Self-assembled multilayers of CdSe nanocrystals and carboxylate-handled phenylene-based molecules: Optical, electrochemical and photoconductive properties

Article abstract of DOI:10.1039/c0jm04214e

A series of linear phenylene-based molecules, bearing carboxylic acid or carboxylate functionalities at both ends, were reacted with (hexadecylamine/ stearate)-capped CdSe nanocrystals (7.5 nm diameter) dispersed in CHCl ₃ to form multilayers on ITO-glass surfaces via layer-by-layer (LBL) alternation. The new multilayered materials were investigated by UV-Vis spectroscopy, cyclic voltammetry, photoelectrochemistry and photoconductivity. The conjugated linkers, with reversible reduction potentials in the range -1.7 to -2.8 V vs. Ag/Ag⁺, ease charge transport between dots (LUMO at -1.4 V). Photoconduction is dominated by interdot distance and linker characteristics rather than by LUMO energy levels. The logarithm of photoconductivity and linker length are linearly related, as in conduction of molecular wires, and the tunnelling attenuation factor β value is 2.4 nm^-1. Comparison of n-type aromatics and p-type oligothiophenes is discussed.

Full text of DOI:10.1039/c0jm04214e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2011, 21, 8645
www.rsc.org/materials
PAPER
Self-assembled multilayers of CdSe nanocrystals and carboxylate-handled
phenylene-based molecules: optical, electrochemical and photoconductive
properties†
a
a
Barbara Vercelli, Gianni Zotti, Anna Berlin,b Mariacecilia Pasinic and Marco Natalid
*
Received 3rd December 2010, Accepted 28th March 2011
DOI: 10.1039/c0jm04214e
A series of linear phenylene-based molecules, bearing carboxylic acid or carboxylate functionalities at
both ends, were reacted with (hexadecylamine/stearate)-capped CdSe nanocrystals (7.5 nm diameter)
dispersed in CHCl3 to form multilayers on ITO-glass surfaces via layer-by-layer (LBL) alternation. The
new multilayered materials were investigated by UV-Vis spectroscopy, cyclic voltammetry,
photoelectrochemistry and photoconductivity. The conjugated linkers, with reversible reduction
potentials in the range ꢀ1.7 to ꢀ2.8 V vs. Ag/Ag+, ease charge transport between dots (LUMO
at ꢀ1.4 V). Photoconduction is dominated by interdot distance and linker characteristics rather than by
LUMO energy levels. The logarithm of photoconductivity and linker length are linearly related, as in
conduction of molecular wires, and the tunnelling attenuation factor b value is 2.4 nmꢀ1. Comparison
of n-type aromatics and p-type oligothiophenes is discussed.
The tunnelling rate G between two orbitals of NC neighbors
1. Introduction
can be approximately controlled by W and d, which are the
height of the tunnelling barrier and the shortest edge-to-edge
distance between the NCs (in fact in our case the length of the
rigid-rod linker molecule) respectively2,3 according to eqn (1):
In a previous investigation1 we have considered photogenerated
charge transport in layer-by-layer (LBL) structures of cadmium
selenide nanocrystals (CdSe-NCs) linked by two-ended carboxy-
lated oligothiophenes, where we found that the presence of such
p-type polyconjugated chains is, against expectations, in fact an
obstacle to the transport of photogenerated carriers.
G ¼ (e2/h) exp [ꢀ8p/h)(2mW)1/2d]
(1)
The scope of this investigation is then the selection of two-
ended carboxylic conjugated linkers with n-type conductive
properties to check whether they can effectively provide an effi-
cient charge transport while bridging CdSe-NCs.
The equation indicates that the tunnelling rate drops expo-
nentially with increasing separation between the NCs and
depends in a somewhat weaker way on the barrier height.
Reducing both parameters by proper design of the interparticle
molecular linkers will increase the strength of electronic coupling
between the NCs.4
In fact photoconduction in NCs should occur through
tunnelling of carriers. The efficiency of this tunnelling process
depends on the architecture of the molecular backbone, and
there are many intriguing effects that remain to be resolved
concerning the role of contacts, the electronic structures of the
molecules and the influence of the intermolecular interactions.
In this study we have selected short linkers with LUMO values
close to those of the CdSe-NCs.
The LUMO of CdSe-NCs (7.5 nm diameter) is located at ca.
ꢀ0.8 V vs. NHE5 (ꢀ3.7 eV vs. vacuum, ca. ꢀ1.4 V vs. Ag/Ag+).
We must recall that the presence of carboxylic moieties in the
linkers causes a lowering of the HOMO–LUMO system, as for
e.g. dihexylterthiophene the oxidation potential of which is posi-
tively shifted by 0.3 V by the presence of two carboxylic ends.1 It
will be anyway shown here that the introduction of two terminal
carboxylic heads does not dramatically shift the reduction
potential. Moreover, to ease electron mobility, the linkers should
not in principle bear electron acceptors too deeply localized.
