (Figure 2).
Figure 2. Evolution of the loadings following time-controlled SI-ATRP
&
^
and subsequent postmodification. : PHEMA repeat units; : grafted
!
dpa 3; and total palladium obtained from the thickness analysis ( ) or
~
from ICP-MS ( ). The lines are computed from the fits shown in
Figure 3. XPS spectra of 80ACTHUNRGTNE[NUG Si-PHEMA], 80[Si-PHEMA-dpa], and
80[Si-PHEMA-dpa-Pd]. The inset is an enlargement of the XPS palladi-
Figure 1. The dotted line fits the palladium content determined from
ICP-MS measurements.
3
um 3d region of the 80[Si-PHEMA-dpa-Pd] brush showing the (3d =
;
2
II
3d = ) doublets of Pd and Pd0 in the sample.
5
2
XPS analysis of the thicker 80ACTHNUGTRNEUNG[Si-PHEMA] brush, which
completely screens the substrate, provides an O/C ratio of
0.5, in good agreement with the brush composition. Incorpo-
ration of dpa ligand 3 results in the appearance of a nitrogen
signal (ca. 400 eV) in the spectrum (Figure 3) with an exper-
imental N/O ratio of 0.39 for the thicker 80[Si-PHEMA-
dpa], which corresponds to 32% attachment of dpa 3 at the
surface of the brush. The increment in the thickness follow-
ing the grafting is 54 nm (Figure 1), which, upon division by
the molar volume of the grafted unit (204.7 cm3 molÀ1, esti-
mated by the method of Fedors et al.),[14] provides
26.5 nmolcmÀ2 of dpa 3 (Figure 2), which indicates a 35%
average grafting over the entire brush thickness, in good
agreement with the XPS evaluation. The observed threshold
of functionalization arises from the congestion caused by
the covalent bonding of dpa 3.[15] Thinner brushes, less than
40 nm, tend to have a lower grafting ratio of dpa 3, probably
due to the greater steric hindrance in the lower part of the
brush. However, the accuracy of the measurements in these
systems is also generally lower.
coordinating nature of the PHEMA-dpa backbone, which
bears a plethora of ester and carbamate donors in close
proximity. A closer examination of the palladium 3d XPS
3
5
region reveals the presence of two doublets (3d = ; 3d = ) cor-
2
2
responding to PdII (343.3 and 338.1 eV) and Pd0 (340.9 and
335.7 eV) in a PdII/Pd0 ratio of around 1.9 (inset of
Figure 3). Identical patterns were found for [Si-PHEMA-
dpa-Pd] brushes of various thicknesses (see Table S2 and
Figure S4 in the Supporting Information). Because the same
observation was made in the starting [Pd
1.8, see Table S2 and Figure S4) we ascribe the presence of
Pd0 in the brush to the purity of the commercial [Pd
(OAc)2],
(OAc)2] (PdII/Pd0 =
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
as has recently been reported for other commercial palladi-
um compounds.[16] Supplementary experiments ensured that
XPS-induced reduction, although unlikely but observed in
some rare cases,[17] did not occur at all in our analysis and
thus may not be imputed to the presence of Pd0 in our sam-
ples. In addition, comparison of the experimental elementa-
ry ratios Pd/C=O (ca. 0.83) and PdII/C=O (ca. 0.53) in
For the thicker brush, complexation with [Pd
forded a Pd/N value of around 0.43, which is larger than the
theoretical value (0.25) expected for the formation of a 1:1
(OAc)2] af-
[PdACHTUGNTREN(NUG OAc)2] (theoretical Pd/C=O=0.5) confirmed the coex-
2
1
istence of around = [Pd
G
3
3
cially available sample.
complex between the pendant dpa ligand and [Pd
This excess palladium presumably reflects the particularly
(OAc)2].
Taking these data into consideration, the refined
PdII/N ratio is 0.28, which indicates a stoichiometric com-
16228
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16226 – 16233

Polymer Brushes
FULL PAPER
plexation of the pendant dpa units of the brush. The pres-
ence of ClÀ in the XPS (ca. 200 eV) suggests a possible but
limited exchange of acetate by chlorine anions (ClÀ/PdII =
0.48) for around 25% of the PdII atoms.
