Layer-by-Layer Assemblies of Catechol-Functionalized TiO2 Nanoparticles and Porphyrins through Electrostatic Interactions

Article abstract of DOI:10.1002/chem.201405039

In the current work, we present the successful functionalization and stabilization of P-25 TiO₂ nanoparticles by means of N1,N7-bis(3-(4-tert-butyl-pyridium-methyl)phenyl)-4-(3-(3-(4-tert-butyl-pyridinium-methyl)phenylamino)-3-oxopropyl)-4-(3,4

Full text of DOI:10.1002/chem.201405039

DOI: 10.1002/chem.201405039
Full Paper
&
Solar Energy Conversion
Layer-by-Layer Assemblies of Catechol-Functionalized TiO2
Nanoparticles and Porphyrins through Electrostatic Interactions
Alexandra Burger,[a] Rubꢀn D. Costa,[b] Volodymyr Lobaz,[c] Wolfgang Peukert,[c]
Dirk M. Guldi,[b] and Andreas Hirsch*[a]
Abstract: In the current work, we present the successful
functionalization and stabilization of P-25 TiO2 nanoparticles
by means of N1,N7-bis(3-(4-tert-butyl-pyridium-methyl)phen-
yl)-4-(3-(3-(4-tert-butyl-pyridinium-methyl)phenylamino)-3-ox-
opropyl)-4-(3,4-dihydroxybenzamido)heptanediamide tribro-
mide (1). The design of the latter is aimed at nanoparticle
functionalization and stabilization with organic building
blocks. On one hand, 1 features a catechol anchor to enable
its covalent grafting onto the TiO2 surface, and on the other
hand, positively charged pyridine groups at its periphery to
prevent TiO2 agglomeration through electrostatic repulsion.
The success of functionalization and stabilization was corro-
borated by thermogravimetric analysis, dynamic light-scat-
tering, and zeta potential measurements. As a complement
to this, the formation of layer-by-layer assemblies, which are
governed by electrostatic interactions, by alternate deposi-
tion of functionalized TiO2 nanoparticles and two negatively
charged porphyrin derivatives, that is, 5,10,15,20-(phenoxy-
acetic acid)-porphyrin (2) and 5,10,15,20-(4-(2-ethoxycarbon-
yl)-4-(2-phenoxyacetamido)heptanedioic acid)-porphyrin (3),
is documented. To this end, the layer-by-layer deposition is
monitored by UV/Vis spectroscopy, scanning electron mi-
croscopy, ellipsometry, and profilometry techniques. The re-
sulting assemblies are utilized for the construction and test-
ing of novel solar cells. From stable and repeatable photo-
currents generated during several “on-off” cycles of illumina-
tion, we derive monochromatic incident photo-to-current
conversion efficiencies of around 3%.
Introduction
multilayer assemblies are realized by employing electrostatic
attractions between oppositely charged building blocks. The
LbL technique was first used by Decher et al. to build up multi-
layers of charged polymers by alternate dipping of substrates
into solutions of either positively or negatively charged poly-
mers. Repeatedly performed cycles of this procedure lead to
multiple polymer layers, which are stabilized by electrostatic
interactions.[7] In light of the latter, the LbL technique consti-
tutes a very simple, but efficient method for the consecutive
deposition of individual molecular layers onto planar solid sup-
ports to build up multilayer architectures.
The supramolecular construction of hierarchically ordered, lay-
ered architectures is currently an emerging field. Most notably,
it opens up facile access to new functional hybrid devices with
potential applications in, for example, solar energy conver-
sion,[1] drug delivery,[2] electrochromic films,[3] membranes,[4] or
hydrogen generation.[5] In particular, the development of lay-
ered organic-hybrid components represents a tremendous
challenge, because the characteristics of both classes of com-
pounds may be combined synergetically, giving more than the
simple sum of the individual components.
In recent years, a myriad of functional architectures with
unique mechanical, optical, electrical, and biological properties
have been designed in this way.[4,6b,8] In general, either poly-
electrolytes[9] or charged organic molecules[10] have been used
as components for the construction of LbL multilayers. Com-
bined LbL assemblies of organic electrolytes with charged inor-
ganic nanoparticles (NPs) have also been constructed.[2b,11] For
example, Pichon et al. developed magneto-tunable hybrid
films of iron oxide nanoparticles and alternating PAH/PSS
layers,[12] Kotov et al. reported polyelectrolyte/semiconductor
nanoparticle composites,[13] and Schmitt et al. focused on poly-
electrolyte/Au nanoparticle composites.[14] Examples of LbL as-
semblies of NPs and graphene oxide have also been report-
ed,[15] but LbL assemblies of organic molecules and nanoparti-
cles have so far been elusive. This is where our current work
makes a significant contribution.
One possible method for building up layered architectures
relies on the layer-by-layer (LbL) technique,[6] through which
[a] A. Burger, Prof. Dr. A. Hirsch
Department of Chemistry and Pharmacy
Friedrich-Alexander Universitꢀt Erlangen-Nꢁrnberg
Institute of Organic Chemistry
Henkestrasse 42, 91054 Erlangen (Germany)
E-mail: andreas.hirsch@fau.de
[b] Dr. R. D. Costa, Prof. Dr. D. M. Guldi
Department of Chemistry and Pharmacy
Friedrich-Alexander Universitꢀt Erlangen-Nꢁrnberg
Interdisciplinary Center for Molecular Materials
Egerlandstrasse 3, 91058 Erlangen (Germany)
[c] Dr. V. Lobaz, Prof. Dr. W. Peukert
Friedrich-Alexander Universitꢀt Erlangen-Nꢁrnberg
Institute of Particle Technology
Cauerstrasse 4, 91058 Erlangen (Germany)
Chem. Eur. J. 2015, 21, 1 – 15
1
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
One of the most prominent inorganic building blocks for the
construction of hybrid materials are TiO2 NPs, which exhibit
a variety of favorable chemical and physical properties as well
as chemical inertness, negligible solubility in water or protic
solvents, and pronounced pH tolerance.[16] TiO2 is a wide-
bandgap semiconductor material, which has been used as
a photocatalyst for the degradation of pollutants, for photoca-
talytic oxidation, and for water splitting. Most importantly, it
represents the inorganic semiconductor component in dye-
sensitized solar cells.[17] Porphyrins are widely used as organic
sensitizers in the latter.[18] TiO2/porphyrin composites range
from porphyrins bound to TiO2 through one single anchor
group, to anionic porphyrins adsorbed on porous TiO2 electro-
des through multiple anchor groups.[19] Additionally, metallo-
porphyrins can be coordinated to TiO2 through axial metal–
ligand coordination to surface-bound isonicotinic acid and its
derivatives.[20] Moreover, porphyrins have been assembled to
TiO2 by using the Langmuir–Blodgett technique,[21] as well as
through the construction of metal-organic frameworks.[22] Our
work describes the formation of TiO2 NPs/porphyrin compo-
sites built up through electrostatic interactions, using the LbL
technique.
We also report for the first time the functionalization of P25-
TiO2 NPs with a positively charged catechol-based building
block, that is, N1,N7-bis(3-(4-tert-butyl-pyridium-methyl)phen-
yl)-4-(3-(3-(4-tert-butyl-pyridinium-methyl)phenylamino)-3-oxo-
propyl)-4-(3,4-dihydroxybenzamido)heptanediamide tribromide
(1). This paves the way for the formation of stable NPs
(Scheme 1a). The functionalization of NPs with a single func-
tional building block such as 1 allows the control of the
number of charges introduced onto the surface, as well as the
thickness of the organic layer on the particle surface. The latter
features are in contrast to those of previously reported NPs,
which are functionalized with just polymers. The dendritic
branching in 1 places steric hindrance as well as three positive
charges, formed by quaternized pyridines, onto the periphery
of the NPs.
As a complement to the functionalized, positively charged
TiO2 NPs, negatively charged porphyrins have been synthe-
sized. With porphyrins featuring characteristic absorption
bands, they can function as ideal reporters in the successive
analysis of the LbL assembly. In this regard, we focused on the
synthesis of two porphyrins, namely 5,10,15,20-(phenoxyacetic
acid)-porphyrin (6) and 5,10,15,20-(4-(2-ethoxycarbonyl)-4-(2-
phenoxyacetamido)heptanedioic acid)-porphyrin (8). Both are
free-base tetraphenyl porphyrin derivatives, which differ in the
number of carboxylic acids at the phenyl rings: 4 versus 12 in
6 and 8, respectively. These oligo-carboxylic acid porphyrins
are transformed easily into the corresponding oligo-carboxy-
lates after deprotonation (see Scheme 1b, porphyrins 2 and 3,
respectively). The formation of multilayer architectures was
demonstrated by using positively charged TiO2 NPs and the
negatively charged porphyrins 2 and 3. (Scheme 1c).