Usually hole-transporting (p-type) semiconductors are
prepared from electron-donating p-systems, whereas electron-
transporting (n-type) semiconductors are prepared from elec-
tron-accepting ones. In recent reviews on charge transport in
aIstituto CNR per l0 Energetica e le Interfasi, C.o Stati Uniti 4, 35127
Padova, Italy. E-mail: g.zotti@ieni.cnr.it; Fax: (+39)49-8295853; Tel:
(+39)49-8295868
bIstituto CNR di Scienze e Tecnologie Molecolari, via C.Golgi 19, 20133
Milano, Italy
cIstituto CNR per lo Studio delle Macromolecole, via E.Bassini 15, 20133
Milano, Italy
dIstituto CNR di Chimica Inorganica e delle Superfici, c.o Stati Uniti 4,
35127 Padova, Italy
† Electronic supplementary information (ESI) available: Scheme (S1)
and FTIR spectrum (S2) of NaphCA4; synthesis of some linkers. See
DOI: 10.1039/c0jm04214e
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 8645–8652 | 8645

View Article Online
organic semiconductors6 it is evidenced that some of the most
widely investigated molecular semiconductors belong to few
families, namely for p-type we have oligoacenes, specially pen-
tacene and tetracene, oligothiophenes and polythiophenes, tri-
phenylamines and tetrathiafulvalenes, whereas n-type are
perylenes and fullerenes.6 Other n-type molecules previously used
in organic electronics are perylene 3,4,9,10-tetracarboxylic dia-
nhydride7 (or diimide8) and naphthalene 1,4,5,8-tetracarboxylic
dianhydride (or diimide9). Also azo-compounds may be candi-
dates, due to their accessible range of reduction potentials and
from the viewpoint of their action as electron bridges, for which
the literature provides good examples of their specific use.10,11
The n-type linkers considered in this report (shown in Charts 1
and 2) are the dicarboxylic or tetracarboxylic derivatives of linear
rigid-rod conjugated molecules, generally of the poly-p-phenyl-
ene structure, which determine in this way the interdot distance
of the substituted nanocrystals. The reduction redox potentials
of the main corresponding non-functionalized aromatics are
summarized in Table 1. The number of phenylene rings in line
varies from
1 (NaphCA4) to 2 (Ph2CA2, NPh2CA2,
Ph2N2CA2), 3 (Ph3F4CA2) and 4 (Ph2FlOCA2, Ph2FlCA2,
Ph4N2CA2). Ph2COCA4 and Ph2HFCA2, in which two phenyl
moieties are connected through an sp2-carbon (carbonyl) and an
sp3-carbon (hexafluoroisopropylidene) moiety respectively (so
that the conjugation within the linker is interrupted), have also
been considered for comparison. The previously investigated
terthiophene and sexithiophene linkers CAHe2T3CA and
CAHe4T6CA,1 here considered for comparison, are shown in
Chart 3.
Chart 2
advantage of the broad absorption of such relatively large NCs.
We used the layer-by-layer (LBL) method to assemble hybrid
organic/CdSe multilayer thin films from nonaqueous solutions of
the linkers and of CdSe-NCs. The multilayer buildup was
monitored by UV-Vis spectroscopy and the new materials were
investigated by cyclic voltammetry, UV-Vis and FTIR spec-
troscopy and photoconductivity.
In general no side alkyl substituent, commonly used to give
solubility to the linker itself, has been introduced so that the
nanocrystal surface density of linkers may be maximized and
conduction improved. Ph2FlCA2 is the exception used to test the
validity of the previous assumption.
2. Experimental
2.1. Chemicals and reagents
Among the linkers we selected also diphenyl- and nitro-
diphenyl-dicarboxylic acids, to check the action of a strong
electron-acceptor (–NO2) moiety in side position relative to the
conjugated backbone. Perylene tetracarboxylic derivatives, due
to their specific optical activity in the visible region of the solar
spectrum, will be the object of a separate investigation.
As in the previous investigation,1 (hexadecylamine/stearate)-
capped CdSe nanocrystals of 7.5 nm diameter were used to take
The solvents used in the reactions (Fluka) were absolute and
stored over molecular sieves. Acetonitrile was reagent grade
(Uvasol, Merck) with a water content <0.01%. The supporting
electrolytes tetrabutylammonium perchlorate (Bu4NClO4),
1,4,5,8-naphthalene tetracarboxylic anhydride, biphenyl-4,
40-dicarboxylic acid (Ph2CA2), 4,40-(hexafluoroisopropylidene)
bis(benzoic
acid)
(Ph2HFCA2),
3,30,4,40-benzophenone
a
Table 1 Reduction redox potentials (E0 ) in acetonitrile + 0.1 M
Bu4NClO4 and maximum absorption wavelength (lmax) of the main used
linkers and relevant non-carboxylated molecules
Compound
E0/V
lmax/nm
Ph2N2CA2
Azobenzene18
NaphCA4
ꢀ1.73
ꢀ1.75
—
ꢀ2.84
ꢀ2.37
ꢀ2.45
ꢀ1.70
ꢀ1.90
—
341b
317
310b
286
257b
252
246d
—
Naphthalene34
Ph2COCA4
Benzophenone
NPh2CA2
Nitrobiphenyl
Ph2CA2
281c
246
Biphenyl35
ꢀ3.05
a
vs. Ag/Ag+. b Dianion form in EtOH. c Acid form in DMF. d Acid form
in EtOH.
Chart 1
8646 | J. Mater. Chem., 2011, 21, 8645–8652
This journal is ª The Royal Society of Chemistry 2011

View Article Online
CdSe-NCs were used ca. 10ꢀ2 M (in CdSe units) in CHCl3.
NaphCA4 as bis(tetrabutylammonium) salt was used 10ꢀ3 M in
EtOH/hexane, Ph2HFCA2 and Ph2CA2 in the acid form 10ꢀ3
M
in EtOH or DMF. Ph2COCA4, NPh2CA2, Ph2N2CA2 were
used as the di(tetrabutylammonium) salt 10ꢀ3 M in EtOH. The
compounds Ph3F4CA2, Ph2FlOCA2, Ph2FlCA2, and
Ph4N2CA2 were used in the acid form 10ꢀ3 M (5 ꢁ 10ꢀ4 M for
Ph4N2CA2) in DMF.