Detailed analysis of the N1s XPS region before and after
the incorporation of palladium also provided evidence for
the complexation of the dpa motifs. A shift of the pyridine
nitrogen atoms of [Si-PHEMA-dpa] (ca. 398.9 eV) towards
higher binding energies (ca. 400.1 eV), and thus towards an
electron-poor state, is consistent with coordination to palla-
dium (see Table S3 and Figure S5 in the Supporting Infor-
mation).[18]
ladium content of the corresponding PBs was approximated
by extrapolation of the ICP-MS calibration curve shown in
Figure 2 (ca. 12.1, 18.6, and 46.9 nmolcmÀ2 respectively).
The activities of the different surfaces were examined in the
fluorogenic deprotection of 1ꢃ10À5 mol of Alloc-coumarin 1
(Scheme 1, Table 1). Furthermore, to access additional,
higher, substrate-to-catalyst ratios (S/C) in a straightforward
manner, surfaces with areas of 1, 0.5, and 0.25 cm2 were
tested for each thickness.
Table 1. Catalytic activity of [Si-PHEMA-dpa-Pd] in the deprotection of
1ꢃ10À5 mol Alloc-coumarin 1.
Based on the molar volumes of [Pd
(102.5 cm3 molÀ1), (44.3 cm3 molÀ1),
PdCl2 and
(8.9 cm3 molÀ1) and on the XPS-determined composition for
[Pd
(OAc)2]/PdCl2/Pd0 of 0.75:0.25:0.53, which corresponds
ACHTUNGTRENNUNG
Sample
Surface area S/C[a] Conversion TON[c]
AHCTUNGTRENNUNG
[cm2] [%][b]
1
2
3
4
5
6
7
8
9
80[Si-PHEMA-dpa-Pd]
1
213 94
200
409
ACHTUNGTRENNUNG
0.5
0.25
1
426 96
852 95
538 96
1076 96
2152 96
826 96
1652 98
3304 96
to an average volume of 60.7 cm3 molÀ1 of added palladium,
the difference of thickness of 28 nm resulting from palladi-
um addition can be converted into 46 nmolcmÀ2 of palladi-
um in the brush, which compares well with the value of
45 nmolcmÀ2 obtained by ICP-MS. Similar good agreement
was found for other brush thicknesses (Figure 2), which
adds support to our analysis.
809
516
40[Si-PHEMA-dpa-Pd]
30[Si-PHEMA-dpa-Pd]
0.5
0.25
1
1033
2066
793
0.5
0.25
1619
3172
[a] Substrate-to-catalyst ratio. [b] Conversion after 20 h, as determined by
1H NMR spectroscopy. [c] Turnover number.
Because ICP-MS provides a direct quantitative analysis of
the total palladium content in the brushes, it will be used in
the following. Analysis of the [Si-PHEMA-dpa-Pd] series
shown in Figure 1 (ca. 20–160 nm) reveals a near-linear cor-
relation between thickness and palladium content
(Figure 4), which indicates that the dpa and palladium moi-
eties diffuse and graft similarly in thin and thick brushes.
This correlation demonstrates the robustness of the post-
modification sequence presented herein and the possibility
of relatively precisely adjusting the catalytic performances
of PB-supported catalysts.[7d,g,h] It also confirms the coordi-
nating nature of the PHEMA-dpa core as being responsible
for excessive palladium incorporation in comparison with
the amount of dpa ligand in the brush.
All the samples showed remarkable performances with
the quantitative formation of free coumarin 2. A maximal
turnover number (TON) of around 3172 was recorded with
the 0.25 cm2 30[Si-PHEMA-dpa-Pd] sample (Table 1,
entry 9).
To assess the full potential of the catalytic brushes, the
catalyzed reactions were also conducted with 1ꢃ10À4 mol of
Alloc-coumarin 1 (Table 2). A progressive decrease in activ-
ity was recorded and directly related to the diminution of
the surface area and PB thickness, and thus to the palladium
content engaged in the reactions. Interestingly, the 0.25 cm2
40[Si-PHEMA-dpa-Pd] brushes afforded the highest TON
of around 4086 (entry 7) whereas the 1 cm2 80[Si-PHEMA-
dpa-Pd] surface gave the best conversion of around 51%
(entry 1). In addition, the samples prepared by the direct in-
corporation of palladium into nonfunctionalized PHEMA
brushes ([Si-PHEMA-Pd], Table 2, entries 4, 8, and 12)
showed limited reactivity compared with their [Si-PHEMA-
dpa-Pd] counterparts, with no particular influence of PB
thickness. Palladium may thus be poorly adsorbed on to the
pristine PHEMA brushes and only complexes (and entraps)
dpa-modified PBs, as previously noted.