With NPs featuring different, size-dependent properties com-
pared with those of bulk materials, one setback is their strong
tendency to form agglomerates, which has to be overcome to
preserve their unique properties. To this end, steric and/or
electrostatic repulsions through chemical surface functionaliza-
tion constitute a powerful approach.[8b,23] Chemical surface
functionalization of NPs with, for example, organic building
blocks through chemisorption has a number of conceptual ad-
vantages over the physical immobilization of molecular build-
ing blocks onto the surface of NPs by means of electrostatic-
and/or dipole–dipole interaction driven physisorption.[24] Nota-
bly, depending on the stabilizers, agglomeration processes, Results and Discussion
and the solubility of the NPs are controlled, and their electron-
Synthesis and characterization of 5,10,15,20-(phenoxyacetic
acid)-porphyrin (6) and 5,10,15,20-(4-(2-ethoxycarbonyl)-4-
(2-phenoxyacetamido)heptanedioic acid-)porphyrin (8)
ic, magnetic, and optical properties are tuned as they are at-
tached covalently to the surface.[25] In terms of organic stabiliz-
ers, polymers, oligomers, and also small molecules have been
probed, which introduce either steric or electrostatic repulsions
or a combination of both.[7,26]
For the development of oligo-carboxylated porphyrins, which
can be transformed easily into oligo-anions after deprotona-
tion, we focused on the synthesis of two systems, namely
5,10,15,20-(phenoxyacetic acid)-porphyrin (6) and 5,10,15,20-(4-
(2-ethoxycarbonyl)-4-(2-phenoxyacetamido)heptanedioic acid)-
porphyrin (8) (Scheme 2). Both are free-base tetraphenylpor-
phyrin derivatives, which differ in the number of carboxylic
groups connected to the phenyl rings. Compound 6 contains
four, whereas 8 contains twelve carboxylic acid groups. In
basic media, the carboxylic groups are present in their depro-
tonated form, affording carboxylates 2 and 3. Hence, they can
be considered as excellent building blocks for LbL assemblies
through electrostatic interactions, as shown in the next sec-
tion. The synthesis of 6 starts from tert-butyl-2-(4-formylphe-
noxy)acetate (4) and pyrrole (Scheme 2). Benzaldehyde 4 was
synthesized from 4-hydroxybenzaldeyde and tert-butyl 2-(4-for-
mylphenoxy)acetate according to a procedure developed by
McKillop et al.[29] Under modified Lindsey conditions,[30] the
10,15,20-(tert-butyl-2-phenoxyacetato)-porphyrin 5 is obtained
For the covalent functionalization of metal oxide NPs, cate-
chol groups have emerged recently as a very potent anchor,
yielding highly stable and soluble coupling products.[27] Never-
theless, the preparation of catechol derivatives bearing
charged groups to enable the functionalization and stabiliza-
tion of NPs in solutions has barely been explored. A notable
exception is our recent report on the synthesis of cationic den-
drons equipped with catechol anchors.[28] Here, besides the
demonstration of the successful functionalization of ZnO NPs,
photoelectrochemical cells were fabricated, in which mesopo-
rous ZnO films were functionalized with these cationic den-
drons and subsequently sensitized with an anionic porphyrin.
In the current work, we demonstrate not only the formation
of double-layer architectures, but also the construction of mul-
tiple-layer architectures governed by electrostatic interactions
between catechol-functionalized P25-TiO2 NPs and anionic por-
phyrins through the LbL technique.
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
2
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 1. a) Schematic representation of P25-TiO2 NPs functionalized with 1. b) Negatively charged porphyrins 2 and 3. c) Schematic representation of the
corresponding LbL assembly process. d) Representation of the resulting multilayer architectures.
as a purple powder. Subsequent treatment with formic acid at
258C leads quantitatively to the desired 5,10,15,20-(phenoxy-
acetic acid)-porphyrin 6. Under modified Steglich conditions,[31]
6 was coupled with the first-generation Newkome-Dendri-
mer,[32] DCC, and HOBT in DMF to yield 5,10,15,20-(di-tert-butyl-
4-(3-(tert-butoxy)-3-oxopropyl)-4-(2-phenoxy-acetamido)hepta-
nedionato)-porphyrin 7 in 79%. Further treatment of 7 with
formic acid at 258C afforded quantitatively the second target
porphyrin, 5,10,15,20-(4-(2-ethoxycarbonyl)-4-(2-phenoxyaceta-
mido)heptanedioic acid)-porphyrin 8. All products and inter-
mediates were characterized by 1H-, and 13C NMR as well as
HRMS spectroscopy (for details see the Experimental Section).
Owing to the large number of carboxylic acid groups, 6 and
8 are more or less insoluble in organic solvents and only
poorly soluble in water or ethanol at neutral pH. To circumvent
this, basic media of pH > 7 were utilized to guarantee com-
plete deprotonation of all carboxylic acid groups. Here, aque-
ous TRIS buffer solutions represent appropriate solvents.
In Figure 1, the absorption and fluorescence spectra of
2.1 mm tris(hydroxymethyl) aminomethane (TRIS) buffered solu-
tions (pH 9) of 6 and 8 are presented.
The absorption spectra of 6 and 8 reveal sharp Soret band
absorptions at 424 nm and four Q-band absorptions at 520,
560, 599, and 641 nm. The shape of the Soret band is com-
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
3
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 2. Schematic representation of the synthesis of 5,10,15,20-(phenoxyacetic acid)-porphyrin (6) and 5,10,15,20-(4-(2-ethoxycarbonyl)-4-(2-phenoxyaceta-
mido)heptanedioic acid)-porphyrin (8).
Figure 1. Absorption (solid) and fluorescence (dashed) spectra of 2.1 mm TRIS buffer solutions (pH 9) of 6 (a) and 8 (b).
monly observed in symmetrical porphyrins bearing bulky
groups, which hamper aggregation in solution. The latter is
also reflected in the high molar extinction coefficients of 3.40ꢂ
105 and 3.17ꢂ105 LmolcmÀ1 for 6 and 8, respectively. This
points to weakly or non-aggregated species as a result of the
higher solubility promoted by the carboxylic acid groups. The
excited-state characteristics of 6 and 8 were determined by
fluorescence spectroscopy in TRIS buffer solutions. As illustrat-
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
4
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
ed in Figure 1, upon excitation at 550 nm, broad fluorescence
bands evolved, centered at 614 and 656 nm for 6 and 611 and
654 nm for 8. Compounds 6 and 8 feature fluorescence quan-
tum yields of 0.12 and 0.14, respectively.
Preparation and characterization of positively functionalized
TiO2 nanoparticles
The positively charged dendron 1,1’-((((4-(3-((3-((4-(tert-butyl)-
pyridin-1-ium-1-yl)methyl)phenyl)amino)-3-oxopropyl)-4-(3,4-di-
hydroxybenzamido)heptanedioyl)bis(azanediyl))bis(3,1-phenyl-
ene))bis(methylene))bis(4-(tert-butyl)pyridin-1-ium) (1) was pre-
pared following the procedure that we published previously.[28]
In the current work, 1 was used for the first time for the sur-
face functionalization of TiO2 NPs.
Figure 2. TiO2 NPs functionalized with different amounts of 1. From left to
right: 16.27ꢂ10À7 molmgÀ1 (T1), 4.07ꢂ10À7 molmgÀ1 (T2),
2.03ꢂ10À7 molmgÀ1 (T3), 1.02ꢂ10À7 molmgÀ1 (T4), 0.51ꢂ10À7 molmgÀ1
(T5), 0.25ꢂ10À7 molmgÀ1 (T6), and 0 molmgÀ1 (T7).
For the formation of stable suspensions of positively
charged nanoparticles (Scheme 3), P-25 TiO2 NPs were func-
change from white to light yellow sets in for the NP suspen-
sions. Moreover, the intensity of the color increases from T1 to
T6, whereas reference T7 remains white, as shown in Figure 2.
A color change of TiO2 upon mixing with catechol was already
observed by Moser et al.[33] In our case, no significant changes
in the absorption spectra of the functionalized TiO2 NPs were
detected.
Next, we compared the colloidal behavior of TiO2 NPs with
4.07ꢂ10À7 molmgÀ1 (T2) with the reference NP suspension
(T7) at a concentration of 29ꢂ10À3 mgmLÀ1 (Figure 3).
Although T7 and T2 form slightly turbid suspensions upon
sonication, NPs in reference T7 sediment after 24 h, whereas
the functionalized TiO2 nanoparticles T2 remain stable in sus-
pension (see Figure 3b). This is a first proof that agglomeration
of TiO2 NPs is suppressed and that stable suspensions are ob-
tained upon functionalization with 1.
Scheme 3. Schematic representation of the functionalization of TiO2 nano-
particles with 1.
tionalized with 1 by mixing an ethanolic suspension of the
nanoparticles with an ethanolic solution of 1, followed by soni-
cation for 50 min at room temperature (for more details see
Experimental Section).
