Catalytic self-assembled monolayers on Au nanoparticles: The source of catalysis of a transphosphorylation reaction

Article abstract of DOI:10.1002/chem.201002590

The catalytic activity of a series of Au monolayer protected colloids (Au MPCs) containing different ratios of the catalytic unit triazacyclononane Zn^II (TACNZn^II) and an inert triethyleneglycol (TEG) unit was measured. The catalytic

Full text of DOI:10.1002/chem.201002590

FULL PAPER
DOI: 10.1002/chem.201002590
Catalytic Self-Assembled Monolayers on Au Nanoparticles: The Source of
Catalysis of a Transphosphorylation Reaction
Giovanni Zaupa, Claudia Mora, Renato Bonomi, Leonard J. Prins,* and
Paolo Scrimin*[a]
Abstract: The catalytic activity of a
series of Au monolayer protected col-
loids (Au MPCs) containing different
ratios of the catalytic unit triazacyclo-
nonane·ZnII (TACN·ZnII) and an inert
triethyleneglycol (TEG) unit was mea-
sured. The catalytic self-assembled
monolayers (SAMs) are highly efficient
in the transphosphorylation of 2-hy-
droxy propyl 4-nitrophenyl phosphate
(HPNPP), an RNA model substrate,
exhibiting maximum values for the
Michaelis–Menten parameters kcat and
KM of 6.7ꢀ10À3 sÀ1 and 3.1ꢀ10À4 m, re-
spectively, normalized per catalytic
unit. Despite the structural simplicity
of the catalytic units, this renders these
nanoparticles among the most active
catalysts known for this substrate. Both
kcat and KM parameters were deter-
mined as a function of the mole frac-
tion of catalytic unit (x1) in the SAM.
Within this nanoparticle (NP) series,
kcat increases up till x1 ꢀ0.4, after which
it remains constant and KM decreases
exponentially over the range studied.
causes an increase of the number of
potential dimeric catalytic sites com-
posed of two catalytic units as a func-
tion of the x1, which causes an apparent
increase in binding affinity (decrease in
KM). Simultaneously, the kcat value is
determined by the number of substrate
molecules bound at saturation. For
values of x1 >0.4, isolated catalytic
units are no longer present and all cat-
alytic units are involved in catalysis at
saturation. Importantly, the observed
trends are indicative of a random dis-
tribution of the thiols in the SAM. As
indicated by the theoretical analysis,
and confirmed by a control experiment,
in case of clustering both kcat and KM
values remain constant over the entire
range of x1.
A
theoretical analysis demonstrated
that these trends are an intrinsic prop-
erty of catalytic SAMs, in which cataly-
sis originates from the cooperative
effect between two neighboring catalyt-
ic units. The multivalency of the system
Keywords: gold
·
monolayers
·
nanoparticles · self-assembly · sup-
ported catalysts · transphosphoryla-
tion
Introduction
the advantages of homogeneous and heterogeneous cata-
lysts.[2] Examples include the incorporation of molecular
complexes, or clusters on surfaces or in pores of otherwise
inert supports, such as silica,[3,4] clays,[5] zeolites,[6] and
resins.[7] In most cases, the inert support serves only to facili-
tate catalyst recovery. A next step towards catalytic architec-
tures are those systems in which the support also plays and
additive role in the catalytic performance. An example are
the so-called metal complex@metal systems obtained by im-
mobilization of an organometallic complex on a metal nano-
particle.[8] A second example are self-assembled monolayers
(SAMs) of catalysts on inorganic nanoparticles.[9–11] Here,
the nanoparticle serves as a large structural unit on which
multiple catalytic complexes are assembled. Also here, the
inorganic core allows recovery of the catalytic system
through filtration,[12] or magnetic separation,[13,14] but the ad-
vantages go far beyond that. As illustrated by us[15–17] and
others,[18] the multivalent nature of Au nanoparticles can
give rise to cooperative effects between catalytic units. In
the past, similar effects have also been observed in micelles
and vesicles, but the much lower stability of these aggregates
poses limits to their application and in most cases does not
permit catalysis under turnover conditions.[19–21] A related
class of multivalent platforms are dendrimers, which have
gained an enormous interest, amongst others for their ability
The increasing demand for highly atom-efficient chemical
transformations with a low environmental impact makes the
development of new catalysts one of the crucial challenges
for chemists.[1] Traditionally, catalysts are classified as either
homogeneous or heterogeneous depending whether the cat-
alyst is solubilized in the reaction medium or supported on
a solid phase. The main advantage of homogeneous catalysts
is that specificity, activity, and selectivity is determined by
their molecular architecture, which allows these properties
to be tuned in a rational fashion. On the other hand, the re-
covery and re-use of homogeneous catalysts is complicated,
and for that reason large industrial processes rely for a large
part on heterogeneous catalysts. A current trend in catalyst
design is the development of hybrid systems that combine
[a] Dr. G. Zaupa, C. Mora, Dr. R. Bonomi, Dr. L. J. Prins,
Prof. Dr. P. Scrimin
Department of Chemical Sciences and CNR-ITM
Padova Section, Via Marzolo 1, 35131 Padova, (Italy)
Fax : (+39)0498275239
E-mail: leonard.prins@unipd.it
paolo.scrimin@unipd.it
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002590.
Chem. Eur. J. 2011, 17, 4879 – 4889
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4879

to induce cooperative effects between peripheral catalytic
units.[22] In fact, the term “dendritic effect” describes the in-
crease in dendrimer performance as a function of its valen-
cy[23–26] Compared to Au monolayer protected colloids (Au
MPCs), however, their synthesis and purification is much
more cumbersome.
Recently, it has been shown that in mixed monolayers on
Au NPs the properties of the catalytic units may be steered
by the surrounding inert units.[27] This opens up the exciting
perspective of modulating the catalytic properties of the
system simply by changing the composition of the mono-
the monolayer composition and it has been shown that this
allows virtually all ratios of thiols in a binary mixture to be
addressed.[32,38]
In this work, we prepared mixed SAMs composed of thiol
1,[32] containing a triazacyclononane (TACN) head group
that forms a stable complex with ZnII-ions, and thiol 2,[39]
containing a TEG head group that renders the NPs soluble
in aqueous solution (Scheme 1). Previously, we have shown
ACHTUNGTRENNUNGlayer. Clearly, in such a system the overall catalytic perfor-
mance is strongly determined by the structural order of the
SAM, that is, the orientation of the catalytic complexes with
respect to each other or other functionalities. Recently, ob-
servations indicating the formation of homodomains or clus-
ters in mixed SAMs have been reported.[28–30] Evidently,
from the perspective of controlling the properties of cata-
lysts within a mixed monolayer, it is of crucial importance to
know the structural order of the SAM.[31] However, studies
that correlate the catalytic performance to the composition
of the SAM are hitherto nearly absent.[11] Here, we perform
such an in-depth study on a catalytic Au-NP-based system
that we had already earlier shown to be highly efficient in
catalyzing the transphosphorylation of 2-hydroxypropyl 4-
Scheme 1. Au MPCs functionalized with thiols 1 and 2.