To study the effect of PB thickness and palladium loading
on catalytic activity, a series of [Si-PHEMA] samples with
thicknesses of around 30, 40, and 80 nm were prepared, with
a good thickness distribution within the series (see Figure S1
in the Supporting Information), and functionalized. The pal-
Figure 5 allows a comprehensive visualization of the cata-
lytic properties of our PBs. Quantitative conversions can be
attained below a threshold S/C ratio of around 3500, above
which conversions decrease as S/C increases. Similarly, the
evolution of TONs against S/C ratios confirmed an asymp-
totic limit TON of around 3500.
Intuitively and experimentally, the thickest surfaces, that
is, those having the highest palladium contents, prove to be
the most active in terms of conversion. However, the TONs
Figure 4. Correlation between [Si-PHEMA-dpa-Pd] thickness and palla-
dium content (as determined by ICP-MS).
Chem. Eur. J. 2012, 18, 16226 – 16233
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16229

O. Riant, A. M. Jonas et al.
Table 2. Catalytic activity of [Si-PHEMA-dpa-Pd] in the deprotection of 1ꢃ10À4 mol
Alloc-coumarin 1.
chitectural effects (swollen/collapsed PBs) particu-
larly complicated.
Sample
Surface area
[cm2]
S/C[a]
Conversion
[%][b]
TON[c]
Finally, the recyclability of our supported cata-
lysts was explored by physically transferring the
surface (after copious rinsing with CH2Cl2) to a
new reaction (Table 3). For iterative runs, the ac-
tivities of the catalytic PBs progressively drop. This
tendency is even more pronounced when working
at larger S/C ratios, that is, for the smallest and
thinnest [Si-PHEMA-dpa-Pd] surfaces. Total TONs
are finally 93500, consistent with the observed
limit of our catalyst. As previously observed,[9g] the
progressive deactivation of the catalytic PBs is
mainly attributed to the gradual leaching of palla-
dium. A simple regeneration process allowed used
samples, after being reloaded by immersion in
AHCTUNGTRENNUNG
1
2
3
4
5
6
7
8
80[Si-PHEMA-dpa-Pd]
1
2132
4264
8528
51
39
29
7
40
25
19
9
1087
1663
2473
n.d.[d]
2150
2688
4086
n.d.[d]
2066
2479
2644
n.d.[d]
0.5
0.25
1
80[Si-PHEMA-Pd]
40[Si-PHEMA-dpa-Pd]
n.d.[d]
1
5376
10752
21504
0.5
0.25
1
40[Si-PHEMA-Pd]
30[Si-PHEMA-dpa-Pd]
n.d.[d]
9
1
8264
16528
33056
25
15
8
10
11
12
0.5
0.25
1
30[Si-PHEMA-Pd]
n.d.[d]
4
[a] Substrate-to-catalyst ratio. [b] Conversion after 20 h, as determined by 1H NMR
spectroscopy. [c] Turnover number. [d] Not determined.
[Pd
ACHTUNGRTENUN(NG OAc)2] solution, to almost recover their initial
potential. XPS analysis of used samples confirmed
palladium leaching as being the main cause of catalyst deac-
tivation.
Homogeneous experiments with commercially available
[Pd
using palladium contents greater than around 0.01 mol%
(0.03 mol% for [Si-PHEMA-dpa-Pd]) and maximum
TON of around 12000 (given the XPS-determined composi-
tion of [Pd(OAc)2]; Figure 5). Thus immobilization of palla-
ACHTUNGRTEN(NUNG OAc)2] revealed quantitative transformations when
a
AHCTUNGTRENNUNG
dium catalysts on to polymer brushes appears at first sight
detrimental to their activity owing presumably to accessibili-
ty and diffusion issues.
The poor reusability of our surfaces prompted us to eluci-
date the exact nature of the catalytic species involved, that
is, whether the brushes work as purely heterogeneous cata-
lysts or as catalytic reservoirs.[7h] With this aim,
[Si-PHEMA-dpa-Pd] samples were first treated with PhSiH3
in CH2Cl2/EtOH to generate the Pd0 species. After 20 h, the
surfaces were removed and engaged in the catalytic depro-
tection of Alloc-coumarin 1 (1ꢃ10À4 mol, 10 equiv PhSiH3).
In addition, the activities of the remaining solutions were
tested similarly (Table 4).
*
Figure 5. Conversion
PHEMA-dpa-Pd] surfaces.
the homogeneous [Pd
line: evolution of the conversion for [Si-PHEMA-dpa-Pd]; grey line: evo-
(
)
and TON (&) versus S/C ratios for [Si-
~
represent conversions obtained by using
ACHTUNGTRENNUNG(OAc)2] catalyst under similar conditions. Black
lution of conversion for homogeneous [Pd
evolution of TON for [Si-PHEMA-dpa-Pd].