Thermogravimetric analysis (TGA) was performed to gain in-
formation about the equilibrium adsorption of 1 on the TiO2
For the determination of the required amount of 1 to obtain
colloidal stable suspensions, different aliquots of 1 were mixed
with the same amount of TiO2 NPs, keeping the overall volume
of ethanol constant. The colloidal properties of pristine TiO2 NP
suspensions were used as a reference. In past works, we calcu-
lated the average surface area of 1 to be 2.32 nm2.[28] Nitrogen
adsorption experiments showed that the used P-25 TiO2 NPs
feature a surface area of 51.0 m2 gÀ1. As a consequence, a mon-
olayer formation of 1 on the NP surface should be obtained
upon binding of 0.37ꢂ10À7 molmgÀ1 of 1 to the surface. In
this study, ethanolic TiO2 NP suspensions containing different
concentrations of 1, within the range 0.25ꢂ10À7 molmgÀ1 (T6)
to 16.27ꢂ10À7 molmgÀ1 (T1), were tested. An excess of 0.37ꢂ
10À7 molmgÀ1, which corresponds to the theoretically necessa-
ry amount for a monolayer formation of 1, was used to study
the equilibrium adsorption and to determine the concentration
of 1 to obtain complete surface coverage. For the functionali-
zation, the overall TiO2 concentration was 2.08 mgmLÀ1. After
functionalization, all suspensions were washed with ethanol to
remove any non-bound surfactants (for details see the Experi-
mental Section). Immediately after mixing, a rapid color
Figure 3. Ethanolic suspensions of pristine TiO2 NPs T7 (left) and functional-
ized TiO2 NPs T2 (right) with a concentration of 29ꢂ10À3 mgmLÀ1, directly
after sonication (a) and after aging for 24 h (b).
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
5
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
surface. The difference in mass losses between functionalized
and pristine TiO2 NPs allows the determination of the exact
amount of chemisorbed catechol. The TiO2 NPs were function-
alized with different amounts of 1, so an adsorption isotherm
could be derived from the TGA data.
Figure 4 shows that upon functionalization of TiO2 NPs with
different amounts of 1, ranging from 0.25ꢂ10À7 molmgÀ1 (T1)
Figure 5. Zeta potential values of ethanolic suspensions as a function of dif-
ferent quantities of 1 added to TiO2 NPs.
owing to the absence of repulsive forces.[23a] If dispersed in liq-
uids, TiO2 NPs may have OH-terminated surfaces, which are de-
protonated in protic solvents such as water or ethanol, result-
ing in negative zeta potentials. Figure 5 shows the zeta poten-
tials of ethanolic suspensions of TiO2 NPs as a function of the
quantity of 1. Overall, the zeta potentials increase with increas-
ing amount of 1. Nonfunctionalized particles T7 display a zeta
potential of À22.1Æ0.53 mV. Samples treated only with small
quantities of 1 show increased, but low, zeta potential values.
For T6 and T5, we recorded À9.9Æ0.23 and 8.9Æ0.23 mV, re-
spectively, which indicate the low colloidal stabilities of the
suspensions. Furthermore, the zeta potential rises to +26.8Æ
0.91 (T4), +32.6Æ0.59 (T3), +34.3Æ0.60 (T2), and +34.2Æ
0.03 mV (T1). With regard to the highly negative zeta potential
values of the nonfunctionalized TiO2 NPs, we also considered
electrostatic interactions of the positively charged catechol
with the NP surface. Taking into account that highly positive
values for the zeta potential are obtained for suspensions T1
to T4, even after three washing steps with pure ethanol, pure
electrostatic coordination are negligible. A strong covalent
binding of the catechol to the NP surface is approved, because
in the case of pure electrostatic coordination of 1 to the sur-
face, zeta potential values would decrease again after washing
of the NPs.
Figure 4. Adsorption isotherm for the functionalization of TiO2 NPs with 1 as
a function of the added amount of 1 with an overall TiO2 concentration of
2.08 mgmLÀ1
.
to 16.27ꢂ10À7 molmgÀ1 (T6), the adsorption increases. The
exact values for all the different concentrations are summar-
ized in Table 1. The highest loading of 8.16ꢂ10À7 molmÀ2 was
Table 1. Chemisorbed amount of 1 on TiO2 NPs depending on the quan-
tity of 1 added for the functionalization.
Sample
Quantity of added
catechol
Amount of
chemisorbed
catechol [molmÀ2
Number of
adsorbed
molecules per m2
[molmgÀ1
]
]
T1
T2
T3
T4
T5
T6
16.27ꢂ10À7
4.07ꢂ10À7
2.03ꢂ10À7
1.02ꢂ10À7
0.51ꢂ10À7
0.25ꢂ10À7
8.16ꢂ10À7
5.84ꢂ10À7
4.61ꢂ10À7
2.95ꢂ10À7
2.21ꢂ10À7
1.87ꢂ10À8
4.9ꢂ1017
3.5ꢂ1017
2.8ꢂ1017
1.8ꢂ1017
1.3ꢂ1017
0.1ꢂ1017
The presence of larger quantities of 1 leads to higher values
of the zeta potential and to higher stabilities of the suspen-
sions owing to increased electrostatic repulsions between NPs.
From the zeta potential measurements we derive that stable
ethanolic suspensions were obtained not only for NPs featur-
ing a monolayer of 1 (T1), but also for NPs functionalized with
4.07ꢂ10À7 (T2), 2.03ꢂ10À7 (T3), and 1.02ꢂ10À7 molmgÀ1 (T4).
The stability of the suspensions was evaluated from the par-
ticle size distributions found through dynamic light scattering
at different time delays. To this end, measurements were taken
directly after sonication (time=0 h) and after ageing for 0.5, 2,
and 24 h. Figure 6 shows the average hydrodynamic diameter
of the TiO2 NPs in ethanol with time.
obtained upon functionalization of NPs with 16.27ꢂ
10À7 molmgÀ1 (T1) in an overall TiO2 concentration of
2.08 mgmLÀ1. Taking the theoretically calculated maximum
amount of 1 into account, this corresponds to 100% coverage
of the surface, and as a consequence, to monolayer formation.
Next, the zeta potential and hydrodynamic diameter were
measured to obtain additional information about the colloidal
stability of the functionalized TiO2 NPs. In general, the zeta po-
tential assists in drawing conclusions about the colloidal stabil-
ity of NPs.[34] For example, high negative or positive values cor-
respond to high charge densities on the particle surface, and,
in turn, to stable suspensions owing to electrostatic repulsions.
Low potentials, on the contrary, indicate moderate or low sur-
face charge densities. In general, this may result in aggregation
Directly after sonication, all the suspensions, with the excep-
tion of T6, feature hydrodynamic diameters similar to that of
unmodified TiO2. T6 features slightly higher values for the hy-
drodynamic diameter and size distribution from the beginning,
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
6
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Figure 7. SEM images of a) pristine TiO2 NPs and b) TiO2 NPs functionalized
with 1.
films prepared with pristine and functionalized TiO2 nanoparti-
cles. TiO2 NPs were functionalized with 1.84ꢂ10À7 molmgÀ1
(T8), which is, according to the previously discussed experi-
ments, a concentration sufficient to ensure the successful func-
tionalization of TiO2 NPs.
Figure 6. a) Hydrodynamic diameter of TiO2 NPs T1 (&, solid), T2 (*, solid),
^
T3 (~, solid), T4 (!, solid), T5 (3, solid), T6 (", solid), and T7 ( , dotted)
as a function of time, and b) particle size distributions after aging of TiO2
suspensions T1–T7 for 24 h.
which gives reason to assume that nanoparticles functionalized
under these specific conditions form unstable suspensions.
During ageing at 258C, poorly functionalized NPs form ag-
glomerates (T5 and T6), whereas the particle size distributions
of successfully functionalized TiO2 NPs remain unchanged (T1
to T4). These observations are in good agreement with the
zeta potential measurements. On one hand, TiO2 nanoparticles
of T1 to T4 carry a sufficient number of positive charges, intro-
duced by functionalization with 1, to form stable colloidal sus-
pensions. On the other hand, T5 and T6 seem to agglomerate
with time owing to insufficient functionalization and a lack of
electrostatic repulsions. T7 agglomerates as no stabilizing sur-
factants are present. Comparing these results with those from
TGA measurements, we observe that complete monolayer for-
mation on the particle surface is not necessarily required to
obtain stable nanoparticle suspensions. Monolayer formation
was achieved upon functionalization of the NPs with 16.27ꢂ
10À7 molmgÀ1 (T1) of 1. However, we derived from zeta poten-
tial and DLS measurements that nanoparticles functionalized
with 1.02ꢂ10À7 molmgÀ1 (T4) already form stable ethanolic
suspensions. Thus, NPs carrying 2.95ꢂ10À7 molmÀ2 of 1 on the
surface (T4), which corresponds to 41% of the amount re-
quired for the formation of a monolayer, already form stable
ethanolic suspensions.
Importantly, the SEM images confirm the presence of TiO2
NPs. Whereas pristine TiO2 NPs form large agglomerates on
silica wafers during drying, a lack of the latter is observed for
any of the functionalized TiO2 NPs. In fact, the latter are best
described as small islands, which are homogenously distribut-
ed throughout the wafer. Again, we take this as support for
the successful stabilization of TiO2 NPs in suspension (Figure 7).