that multivalent systems, such as tripodal ligands,[40] den-
drimers,[41–43] and Au NPs,[15] that contain multiple copies of
the TACN·ZnII-complex are highly efficient catalysts for the
transphosphorylation of HPNPP. For all systems, kinetic
ZnII titrations revealed that catalysis requires the coopera-
tive action of two TACN·ZnII complexes. The isolated
TACN·ZnII complex is practically inactive in catalyzing this
reaction.[15] This feature makes this catalytic system intrinsi-
cally very well-suited to be implemented in multivalent sys-
tems, similar to imidazoles for ester hydrolysis[26,44,45] or the
Jacobsen salen complexes (salen=N,N’-ethylenebis(salicyl-
nitroACHTUNGTRENNUNGphenyl phosphate (HPNPP), an RNA model sub-
strate.[15] The study of a series of Au MPCs with various cat-
alyst loadings on the surface gives for the first time a corre-
lation between catalyst loading and overall catalytic efficien-
cy. We address the important question whether the catalytic
sites, once formed, change upon increasing the mole fraction
of catalytic units and whether “dendritic effects” occur. In
addition, it is shown that the obtained correlation can be
used to gain information on the structural order within
mixed SAMs as well as on the source of catalysis in these
systems.
AHCTUNGTRENNUNG
imine)) for asymmetric epoxidation reactions.[46] Compared
to our previous studies,[15] we have modified the TACN-
functionalized thiol, eliminating the terminal amide group.
This modification facilitated the synthesis and also had
some beneficial effect on the catalytic activity.
To study the correlation between surface coverage and
catalytic activity we prepared a series of Au NPs (I–VII)
protected with thiols 1 and 2 in the ratios indicated in
Table 1. To ensure an identical core size for all Au NPs the
Results and Discussion
The synthetic protocol for the preparation of mixed thiol
monolayers on Au NPs follows a modified literature proce-
dure[32] originally developed by Jana and Peng.[33] This proto-
col relies on the initial formation of Au NPs transiently sta-
bilized with secondary amines (a synthetic scheme is given
in the Supporting Information). Here, the ratio of dioctyl-
Table 1. Monolayer composition of Au MPCs I–VIII determined by
1H NMR spectroscopy (LEDbp and after I2 oxidation) and UV/Vis spec-
troscopy.
[a]
NP batch
x1
ACHTUNGTRENNUNG
Expected
NMR
LEDbp
NMR
after I2
UV/Vis
CuII
av[b]
ACHTUNGTRENNUNG
I
II
1.00
0.80
0.60
0.40
0.20
0.10
0.05
0
–
–
–
1.00
0.74
0.56
0.35
0.17
0.10
0.05
0
sired thiols to give the final products. This two-step protocol
offers an important advantage over other synthetic proto-
cols. Only the minimal amount of thiol necessary to cover
the Au NPs needs to be added, which means that an excess
of thiol is hardly present. Accordingly, the composition of
mixed monolayers neatly reflects the composition of the
mixture of thiols added. Compared to other methods based
on thiol exchange,[34–37] this gives a much larger control over
0.79
0.60
0.40
0.15
0.09
0.04
–
0.72
0.54
0.34
0.15
N.D.[c]
N.D.
–
0.72
0.56
0.31
0.20
0.11
0.06
–
III
IV
V
VI
VII
VIII
[a] x1 is the mole fraction of 1 on the NP surface. [b] We estimate the
final error in the x1 values to be <0.05. [c] N.D.=not determined.
4880
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4879 – 4889

Catalytic Self-Assembled Monolayers
FULL PAPER
initial batch of dioctylamine-stabilized Au NPs was split into
eight batches to which thiols 1 and 2 were added in the de-
sired ratios. All nanoparticle batches were completely solu-
ble in water. Before assessing their catalytic activity, a varie-
ty of analytical techniques was used to characterize the Au
MPCs and to determine the obtained ratio of thiols 1 and 2
in the monolayer.
Monolayer composition: Of fundamental importance in this
study is the composition of the monolayer in terms of the
thiol ratio 1:2. This parameter was assessed in a variety of
ways. Diffusion-ordered 1H NMR spectra were recorded
using the low eddy currents distortion bipolar gradients
(LEDbp) sequence,[47] which allows differentiation of mole-
cules based on their diffusion coefficients.[48] Not only does
this provide unequivocal confirmation that the thiols are
bound to the Au NP surface, but also provides a way to
assess the purity of the samples. The obtained NMR spectra
with and without the diffusion filter showed that only mini-
mal amounts of unbound additives were present in the puri-
fied samples (see Supporting Information; for an example,
see Figure 2a and 2b). For each batch, the thiol ratio 1:2
was determined by integration of characteristic signals for
each of the thiols (1: d=2.9 ppm; 2: d=3.4–3.6 ppm) cali-
brated on a reference signal common to both thiols (d=
2.13 ppm) (for a representative example, see Figure 2c). To
verify that integration of the broad signals did not cause
large errors, all integrations were repeated after treatment
of the Au MPCs with I2. Addition of I2 causes a decomposi-
tion of the Au MPCs and the formation of disulfides (both
homomeric and heteromeric). No longer being adsorped on
Size of the Au core: The diameter of the gold nuclei was de-
termined as 1.6Æ0.2 nm by means of HR-TEM (for a repre-
sentative example see Figure 1). The absence of the surface
plasmon resonance band at 520 nm in the UV/Vis spectrum
gave additional proof for the presence of sub 2 nm sized
NPs and, importantly, for the absence of clustered gold
nanoparticles (Figure 1c).
1
the nanoparticle, the H NMR spectra of the organic com-
pounds exhibit sharp signals which can be readily integrated
(see Supporting Information). The 1:2 ratios thus deter-
mined did not substantially differentiate from those previ-
ously determined.
The concentration of TACN head groups was also deter-
mined by means of a spectrophotometric titration of each
batch of nanoparticles with CuII, taking advantage of the
characteristic absorbance of the 1:1 TACN·CuII complex at
264 nm
(e264 =7800 LmolÀ1 cmÀ1
compared
to
800 LmolÀ1 cmÀ1 for free CuII). The result of one such a ti-
tration is given in Figure 3 (for details see the Experimental
Section). The absorbance increases linearly up till a satura-
tion of all TACN ligands with CuII. The slight increase in ab-
sorbance after the addition of one equivalent of CuII corre-
sponds to free CuII in solution. The obtained concentrations
of TACN were found to be in good agreement with those
1
obtained from the H NMR studies. Finally, kinetic titrations
using ZnII (see next paragraph) gave additional proof for
the monolayer composition.