ACHTUNGRTEN(NUNG OAc)2]; black dashed line:
are more difficult to apprehend. For S/C<3500, the TONs
naturally follow the increase in the S/C ratios directly linked
to the modification of the surface area and PB thickness.
For example, increasing the S/C ratio by a factor
of two results in a proportional augmentation in
the TONs (Table 1). When pushing the system
above its limits (S/C>3500, Table 2), although a
TON of around 3500 should theoretically be meas-
ured in every case, a discrepancy is observed and,
for an identical surface area, higher TONs are re-
corded for the thinnest PBs. These observations
suggest that only a portion of the thickest catalytic
brushes seems available for the reaction to take
place. Facilitated diffusion of reactants and prod-
ucts in the thinnest PBs also has to be considered.
The influence of the solvent was not examined
in this work as we have previously observed[9g] that
the outcome of our model reaction is strongly de-
pendent on the nature of the solvent. Thus, this
renders the approximation of solvent-triggered ar-
Table 3. Recyclability of [Si-PHEMA-dpa-Pd] catalytic brushes.
Amount
Sample
Surface S/C[a] Conversion [%][b] for Total
of 1
area
run
3
TON[c]
A
E
1
2
4
5
1
2
3
4
5
6
7
8
9
1ꢃ10À5 80[Si-PHEMA-dpa-Pd]
213 94 96 30 15 11
426 96 97 32
852 95 74 12
538 96 24 29 17 19
1076 96 15 11
2152 96 19
826 96 17 12
1652 98 15 12
3304 96 10
2132 51 10
524
993
1602
995
1313
2668
1033
2065
3502
1298
1958
3149
0.5
0.25
1
0.5
0.25
1
0.5
0.25
1
8
7
–
–
40[Si-PHEMA-dpa-Pd]
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
9
30[Si-PHEMA-dpa-Pd]
–
–
–
–
10 1ꢃ10À4 80[Si-PHEMA-dpa-Pd]
11
12
0.5
0.25
4264 39
8528 29
7
8
[a] Substrate-to-catalyst ratio. [b] Conversion after 20 h, as determined by 1H NMR
spectroscopy. [c] Total turnover number.
16230
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16226 – 16233

Polymer Brushes
FULL PAPER
Table 4. Unraveling the catalytic nature of [Si-PHEMA-dpa-Pd].[a]
The preliminary study reported herein represents, to the
best of our knowledge, the first example of a transition-
metal catalyst immobilized on PBs prepared from single-
crystal silicon surfaces. Although our catalytic system needs
further improvement, the PBs represent an appealing plat-
form for the construction of innovative catalytic devices.
The possibility of reversibly collapsing/stretching PBs in re-
sponse to external stimuli[19] has tremendous potential for
the assembly of switchable catalytic surfaces[20] to achieve
precise control over chemical processes.[21]
Further studies will be dedicated to configuring adapta-
tive, robust, heterogeneous catalytic systems from planar sil-
icon surfaces. The deactivation can be attributed to the
strong reducing media (PhSiH3, used as a nucleophile and
reducing agent) that keep the palladium species in a low ox-
idation state. Decoordination from the weak low-oxidation-
state coordinating ligand dpa is then favored and leads most
probably to the formation of palladium particles. Model re-
actions that do not involve the formation of Pd0 species
(such as Michael addition or nitro-aldol reaction) should
give more robust catalytic systems and will be investigated
in the near future.
Sample
S/C[b] Conversion Conversion Conversion
[%][c]
[%][d]
[%][e]
80[Si-PHEMA-dpa-Pd] 2132
40[Si-PHEMA-dpa-Pd] 5376
30[Si-PHEMA-dpa-Pd] 8264
51 (96)
40 (95)
25 (95)
12
7
7
95
91
21
[a] 1 cm2 [Si-PHEMA-dpa-Pd], 1ꢃ10À4 mol Alloc-coumarin 1, 20 h.
[b] Substrate-to-catalyst ratio. [c] No initial treatment. Conversions in pa-
rentheses correspond to homogeneous controls with [PdACTHUNRGTNEUNG(OAc)2]. [d] Sur-
faces treated with PhSiH3 prior to catalysis. [e] Activity of the remaining
solution resulting from the treatment of [Si-PHEMA-dpa-Pd] with
PhSiH3.