Formation of layer-by-layer assemblies of catechol-function-
alized TiO2 nanoparticles and porphyrins
Having accomplished both, the successful functionalization of
TiO2 NPs to obtain stable ethanolic suspensions of positively
charged NPs and the synthesis of negatively charged porphyr-
ins, the construction of LbL assemblies was investigated. Here,
we constructed LbL assemblies consisting of positively charged
TiO2 NPs and negatively charged porphyrins on three different
substrates, namely soda-lime-glass, silicon wafers, and indium
tin oxide (ITO) electrodes. The assembly formation was per-
formed using a 10À3 m ethanolic suspension of functionalized
TiO2 NPs T8 and 10À3 m solutions of porphyrins 2 and 3 in
aqueous TRIS buffer solution (pH 9). To monitor the LbL forma-
tion, we initially utilized soda-lime-glass and silicon wafer sub-
strates, which allow the performance of absorption, profilome-
try, and SEM studies. Indium tin oxide (ITO) electrodes were
used to perform device assays. Regardless of the type of sub-
strate, the LbL deposition started with the deposition of a thin
layer of positively charged poly-(diallyl-dimethylammonium)
To round off the characterization of catechol-functionalized
TiO2 NPs, scanning electron microscopy (SEM) investigations
with thin films prepared by drop-casting of ethanolic TiO2 NP
suspensions were performed. Figure 7 displays SEM images of
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
7
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
(PDDA+) and/or negatively charged poly-(sodium-4-styrene-
sulfonat) (PSSÀ). To this end, the substrate surface was modi-
fied with either positive or negative charges that allow the
preparation of different films, namely [PDDA+-PSSÀ]xPDDA+
-[porphyrin-TiO2]x and [PDDA+-PSSÀ]x-[TiO2-porphyrin]x, in
which x is the number of sandwich layers. As an example,
Figure 8 shows in detail the procedure for the preparation of
glass substrates with a polymer base consisting of [PDDA+
-PSSÀ]4-PDDA+.[36] Absorption spectra were recorded after the
deposition of each porphyrin layer. On one hand, strong Soret-
and weaker Q-band absorptions are discernible at around
430 nm and in the 520–660 nm range, respectively. On the
other hand, a broad band at around 320 nm, which is ascribed
to the TiO2 NPs, is also clearly noted. At first glance, the inten-
sities of all of these absorption bands increase linearly
(Figure 9). This finding confirms successful LbL deposition.
Figure 8. Formation of layer-by-layer assemblies. a) Deposition of PDDA+,
b) deposition of PSSÀ, c) deposition of positively charged TiO2 NPs, d) depo-
sition of negatively charged porphyrins, e) multiple LbL assemblies of por-
phyrins and TiO2 NPs.
Figure 9. Absorption spectra of LbL assemblies of a) T8 and 2 or b) T8 and
3. The inset describes the dependence of the maxima of the Soret band of
the porphyrins and of the TiO2 band on the number of stacks of porphyrin
deposited on the substrate.
the LbL films. First, a layer of PDDA+ was deposited onto the
surface through dip-coating for 10 min followed by washing
with a mixture of ethanol/water 1:1 (v/v), and drying with
a stream of air. This procedure resulted in substrates that were
covered with positively charged hydrophilic surfaces. Secondly,
the same procedure was followed to deposit PSSÀ onto the
PDDA+-covered substrate, to obtain a negatively charged hy-
drophilic surface. The first two steps were repeated up to four
times to ensure a thin homogenous layer of around 30 nm.
Thirdly, alternate depositions of positively functionalized TiO2
NPs and negatively charge porphyrins achieved by dipping the
modified substrates for 10 min each into the corresponding
suspensions enabled the formation of multilayer assemblies.
The application of these deposition times is important, as
stable and constant layer growth is achieved only after a cer-
tain time period.[35]
However, a closer look reveals subtle differences. For exam-
ple, a direct comparison of the absorption spectra reveals that
utilizing porphyrin 3 for the buildup of LbL assemblies results
in a smaller increase in the Soret-band absorption (17.6%)
compared with those being built up with porphyrin 2 (24.0%).
In contrast, the intensity of the band ascribed to the TiO2
nanoparticles increases by more or less equal values upon uti-
lization of porphyrin 2 (47.1%) or 3 (48.4%). Considering that
both porphyrins feature similar extinction coefficients, that is,
3.40ꢂ105 (2) versus 3.17ꢂ105 LmolcmÀ1 (3), we infer that dif-
ferences in the steric hindrance and the number of charges at
the porphyrin are responsible for the slightly different amounts
of the porphyrins. Different quantities of porphyrins 2 and 3
adsorbed on the substrate provide therewith the same overall
density of negative charges; hence, the same quantities of
functionalized TiO2 NPs are adsorbed.
In a first approach, the LbL growth with T8 and porphyrins 2
and 3, respectively, was monitored by UV/Vis spectroscopy.
Therefore, LbL assemblies have been built up on soda-lime-
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
8
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Regardless of the type of porphyrin, the film thicknesses of
the assemblies, which were prepared on silicon wafers, were
probed by combining ellipsometry and profilometry measure-
ments. The former indicates that each porphyrin layer deposi-
tion increases the overall thickness by about 0.3–0.6 nm. The
latter implies that each deposition of functionalized TiO2 NPs
increases the thickness by around 30–50 nm. Notably, we mon-
itor the sequential increase in the layer thickness by combining
both techniques to obtain a linear relationship that corrobo-
rates the absorption measurements (Figure 10).
Figure 10. Thickness of the LbL assemblies on silicon wafer as a function of
[porphyrin 2-T8] stacks.
For confirmation of the notion of controlled two-dimension-
al film growth, SEM images of the assemblies prepared on sili-
con wafers were obtained (Figure 11). SEM images were re-
corded after one, three, and six deposition cycles. A good cov-
erage with uniformly distributed TiO2 NPs on the substrate was
already obtained after six deposition cycles, indicating the suc-
cessful control of the homogeneous LbL deposition as
a means to integrate functionalized TiO2 NPs and anionic por-
phyrins.
Figure 11. SEM images of a silicon wafers coated with a) [PDDA+-PSSÀ]4-
PDDA+-[2-T8]1, b) [PDDA+-PSSÀ]4-PDDA+-[2-T8]3, and c) [PDDA+-PSSÀ]4-
PDDA+-[2-T8]6.
Photoelectrochemical cell fabrication
Encouraged by the features of the LbL assemblies, we turned
to the fabrication of photoelectrochemical cells with photoac-
tive ITO electrodes modified by sequential LbL deposition of
the functionalized TiO2 NPs and both porphyrins. Importantly,
the assemblies prepared on ITO substrates showed strong fluo-
rescence quenching of both porphyrins upon 420 nm excita-
tion, indicating a qualitative electron and/or energy transfer
from the porphyrin to the TiO2 NPs (Figure 12).
As reference electrodes, the porphyrins were replaced with
PSSÀ to give ITO-[PDDA+-PSSÀ]-[T8-PSSÀ]x. To complete the de-
vices we utilized polysulfide electrolytes (i.e., equimolar 3m
Na2S/S/NaOH) and Cu2S-based counter electrodes (for more
details see Experimental Section).
Figure 12. Fluorescence spectra of ITO-[PDDA+-PSSÀ]4-PDDA+-[2] (gray) and
ITO-[PDDA+-PSSÀ]4-PDDA+-[2-T8] (black) films at 420 nm excitation.
aforementioned trends were validated in complementary I–V
measurements under AM 1.5 conditions. On one hand, maxi-
mum short-circuit current densities of around 25 mAcmÀ2 were
derived for both assemblies, featuring, however, only a single
stack layer, that is, ITO-[PDDA+-PSSÀ]4-PDDA+-[porphyrin-T8]1.
The main differences between the assemblies are seen in the
As shown in Figure 13, assemblies prepared with porphyr-
ins 2 and 3 reveal promising and stable photoresponse profiles
during several “on-off” cycles of illumination under AM 1.5 con-
ditions. In line with the absorption features, the photocurrents
for 2 and 3 in the “on-off” experiments are rather similar. The
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
9
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
the open-circuit voltage parameters (Figure 13). For 3, the
device performance is reduced dramatically upon increasing
the number of layers. Such behavior is ascribable to the thick-
ness of the photoelectrode, which increases after each sand-
wich layer deposition to reach a value of around 400 nm. Al-
though light-harvesting by the electrodes is enhanced with
each deposition, the loss of open-circuit voltage with thickness
implies that the electron transport and electron collection at
the electrode are interrupted because of recombination pro-
cesses with the electrolyte. In fact, the overall photogenerated
currents decrease.