Catalytic activity: The catalytic activity of Au MPCs I–VII
(TEG-functionalized Au MPC VIII is not catalytically
active) in the transesterification of HPNPP[49] was first eval-
uated by means of kinetic titrations with ZnII. In these ex-
periments the amount of ZnII is gradually increased at con-
stant Au MPC concentration and the initial rate is deter-
mined by measuring the amount of released p-nitropheno-
late ion in time (Scheme 2). It should be noted that all ki-
netic studies that will be described in this work were
performed at the same nominal concentration of TACN
(20 mm). This allows for a direct comparison of the catalytic
activity of each NP batch, since numerical contributions are
eliminated. Evidently, this implies that the concentration of
Figure 1. a) Representative example of a HR-TEM image (sample I). b)
Size distribution of the Au NPs. c) Representative UV/Vis spectrum indi-
cating the absence of the surface plasmon band at 520 nm.
Chem. Eur. J. 2011, 17, 4879 – 4889
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4881

L. J. Prins. P. Scrimin et al.
without ZnII. Second, the cata-
lytic activity continues to in-
crease until TACN and ZnII are
present in a 1:1 ratio, indicating
that the TACN·ZnII complex is
the catalytic unit. Additionally,
this also provides final evidence
that the TACN surface loadings
as determined in the previous
paragraph are correct. Third,
analogous to previously studied
multivalent systems,[15,41] sig-
moidal curves are obtained in-
dicating that the process is co-
operative and the catalytic site
is composed of two TACN·ZnII
complexes.[50] Finally, a simple
glance on the curves tells imme-
diately that the catalytic activity
strongly increases as a function
of the surface loading of TACN.
Subsequently, the catalytic
performances of each nanopar-
ticle batch I–VII were assessed
separately (in the presence of
Figure 2. a) Part of the 1H NMR spectrum of NP batch I. b) The same sample using the LEDbp sequence. c)
Part of the H NMR spectrum of NP batch II indicating the signals used for quantification. All spectra are re-
corded in [D4]MeOD (300 MHz, 301 K).
1
Figure 3. Changes in the absorbance at 264 nm of Au NP batch I upon
the addition of CuII. For a detailed description see the Experimental Sec-
tion.
Scheme 2. Transphosphorylation of HPNPP.
&
Figure 4. a) Initial rates for the cleavage of HPNPP by NP batch I ( ), II
NPs increases as the surface loading decreases. Potentially
this might affect the catalytic parameters, which is an issue
that will be addressed separately. Plots of the initial rates
against the amount of added ZnII already provide much in-
formation on the behavior of these systems, which bear
much resemblance to the TACN-based systems that we have
studied earlier (Figure 4a).[15,41] First, the pivotal role of ZnII
is evident from the absence of catalytic activity for the NPs
~
*
*
^
(
), III ( ), IV (ꢀ), VI ( ), and VII ( ) as a function of the equivalents
of ZnII added. Conditions: [TACN]=2ꢀ10À5 m, [HPNPP]=2ꢀ10À4 m,
[HEPES]=1ꢀ10À2 m, pH 7.5, T=408C). For 4 batches trendlines are
added to indicate the sigmoidal behavior. b) Initial rates for the cleavage
~
&
*
*
of HPNPP by NP batch I ( ), II ( ), III ( ), IV (ꢀ), V (*), VI ( ), and
^
VII
(
)
as
a
function of the substrate concentration. Conditions:
[TACN]=2ꢀ10À5 m,
[Zn
ACHTUNGTRENNUNG
pH 7.5, T=408C. Solid lines indicate the best fits to the Michaelis–
Menten equation. Errors are typically ꢁ10%.
4882
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4879 – 4889

Catalytic Self-Assembled Monolayers
FULL PAPER
one equivalent of ZnII per TACN) by measuring the initial
reaction rates in the presence of increasing amounts of sub-
strate HPNPP, that is, under turn-over conditions (experi-
ment series A, Figure 4b). Enzyme-like saturation curves
were obtained, which were fitted to the Michaelis–Menten
equation vinit =kcat[TACN·ZnII][S]/
ACTHNUTRGEN(UNG KM+[S]) yielding the Mi-
chaelis–Menten parameters kcat and KM for each nanoparti-
cle batch I–VII.
In experiment series B, increasing amounts of TEG-func-
tionalized nanoparticles VIII were mixed with a constant
amount of nanoparticles I (and one equivalent of ZnII per
TACN). Thus, in this series exclusively homofunctionalized
nanoparticles (containing either TACN- or TEG-head
groups) were used.[51] For each ratio I:VIII the Michaelis–
Menten parameters were obtained from the respective satu-
ration profiles. The resulting Michaelis–Menten parameters
for both experiment series A and B are given in Tables 2
and 3 and plotted in Figure 5 as a function of the mole frac-
Table 2. Michaelis–Menten parameters for Au MPCs I-VII (experiment
series A). Parameters for multivalent systems 3–5 are added for compari-
son. These last values are normalized for the number of ZnII ions pres-
ent.
[b]
System
kcat
KM
[ꢀ10À3 m]
kcat/KM
kcat/kuncat
[ꢀ10À3 sÀ1
]
N
[LmolÀ1 sÀ1
]
I[a]
6.7Æ0.6
7.9Æ0.8
6.1Æ0.6
5.9Æ0.6
2.6Æ0.3
1.5Æ0.2
0.8Æ0.1
0.6
0.3Æ0.1
0.5Æ0.1
0.4Æ0.1
0.4Æ0.1
0.7Æ0.1
1.2Æ0.1
2.8Æ0.4
1.2
21.7Æ4.7
14.8Æ2.9
15.8Æ3.1
13.7Æ2.7
3.8Æ0.8
1.2Æ0.3
0.3Æ0.1
0.5
33587
39313
30529
29327
13095
7586
3842
3000
500
42500
II[a]
III[a]
IV[a]
V[a]
VI[a]
VII[a]
3[c,f]
4[d,f]
5[e,f]
0.1
8.5
3.9
0.32
0.03
26.6
Figure 5. Experimentally obtained Michaelis–Menten parameters a) kcat
,
[a] Conditions as given in the legend of Figure 5. [b] kuncat =2ꢀ10À7 sÀ1
[c] Taken from reference [42]. [d] Taken from reference [58]. [e] Taken
from reference [52]. [f] The structures of 3, 4 and 5 are given in Figure 8.
.
&
b) KM, and c) kcat/KM as a function of x1 for experiment series A ( ) and
*
B ( ). The solid lines in a) and b) are the calculated values for randomly
distributed thiols in the SAM according to the model (see Figure 7).
Table 3. Michaelis–Menten parameters for mixtures of Au MPCs I and
VIII in different ratios (experiment series B).
list of catalysts see references [53,54]). It has previously
been shown that the single TACN·ZnII complex is a very
poor catalyst, exhibiting an apparent second-order rate con-
stant of 0.007mÀ1 sÀ1 (pH 7.4, 408C).[15] Second, for series A
and B a different correlation is observed between the pa-
rameters kcat and KM and the mole fraction x1 (Figure 5a,b).