Accordingly, surfaces treated (PhSiH3) as stated above
showed no or poor activity (7–12%), which suggests that all
the catalytic species had leached into the reducing solution.
Indeed, the solutions resulting from the initial treatment of
the [Si-PHEMA-dpa-Pd] samples were found to be highly
active, with conversions almost similar to those of homoge-
neous controls with [PdACTHNUTRGNE(NUG OAc)2]. The difference between
treated/untreated [Si-PHEMA-dpa-Pd] and homogeneous
controls suggests that the reaction most probably occurs
within or in the vicinity of the brushes, which thus restrains
the activity to a level comparable to the homogeneous
[PdACHTUNGTRENNUNG(OAc)2] counterparts.
Therefore these experiments together with the recyclabili-
ty tests (Table 3) and XPS studies have revealed that the
catalytic brushes presented herein function as forms of cata-
lytic reservoirs releasing active palladium species into solu-
tion as a result of the poor stability of [dpa–Pd0] complexes
compared with [dpa-PdII].
Experimental Section
General: Reagents were obtained from commercial sources and used
without further purification. All reactions were carried out under N2 or
argon. ROCC 60 granular silica gel (40–63 mm, 230–400 mesh ASTM)
was employed for column chromatography. Single-side-polished silicon
wafers were purchased from ACM (France) with (100) orientation. Milli-
Q water (resistivity 18.2 MWcm) was obtained from the Millipore system
(Elga Purelab Ultra). All reaction vessels were cleaned prior to use by
immersion in a hot, freshly prepared piranha solution (H2SO4 (98%)/
H2O2 (30%)) and then extensively washed with Milli-Q water and dried
in an oven. (Caution! Piranha solution is an extremely strong oxidant
and should be handled only with the proper equipment.)
Conclusion
PHEMA polymer brushes have been used for the immobili-
zation of dpa–palladium complexes on to flat silicon oxide
to prepare catalytically active surfaces.
We have demonstrated that the palladium concentration
can be well adjusted by using PBs of various thicknesses at-
tainable by time-controlled SI-ATRP. Our system relies on
the simple and robust postfunctionalization of the pendant
hydroxy groups of the PHEMA PBs by conventional activa-
tion and coupling chemistry.
A systematic study of our catalytic brushes (thickness, S/C
screenings, recyclability, leaching test) has unambiguously
revealed that our system functions as a reservoir that pro-
gressively releases catalytically active species into solution.
The PB reservoir shows a limiting TON of around 3500, less
Ellipsometry (Uvisel, Horiba-Jobin-Yvon, France) was performed at an
incidence angle of 708 and in a wavelength range of 400–850 nm. The el-
lipsometric data were fitted by using the DeltaPsi 2 software accompany-
ing the measuring apparatus. The ellipsometric model consists of three
layers: Silicon (bulk), native silicon oxide (1.5 nm thickness), and a poly-
mer brush. The complex indices of refraction of silicon and native SiO2
were taken from tabulated data provided by the manufacturer. The com-
plex indices of refraction of the brushes, nÀjk, was modeled by a trans-
parent Cauchy layer with n(l)=A+BlÀ2 +ClÀ4 and k(l)=0, with A, B,
and C three fitted parameters. X-ray reflectivity (XRR) measurements
were carried out with a modified Siemens D5000 2-circle goniometer
(0.0028 positioning accuracy). X-rays of wavelength 0.15418 nm (CuKa1
)
were obtained from a Rigaku rotating anode operated at 40 kV and
300 mA, fitted with a collimating mirror (Osmic, Japan) delivering a
close-to-parallel beam with around 0.00858 vertical angular divergence.
X-ray photoelectron spectroscopy (XPS) measurements were performed
on a SSX 100/206 photoelectron spectrometer from Surface Science In-
struments (USA) equipped with a monochromatized microfocused alumi-
nium X-ray source (hn=1486.6 eV) operated at 20 mA and 10 kV. All
À
than the homogeneous controls with [PdACTHNUTRGENUG(N OAc)2], which in-
dicates that the reaction occurs in the vicinity of the brush
framework and thus analogous activity is hindered. The
thickest surfaces offered the highest conversions due to in-
creased palladium content in the brushes. When working
under “extreme” conditions (S/C>3500), the thinnest surfa-
ces proved to be the most effective in terms of TON as a
result of catalyst release by enhanced diffusion out of the
brushes.