Figure 14 reveals that measurements under monochromatic
illumination to determine the incident photon conversion effi-
Figure 13. Top: Photocurrent generated under “on-off” illumination cycles
for photoelectrodes ITO-[PDDA+-PSSÀ]4-PDDA+-[3/T8]1 (black) and ITO-
[PDDA+-PSSÀ]4-PDDA+-[2-T8]1 (gray). Center: Photocurrent–voltage charac-
teristics of ITO-[PDDA+-PSSÀ]4-PDDA+-[2-T8]x photoelectrodes with increas-
ing number of [2-T8] sandwich layers [1 (black), 3 (gray), and 6 (light gray)]
under AM 1.5 conditions. Bottom: Photocurrent–voltage characteristics of
ITO-[PDDA+-PSSÀ]4-PDDA+-[3-T8]x photoelectrodes with increasing number
of [3-T8] sandwich layers [1 (black), 3 (gray), and 6 (light gray)] under AM 1.5
conditions.
Figure 14. Top: Photocurrent–voltage characteristics of ITO-[PDDA+-PSSÀ]4-
[T8-PSSÀ]x photoelectrodes with increasing number of [T8-PSSÀ] sandwich
layers [1 (black) and 3 (gray)] under AM 1.5 conditions. Bottom: IPCE spectra
of ITO-[PDDA+-PSSÀ)4-PDDA+-2 (&, black), ITO-[PDDA+-PSSÀ]4-T8 (!, light
gray), ITO-[PDDA+-PSSÀ]4-PDDA+-[2-T8]1 (*, dark gray), and ITO-[PDDA+-
PSSÀ]4-PDDA+-[2-T8]3 (~, gray) photoelectrodes under AM 1.5 conditions.
Inset graph is a magnification to show the contribution from the Q-bands.
open-circuit voltages, which are 0.12 and 0.03 V for assemblies
with 2 and 3, respectively. As a consequence, photoelectrodes
prepared with 2 feature efficiencies four times higher than
those prepared with 3: 1ꢂ10À3 versus 2.5ꢂ10À4 %. On the
other hand, the increase in the number of sandwich layers has
different impacts in the two assemblies. For 2, increasing the
number of layers to three enhances the photocurrent values
from 25 to 50 mAcmÀ2, and the open-circuit voltage is reduced
from 0.12 to 0.07 V. A further increase in the number of layers
leads to a significant reduction in both the photocurrent and
ciency (IPCE) spectrum show a good resemblance to the ab-
sorption spectrum. From this, we hypothesize that the photo-
sensitizing attributes of both porphyrins are the inception for
electron injection into functionalized TiO2 NPs. To corroborate
this hypothesis, we prepared devices lacking any porphyrins,
that is, ITO-[PDDA+-PSSÀ]-[T8-PSSÀ]x devices. The lack of appre-
ciable photocurrents (<4 mAcmÀ2) upon increasing the
number of [TiO2-PSSÀ] sandwich layers clearly confirmed that
the porphyrins are responsible for the photogenerated cur-
rents (Figure 14).
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
10
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Thermogravimetric measurements of dried powders were per-
formed on a TGA Q50 device (TA Instruments) in ramp regime
(58CminÀ1) under a nitrogen flow over the temperature range 20–
9508C. The amounts of chemisorbed compounds were calculated
from the weight loss in the temperature range 100–845.88C.
Conclusion
P25-TiO2 NPs were successfully functionalized with N1,N7-bis(3-
(4-tert-butyl-pyridium-methyl)phenyl)-4-(3-(3-(4-tert-butyl-pyri-
dinium-methyl)phenylamino)-3-oxopropyl)-4-(3,4-dihydroxy-
benzamido)heptanediamide tribromide (1). From zeta potential
and DLS measurements we derived that stable colloidal sus-
pensions are realized if a minimum concentration of 1.02ꢂ
10À7 molmgÀ1 (T4) is used for the NP functionalization. In par-
ticular, strongly positive values for the zeta potential and small
average hydrodynamic diameters were determined. TGA meas-
urements showed that by functionalizing TiO2 NPs with con-
centrations equal to 1.02ꢂ10À7 molmgÀ1, 2.65ꢂ10À7 molmÀ2
of 1 is adsorbed on their surface. This corresponds to 41% of
the amount needed for monolayer formation on the NP sur-
face. The maximum coverage of 8.16ꢂ10À4 mmolmÀ2, which
corresponds to the formation of a monolayer, was obtained
upon functionalization with 16.27ꢂ10À7 molmgÀ1 (T1) and an
overall TiO2 concentration of 2.08 mgmLÀ1. Therefore, we con-
clude that at least 41% of the surface needs to be covered
with 1 to achieve colloidal stability of the TiO2 NPs.
Nitrogen sorption experiments at liquid nitrogen temperature
were performed with
a volumetric gas sorption analyzer
(Nova4200e; Quantachrome, Odelzhausen, Germany). Degassing
under vacuum conditions was performed for 2 h at 1208C. Mass
specific surface areas (Sm) were evaluated in a p/p0 range from 0.02
to 0.35 according to the Brunauer–Emmett–Teller (BET) theory.
Mesopore volumes were calculated from desorption data using the
Barrett–Joyner–Halenda (BJH) model; relative pressures below 0.35
were omitted for the calculations.
Scanning electron microscopy images were recorded with a field
emission scanning electron microscope (ULTRA 55, Carl Zeiss NTS
GmbH, Germany) at an operating voltage of 10 kV using an in-lens
secondary electron detector. TiO2 nanoparticles were functionalized
with 1.84ꢂ10À7 molmgÀ1. The samples were prepared by drop-
casting of ethanolic suspensions on silicon wafers.
Adsorption spectra were recorded with Varian Cary 5000 (charac-
terization of porphyrins) and Shimadzu UV-3102 PC (characteriza-
tion of layer-by-layer assemblies) instruments.
Furthermore, the successful synthesis of two porphyrin de-
rivatives, 5,10,15,20-(phenoxyacetic acid)-porphyrin (6) and
5,10,15,20-(4-(2-ethoxycarbonyl)-4-(2-phenoxyacetamido)hepta-
nedioic acid)-porphyrin (8) was presented. Both porphyrins fea-
ture anionic carboxylates under basic conditions, which make
them excellent building blocks for the formation of layer-by-
layer assemblies with TiO2 NPs 1 through electrostatic interac-
tions. Layer-by-layer assemblies were built up on three differ-
ent substrates. Investigation of layer-by-layer assemblies of
functionalized TiO2 NPs 1 and porphyrins 2 or 3 by UV/Vis
spectroscopy showed that the quantity of adsorbed porphyrin
depends on their structure. SEM studies confirmed the forma-
tion of homogeneous TiO2 layers after six deposition cycles,
and ellipsometry and profilometry assays also indicated
a linear increase in the thickness up to a value of 400 nm upon
the deposition of 14 sandwich layers. Finally, we demonstrated
the performance of photoelectrodes prepared with both LbL
hybrids, which showed stable and repeatable photocurrent
generation during several “on-off” cycles of illumination with
monochromatic incident photo-to-current conversion efficien-
cies of around 3%.
Profilometry measurements were performed with a DektakXT on
silica wafers and ITO for the study of film formation and device
fabrication, respectively. We used a Stylus tip of 12.5 mm with
a force of 3 mg and a resolution of 0.66 mm/pt.
The ellipsometric measurements were performed with a spectro-
scopic ellipsometer HORIBA Smart SE on a silicon wafer using
a stable laser power of 12 mW and an angle of 708. For the fitting
of the characteristics we used the data of the tetraphenylzincpor-
phyrin included in the database of the program.
Emission spectra were recorded with a Shimadzu RF-5301 PC in-
strument.
Mass spectra were recorded with Shimadzu AXIMA Confidence
(MALDI-TOF, matrices: sinapinic acid SIN, 2,5-dihydroxybenzoic acid
DHB) and Bruker Esquire 6000 (ESI-QIT) instruments.
NMR spectra were recorded with Bruker Avance 300 (300 MHz),
Avance 400 (400 MHz), and Jeol EX 400 (400 MHz) spectrometers.
Chemical shifts are given in ppm relative to TMS. The resonance
multiplicities are indicated as s (singlet), d (doublet), and t (triplet),
and broad resonances are indicated as bs.
Synthesis of anionic porphyrins 5,10,15,20-(phenoxyacetic
acid)-porphyrin (2) and 5,10,15,20-(4-(2-ethoxycarbonyl)-4-
(2-phenoxyacetamido)heptanedioic acid)-porphyrin (3)
5,10,15,20-(tert-Butyl-2-phenoxyacetato)-porphyrin (5): tert-Butyl-
2-(4-formylphenoxy)acetate (3 g; 12.71 mmol) and pyrrole
(0.89 mL; 12.71 mmol) were dissolved in CH2Cl2 (1.9 L). EtOH
(15 mL) and BF3·OEt2 (0.16 mL; 1.27 mmol) were added. The solu-
tion was stirred at room temperature for 1.5 h. DDQ (2.16 g;
9.53 mmol) was added and the solution was stirred at room tem-
perature for 3.5 h. Afterwards, the volume of the reaction media
was reduced to 400 mL and the solvent was removed on a rotary
evaporator. The residual was filtered over a silica plug with CH2Cl2
as eluent, followed by further purification through column chroma-
tography over silica with a mixture of CH2Cl2/EtOAc (98:2) as
eluent. The solvent was removed under reduced pressure on
a rotary evaporator and the product was dried under vacuum.