2.3. Apparatus and procedure
Spectroscopy and electrochemistry. Electronic spectra were
Chart 3
obtained from
a Perkin-Elmer Lambda 15 spectrometer.
Absorbance data are given as measured, i.e. for the sum of the
ITO and glass sides, the same level of functionalization being
obtained at both sides of the electrode.
tetracarboxylic acid (Ph2COCA4), poly(acrylic acid) (PAAH)
and all other chemicals were reagent grade (Aldrich) and used as
received.
FTIR spectra were taken in transmission on KBr pellets using
a Perkin Elmer 2000 FTIR spectrometer. Electrochemistry was
performed at room temperature under nitrogen in three electrode
cells using 0.1 M Bu4NClO4 or NaClO4 as a supporting elec-
trolyte. The counter electrode was platinum; the reference elec-
trode was a silver/0.1 M silver perchlorate in acetonitrile (0.34 V
vs. SCE). The voltammetric apparatus (AMEL, Italy) included
a 551 potentiostat modulated by a 568 programmable function
generator and coupled to a 731 digital integrator. The working
electrode for cyclic voltammetry was a platinum or a glassy
carbon minidisc electrode (0.003 or 0.06 cm2 respectively). CV of
multilayers was performed on 1 ꢁ 4 cm2 ITO-glass samples with
Soluble CdSe-NCs with the surface capped by hexadecylamine
and stearic acid were produced as previously reported.12 Used
NCs display the first exciton peak at 645 nm, corresponding to an
average size of 7.5 nm.13
The compounds dimethyl 2-nitrobiphenyl-4,40-dicarboxy-
late,14 azobenzene-4,40-dicarboxylic acid (Ph2N2CA2),15 and
4,40-dibromoazobenzene16 were prepared according to literature
prescriptions. Poly(vinylbenzoic acid) (PVBH) was produced as
previously reported.12
1,4,5,8-Naphthalene tetracarboxylic anhydride was trans-
formed into the corresponding acid as bis(tetrabutylammonium)
salt (NaphCA4) (see Scheme S1† and FTIR spectrum S2†) by
reaction with a solution of tetrabutylammonium hydroxide in
EtOH. 2-Nitrobiphenyl-4,40-dicarboxylic acid (NPh2CA2) was
prepared by nitration of dimethyl 2-nitrobiphenyl-4,40-dicar-
boxylate,14 followed by hydrolysis (see ESI†). Dimethyl
an exposed area of 1 cm2 and at a scan rate of 0.02 V sꢀ1
.
Photoelectrochemistry. Photoelectrochemical (PEC) experi-
ments were performed in O2-saturated aqueous 0.1 M NaClO4
solution using a conventional three-electrode cell with a satu-
rated calomel electrode (SCE) as a reference and a platinum
counter electrode. The working electrode was illuminated on the
solution side with a water-filtered 100 W halogen lamp, spotted
over ca. 10 cm2. The resulting light power, calibrated with a Si
photodiode, was ca. 100 mW cmꢀ2. Light was chopped with
a manually driven shutter.
20,30,50,60-tetrafluoro-p-terphenyl-4,400-dicarboxylate,
2,7-bis-
2,7-bis-
(4-methoxycarbonyl-phenyl)-fluoren-9-one, and
(4-methoxycarbonyl-phenyl)-9,9-dioctyfluorene were obtained
by Suzuki coupling reaction of 4-methoxycarbonylphenylbor-
onic acid pinacol ester with the proper dibromo-derivative,
catalyzed by PdCl2(PPh3)2 or Pd(PPh3)4 (see ESI†). The acids
Ph3F4CA2 and Ph2FlOCA2 were prepared from the corre-
sponding dimethyl esters by hydrolysis with tetrabutylammo-
nium hydroxide in EtOH, while Ph2FlCA2 by hydrolysis of the
ester with KOH in dioxane/water (see ESI†). Finally 4,40-bis-
(4-carboxyphenyl)-azobenzene (Ph4N2CA2) was obtained by
Suzuki coupling reaction of 4-carboxyphenylboronic acid pina-
col ester with 4,40-dibromoazobenzene catalyzed by Pd(PPh3)4
(see ESI†).
Photoconductivity. Photoconductivity measurements of
multilayers were performed with a special Hg electrode con-
tacting the multilayer-covered ITO as described previously.12
Bias was applied to ITO vs. Hg electrode. Illumination
(100 mW cmꢀ2) was performed as described above on the back
glass side of the ITO/multilayer.
SEM and profilometry. Scanning electron microscopy (SEM)
was performed by means of a Philips Model XL40 recording
secondary electron images with a 10 KeV electron beam.
Multilayer thicknesses were determined with an Alphas-step IQ
profilometer from KLA Tencor.
2.2. Substrates and multilayer film formation
Transparent conducting surfaces were prepared from glass sheets
or indium-tin-oxide (ITO)-glass electrodes (20 U sqꢀ1 from
Kintec, Hong Kong). The ITO-glass electrodes were first modi-
fied with a monolayer of PVBH (or PAAH when specified)
10ꢀ3 M in EtOH, as reported previously.12
3. Results and discussion
3.1. Carboxylate-capping of CdSe-NCs
The build-up of multilayers was performed by the layer-by-
layer (LBL) methodology, i.e. by dipping the modified ITO-glass
electrodes alternatively into the solutions of the CdSe-NCs and
linker. Exposing time was 5 minutes.