For series B both parameters are nearly constant over the
range of mole fractions studied. For series A, however, kcat
increases in a linear manner up till x1 ꢀ0.4 after which it
levels off. KM shows an exponential decay, indicative of a
significant increase in substrate affinity as the catalyst sur-
face loading increases. The combined positive effects of kcat
and KM result in an almost linear increase of the second-
order rate constant kcat/KM as a function of x1. In terms of
the second-order rate constant, the single TACN·
ZnII complex is almost 80-fold more active in batch I (x1 =1)
compared to batch VII (x1 =0.05). For series B the differ-
ence is negligible (1.5-fold). This is strongly indicative of a
positive dendritic effect, similar to that observed in den-
Ratio
I/VIII
kcat
KM
[ꢀ10À3 m]
kcat/KM
[ꢀ10À3 sÀ1
]
N
[LmolÀ1 sÀ1
]
1.00
0.50
0.17
0.09
6.7Æ0.6
5.0Æ0.6
5.7Æ0.6
4.8Æ0.5
0.3Æ0.1
0.3Æ0.1
0.3Æ0.1
0.4Æ0.1
21.7Æ4.7
17.6Æ4.6
17.3Æ3.9
13.0Æ2.4
[a] Conditions as given in the legend of Figure 5. [b] kuncat =2ꢀ10À7 sÀ1
.
tion x1. The graphs will be discussed extensively in the final
section of this manuscript, but at first glance immediately
some trends can be evidenced. First, nanoparticles with a
high surface loading of catalyst (x1 >0.5) are characterized
by a high catalytic activity (kcat/kuncat >30000) and high affin-
ity for the substrate (KM in the sub-mm range), which com-
bine to give a kcat/KM value of 21.7 LmolÀ1 sÀ1 for x1 =1.
This makes this system among the most potent catalysts re-
ported for this reaction (the best-performing dinuclear cata-
lyst 5[52] is added as a comparison in Table 2; for a detailed
Chem. Eur. J. 2011, 17, 4879 – 4889
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4883

L. J. Prins. P. Scrimin et al.
drimers. However, before interpreting these correlations it is
important to first analyze which kind of trends in kcat and
KM are expressed intrinsically by this kind of multivalent
systems.
local pH, etc.) will additionally affect the “overall” Michae-
lis–Menten parameters (either positively or negatively). To
correctly interpret the data of the NP-based systems pre-
sented in this study we need first to establish the intrinsic
changes of the “overall” Michaelis–Menten parameters, KM
and kcat, as a function of the composition of the mixed mon-
olayers. It is evident that these changes will also depend on
the structural order of the monolayer, that is, whether the
two thiols 1 (catalytic unit) and 2 (inert unit) are randomly
distributed or present in clusters or homodomains. These
boundary conditions will both be addressed. To evaluate the
correlation between KM, kcat, and the mole fraction x1 on the
surface, we took a truncated icosahedron as a simple model
for a spherical nanoparticle.[55] A truncated icosahedron is
an Archimedean solid composed of 12 pentagonal and 20
hexagonal faces. It was assumed that each face can accom-
modate one unit of 1 or 2. This implies that at full surface
coverage 32 units are bound. In this model each unit has
five or six neighboring units, which corresponds reasonably
to the packing of SAMs of thiols on Au surfaces. The only
catalytically relevant species is the dimeric site 1–1 formed
between two neighboring units 1. In all simulations the 1–1
site has constant KM and kcat values. Contributions from the
single catalytic unit 1 and the dimeric site 1–2 to catalysis
were not included.
A statistical distribution of units of 1 on the surface was
simulated by the stepwise random insertion of 1 on the
facets of the model completely covered with 2 (x1 =0). This
was repeated until all facets were occupied with 1 (x1 =1).
After each addition, the number of catalytic dimeric sites 1–
1 was counted and also the number of dimeric sites 1–1 that
could be occupied simultaneously. It is evident that the first
number is higher than the second, because a single catalytic
unit 1 can potentially form a catalytic site with each of its 1
neighbors, but effectively can participate in the binding of
one substrate molecule only. This is exemplified in Fig-
ure 6a, which depicts the number of potential catalytic sites
and the effective catalytic sites for x1 =0.5. Assuming that
for a given amount of 1 and 2 SAMs are formed according
to a binomial distribution and normalizing for the amount
of 1 present (meaning that the concentration of 1 is a con-
stant) gives the profiles for the number of initial binding
sites and for the maximum number of substrate molecules
bound as a function of the mole fraction x1 as depicted in
Theoretical analysis: Enzyme-like behavior of a catalyst im-
plies the occurrence of an initial binding event between the
catalyst and the substrate, characterized by a dissociation
constant KM, followed by the conversion of substrate into
product, characterized by a first-order rate constant kcat. The
consequence is a saturation profile when the initial rate of
reaction (ninit) is plotted as a function of substrate concentra-
tion, [S]. This graph is defined by a second-order regime at
low substrate concentration ([S]!KM for which ninit =(kcat
/
KM)[E][S]; [E]=“enzyme” or catalyst concentration) and a
first-order regime at high substrate concentrations ([S]@KM
for which ninit =kcat[E]). The Michaelis–Menten equation is
designed to describe the behavior of a single substrate–
enzyme interaction. The interpretation of the Michaelis–
Menten parameters for multivalent enzyme-like catalysts is
not straightforward, because in such systems multiple sub-
strate–enzyme interactions take place simultaneously. The
experimentally obtained values are therefore average values
of all single binding and catalytic events. To have a meaning-
ful discussion of these “averaged” values, one has to know
the intrinsic effect originating from the multivalency of the
system. Previously we have performed such an analysis on a
series of homofunctionalised dendrimers of increasing valen-
cy.[42] Rather interestingly, we observed that the clustering of
catalytic units on a multivalent scaffold gives spontaneously
rise to a positive dendritic effect; that is, an increased cata-
lytic efficiency (kcat/KM) as a function of the dendrimer va-
lency on the condition that catalysis requires the simultane-
ous action of two catalytic units. A theoretical analysis
showed that this intrinsic dendritic effect originates from the
fact that the concentration of catalytic sites (composed of
two units) increases more than linearly as a function of the
valency of the system. In other words, the same number of
catalytic units can create more catalytic sites in case the va-
lency is higher. Consequently, the “apparent” concentration
of catalytic sites is much higher than the nominal concentra-
tion of catalytic units. This implies that fitting of the satura-
tion profile using the classical Michaelis–Menten equation
(taking the concentration of catalytic units as reference)
yields an “overall” value for KM that decreases as a function
of the valency (suggesting stronger binding). This decrease
in the “overall” KM value is an intrinsic property of multiva-
lent systems and has nothing to do with the actual binding
affinity between the catalytic site and the substrate (which
remains constant). Additionally, we showed that the catalyt-
ic efficiency of the multivalent system at saturation (where
&
Figure 7 ( ). Next, homodomain formation of 1 was simulat-
ed by inserting new 1 units on the facets of the model on po-
sitions neighboring other 1 units (Figure 6b). Also here, it
was assumed that SAMs are formed according to a binomial
distribution of 1 and 2 on different nanoparticles. This as-
sumption implies that clustering occurs after monolayer for-
mation. The resulting profiles are also depicted in Figure 7
(&). Clustering during monolayer formation effectively
means the tendency of the thiols to self-sort preferentially
on different nanoparticles, which in an extreme case means
that only homomeric 1 and 2 nanoparticles are present. Nor-
malized on the concentration of 1 would give constant KM
and kcat values independent of x1. This situation is added to
v
init =kcat[E]) is determined by the actual number of catalytic
units that are participating in the formation of catalytic sites.