Yield: 0.875 g (0.77 mmol), 6%.
Experimental Section
Chemicals and instruments
All chemicals were purchased from chemical suppliers and used
without further purification. All analytical reagent-grade solvents
were purified by distillation.
Average hydrodynamic diameters and zeta potentials of nanoparti-
cles in ethanolic dispersions were measured with a Zetasizer Nano
(Malvern Instruments Ltd. UK) dynamic light scattering (DLS) ana-
lyzer. Intensity distributions were recorded for TiO2 nanoparticle
suspensions in ethanol with a particle concentration of 16ꢂ
10À3 mgmLÀ1 at 258C. The suspensions were sonicated for 1 h at
258C (time=0).
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
11
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
1H NMR (400 MHz, 258C, CDCl3): d=À2.81 (bs, 2H, NH), 1.59 (s,
1H NMR (400 MHz, 258C, DMSO): d=À2.92 (bs, 2H, NH), 1.99 (t,
3J=8.0 Hz, 24H, CH2CH2CO), 2.25 (t, 3J=8.0 Hz, 24H, CH2CH2CO),
3
36H, CH3), 4.80 (s, 8H, CH2), 7.26 (d, J=8.5 Hz, 8H, CH-benz), 8.10
(d, 3J=8.5 Hz, 8H, CH-benz), 8.83 ppm (s, 8H, CH-pyrr); 13C NMR
(400 MHz, 258C, CDCl3): d=14.48 (s, 12C, CH3), 52.38 (s, 4C, CH2),
68.91 (s, 4C, CCH3), 99.22 (s, 8C, CH-benz), 105.85 (s, 4C, C-por),
121.75 (s, 4C, C-C-benz), 121.83 (s, 8C, CH-benz), 144.14 (s, 4C, C-O-
benz), 154.48 ppm (s, 4C, COOCCH3); MS (MALDI, sin): m/z: 1135
[M]+, 1136 [M+H]+, 1137 [M+2H]+; HRMS (ESI-TOF, MeOH-ACN):
m/z calcd for C68H70N4O12: 1135.5063 [M]+; found: 1135.5049 [M]+.
4.77 (s, 8H, CH2), 7.38 (d, J=8.4 Hz, 8H, CH-benz), 7.58 (s, 4H, 6),
3
3
8.12 (d, J=8.4 Hz, 8H, CH-benz), 8.87 (s, 8H, CH-pyrr), 12.15 ppm
(bs, 12H, COOH); 13C NMR (400 MHz, 258C, DMSO): d=28.10 (s,
12C, CH2CH2CO), 29.24 (s, 12C, CH2CH2CO), 56.96 (s, 4C, CONHC),
67.03 (s, 4C, CH2), 113.14 (s, 8C, CH-benz), 119.57 (s, 7C, C-por),
131.47 (m, 8C, CH-pyrr), 133.99 (s, 4C, C-C-benz), 135.33 (s, 8C, CH-
benz), 157.79 (s, 4C, C-O-benz), 167.23 (s, 4C, CONH), 174.47 ppm
(s, 12C, COOH); HRMS (ESI-TOF, ACN-MeOH-HCOOH): m/z calcd for
5,10,15,20-(Phenoxyacetic acid)-porphyrin (6): 5,10,15,20-(tert-
Butyl-2-phenoxyacetato)-porphyrin (334 mg; 0.29 mmol) was dis-
solved in formic acid (25 mL). The solution was stirred for 18 h at
room temperature. The formic acid was removed on a rotary evap-
orator. Subsequently, the product was transferred to toluene and
evaporated twice to remove any residual formic acid. The product
was finally dried under reduced pressure. Yield: 260 mg
(0.28 mmol), 99%.
C92H98N8O32: 936.3036 [M+2Na]2+
,
1826.6282 [M]+, 1827.6360
[M+H]+, 1849.6179 [M+Na]+; found: 936.3019 [M+2Na]2+
1827.6279 [M]+, 1827.6279 [M+H]+, 1849.6154 [M+Na]+.
,
Functionalization of TiO2 nanoparticles with N1,N7-bis(3-(4-
tert-butylpyridium-methyl)phenyl)-4-(3-(3-(4-tert-butylpyridi-
nium-methyl)phenylamino)-3-oxopropyl)-4-(3,4-dihydroxy-
benzamido)heptanediamide tribromide (1)
1H NMR (400 MHz, 258C, DMSO): d=À2.91 (s, 2H, NH), 4.97 (s, 8H,
3
3
An ethanolic suspension (16.67 mgmLÀ1) of P-25 TiO2 nanoparticles
(purchased from Evonic) was prepared. For investigation of the
functionalization of the TiO2 surface, a suspension of TiO2 nanopar-
ticles (washed once with EtOH) was mixed with different quantities
of 1 in ethanol [16.272ꢂ10À7 (T1), 4.068ꢂ10À7 (T2), 2.034ꢂ10À7
(T3), 1.017ꢂ10À7 (T4), 0.508ꢂ10À7 (T5), 0.254ꢂ10À7 (T6), and
0 molmgÀ1 (T7)]. The overall TiO2 concentration for the functionali-
zation was adjusted to 2.08 mgmLÀ1. The mixture was sonicated
for 50 min at room temperature. Subsequently, the suspensions of
TiO2 nanoparticles were centrifuged until the TiO2 nanoparticles
settled completely. The supernatant was removed and the particles
were washed three times by the addition of 5 mL ethanol followed
by sonication until redispersion was completed, followed by centri-
fugation until the TiO2 nanoparticles settled completely. After the
last washing step, the TiO2 nanoparticles were dispersed in ethanol
CH2), 7.35 (d, J=8.4 Hz, 8H, CH-benz), 8.12 (d, J=8.4 Hz, 8H, CH-
benz), 8.84 (s, 8H, CH-pyrr), 13.16 ppm (bs, 4H, COOH); 13C NMR
(400 MHz, 258C, DMSO): d=64.84 (s, 4C, CH2), 113.04 (s, 8C, CH-
benz), 119.57 (s, 4C, C-por), 133.94 (s, 4C, C-C-benz), 135.31 (s, 8C,
CH-benz), 157.73 (s, 4C, C-O-benz), 170.32 ppm (s, 4C, COOH); MS
(MALDI, om): m/z: 911 [M]+, 912 [M+H]+, 913 [M+2H]+; HRMS
(ESI-TOF, MeOH-THF): m/z calcd for C52H38N4O12: 910.248074 [M]+;
found: 910.244958 [M]+; calcd for C52H39N4O12: 911.255899 [M+H]+;
found: 911.252284 [M+H]+.
5,10,15,20-(Di-tert-butyl-4-(3-(tert-butoxy)-3-oxopropyl)-4-(2-phe-
noxyacetamido)heptanedionato)-porphyrin (7): 5,10,15,20-(Phe-
noxyacetic acid)-porphyrin (140 mg; 0.15 mmol) was dissolved in
DMF (25 mL) and the solution was cooled to 08C. Di-tert-butyl-4-
amino-4-(3-(tert-butoxy)-3-oxopropyl)heptanedionate (461.32 mg;
1.11 mmol) was added, then DCC (216.37 mg; 1.05 mmol) and
HOBT (141.88 mg; 1.05 mmol) were added and the solution was
stirred for three days at room temperature. The solvent was re-
moved under reduced pressure on a rotary evaporator and the
product was purified through column chromatography over silica
with a mixture of CH2Cl2 and EtOAc (4:1) as eluent. Yield: 273 mg
(0.11 mmol), 79%.
to obtain a particles suspension of 4.81 mgmLÀ1
.
Preparation of tris(hydroxymethyl) aminomethane buffer
solutions
A 0.1m aqueous solution of tris(hydroxymethyl) aminomethane
was prepared and subsequently adjusted to pH 9 with 0.1m HCl
solution.
1H NMR (400 MHz, 258C, CDCl3): d=À2.79 (s, 2H, NH), 1.44 (s,
3
3
108H, CH3), 2.14 (t, J=8.6 Hz, 24H, CH2CH2CO), 2.34 (t, J=8.6 Hz,
3
24H, CH2CH2CO), 4.69 (s, 8H, CH2), 6.84 (s, 4H, CONH), 7.34 (d, J=
8.6 Hz, 8H, CH-benz), 8.16 (d, J=8.6 Hz, 8H, CH-benz), 8.84 ppm (s,
3
Preparation of poly-(diallyl-dimethylammonium) (PDDA) and
poly-(sodium-4-styrene-sulfonato) (PSS) solutions
8H, CH-pyrr); 13C NMR (400 MHz, 258C, CDCl3): d=28.00 (s, 36C,
CH3), 29.66 (s, 12C, CH2CH2CO), 29.97 (s, 12C, CH2CH2CO), 57.68 (s,
4C, CONHC), 67.54 (s, 4C, CH2), 80.62 (s, 12C, CCH3), 112.96 (s, 8C,
CH-benz), 119.30 (s, 4C, C-pyr), 135.58 (s, 8C, C-C-benz), 135.88 (s,
4C, CH-benz), 156.88 (s, 4C, C-O-benz), 167.25 (s, 4C, CONH),
172.48 ppm (s, 12C, COOCCH3); MS (MALDI, dctb): m/z: 2501 [M]+,
2502 [M+H]+; HRMS (ESI-TOF, ACN-MeOH): m/z calcd for
0.5% PDDA and PSS solutions at pH 3 and pH 4, respectively, were
prepared from a 20 wt% solutions in water (Mw =400000–500000;
PDDA) and an 18 wt% solution in water (Mw =1000000; PSS), pur-
chased from Aldrich.