The occurrence of extensive surface ligand substitution has been
checked with poly(acrylic acid) (PAAH).1 The polymer (1 mg)
dissolved in THF (5 ml) was added to 8 mg of CdSe-NCs in 5 ml
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 8645–8652 | 8647

View Article Online
of CHCl3 and reacted for some hours. After evaporation of the
solvent and washing with ethanol, the residue had become
insoluble in CHCl3, indicating the cross-linking of the NCs
occurred.
functionalized linker molecule and, since two carboxylic acid
moieties do shift oxidation potentials, it appears that the negative
charge on the carboxylate compensates the shift.
The FTIR spectrum of the product (Fig. 1) shows the almost
complete loss of the amine bands at 3300 and 1640 cmꢀ1 and the
strong enhancement of the band at 1560 cmꢀ1, due to the COOꢀ
asym stretching and of the corresponding weaker sym stretching
band at 1407 cmꢀ1, whereas a band at 1720 cmꢀ1, because of free
COOH moieties of PAAH, has appeared. In any case the
persistence of strong CH stretching bands around 2900 cmꢀ1
indicates that the stearate ligands, originally present along with
the hexadecylamine ligands, are essentially preserved. Through
the use of oligothiophene-based carboxylate-ended molecules it
has also been possible to evaluate the ratio (ca. 2 : 3) of initially
surface-bound stearate and new carboxylate caps.17 It can be
argued that the same amine displacement happens also with the
carboxyl-functionalized molecules used throughout this study.
3.3. Multilayers
Multilayers are built on PVBH-primed ITO via alternation of
CdSe-NCs and linker. UV-visible spectroscopy was used to
monitor the assembling process of such LBL films. In any case
the same results could be obtained from either PVBH- or PAAH-
primed ITO. Investigated multilayers were formed by 10 bilayers
and hereafter named simply (CdSe/linker)10.
Previous literature on the layer-by-layer alternation of semi-
conductor nanocrystals and organic oligomers or polymers was
given in part 1 of this series.12
Multilayer buildup. Considering (CdSe/Ph2N2CA2)n multi-
layers as a typical example, the growth, monitored by UV-vis
spectroscopy, is reproduced in Fig. 2 and it is apparently linear
with an optical differential growth of 7 ꢁ 10ꢀ3 au at 645 nm
(CdSe exciton) and 50 ꢁ 10ꢀ3 au at 330 nm (Ph2N2CA2). From
the extinction coefficients of CdSe-NCs13 and of azobenzene
(23 000 Mꢀ1 cmꢀ118) these data correspond to ca. 1.5 ꢁ 10ꢀ8 mol
(CdSe units) cmꢀ2 layerꢀ1 and 1 ꢁ 10ꢀ9 mol cmꢀ2 layerꢀ1
respectively, i.e. with a CdSe/Ph2N2CA2 molar ratio of 15.
Compared with the theoretical ratio of bulk CdSe (Nb) and
surface ferrocene (Ns) molecules, Nb/Ns ¼ ca. 10 for a 7.5 nm
CdSe-NC,1,12 this corresponds to a dense coverage of the CdSe-
NCs. Analogous results are obtained with Ph4N2CA2, which
shows its maximum red shift to 345 nm.
3.2. Linkers
Depending on the linker solubility, we have used either the acid
or the tetrabutylammonium salt for the multilayer build-up (see
Experimental section for details). In fact when both the acid and
the dianion forms could be used for the same linker, results in the
multilayer buildup were essentially the same.
In this and in the following sections we will designate with the
same denomination both the acid and the anion forms of the
linkers.
Reduction CV of the acid form of the linkers is not possible
either for insolubility or because the CV is dominated by the
reduction of the acidic protons. On the contrary the dianion
tetrabutylammonium salt is perfectly reducible.
Reduction potentials (E0) in acetonitrile + 0.1 M Bu4NClO4 of
the main used linkers (as bis(tetrabutylammonium) salt) and of
the relevant non-carboxylated molecules are summarized in
Table 1 along with their maximum absorption wavelengths.
These data are required for the calculation of HOMO and
LUMO energy levels.
In the absence of a specific layer-by-layer tracking, which is
out of the scope of the work, the analysis conducted here is to be
considered a purely phenomenological one. Therefore, no final
conclusion on the structure of the multilayer build-up of the films
can be drawn.
Results in terms of optical growth, i.e. two-side CdSe exciton-
band differential absorbance, and thickness of multilayers are
summarized in Table 2.
Multilayer thickness. Under SEM examinations the multi-
layers appear uniform; moreover they are robust enough to
withstand a standard sticky-tape test without appreciable losses.
A rough estimation of the thickness of (CdSe/Ph2N2CA2)10
multilayers has been performed by profilometry and a value of
The CV results show that the two carboxylate ends of the
anions do not modify strongly the energy levels of the non-
Fig.
2 UV-Vis spectra of ITO/(PVBH/CdSe)/(Ph2N2CA2/CdSe)n
Fig. 1 FTIR spectrum of CdSe-NCs (a) before and (b) after PAAH
multilayers (n ¼ 0–4). Spectra are background-corrected.
treatment.