binding energies are referenced to the C ACHTNUTGRNEUNG(C,H) component of the carbon
1s peak fixed at 284.8 eV. The base pressure in the spectrometer was in
the low 10À8 Torr range. Quantitative information was obtained from the
photoemission peak areas of each element normalized according to ac-
quisition parameters and sensitivity factors provided by the manufactur-
er. Peak decomposition was achieved with the Casa XPS software (Casa
Software Ltd., UK). Inductively coupled plasma mass spectroscopy (ICP-
Chem. Eur. J. 2012, 18, 16226 – 16233
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16231

O. Riant, A. M. Jonas et al.
MS) was performed on an Agilent Quadrupole inductively coupled
plasma mass spectrometer. The samples were treated with ultrapure aqua
regia (3:1 concentrated HCl/HNO3) and the resulting solution was dilut-
ed 10 times prior to analysis. (Caution! Aqua regia is a very corrosive ox-
idizing agent, which should be handled with great care.) NMR spectra
were recorded on a Bruker-300 spectrometer. Mass spectra were ob-
tained by using a ThermoFinnigan LCQ Quantum spectrometer.
[3] L. Li, W. Hu, L. Chi, H. Fuchs, J. Phys. Chem. B 2010, 114, 5315–
5319.
[4] W. Senaratne, L. Andruzzi, C. K. Ober, Biomacromolecules 2005, 6,
2427–2448.
[5] a) S. Tugulu, A. Arnold, I. Sielaff, K. Johnsson, H.-A. Klok, Bioma-
cromolecules 2005, 6, 1602–1607; b) K. Glinel, A. M. Jonas, T. Jou-
enne, J. Leprince, L. Galas, W. T. S. Huck, Bioconjugate Chem. 2009,
20, 71–77; c) X. Laloyaux, E. Fautrꢁ, T. Blin, V. Purohit, J. Leprince,
T. Jouenne, A. M. Jonas, K. Glinel, Adv. Mater. 2010, 22, 5024–5028.
[6] S. Samanta, J. Locklin, Langmuir 2008, 24, 9558–9565.
[7] a) J. Huang, X. Li, Y. Zheng, Y. Zhang, R. Zhao, X. Gao, H. Yan,
Macromol. Biosci. 2008, 8, 508–515; b) B. Zhao, X. Jiang, D. Li, X.
Jiang, T. OꢀLenick, B. Li, C. Y. Li, J. Polym. Sci. Part A 2008, 46,
3438–3446; c) C. S. Gill, K. Venkatasubbaiah, N. T. S. Phan, M.
Weck, C. W. Jones, Chem. Eur. J. 2008, 14, 7306–7313; d) F. Costan-
tini, W. P. Bula, R. Salvio, J. Huskens, H. J. G. E. Gardeniers, D. N.
Reinhoudt, W. Verboom, J. Am. Chem. Soc. 2009, 131, 1650–1651;
e) C. S. Gill, W. Long, C. W. Jones, Catal. Lett. 2009, 131, 425–431;
f) A. Schꢄtz, T. R. Long, R. N. Grass, W. J. Stark, P. R. Hanson, O.
Reiser, Adv. Funct. Mater. 2010, 20, 4323–4328; g) F. Costantini,
E. M. Benetti, D. N. Reinhoudt, J. Huskens, G. J. Vancso, W. Ver-
boom, Lab Chip 2010, 10, 3407–3412; h) F. Costantini, E. M. Benet-
ti, R. M. Tiggelaar, H. J. G. E. Gardeniers, D. N. Reinhoudt, J. Husk-
ens, G. J. Vancso, W. Verboom, Chem. Eur. J. 2010, 16, 12406–
12411; i) M. Zeltner, A. Schꢄtz, M. L. Hefti, W. J. Stark, J. Mater.
Chem. 2011, 21, 2991–2996; j) W. Long, C. W. Jones, ACS Catal.
2011, 1, 674–681.
Silanization: The silicon wafers were cleaned by plasma etching and then
immediately installed in a Teflon holder and transferred into a Schlenk
tube. The whole system was sealed and heated at 808C. Three cycles of
vacuum/N2 were performed. The ATRP initiator silane (15 mL) was then
injected and the reaction was carried out at 808C for 2 h. The silanized
silicon wafers were removed, rinsed with toluene, and dried under a flow
of nitrogen.
Growth of poly(2-hydroxyethyl methacrylate) (PHEMA) polymer brush-
es by surface-initiated atom-transfer radical polymerization (SI-ATRP):
HEMA (8 mL) monomer was dissolved in deionized water (8 mL) and
CuCl (110 mg), CuBr2 (72 mg), and bipyridine (488 mg) were added. The
vessel was then sealed and N2 was bubbled. The solution was stirred
under N2 for 45 min until a homogeneous dark-brown solution formed.