The “overall” kcat value is lowered by the presence of isolat-
ed catalytic units unable to form a dimeric catalytic site.
Furthermore, an alteration of the catalytic site as a function
of valency (distance between the catalytic units, polarity,
4884
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4879 – 4889

Catalytic Self-Assembled Monolayers
FULL PAPER
sidering the nearest-neighbor
recognition in vesicular aggre-
gates.[56,57] This difference arises
from the fact that, contrary to
Regensꢂ system, here the
heteroACHTNUTRGNEUNGdimeric site 1–2 cannot
bind a substrate molecule. At
lower surface loadings, isolated
catalytic units 1 are present,
which are unable to participate
in the formation of a catalytic
site 1–1. Evidently, in the case
of cluster formation, isolated
units of 1 are never present and
a maximum value for kcat is al-
ready reached at x1 ꢀ0.15 (Fig-
ure 7c). Below this value, de-
creased kcat values result from
the presence of small clusters
with an odd number of units 1
(e.g., 3 or 5), which implies that
at saturation some units do not
participate in catalysis.
Figure 6. Simulation of catalytic units 1 (grey areas) on a truncated icosahedron for x1 =0.5: a) random distri-
bution or b) homodomain formation. Solid lines indicate potential catalytic sites (dimers 1–1) for the first sub-
strate. Solid lines terminated with dots indicate catalytic sites at saturation.
Summarizing, Figure 7b and
7c are indicative of the intrinsic
behavior of catalytic SAMs
with regard to the apparent Mi-
*
Figure 7 ( ) and represents an extreme case of clustering.
Together, both profiles represent the boundaries for cluster-
ing.
chaelis–Menten parameters kcat and KM and form the refer-
ence framework against which to analyze the experimental
data. This implies that identical experimental profiles do not
require any chemical explanation, but can be attributed
merely to statistics. Alternatively, a deviation from the an-
ticipated intrinsic behavior implies that the Michaelis–
Menten parameters kcat and KM for the individual catalytic
sites are not constant. This indicates a change in the catalyt-
ic site or local environment as a function of the surface cov-
erage.
The dissociation constant KM: In case of a statistical distribu-
tion of 1, the concentration of dimeric sites 1–1 increases
strictly linearly as a function of the mole fraction x1. This
implies that the same number of units of 1 is able to gener-
ate more potential catalytic sites 1–1 as the surface loading
increases. If 1 has a tendency to cluster, the amount of cata-
lytic sites on the surface is much less dependent on the mole
fraction x1. The concentration of 1–1 rapidly increases up to
x1 ꢀ0.2, after which it slowly increases to the final value.
The lower concentration of 1–1 compared to the homomeric
SAMs (x1 =1) results from units of 1 localized on the border
of the clusters. In fact, clustering resulting from self-sorting
during monolayer formation results in a concentration of 1–
1 independent from x1. In all cases, the higher apparent con-
centration of catalytic sites will result in an apparent in-
crease in binding affinity (lower KM values). The anticipated
correlation between KM and mole fraction x1 for the three
situations is given in Figure 7b (which is the inverse of the
concentration of [1–1] normalized on 1 for x1 =1).
Analysis of the experimental data: Experimentally we have
varied the mole fraction of the catalytic unit TACN·ZnII (x1)
in the system in two different ways. In experiment series A,
Au MPCs were prepared with mixed monolayers composed
of thiols 1 and 2 in various ratios. In series B, increasing
amounts of TEG-functionalized nanoparticle VIII were
added to a constant amount of TACN-functionalized nano-
particle I in the presence of one equivalent of ZnII per
TACN. We were interested in the following key questions:
1) Are the experimentally observed correlations between
the kcat and KM values in the mixed monolayer system an
intrinsic consequence of the multivalency of the system
or can they be ascribed to a “true” dendritic effect (im-
plying actual improvements in the catalytic site)?
2) Do the correlations provide information on the structural
order of the mixed monolayers in terms of clustering or
statistical distribution?
The catalytic constant kcat: The number of substrate mole-
cules bound at saturation is directly correlated to kcat. For a
random distribution of 1, kcat reaches a maximum at x1 ꢀ0.4
(Figure 7c). Thus, kcat does not increase linearly with x1 as
would be expected in a non-confined environment. This sit-
uation is different from that reported by Regen when con-
Chem. Eur. J. 2011, 17, 4879 – 4889
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4885

L. J. Prins. P. Scrimin et al.
based on the theoretical analysis. It illustrates that thiol 2 is
an inert thiol with regard to catalysis. Furthermore, it illus-
trates that TEG-based monolayers do not have any affinity
for the substrate. If nonspecific binding of HPNPP to TEG-
monolayers had taken place, this would have resulted in an
increase of the apparent KM value (indicating weaker bind-
ing) upon lowering x1, since higher HPNPP concentrations
would be required to saturate the catalyst.
Mixed monolayers composed of thiols 1 and 2 (series A):
Analysis of the obtained kcat values as a function of the
mole fraction of 1 shows a nearly perfect linear increase up
to x1 =0.4 and an intersect at origin when extrapolated to
&
x1 =0 (Figure 5a: ). In our experimental set up, a mole frac-
tion of x1 =0 corresponds to the dilution of a constant
amount of 1 (20 mm) in an infinite amount of thiol 2. In fact,
the intersect close to 0 confirms that isolated TACN·ZnII
units surrounded by TEG have a negligible catalytic activity.