8648 | J. Mater. Chem., 2011, 21, 8645–8652
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Table 2 Two-side CdSe-NC exciton-band differential absorbance (DA)
and total thickness (d) of (CdSe/linker)10 multilayers on PVBH-primed
ITO using different linkers. Linker coverage of CdSe-NCs (G) is also
given
DA/10ꢀ3
au layerꢀ1
G/10ꢀ10 mol
cmꢀ2 layerꢀ1
Linker
d/nm
NaphCA4
Ph2CA2
3.0
4.0
5.0
7.0
6.0
5.0
4.0
5.0
5.0
6.0
25
35
40
55
50
55
40
55
70
50
—
—
9–13
10
7
20
—
3
NPh2CA2
Ph2N2CA2
Ph2COCA4
Ph2HFCA2
Ph3F4CA2
Ph2FlOCA2
Ph2FlCA2
Ph4N2CA2
—
10
Fig. 3 Single sweep reduction voltammograms of ITO/(PVBH/CdSe/
(NPh2CA2/CdSe)n multilayers (n ¼ 1–4) in acetonitrile + 0.1 M NaClO4.
Scan rate: 0.02 V sꢀ1
.
The above procedures give the total amount of the linkers
included in the films. The reported linker contents per layer are to
be considered as average estimates.
55 ꢂ 5 nm has been found, in good agreement with that expected
from particle size and packing mode.12 For the multilayers here
reported in general the thickness is in the range 25–55 nm, with
a fair correspondence of thickness and CdSe-NC exciton band
absorbance. The only exception is constituted by the (CdSe/
Ph2FlCA2)10 multilayer, with a record thickness of 70 nm,
evidence of a high content of ordered linkers, probably packed
by self-assembly promoted by the dialkyl chains central to the
linker molecule.
After a comparison of the results in Table 2, it is clear that short
linkers produce dense layers, between 7 and 20 ꢁ 10ꢀ10 mol cmꢀ2
layerꢀ1, whereas long linkers may form less dense layers, ca.
3 ꢁ 10ꢀ10 mol cmꢀ2 layerꢀ1, as a consequence of the higher steric
demand of their longer conjugated chain.
3.4. Photoconductivity
Linker content of multilayers. Different from (CdSe/
Ph2N2CA2)10 and (CdSe/Ph4N2CA2)10 multilayers, the other
linkers do not show up in the UV-vis spectrum, due to their
higher energies of absorption and the limitations given by the
The original photoluminescence of the CdSe-NCs is almost
completely quenched in all our multilayers, which in principle
favors photocarrier transport.
ITO/glass substrate. Yet the linker coverage, as mol cmꢀ2 layerꢀ1
,
The photoconductivity in thin films of CdSe-NCs has been
extensively studied in the past.20 In close-packed NC films a high
surface trap density inhibits charge transport,20b which is
expected to be limited by tunneling between dots and therefore to
have a strong dependence on the interdot spacing. In fact the
photoconductivity in CdSe-NCs increases by decreasing inter-
particle spacing and improving passivation of surface trap
sites.20c
could be obtained in general by dissolving the multilayer in
concentrated hydrochloric acid (1 ml, ca. 10ꢀ5 M concentration
of the linker) and measuring the absorption at the maximum
wavelength specific to the linker, after subtraction of the minor
contribution to absorbance arising from the dissolved ITO and
PVBH primer layer. Table 2 summarizes these results. Such
determination could not be performed with linkers such as
NaphCA4, Ph2CA2, Ph2FlCA2 and Ph3F4CA2 due to forma-
tion of aggregates.
Under illumination and applied bias the investigated multi-
layers display photocurrents attributable to field-assisted ioni-
zation of the hole–electron pair and tunnelling of the charge
carrier (electron) from the NC, in which it is created, to
a neighboring one.21 Photocurrents of the various (CdSe/
linker)10 multilayers were recorded at 1 V bias and 100 mW cmꢀ2
illumination (Fig. 4). Slow response times of some seconds are
evident. Table 3 summarizes the solid-state photocurrent values,
both absolute and, given the proportionality with thickness (i.e.
CdSe-NC content), referred to a multilayer of thickness 55 nm,
for multilayers of the different linkers.
The linker content of the NPh2CA2 layers has been evaluated
also electrochemically by means of the NO2-reduction process.
The compound is in fact reduced reversibly at E0 ¼ ꢀ1.70 V when
dissolved in acetonitrile + 0.1 M Bu4NClO4, yet no reduction is
revealed in the multilayers. This is possibly due to the hindered
penetration of the bulky tetrabutylammonium cation with
reduction of the nitro moiety to anion. In fact passing to 0.1 M
NaClO4 as a supporting electrolyte, the layers are on the contrary
reduced at Ep ¼ ꢀ1.4 V (anticipated by anion stabilization
operated by the sodium cation and follow-up protonation–
reduction reactions) with a charge proportional with the number
of layers (Fig. 3). The process is irreversible due to dissolution of
the structure. The passed charge is ca. 250 mC cmꢀ2 layerꢀ1, which,
with the reasonable assumption of a two-electron reduction
process,19 would correspond to ca. 13 ꢁ 10ꢀ10 mol cmꢀ2 layerꢀ1, i.e.
a value comparable with that measured spectrophotometrically
(9 ꢁ 10ꢀ10 mol cmꢀ2 layerꢀ1) on account of some background
current contributions to the electrochemical result.