A series of Schlenk tubes containing one [Si-Initiator] sample each were
sealed and subjected three vacuum/N2 cycles. The growth of the polymer
brushes was initiated by the injection of the previously described solution
(2.5 mL solution per sample). The reactions were performed for deter-
mined periods of time to give PBs of various thicknesses. The [Si-
PHEMA] surfaces were recovered, rinsed twice with methanol and de-
ionized water, and dried under a flow of nitrogen.
[8] a) R. Barbey, L. Lavanant, D. Paripovic, N. Schꢅwer, C. Sugnaux, S.
Tugulu, H.-A. Klok, Chem. Rev. 2009, 109, 5437–5527; b) S. V.
Orski, K. H. Fries, S. K. Sontag, J. Locklin, J. Mater. Chem. 2011, 21,
14135–14149.
Incorporation of the dpa ligand 3: Reactions were performed in a glove
box. [Si-PHEMA] was first activated prior to coupling with 3. The [Si-
PHEMA] surfaces were immersed in a solution of N,N’-disuccinimidyl
carbonate (DSC; 0.10 mmol per sample) and Et3N (0.20 mmol per
sample) in DMF (2 mL per sample) for 24 h. The samples were then re-
moved, washed with DMF, and immersed into a solution containing dpa
ligand 3 (0.03 mmol per sample) and Et3N (0.10 mmol per sample) in
DMF (1.5 mL per sample). After 24 h of reaction, the samples were re-
moved and thoroughly washed with CH2Cl2 to give [Si-PHEMA-dpa] sur-
faces.
[9] a) M. G. L. Petrucci, A. K. Kakkar, J. Chem. Soc., Chem. Commun.
1995, 1577–1578; b) M. G. L. Petrucci, A. K. Kakkar, Chem. Mater.
1999, 11, 269–276; c) A. K. Kakkar, Chem. Rev. 2002, 102, 3579–
3588; d) B. A. Harper, D. A. Knight, C. George, S. L. Brandow, W. J.
Dressick, C. S. Dalcey, T. L. Schull, Inorg. Chem. 2003, 42, 516–524;
e) K. Hara, S. Tayama, H. Kano, T. Masuda, S. Takakusagi, T.
Kondo, K. Uosaki, M. Sawamura, Chem. Commun. 2007, 4280–
4282; f) K. Hara, R. Akiyama, S. Takakusagi, K. Uosaki, T. Yoshino,
H. Kagi, M. Sawamura, Angew. Chem. 2008, 120, 5709–5712;
Angew. Chem. Int. Ed. 2008, 47, 5627–5630; g) A. Fernandes, P.
Hensenne, B. Mathy, W. Guo, B. Nysten, A. M. Jonas, O. Riant,
Chem. Eur. J. 2012, 18, 788–792.
Incorporation of [PdACHTUNGTRENNUNG(OAc)2]: Reactions were performed in a glove box.
Prior to complexation, the backsides of the samples were covered with a
layer of poly(vinylidene difluoride) (PVdF). The [Si-PHEMA-dpa] surfa-
ces were immersed in a solution of [PdACTHNUGTRNEUNG(OAc)2] (0.05 mmol per sample)
in CH2Cl2 (3 mL per sample) for 2 h. The samples were then removed
and washed with CH2Cl2 in a Soxhlet apparatus overnight to obtain the
[Si-PHEMA-dpa-Pd] catalysts.
[10] D. Kisailus, Q. Truong, Y. Amemiya, J. C. Weaver, D. E. Morse,
Proc. Natl. Acad. Sci. USA 2006, 103, 5652–5657.
Catalysis: The reactions were performed in freshly cleaned (piranha)
Schlenk tubes. Appropriate stock solutions of Alloc-coumarin 1 were
prepared prior to catalysis (CH2Cl2/EtOH, 1:1, v/v). A [Si-PHEMA-dpa-
Pd] surface was immersed in coumarin 1 (1 mL) and PhSiH3 (10 equiv)
was added. The system was kept under argon for 20 h without stirring.
[11] a) M. R. Buchmeiser, S. Lubbad, M. Mayr, K. Wurst, Inorg. Chim.
Acta 2003, 345, 145–153; b) S. Fakih, W. C. Tung, D. Eierhoff, C.
Mock, B. Krebs, Z. Anorg. Allg. Chem. 2005, 631, 1397–1402; c) B.