C140H194N8Na2O32: 1272.6792 [M+2Na]2+
[M+2Na]2+, 856.4538 [M+3Na]3+
;
found: 1272.6822
Formation of layer-by-layer assemblies
.
For the buildup of layer-by-layer assemblies on different substrates,
the following sequences were used: 1) dipping of the substrate
into a solution of 0.5% PDDA+ at pH 3 for 10 min; 2) rinsing with
ethanol/water 1:1 (v/v); 3) dipping into a solution of 0.5% PSSÀ at
pH 4 for 10 min; 4) rinsing with ethanol/water 1:1 (v/v); 5) dipping
into a 10À3 m ethanolic suspension of TiO2 nanoparticles for
10 min; 6) rinsing with ethanol; 7) dipping into a 10À3 m TRIS buffer
solution at pH 9 of 2 or 3, respectively, for 10 min; 8) rinsing with
ethanol/water 1:1 (v/v). Sequences 1–4) were repeated up to four
times. Depending on whether the first deposited layer should con-
5,10,15,20-(4-(2-Ethoxycarbonyl)-4-(2-phenoxyacetamido)hepta-
nedioic acid)-porphyrin (8): 5,10,15,20-(Di-tert-butyl-4-(3-(tert-
butoxy)-3-oxopropyl)-4-(2-phenoxy-acetamido)heptanedionato)-
porphyrin (140 mg; 0.056 mmol) was dissolved in formic adic
(40 mL) and stirred for 4 h at room temperature. The formic acid
was removed on a rotary evaporator. Subsequently, the product
was transferred to toluene and evaporated twice to remove any re-
sidual formic acid. The product was finally dried under reduced
pressure. Yield: (0.056 mmol), 100%.
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
12
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
[16] X. Chen, S. S. Mao, Chem. Rev. 2007, 107, 2891–2959.
sist of porphyrins or TiO2 nanoparticles, the final polymer layer is
PDDA+ or PSSÀ, respectively.
[17] a) D. Chen, R. A. Caruso, Adv. Funct. Mater. 2013, 23, 1356–1374; b) M.
Grꢆtzel, J. Photochem. Photobiol. A 2004, 164, 3–14; c) M. M. Byranvand,
A. N. Kharat, M. H. Bazargan, Nano-Micro Lett. 2012, 4, 253–266; d) Y.
Horiuchi, T. Toyao, M. Takeuchi, M. Matsuoka, M. Anpo, Phys. Chem.
Chem. Phys. 2013, 15, 13243–13253.
[18] W. M. Campbell, K. W. Jolley, P. Wagner, K. Wagner, P. J. Walsh, K. C.
Gordon, L. Schmidt-Mende, M. K. Nazeeruddin, Q. Wang, M. Gratzel,
D. L. Officer, J. Phys. Chem. C 2007, 111, 11760–11762.
Device fabrication
ITO substrates (purchased from Naranjo substrates) were cleaned
with water and soap and isopropanol solutions for 15 min in an
ultrasonication bath. Immediately, the PDDA+, and PSSÀ layers as
well as the functionalized TiO2 nanoparticles and porphyrins were
deposited as explained. To complete the devices, we utilized poly-
sulfide electrolytes (i.e., equimolar 3m Na2S/S/NaOH) and Cu2S-
based counter electrodes.
[19] a) H. Tokuhisa, P. T. Hammond, Adv. Funct. Mater. 2003, 13, 831–839;
b) T. Hasobe, S. Hattori, P. V. Kamat, S. Fukuzumi, Tetrahedron 2006, 62,
1937–1946; c) T. Hasobe, S. Hattori, P. V. Kamat, Y. Urano, N. Umezawa,
T. Nagano, S. Fukuzumi, Chem. Phys. 2005, 319, 243–252; d) C. Yang, Z.
Yang, H. Gu, C. K. Chang, P. Gao, B. Xu, Chem. Mater. 2008, 20, 7514–
7520; e) M. Funes, P. Zabel, T. Dittrich, E. N. Durantini, L. Otero, Phys.
Status solidi C 2010, 7, 280–283; f) J. E. Kroeze, T. J. Savenije, J. M.
Warman, J. Photochem. Photobiol. A 2002, 148, 49–55; g) C.-F. Lo, L.
Luo, E. W.-G. Diau, I.-J. Chang, C.-Y. Lin, Chem. Commun. 2006, 1430.
[20] a) J. Cao, D.-C. Hu, J.-C. Liu, R.-Z. Li, N.-Z. Jin, Org. Electron. 2014, 15,
509–516; b) M. T. Brumbach, A. K. Boal, D. R. Wheeler, Langmuir 2009,
25, 10685–10690; c) J. Cao, D.-C. Hu, J.-C. Liu, R.-Z. Li, N.-Z. Jin, Inorg.
Chim. Acta 2014, 414, 165–169; d) Y. Wu, J.-C. Liu, J. Cao, R.-Z. Li, N.-Z.
Jin, Res. Chem. Intermediates 2014, 1–10; e) H. B. Gobeze, S. K. Das, F.
D’Souza, J. Phys. Chem. C 2014, 118, 16660–16671; f) L. A. Martini, G. F.
Moore, R. L. Milot, L. Z. Cai, S. W. Sheehan, C. A. Schmuttenmaer, G. W.
Brudvig, R. H. Crabtree, J. Phys. Chem. C 2013, 117, 14526–14533;
g) C. F. A. Negre, R. L. Milot, L. A. Martini, W. Ding, R. H. Crabtree, C. A.
Schmuttenmaer, V. S. Batista, J. Phys. Chem. C 2013, 117, 24462–24470.
[21] J. H. Yang, Y. M. Chen, Y. B. Bai, M. Xian, D. F. Shen, Y. Q. Wang, S. R. Du,
R. Lu, T. J. Li, Y. Wu, W. Q. Xu, Supramol. Sci. 1998, 5, 599–602.
Acknowledgements
We gratefully acknowledge funding from the German Research
Council (DFG), which, within the framework of its “Excellence
Initiative”, supports the Cluster of Excellence “Engineering of
Advanced Materials” at the University of Erlangen-Nꢃrnberg.
Keywords: catechol
·
layer-by-layer
assemblies
·
nanoparticles · solar energy conversion · surface chemistry ·
surface functionalization
[1] a) I. Ahmed, R. Farha, M. Goldmann, L. Ruhlmann, Chem. Commun.
2013, 49, 496–498; b) S. Choi, H. Jin, J. Bang, S. Kim, J. Phys. Chem. Lett.
2012, 3, 3442–3447; c) G. Zhang, K. Pan, W. Zhou, Y. Qu, Q. Pan, B.
Jiang, G. Tian, G. Wang, Y. Xie, Y. Dong, X. Miao, C. Tian, Dalton Trans.
2012, 41, 12683–12689.
[2] a) E. M. Shchukina, D. G. Shchukin, Adv. Drug Delivery Rev. 2011, 63,
837–846; b) K. Sato, J.-i. Anzai, Molecules 2013, 18, 8440–8460.
[3] D. M. DeLongchamp, P. T. Hammond, Adv. Funct. Mater. 2004, 14, 224–
232.
[4] T. R. Farhat, P. T. Hammond, Adv. Funct. Mater. 2005, 15, 945–954.
[5] N. Dal’Acqua, F. R. Scheffer, R. Boniatti, B. V. M. da Silva, J. V. de Melo, J.
da Silva Crespo, M. Giovanela, M. B. Pereira, D. E. Weibel, G. Machado, J.
Phys. Chem. C 2013, 117, 23235–23243.
[6] a) G. Decher, J.-D. Hong, Makromol. Chem. Macromol. Symp. 1991, 46,
321–327; b) G. Decher, Science 1997, 277, 1232–1237.
[7] H. Zhang, J. Rꢃhe, Macromolecules 2003, 36, 6593–6598.
[8] a) N. G. Hoogeveen, M. A. Cohen Stuart, G. J. Fleer, Langmuir 1996, 12,
3675–3681; b) M.-A. Neouze, J. Mater. Sci. 2013, 48, 7321–7349; c) G.
Wu, X. Zhang, Polymer Thin Films, Abbass A Hashim (Ed.) 2010.
[9] a) Y. M. Lvov, G. Decher, H. Mçhwald, Langmuir 1993, 9, 481–486;
b) C. C. Buron, C. Filiꢄtre, J. Colloid Interface Sci. 2014, 413, 147–153;
c) Y. Tan, U. H. Yildiz, W. Wei, J. H. Waite, A. Miserez, Biomacromolecules
2013, 14, 1715–1726; d) S. Fujita, S. Shiratori, Nanotechnology 2005, 16,
1821–1827; e) V. V. Tsukruk, F. Rinderspacher, V. N. Bliznyuk, Langmuir
1997, 13, 2171–2176.