It is interesting to note that substantial photocurrent levels (in
the range 1–25 mA cmꢀ2) are obtained for all the investigated
linkers, with the only significant exception of the conjugation-
interrupted Ph2HFCA2 layers, for which no measurable signal
was obtained. The intensity does not appear to be in any case
correlated with the linker LUMO level, such as in the case of the
diphenyl linker Ph2CA2, with the highest LUMO level (above
1 eV higher than the CdSe-NC LUMO) and a response among
the strongest. The reason for this may be found in an electric-field
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 8645–8652 | 8649

View Article Online
Fig. 4 Photocurrent transients of a (Ph2N2CA2/CdSe)10 multilayer at
1 V applied voltage and 100 mW cmꢀ2 illumination.
Fig. 5 ln of photocurrent vs. calculated linker length. Line: linear fit with
slope ¼ ꢀ2.4(ꢂ0.2) nmꢀ1; R ¼ ꢀ0.978.
and for photoinduced electron transfer.25 Below the critical
length, which is also the condition of our linkers, the measured
resistance R increases exponentially with the molecular length d23
as described by nonresonant tunnelling equation (2):
Table 3 Absolute solid-state photocurrents iph and relative solid-state
photocurrents iph* (i.e. referred to a compact multilayer of thickness
55 nm, see text) for (CdSe/linker)10 multilayers of thickness d under
100 mW cmꢀ2 illumination (at 1 V applied potential on 0.023 cm2 ITO–
Hg contacts). Calculated linker length (l) is also given
R ¼ R0 exp (bd)
(2)
Linker
d/nm
iph/mA cmꢀ2
iph*/mA cmꢀ2
l/nm
where R0 is the effective contact resistance and b is the tunnelling
attenuation factor. The b value obtained from a linear fit is
2.2 nmꢀ1, i.e. very close to the value we obtained for our linkers,
suggesting a very efficient electron tunnelling through the all-
aromatic backbones of the molecules. Analogous conclusions
were drawn by a subsequent paper on oligofluorene chains,
where the attenuation factor was 2.0, i.e. practically the same.26
In conclusion, passing from thiophene- to phenylene-based
chains, i.e. from p- to n-type semiconductors, has given the
hoped positive effect on the photoconductive properties of these
hybrid structures. Clear evidence of this is given by Scheme 1:
molecular modelling gives a comparable length to Ph2N2CA2
(top) and CAHe2T3CA (bottom), but the latter blocks photo-
carrier transport, which on the contrary is easy with the former.
CAHe4T6CA1
CAHe2T3CA1
NaphCA4
Ph2CA2
NPh2CA2
Ph2N2CA2
Ph2COCA4
Ph2HFCA2
Ph3F4CA2
Ph2FlOCA2
Ph2FlCA2
Ph4N2CA2
35
30
25
35
40
55
50
55
40
55
70
50
<0.05
<0.05
10
5
4
3
1.5
<0.1
<0.1
25
8
5.5
3
1.5
<0.1
3
2.53
1.34
0.58
1.02
1.02
1.21
1.07
0.99
1.46
1.85
1.85
1.92
<0.05
2
1
1
1
1
1
1
dependence of the barrier height which tends to level off the
barrier height contribution to the tunneling rate, or more simply
in an independence of the barrier height from the LUMO levels.
In fact, if we consider the calculated linker lengths, given in the
last column of Table 3, the dominant factor appears to be the
interdot distance. It can thus be observed that the side nitro
group (NPh2CA2) decreases conductivity only moderately, the
carbonyl group (Ph2COCA4) creating instead a higher obstacle
to charge transport. The azo-group acts simply as an additional
spacing element like a phenylene moiety.
3.5. Photoconduction in NC solids
Primary and secondary photocurrents. Photoelectrochemistry
is a powerful tool for the investigation of the charge transfer
ability of semiconductor (including CdSe-NCs) layers.
In part 1 of this series,12 the photoelectrochemical properties of
multilayers constituted of CdSe-NCs and different polymers
In Fig. 5 the photocurrent values (as natural logarithm) are
plotted vs. the calculated linker length d (the anomalous cases of
Ph2HFCA2 and Ph2COCA4 are excluded from the plot). The
relationship is fairly linear, thus corresponding to the theoretical
expectation of eqn (1) (with a constant barrier height W).
The slope of the plot (ꢀ2.4 nmꢀ1) is in fact somewhat lower
than the attenuation factor b in molecular electronic conduction,
generally found in the past in the range 4–6 nmꢀ1 for oligophe-
nylenes.22 Yet the length-dependent conduction of very long (up
to 40 nm) phenylene-based conjugated molecular wires con-
nected to metal electrodes has been more recently investi-
gated.23,24 Above a critical length around 4 nm the conduction
mechanism switches from tunnelling to hopping24 as from theory
Scheme 1
8650 | J. Mater. Chem., 2011, 21, 8645–8652
This journal is ª The Royal Society of Chemistry 2011

View Article Online
were investigated by use of a liquid junction, containing dioxygen
as a photoelectron scavenger. The main result was that the
transport ability depends on the type of coordinating ligand
present in the polymer.
4. Conclusions
Layer-by-layer structures constituted by CdSe quantum dots
linked by electron-transporting conjugated molecules, of various
lengths and conjugation patterns, have been produced and their
photoconductive properties investigated.
The linkers used in the present investigation bear carboxylic
functions and, as expected, following the same experimental
conditions of the previous study,12 provide the same photocur-
rent values. Values of 10–20 mA cmꢀ2 (measured at ꢀ0.5 V from
the onset potential) were found.