Antonioli, D. J. Bray, J. K. Clegg, K. Gloe, K. Gloe, A. Jꢄger, K. A.
Jolliffe, O. Kataeva, L. F. Lindoy, P. J. Steel, C. J. Sumby, M. Wenzel,
Polyhedron 2008, 27, 2889–2898.
1
Conversions were determined by H NMR spectroscopy.[9g]
[12] a) B. S. Lee, Y. S. Chi, K.-B. Lee, Y.-G. Kim, I. S. Choi, Biomacro-
molecules 2007, 8, 3922–3929; b) S. Diamanti, S. Arifuzzaman, A.
Elsen, J. Genzer, R. A. Vaia, Polymer 2008, 49, 3770–3779; c) S. M.
Lane, Z. Kuang, J. Yom, S. Arifuzzaman, J. Genzer, B. Farmer, R.
Naik, R. A. Vaia, Biomacromolecules 2011, 12, 1822–1830.
[13] N. A. Peppas, H. J. Moynihan, L. M. Lucht, J. Biomed. Mater. Res.
1985, 19, 397–411.
[14] R. F. Fedors, Polym. Eng. Sci. 1974, 14, 147–154.
[15] V. A. Baulin, E. B. Zhulina, A. Halperin, J. Chem. Phys. 2003, 119,
10977–10989.
[16] S. S. Zalesskiy, V. P. Ananikov, Organometallics 2012, 31, 2302–2309.
[17] a) J. El Fallah, L. Hilaire, M. Romꢁo, F. Le Normand, J. Electron
Spectrosc. Relat. Phenom. 1995, 73, 89–103; b) C. C. Chusuei, M. A.
Brokkkshier, D. W. Goodman, Langmuir 1999, 15, 2806–2808.
[18] a) J. P. Mathew, M. Srinivasan, Eur. Polym. J. 1995, 31, 835–839;
b) A. K. Zharmagambetova, V. A. Golodov, Y. P. Saltykov, J. Mol.
Catal. 1989, 55, 406–414; c) G. B. Onoa, V. Moreno, M. Font-
Acknowledgements
The authors gratefully acknowledge access to the SUCH and ASM plat-
forms of IMCN. BN is a Senior Research Associate of the FRS-FNRS.
Financial support was provided by the Communautꢁ FranÅaise de Belgi-
que (ARC 06-11/339), the Belgian Federal Science Policy (IAP-PAI P6/
27), and the FRS-FNRS. Michel Genet is gratefully acknowledged for
discussions on the XPS data.
[1] S. T. Milner, Science 1991, 251, 905–914.
[2] a) S. Morsch, W. C. Schofield, J. P. Badyal, Langmuir 2010, 26,
12342–12350; b) Y. Roiter, I. Minko, D. Nykypanchuk, I. Tokarev,
S. Minko, Nanoscale 2012, 4, 284–292.
16232
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16226 – 16233

Polymer Brushes
FULL PAPER
Bardia, X. Solans, J. M. Pꢁrez, C. Alonso, J. Inorg. Biochem. 1999,
75, 205–212.
Ed. 2010, 49, 9254–9257; c) J. Wang, B. L. Feringa, Science 2011,
331, 1429–1432; d) V. Blanco, A. Carlone, K. D. Hꢄnni, D. A. Leigh,
B. Lewandowski, Angew. Chem. 2012, 124, 5256–5259; Angew.
Chem. Int. Ed. 2012, 51, 5166–5169; e) D. Wilson, N. R. Branda,
Angew. Chem. 2012, 124, 5527–5530; Angew. Chem. Int. Ed. 2012,
51, 5431–5434.
[19] a) X. Laloyaux, B. Mathy, B. Nysten, A. M. Jonas, Macromolecules
2010, 43, 7744–7751; b) X. Laloyaux, B. Mathy, B. Nysten, A. M.
Jonas, Langmuir 2010, 26, 838–847.
[20] R. S. Stoll, S. Hecht, Org. Lett. 2009, 11, 4790–4793.
[21] a) H. J. Yoon, J. Kuwabara, J.-H. Kim, C. A. Mirkin, Science 2010,
330, 66–69; b) Y. Sohtome, S. Tanaka, K. Takada, T. Yamaguchi, K.
Nagasawa, Angew. Chem. 2010, 122, 9440–9443; Angew. Chem. Int.
Received: July 16, 2012
Published online: October 2, 2012
Chem. Eur. J. 2012, 18, 16226 – 16233
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16233