[22] M. C. So, S. Jin, H.-J. Son, G. P. Wiederrecht, O. K. Farha, J. T. Hupp, J. Am.
Chem. Soc. 2013, 135, 15698–15701.
[23] a) F. Hasani Bijarbooneh, Y. Zhao, J. H. Kim, Z. Sun, V. Malgras, S. H.
Aboutalebi, Y.-U. Heo, M. Ikegami, S. X. Dou, G. Franks, J. Am. Ceram.
Soc. 2013, 96, 2636–2643; b) R. A. French, A. R. Jacobson, B. Kim, S. L.
Isley, R. L. Penn, P. C. Baveye, Environ. Sci. Technol. 2009, 43, 1354–1359.
[24] M.-A. Neouze, U. Schubert, Monatsh. Chem. 2008, 139, 183–195.
[25] a) N. Erathodiyil, J. Y. Ying, Acc. Chem. Res. 2011, 44, 925–935; b) F. Is-
kandar, Adv. Powder Technol. 2009, 20, 283–292; c) S. T. Kim, K. Saha, C.
Kim, V. M. Rotello, Acc. Chem. Res. 2013, 46, 681–691; d) R. Mout, D. F.
Moyano, S. Rana, V. M. Rotello, Chem. Soc. Rev. 2012, 41, 2539–2543;
e) T. Banerjee, S. Kaniyankandy, A. Das, H. N. Ghosh, J. Phys. Chem. C
2013, 117, 3084–3092.
[26] a) A. Vaccaro, J. Hierrezuelo, M. Skarba, P. Galletto, J. r. Kleimann, M. Bor-
kovec, Langmuir 2009, 25, 4864–4867; b) E. Rampazzo, E. Brasola, S.
Marcuz, F. Mancin, P. Tecilla, U. Tonellato, J. Mater. Chem. 2005, 15,
2687–2695; c) Y. Kim, S. An, S. Kim, J. H. Park, H. A. Son, H. T. Kim, K.-D.
Suh, J. W. Kim, Polymer 2013, 54, 5609–5614; d) H. Jin, J. Nam, J. Park,
S. Jung, K. Im, J. Hur, J.-J. Park, J.-M. Kim, S. Kim, Chem. Commun. 2011,
47, 1758–1760.
[27] a) H. Lee, S. M. Dellatore, W. M. Miller, P. B. Messersmith, Science 2007,
318, 426–430; b) H. Lee, N. F. Scherer, P. B. Messersmith, Proc. Natl.
Acad. Sci. USA 2006, 103, 12999–13003; c) B. Malisova, S. Tosatti, M.
Textor, K. Gademann, S. Zꢃrcher, Langmuir 2010, 26, 4018–4026; d) Q.
Ye, F. Zhou, W. Liu, Chem. Soc. Rev. 2011, 40, 4244–4258; e) H. Gulley-
Stahl, P. A. Hogan, W. L. Schmidt, S. J. Wall, A. Buhrlage, H. A. Bullen, En-
viron. Sci. Technol. 2010, 44, 4116–4121; f) J. Saiz-Poseu, J. Sedꢇ, B.
Garcꢈa, C. Benaiges, T. Parella, R. Alibꢀs, J. Hernando, F. Busquꢀ, D. Ruiz-
Molina, Adv. Mater. 2013, 25, 2066–2070; g) J. Sedꢇ, J. Saiz-Poseu, F.
Busquꢀ, D. Ruiz-Molina, Adv. Mater. 2013, 25, 653–701; h) C. Creutz,
[10] K. Rosenlehner, T. Schunk, N. Jux, M. Brettreich, A. Hirsch, Org. Biomol.
Chem. 2008, 6, 2697–2705.
[11] a) E. R. Kleinfeld, G. S. Ferguson, Science 1994, 265, 370–373; b) Y. M.
Lvov, H. Haas, G. Decher, H. Mçhwald, Langmuir 1994, 10, 4232–4236;
c) V. V. Tsukruk, Adv. Mater. 1998, 10, 253–257; d) J. H. Fendler, Chem.
Mater. 1996, 8, 1616–1624; e) S. Srivastava, N. A. Kotov, Acc. Chem. Res.
2008, 41, 1831–1841; f) M. Shao, J. Han, W. Shi, M. Wei, X. Duan, Electro-
chem. Commun. 2010, 12, 1077–1080; g) A. R. G. Smith, J. L. Ruggles, A.
Yu, I. R. Gentle, Langmuir 2009, 25, 9873–9878.
[12] B. P. Pichon, P. Louet, O. Felix, M. Drillon, S. Begin-Colin, G. Decher,
Chem. Mater. 2011, 23, 3668–3675.
[13] N. A. Kotov, I. Dꢀkꢅny, J. H. Fendler, J. Phys. Chem. 1995, 99, 13065–
13069.
´
M. H. Chou, Inorg. Chem. 2008, 47, 3509–3514; i) I. A. Jankovic, Z. V.
ˇ
ˇ
´
´
Saponjic, M. I. Comor, J. M. Nedeljkovic, J. Phys. Chem. C 2009, 113,
12645–12652; j) D. S. Hwang, M. J. Harrington, Q. Lu, A. Masic, H. Zeng,
J. H. Waite, J. Mater. Chem. 2012, 22, 15530–15533.
[28] J.-F. Gnichwitz, R. Marczak, F. Werner, N. Lang, N. Jux, D. M. Guldi, W.
Peukert, A. Hirsch, J. Am. Chem. Soc. 2010, 132, 17910–17920.
[29] A. McKillop, J.-C. Fiaud, R. P. Hug, Tetrahedron 1974, 30, 1379–1382.
[30] a) J. S. Lindsey, I. C. Schreiman, H. C. Hsu, P. C. Kearney, A. M. Margueret-
taz, J. Org. Chem. 1987, 52, 827–836; b) J. S. Lindsey, R. W. Wagner, J.
Org. Chem. 1989, 54, 828–836; c) J. S. Lindsey, K. A. MacCrum, J. S. Ty-
honas, Y.-Y. Chuang, J. Org. Chem. 1994, 59, 579–587.
[14] J. Schmitt, G. Decher, W. J. Dressick, S. L. Brandow, R. E. Geer, R. Shashid-
har, J. M. Calvert, Adv. Mater. 1997, 9, 61–65.
[15] a) W.-H. Khoh, J.-D. Hong, Colloids Surf. A 2013, 436, 104–112; b) B. H. R.
Suryanto, X. Lu, C. Zhao, J. Mater. Chem. A 2013, 1, 12726–12731.
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
13
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[31] a) B. Neises, W. Steglich, Angew. Chem. 1972, 105, 1368–1376; b) B.
Neises, W. Steglich, Angew. Chem. 1978, 90, 556–557.
[35] V. V. Tsukruk, V. N. Bliznyuk, D. Visser, A. L. Campbell, T. J. Bunning, W. W.
Adams, Macromolecules 1997, 30, 6615–6625.
[32] a) G. R. Newkome, R. K. Behera, C. N. Moorefield, G. R. Baker, J. Org.
Chem. 1991, 56, 7162–7167; b) G. R. Newkome, A. Nayak, R. K. Behera,
C. N. Moorefield, G. R. Baker, J. Org. Chem. 1992, 57, 358–362; c) M.
Brettreich, A. Hirsch, Synlett 1988, 1396–1398.
[36] D. M. Guldi, F. Pellarini, M. Prato, C. Granito, L. Troisi, Nano Lett. 2002, 2,
965–968.
[33] J. Moser, S. Punchihewa, P. P. Infelta, M. Grꢆtzel, Langmuir 1991, 7,
3012–3018.
Received: August 29, 2014
[34] a) T. Imae, K. Muto, S. Ikeda, Colloid Polym. Sci. 1991, 269, 43–48;
b) D. L. Liao, G. S. Wu, B. Q. Liao, Colloids Surf. A 2009, 348, 270–275.
Published online on && &&, 0000
&
&
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
14
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
&
Solar Energy Conversion
A. Burger, R. D. Costa, V. Lobaz,
W. Peukert, D. M. Guldi, A. Hirsch*
&& – &&
Layer-by-Layer Assemblies of
Catechol-Functionalized TiO2
Nanoparticles and Porphyrins through
Electrostatic Interactions
Multilayer assemblies: Layer-by-layer
assemblies of positively charged TiO2
nanoparticles (NPs) and anionic por-
phyrins are applied in solar-energy con-
version schemes. TiO2 NPs are function-
alized with permanent positively
stable ethanolic NP suspensions. Alter-
nate deposition of these NPs and anion-
ic porphyrins through simple dip coat-
ing results in multilayer assemblies
formed through electrostatic attraction
(see figure).
charged catechol molecules, leading to
Chem. Eur. J. 2015, 21, 1 – 15
www.chemeurj.org
15
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