The KM values decrease exponentially as a function of the
&
mole fraction x1 (Figure 5b: ). From these observations, and
taking into account the theoretical analysis, we can draw
two conclusions for this mixed thiol system. First, clustering
of thiol 1 does not appear to occur. In case of domain for-
mation both parameters kcat and KM would have been practi-
cally constant already at very low x1 values as confirmed by
experiment series B. Rather, both the behavior of kcat and
KM are fully compatible with the expected behavior for a
random distribution of thiols. This is supported by the corre-
spondence of the experimental data points representing the
dependence of kcat and KM on x1 with the simulated behavior
for a random distribution of thiols (solid lines of Figure 5a
and 5b). Importantly, independent proof for the statistical
distribution of thiols 1 and 2 was very recently obtained
from a study based on the quenching of anionic fluorescence
probes bound to the Au MPC surface.[58] The results of
series A also leads towards the second conclusion. Although
it appears on first sight that the catalytic sites on the surface
Figure 7. Simulated correlations between a) the concentration of dimeric
sites 1–1, b) KM and c) kcat as a function of the mole fraction x1. The solid
change as a function of the mole fraction x (increase kcat
;
1
&
line marked with ( ) represents the correlation in case 1 is randomly dis-
decrease KM), the theoretical analysis shows that this is an
intrinsic behavior originating from the multivalency of the
catalytic system. There is no reason to offer chemical ex-
planations for these changes, which is similar to what we
had observed for a related dendrimer study.[42] Importantly,
it shows the intrinsic advantage of using a multivalent scaf-
fold.
tributed on the surface. The dashed line marked with (&) represent ho-
modomain formation of 1 after monolayer formation. The dashed line
*
marked with ( ) represent self-sorting of thiols on different nanoparticles
during monolayer formation.
3) What is the origin of the much higher catalytic activity of
these nanoparticles compared to related multivalent sys-
tems based on the identical TACN·ZnII complex?
Evaluation of catalytic SAMs on NPs: Previously, we have
extensively studied a wide variety of transphosphorylation
catalysts based on the same TACN·ZnII catalytic unit
(Figure 8). In all cases we have shown that catalysis requires
the cooperative action of two catalytic units. For systems
3[42] and 4,[59] we were able to determine the Michaelis–
Menten parameters, which are added in Table 2. By them-
selves these are reasonable catalysts in terms of rate acceler-
ation (kcat/kuncat), but their activity becomes insignificant
when compared to the NP-based system. The gain of the NP
system lies mainly in a strongly increased kcat value (>13
The results of experiment series B represent what would
be observed in case the two thiols 1 and 2 would perfectly
separate in two phases on the nanoparticle surface, that is,
addition of the two thiols to the amine-stabilized nanoparti-
cles results in the formation of exclusively homofunctional-
ized nanoparticles. From the observed trend for this series it
appears that the catalytic properties, both in terms of kcat
and KM, of the catalytic unit TACN·ZnII are rather unaffect-
*
ed by the presence of the inert thiol 2 (Figure 5: ). This cor-
responds to the expectation based on intuition, and also,
4886
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4879 – 4889

Catalytic Self-Assembled Monolayers
FULL PAPER
Normalized for the number of metal centers, these systems
are among the best catalysts known for the transphosphory-
lation of HPNPP exhibiting kcat/KM values of more than
20 LmolÀ1 sÀ1. This is quite remarkable, considering the
structural simplicity of the TACN·ZnII catalytic unit. The
fact that the same complex in other (multivalent) systems is
much less active, indicates that the use of nanoparticles as
scaffolds carries particular advantages. The reduced freedom
of the complexes and the desolvation of the reacting nucleo-
phile could be possible explanations,[45,60] but the forced
close packing of the TACN·ZnII head groups is very likely
to play an important role. The theoretical analysis showed
the intrinsic advantage of the multivalent system in generat-
ing catalytic sites composed of two TACN·ZnII complexes.
The clustering of multiple catalytic units on the NP surface
creates a large number of potential binding sites for the sub-
strate, which increases as a function of the surface catalyst
loading. Consequently, this gives the impression that the af-
finity of the substrate for the catalyst increases as the sur-
face loading goes up (exponential decrease in KM). Like-
wise, the apparent catalytic constant kcat is smaller for low
catalyst loadings, because individual residues cannot partici-
pate in the formation of catalytic sites resulting in a lower
activity at saturation. Our experimental data indicate an in-
crease in the apparent kcat up till a molar fraction of ꢀ0.4
after which it remains practically constant. The apparent KM
value on the other hand shows an exponential decay as a
function of the surface loading. The result is a continuous
increase in kcat/KM as a function of the catalyst loading.
Without the theoretical analysis, this would have been as-
cribed to an improvement in the catalytic site or to other
reasons of chemical origin. However, the profiles perfectly
match those that are intrinsic to the multivalent system,
which implies that the catalytic site is more or less the same
over the range of mole fractions studied. The experimentally
observed trends for kcat and KM also suggest that the two
thiols (1 and 2) are more or less randomly distributed on
the NP surface. Clustering would have resulted in nearly
constant values for these parameters over the whole range
studied or, as a minimum, would have resulted in a much
faster arrival at a constant kcat value (before x1 ꢀ0.2). This
was confirmed experimentally by performing a mixing ex-
periment in which the catalytic activity of different ratios of
homofunctionalized nanoparticles was measured. This obser-
vation is different from that observed for other systems,[30]
but surface ordering is presumably strongly dependent on
the chemical nature of the thiols.
Figure 8. Structures of selected catalysts for the transesterification of
HPNPP. The Michaelis–Menten parameters (corrected for the amount of
ZnII ions) are given in Table 2.
times compared to 3) and to a lesser extent on an increased
binding affinity (ꢀ4 times compared to 3). In a previous
study,[42] we had already discussed that lysine-based den-
drimers are not optimal scaffolds for surface modifications,
because of the asymmetric branching units. Nanoparticles
are on the opposite side of the spectrum, inducing an excel-
lent alignment of the catalytic units. It is worth considering
that the best dimetallic catalysts known for the transphos-
phorylation of HPNPP are typically equipped with an
alkoxy function in the spacer that coordinates to both metal
centers and brings them in close proximity (for example
5,[52] Figure 8). This makes us postulate that the origin for
the enhanced catalytic activity in the NPs is the close prox-
imity of the ZnII ions induced by the packing of the mono-
layer. In the monolayer all thiols are aligned and the head
groups find themselves in close proximity with a relative low
mobility. Apparently the spatial orientation of two neighbor-
ing catalytic units creates a perfect catalytic site without for-
mation of catalytically inactive m-hydroxo species.
In summary, this first detailed investigation of catalytic
SAMs has given new insights on the intrinsic catalytic be-
havior that can be expected for these multivalent systems. It
clearly evidences that experimental data on catalysis should
be analysed properly, before claiming a real “dendritic”
effect, that is, an increase of the efficiency of the catalyst
with its valency. Additionally, it has allowed us to indirectly
obtain information on the structural order within the SAM.
Based on these studies we now intend to develop a second
generation of catalysts in which the bystander thiol will
Conclusion
We have studied the catalytic activity of a series of Au
MPCs as a function of the catalyst loading on the surface.
Chem. Eur. J. 2011, 17, 4879 – 4889
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4887

L. J. Prins. P. Scrimin et al.
[2] A. Thomas, M. Driess, Angew. Chem. 2009, 121, 1922–1924; Angew.
Chem. Int. Ed. 2009, 48, 1890–1892.