Photoconduction in these structures is high, in contrast with
that encountered in those formed by (p-type) oligothiophenes,
thus evidencing the n-type character of charge transport.
Moreover the photoconduction is dominated by interdot
distance and linker characteristics rather than by LUMO energy
levels. Photoconductivity and linker length are exponentially
related, like in the conduction of molecular wires, and the
tunnelling attenuation is comparable with other recently
produced phenylene-based molecular wires.
It may be surprising that photocurrents generated on the same
multilayer in the photoelectrochemical and in the solid-
state configuration are strikingly different, passing from
10–20 mA cmꢀ2 to 10–15 mA cmꢀ2 (i.e. with a 1000-fold gain)
under the same illumination. The explanation is in the different
contacts and is related to charge trapping.
Such structures, which display optical and electronic proper-
ties useful in photovoltaic devices, are still under our attention
with a particular focus on the linker core, which may be also light
absorbing, and head, the nature of which also may be crucial for
an efficient transfer between quantum dots.
Unpassivated states of the NC surface may trap electrons and
holes and when one charge carrier is trapped, the other carrier
cycles through the circuit until it recombines with the trapped
carrier.21 The enhancement of the photoconductance depends on
the nature of the contacts between the electrodes and the NC
film. If the contacts are blocking, then the maximum photo-
conductive gain is one electron per absorbed photon, and the
photocurrent is called primary. If the contact is ohmic or
injecting, which is the case of our solid-state device, then the gain
can be much greater than unity and the photocurrent is called
secondary.27
References
1 G. Zotti, B. Vercelli, A. Berlin, M. Pasini, T. L. Nelson,
R. D. McCullough and T. Virgili, Chem. Mater., 2010, 22, 1521.
2 T. Mokari, E. Rothenberg, I. Popov, R. Costi and U. Banin, Science,
2004, 304, 1787.
3 C. B. Murray, C. R. Kagan and M. G. Bawendi, Annu. Rev. Mater.
Sci., 2000, 30, 545.
In the case of secondary (trap-induced) photocurrents the
recombination lifetime is limited by the release rate of the trap.
This obviously results in slow response times, frequently higher
than some seconds,28 which is our actual case (Fig. 4).
4 D. Yu, C. Wang and P. Guyot-Sionnest, Science, 2003, 300, 1277.
5 C. J. Wang, M. Shim and P. Guyot-Sionnest, Science, 2001, 291, 2390.
6 (a) A. R. Murphy and J. M. J. Frechet, Chem. Rev., 2007, 107, 1066;
(b) V. Coropceanu, J. Cornil, D. A. da Silva Filho, Y. Olivier,
R. Silbey and J. L. Bredas, Chem. Rev., 2007, 107, 926.
7 J. G. Laquindanum, H. E. Katz, A. Dodabalapur and A. J. Lovinger,
J. Am. Chem. Soc., 1996, 118, 11331.
8 P. R. L. Malenfant, C. D. Dimitrakopoulos, J. D. Gelorme,
L. L. Kosbar and T. O. Graham, Appl. Phys. Lett., 2002, 80, 2517.
9 H. E. Katz, J. Johnson, J. Andrew and W. Li, J. Am. Chem. Soc.,
2000, 122, 7787.
The blocking action of oligo(poly)thiophenes. Organic semi-
conductors are commonly classified as either p-type (hole-con-
ducting) or n-type (electron-conducting) depending on which
type of charge carrier is more efficiently transported through the
material.
The photoconductance in CdSe-NCs is likely due to mobile
electrons rather than holes4,29 so that CdSe-NCs should behave
as n-type semiconductors. Thus it is conceivable that n-type
linkers are mostly efficient in favoring charge transport in CdSe-
NCs. As a consequence the results are negative with (p-type)
oligothiophenes but positive with (n-type) oligophenylenes.
One possibility is that in the case of oligothiophenes photo-
current is mainly dominated by the light induced enhancement
effect of electron tunnelling,30 while in the case of oligopheny-
lenes, both the aforementioned effect and the n-type linkers play
roles in the photoconductivity.
10 S. A. P. Guarin, S. Dufresne, D. Tsang, A. Sylla and W. G. Skene, J.
Mater. Chem., 2007, 17, 2801.
11 D. I. Schuster, K. Li, D. M. Guldi, A. Palkar, L. Echegoyen,
C. Stanisky, R. J. Cross, M. Niemi, N. V. Tkachenko and
H. Lemmetyinen, J. Am. Chem. Soc., 2007, 129, 15973.
12 G. Zotti, B. Vercelli, A. Berlin, P. T. K. Chin and U. Giovanella,
Chem. Mater., 2009, 21, 2258.
13 W. W. Yu, L. Qu, W. Guo and X. Peng, Chem. Mater., 2003, 15, 2854.
14 V. K. Ol’klovik, A. A. Pap, V. A. Vasilevskii, N. A. Galinovskii and
S. N. Tereshko, Russ. J. Org. Chem., 2008, 1172.
15 M. L. Tomlinson, J. Chem. Soc., 1946, 756.
16 T. Yamamoto, I. Nishiyama and H. Yokoyama, Chem. Lett., 2007,
1108.
17 G. Zotti, B. Vercelli, A. Berlin, unpublished.
About the origin of the p-type character of oligothiophenes,
which does not appear to be accounted for in the literature, we
present here a possible explanation. The conductivity of poly-
conjugated polymers in general is expected to be comparable for
p- and n-doping, as e.g. in polyacetylene.31 Yet in polythiophenes
the n-conductivity is ca. 100–300 times lower.32 Localization of
charge in oligothiophenes may in principle be traced to the
different localization of the LUMO (on the sulfur atoms) and of
the HOMO (on the polyacetylene CH backbone) orbitals.33 Thus
the blocking action of oligothiophenes in the photoconduction in
CdSe-NCs may be traced to a strong localization of photo-
generated electrons.