[3] F. M. de Rege, D. K. Morita, K. C. Ott, W. Tumas, R. D. Broene,
Chem. Commun. 2000, 1797–1798.
tailor the properties of the catalytic units. Eventually, this
should lead to truly engineered catalytic nanoarchitectures.
[4] C. Bianchini, P. Barbaro, V. Dal Santo, R. Gobetto, A. Meli, W.
Oberhauser, R. Psaro, F. Vizza, Adv. Synth. Catal. 2001, 343, 41–45.
[5] R. Margalef-Catalꢃ, C. Claver, P. Salagre, E. Fernꢄndez, Tetrahe-
dron: Asymmetry 2000, 11, 1469–1476.
[6] M. J. Alcꢅn, A. Corma, M. Iglesias, F. Sꢄnchez, J. Mol. Catal. A
1999, 144, 337–346.
[7] P. Barbaro, F. Liguori, Chem. Rev. 2009, 109, 515–529.
[8] I. Yosef, R. Abu-Reziq, D. Avnir, J. Am. Chem. Soc. 2008, 130,
11880–11882.
[9] M. Bartz, J. Kuther, R. Seshadri, W. Tremel, Angew. Chem. 1998,
110, 2646–2649; Angew. Chem. Int. Ed. 1998, 37, 2466–2468.
[10] M. C. Daniel, D. Astruc, Chem. Rev. 2004, 104, 293–346.
[11] S. Roy, M. A. Pericas, Org. Biomol. Chem. 2009, 7, 2669–2677.
[12] T. Belser, M. Stohr, A. Pfaltz, J. Am. Chem. Soc. 2005, 127, 8720–
8731.
[13] H. M. R. Gardimalla, D. Mandal, P. D. Stevens, M. Yen, Y. Gao,
Chem. Commun. 2005, 4432–4434.
[14] A. G. Hu, G. T. Yee, W. B. Lin, J. Am. Chem. Soc. 2005, 127, 12486–
12487.
[15] F. Manea, F. B. Houillon, L. Pasquato, P. Scrimin, Angew. Chem.
2004, 116, 6291–6295; Angew. Chem. Int. Ed. 2004, 43, 6165–6169.
[16] P. Pengo, S. Polizzi, L. Pasquato, P. Scrimin, J. Am. Chem. Soc. 2005,
127, 1616–1617.
[17] R. Bonomi, F. Selvestrel, V. Lombardo, C. Sissi, S. Polizzi, F.
Mancin, U. Tonellato, P. Scrimin, J. Am. Chem. Soc. 2008, 130,
15744.
[18] T. Belser, E. N. Jacobsen, Adv. Synth. Catal. 2008, 350, 967–971.
[19] P. Scrimin, G. Ghirlanda, P. Tecilla, R. A. Moss, Langmuir 1996, 12,
6235–6241.
[20] P. Scrimin, P. Tecilla, U. Tonellato, C. A. Bunton, Colloids Surfaces
A. 1998, 144, 71–79.
[21] C. A. Bunton, P. Scrimin, P. Tecilla, J. Chem. Soc. Perkin Trans. 2
1996, 419–425.
[22] J. M. J. Frechet, Science 1994, 263, 1710–1715.
[23] L. Ropartz, R. E. Morris, D. F. Foster, D. J. Cole-Hamilton, Chem.
Commun. 2001, 361–362.
[24] A. Dahan, M. Portnoy, Org. Lett. 2003, 5, 1197–1200.
[25] Y. Ribourdouille, G. D. Engel, M. Richard-Plouet, L. H. Gade,
Chem. Commun. 2003, 1228–1229.
[26] E. Delort, T. Darbre, J. L. Reymond, J. Am. Chem. Soc. 2004, 126,
15642–15643.
[27] C. C. Paluti, E. S. Gawalt, J. Catal. 2009, 267, 105–113.
[28] A. M. Jackson, J. W. Myerson, F. Stellacci, Nat. Mater. 2004, 3, 330–
336.
[29] L. Duchesne, G. Wells, D. G. Fernig, S. A. Harris, R. Levy, Chem-
BioChem 2008, 9, 2127–2134.
Experimental Section
NMR spectra were recorded on a Bruker Avance 300 MHz spectrometer
operating at 301 K. UV/Vis spectra were recorded on a Perkin–Elmer
Lambda 16 spectrophotometer equipped with
holder. The kinetic measurements were performed on a TECAN-Infinite
F200 microplate reader equipped with a 405
a thermostatted cell
A
transparent flat-bottom 96-well plates. TEM images were recorded on a
Jeol 300 PX electron microscope. The synthesis of compounds 1[32] and
2[39] has been reported previously.
Nanoparticle synthesis: HAuCl4·3H2O (102 mg, 0.26 mmol), weighed in a
dry-box, was dissolved in H2O (mQ; 7 mL). Separately, a solution of
TOABr (5.50 g, 3 mmol) in degassed toluene (250 mL) was prepared.
The aqueous solution of AuIII was extracted with the TOABr-solution
(3ꢀ15 mL) causing a transfer of the gold ions to the organic phase (red
color). Di-n-octylamine (4.3 mL, 14 mmol) was then added to the remain-
ing amount of the TOABr solution, and the resulting mixture was vigo-
rously stirred with the AuIII-containing solution for 30 min under N2 re-
sulting in a complete decoloration of the solution. Subsequently, NaBH4
(46 mg, 1.2 mmol) dissolved in H2O (2 mL) was added under vigorous
stirring resulting in the formation of the Au nanoparticles (brown color-
ing). The solution was stirred for an additional 3 h under N2, after which
the aqueous phase was removed. The resulting nanoparticle solution was
stored under an inert atmosphere over night. The nanoparticle solution
was divided in eight batches and thiols 1 and 2 were added in different
ratios from stock solutions in DMF. The two thiols were mixed before
being added to the nanoparticles under vigorous stirring. Within 15 min
the solution became colorless and a brown precipitate was formed. After
5 min H2O (mQ; 5 mL) was added and the solution was stirred for an
hour under N2. Next, the aqueous phase was separated and washed with
diethyl ether (2ꢀ10 mL), toluene (4ꢀ10 mL), ethylacetate (3ꢀ10 mL),
and again diethyl ether (2ꢀ10 mL). Finally, water was removed by evap-
oration under reduced pressure and the Au MPCs were obtained as
brown solids.
Spectrophotometric titrations with CuII: Two cuvettes were prepared con-
taining a 0.1m solution of MES (80 mL) at pH 6.5, the stock solution of
nanoparticles (48 mL) with unknown concentration and H2O (mQ;
672 mL). The titrating solution contained CuCl2 (0.6 mm) and MES
(10 mm) in H2O (mQ). An additional solution containing only MES
(10 mm) in H2O (mQ) was used to correct for the dilution of the refer-
ence cuvette. The absorbance was measured at 264 nm after adding the
CuII-containing solution to one cuvette and the buffer solution to the ref-
erence cuvette.