18 (a) G. H. Aylward, J. L. Garnett and J. H. Sharp, Anal. Chem., 1967,
39, 457; (b) J. L. Sadler and A. J. Bard, J. Am. Chem. Soc., 1968, 90,
1979.
19 H. Lund, in Organic Electrochemistry, ed. H. Lund and M. M. Baizer,
M. Dekker, New York, 1991, p. 401.
20 (a) C. A. Leatherdale, C. R. Kagan, N. Y. Morgan, S. A. Empedocles,
M. A. Kastner and M. G. Bawendi, Phys. Rev. B: Condens. Matter,
2000, 62, 2669; (b) D. S. Ginger and N. C. Greenham, J. Appl.
Phys., 2000, 87, 1361; (c) M. V. Jarosz, V. J. Porter, B. R. Fisher,
M. A. Kastner and M. G. Bawendi, Phys. Rev. B: Condens.Matter
Mater. Phys., 2004, 70, 195327; (d) V. J. Porter, T. Mentzel,
S. Charpentier, M. A. Kastner and M. G. Bawendi, Phys. Rev. B:
Condens. Matter Mater. Phys., 2006, 73, 155303.
21 V. J. Porter, S. Geyer, J. E. Halpert, M. A. Kastner and
M. G. Bawendi, J. Phys. Chem. C, 2008, 112, 2308.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 8645–8652 | 8651

View Article Online
22 A. Helms, D. Heller and G. McLendon, J. Am. Chem. Soc., 1992, 114,
6227.
P. Ascarelli, E. Cappelli and D. Trucchi, Diamond Relat. Mater.,
2003, 12, 696.
23 L. Luo and C. D. Frisbie, J. Am. Chem. Soc., 2010, 132, 8854.
24 (a) S. H. Choi, B. Kim and C. D. Frisbie, Science, 2008, 320, 1482; (b)
S. H. Choi, C. Risko, M. C. Ruiz Delgado, B. Kim, J. L. Bredas and
C. D. Frisbie, J. Am. Chem. Soc., 2010, 132, 4358; (c) N. Tuccitto,
V. Ferri, M. Cavazzini, S. Quici, G. Zhavnerko, A. Licciardello and
M. A. Rampi, Nat. Mater., 2009, 8, 41.
28 R. H. Bube, Photoconductivity of Solids, Wiley, New York, 1960.
29 (a) X. G. Peng, M. C. Schlamp, A. V. Kadavanich and
A. P. Alivisatos, J. Am. Chem. Soc., 1997, 119, 7019; (b) D. Yu,
B. L. Wehrenberg, P. Jha, J. Ma and P. Guyot-Sionnest, J. Appl.
Phys., 2006, 99, 104315; (c) X. C. Yang, C. C. Xu and N. C. Giles,
J. Appl. Phys., 2008, 104, 073727.
25 (a) B. Giese, J. Amaudrut, A. K. Kohler, M. Spormann and
S. Wessely, Nature, 2001, 412, 318; (b) B. Giese, Acc. Chem. Res.,
2000, 33, 631; (c) D. Segal, A. Nitzan, W. B. Davis,
M. R. Wasielewski and M. A. Ratner, J. Phys. Chem. B, 2000, 104,
3817; (d) W. B. Davis, W. A. Svec, M. A. Ratner and
M. R. Wasielewski, Nature, 1998, 396, 60.
26 T. Hines, I. Diez-Perez, J. Hihath, H. Liu, Z. S. Wang, J. Zhao,
G. Zhou, K. Mullen and N. Tao, J. Am. Chem. Soc., 2010, 132, 11658.
27 (a) T. J. McMahon and R. S. Crandall, Phys. Rev. B, 1989, 39, 1766;
(b) M. Yamaguchi, M. Kondo, K. Hayashi, H. Nishio and
Y. Tawada, J. Non-Cryst. Solids, 1991, 137–138, 1305; (c)
M. C. Rossi, F. Spaziani, G. Mazzeo, S. Salvatori, G. Conte,
30 X. Zhang, Y. Wang, Y. Ma, Y. Ye, Y. Wang and K. Wu, Langmuir,
2006, 22, 344.
31 A. G. McDiarmid and A. J. Heeger, Synth. Met., 1979, 1, 101.
32 (a) K. Kaneto, S. Ura, K. Yoshino and Y. Nuishi, Jpn. J. Appl. Phys.,
1984, 23, L109; (b) R. M. Crooks, O. M. R. Chyan and
M. S. Wrighton, Chem. Mater., 1989, 1, 2; (c) G. Zotti and
G. Schiavon, Synth. Met., 1994, 63, 53.
33 A. K. Bakhshi, J. Ladik and M. Seel, Phys. Rev. B, 1987, 35, 704.
€
34 A. C. Aten, C. Buthker and G. J. Hoijtink, Trans. Faraday Soc., 1959,
55, 324.
35 C. K. Mann and K. K. Barnes, in Electrochemical Reactions in
Nonaqueous Systems, M. Dekker, New York, 1970.
8652 | J. Mater. Chem., 2011, 21, 8645–8652
This journal is ª The Royal Society of Chemistry 2011