Kinetic measurements: Kinetics were measured at 408C in aqueous solu-
tion, buffered at pH 7,5 with HEPES 0.01m. Reaction mixtures were pre-
pared by adding in order the solution of nanoparticles, a solution of Zn-
[30] C. Gentilini, P. Franchi, E. Mileo, S. Polizzi, M. Lucarini, L. Pasqua-
to, Angew. Chem. 2009, 121, 3106–3110; Angew. Chem. Int. Ed.
2009, 48, 3060–3064.
[31] C. Gentilini, L. Pasquato, J. Mater. Chem. 2010, 20, 1403–1412.
[32] F. Manea, C. Bindoli, S. Polizzi, L. Lay, P. Scrimin, Langmuir 2008,
24, 4120–4124.
ACHTUNGTRENNUNG(NO3)2, and a solution of HPNPP in water. The cleavage of the phospho-
diester was monitored by measuring the absorbance of p-nitrophenolate
at 405 nm. Initial velocities were calculated by a linear fitting of the ini-
tial part of the kinetics (conversion <10%).
[33] N. R. Jana, X. G. Peng, J. Am. Chem. Soc. 2003, 125, 14280–14281.
[34] A. C. Templeton, M. J. Hostetler, E. K. Warmoth, S. W. Chen, C. M.
Hartshorn, V. M. Krishnamurthy, M. D. E. Forbes, R. W. Murray, J.
Am. Chem. Soc. 1998, 120, 4845–4849.
[35] M. J. Hostetler, A. C. Templeton, R. W. Murray, Langmuir 1999, 15,
3782–3789.
Acknowledgements
Financial support from the European Research Council under the Euro-
pean Communityꢂs Seventh Framework Programme (FP7/2007-2013)/
ERC Starting Grant agreement n8 239898 is acknowledged.
[36] L. Pasquato, F. Rancan, P. Scrimin, F. Mancin, C. Frigeri, Chem.
Commun. 2000, 2253–2254.
[37] R. Shenhar, V. M. Rotello, Acc. Chem. Res. 2003, 36, 549–561.
[38] R. Klajn, K. J. M. Bishop, B. A. Grzybowski, Proc. Natl. Acad. Sci.
USA 2007, 104, 10305–10309.
[1] R. A. Sheldon, I. Arends, U. Hanefeld, Green Chemistry and Cataly-
sis, 1st ed., Wiley-VCH, Weinheim, 2007.
[39] P. Pengo, S. Polizzi, M. Battagliarin, L. Pasquato, P. Scrimin, J.
Mater. Chem. 2003, 13, 2471–2478.
4888
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4879 – 4889

Catalytic Self-Assembled Monolayers
FULL PAPER
[40] A. Scarso, G. Zaupa, F. B. Houillon, L. J. Prins, P. Scrimin, J. Org.
Chem. 2007, 72, 376–385.
[53] F. Avenier, J. B. Domingos, L. D. Van Vliet, F. Hollfelder, J. Am.
Chem. Soc. 2007, 129, 7611–7619.
[41] M. Martin, F. Manea, R. Fiammengo, L. J. Prins, L. Pasquato, P.
Scrimin, J. Am. Chem. Soc. 2007, 129, 6982–6983.
[54] F. Mancin, P. Scrimin, P. Tecilla, U. Tonellato, Chem. Commun.
2005, 2540–2548.
[42] G. Zaupa, P. Scrimin, L. J. Prins, J. Am. Chem. Soc. 2008, 130, 5699–
5709.
[43] G. Zaupa, L. J. Prins, P. Scrimin, Bioorg. Med. Chem. Lett. 2009, 19,
3816–3820.
[44] T. Darbre, J. L. Reymond, Acc. Chem. Res. 2006, 39, 925–934.
[45] P. Pengo, L. Baltzer, L. Pasquato, P. Scrimin, Angew. Chem. 2007,
119, 404–408; Angew. Chem. Int. Ed. 2007, 46, 400–404.
[46] R. Breinbauer, E. N. Jacobsen, Angew. Chem. 2000, 112, 3750–3753;
Angew. Chem. Int. Ed. 2000, 39, 3604–3607.
[55] Simulations using a dodecahedron (12 faces) or icosahedron (20
faces) gave essentially identical results (data not shown). This indi-
cates that, at least from a theoretical point of view, the polydispers-
ity of the nanoparticles is less relevant. This originates from the fact
that the catalytic site is composed of only 12 units, which is far less
than the total number of thiols present in the monolayer. Experi-
mentally, polydispersity may cause a different spacing between two
catalytic units, resulting in different efficiency. However, the similar-
ity of the catalytic parameters of the nanoparticles presented here
[47] D. H. Wu, A. D. Chen, C. S. Johnson, J. Magn. Reson. Ser. A 1995,
115, 260–264.
((1.6Æ0.2) nm) and those of
a
previous study ((2.5Æ0.7) nm,
ref. [15]) indicate that polydispersity does not appear to play an im-
portant role. For that reason, and also considering the relative
narrow size distribution of the nanoparticles here, polydispersity was
not taken into account in the analysis.
[48] The use of large gradient pulses in diffusion-ordered 2D NMR spec-
troscopy (DOSY) results in the induction of eddy currents in the
metal structures of the magnet and probe, which causes a distortion
in the high-resolution spectra. The LEDbp sequence reduces that
problem by using bipolar pulse pairs, since the eddy currents caused
by two gradient pulses with identical areas, but different polarities
tend to cancel.
[56] S. L. Regen, Curr. Opin. Chem. Biol. 2002, 6, 729–735.
[57] S. M. K. Davidson, S. L. Regen, Chem. Rev. 1997, 97, 1269–1279.
[58] R. Bonomi, A. Cazzolaro, L. J. Prins, Chem. Commun. 2011, 47,
445–447.
[49] D. M. Brown, D. A. Usher, J. Chem. Soc. 1965, 6558–6564.
[50] It should be noted that the initial rate obtained for [ZnII]=0 is an
experimental point.
[59] A. Scarso, U. Scheffer, M. Gobel, Q. B. Broxterman, B. Kaptein, F.
Formaggio, C. Toniolo, P. Scrimin, Proc. Natl. Acad. Sci. USA 2002,
99, 5144–5149.
[51] Under the experimental conditions thiol exchange between the
nanoparticle is much slower than the kinetic process studied; see:
M. Zachary, V. Chechik, Angew. Chem. 2007, 119, 3368–3371;
Angew. Chem. Int. Ed. 2007, 46, 3304–3307.
[60] F. Avenier, F. Hollfelder, Chem. Eur. J. 2009, 15, 12371–12380.
[52] G. Q. Feng, D. Natale, R. Prabaharan, J. C. Mareque-Rivas, N. H.
Williams, Angew. Chem. 2006, 118, 7214–7217; Angew. Chem. Int.
Ed. 2006, 45, 7056–7059.
Received: September 7, 2010
Revised: November 22, 2010
Published online: March 14, 2011
Chem. Eur. J. 2011, 17, 4879 – 4889
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4889