Stabilization and activation of Pd nanoparticles for efficient CO2-reduction: Importance of their generation within supramolecular network of tridentate Schiff-base ligands with N,N coordination sites

Article abstract of DOI:10.1016/j.electacta.2021.138550

Tridentate Schiff base ligand L (L = C₁₄H₁₂N₂O₃) reacts with palladium(II) chloride, to give a complex of formula [Pd(C₁₄H₁₂N₂O₃)Cl₂]₂?MeOH. The

Full text of DOI:10.1016/j.electacta.2021.138550

Journal Pre-proof
Stabilization and Activation of Pd Nanoparticles for Efficient
CO -Reduction: Importance of their Generation within
2
Supramolecular Network of Tridentate Schiff-Base Ligands with N,N
Coordination Sites
Anna Wadas , Adam Gorczynski , Iwona A. Rutkowska ,
Ewelina Seta , Ewelina Szaniawska , Maciej Kubicki ,
Adam Lewera , Maciej Gorzkowski , Aneta Januszewska ,
Rafal Jurczakowski , Barbara Palys , Violetta Patroniak ,
Pawel J. Kulesza
PII:
S0013-4686(21)00840-9
DOI:
Reference:
https://doi.org/10.1016/j.electacta.2021.138550
EA 138550
To appear in:
Electrochimica Acta
Received date:
Revised date:
Accepted date:
8 February 2021
16 April 2021
1 May 2021
Please cite this article as: Anna Wadas , Adam Gorczynski , Iwona A. Rutkowska , Ewelina Seta ,
Ewelina Szaniawska , Maciej Kubicki , Adam Lewera , Maciej Gorzkowski , Aneta Januszewska ,
Rafal Jurczakowski , Barbara Palys , Violetta Patroniak , Pawel J. Kulesza , Stabilization and Ac-
tivation of Pd Nanoparticles for Efficient CO -Reduction: Importance of their Generation within
2
Supramolecular Network of Tridentate Schiff-Base Ligands with N,N Coordination Sites, Electrochim-
ica Acta (2021), doi: https://doi.org/10.1016/j.electacta.2021.138550
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2021 Published by Elsevier Ltd.

Highlights
0
II




Pd is generated upon electroreduction of Pd ions in the supramolecular complex
0
Pd catalytic centers are stabilized and activated by ligand N,N coordination sites
Thus generated Pd nanoparticles exhibit high activity during CO -electroreduction
2
CO and H are the main reduction products in CO -saturated aqueous KHCO
3
2
2
1

Stabilization and Activation of Pd Nanoparticles for Efficient CO2-
Reduction: Importance of their Generation within Supramolecular
Network of Tridentate Schiff-Base Ligands with N,N Coordination Sites
a
b
a
a
Anna Wadas , Adam Gorczynski , Iwona A. Rutkowska , Ewelina Seta ,
a
b
a
a
Ewelina Szaniawska , Maciej Kubicki , Adam Lewera , Maciej Gorzkowski ,
a
a
a
b
Aneta Januszewska , Rafal Jurczakowski , Barbara Palys ,Violetta Patroniak *,
a
Pawel J. Kulesza *
a
Faculty of Chemistry and Biological and Chemical Research Centre,
University of Warsaw, Pasteura 1, 02-093 Warsaw, Poland
b
Faculty of Chemistry, Adam Mickiewicz University,
Umultowska 89b, 61-614 Poznan, Poland
_
________________________
*Corresponding Authors:
pkulesza@chem.uw.edu.pl (Pawel J. Kulesza)
violapat@amu.edu.pl (Violetta Patroniak)
2

Abstract
Tridentate Schiff base ligand L (L = C H N O ) reacts with palladium(II) chloride, to give a
1
4
12
2
3
complex of formula [Pd(C H N O )Cl ] ∙MeOH. The spectroscopic, microanalytical and X-
14
12
2
3
2 2
ray crystallographic data are consistent with the view that palladium(II) ion is coordinated by
II
two nitrogen donor atoms, with typical Pd -N distances and two chloride ligands,
0
representing thus N,N instead of possible N,N,O coordination mode. The reduced Pd
catalytic sites are electrochemically generated (by voltammetric potential cycling) within the
film of supramolecular network of tridentate Schiff base ligands. The IR and XPS data of the
electroreduced material are consistent with the existence of interactions between the
palladium catalytic centers and the network of supramolecular ligands with N,N coordination
sites. In comparison to conventional Pd nanoparticles, the CO -reduction current densities,
2
which have been recorded in near-neutral KHCO3 both under voltammetric and
chronoamperometric conditions and normalized against electrochemically active surface area
of palladium catalyst, are significantly higher at the supramolecular-ligand-supported Pd
nanoparticles.
Key-words:
palladium(II)
supramolecular
complex,
tridentate-Schiff-base-ligand,
palladium(0) catalytic centers, carbon dioxide reduction, metal support interactions
3

1
. Introduction
The continuously rising levels and accumulation of emitted carbon dioxide have led to
severe environmental problems and concerns over climate change and possibility of global
warming [1]. Low-temperature electrochemical (electrocatalytic) promotion to reduction of
carbon dioxide is a promising approach to convert CO and H O to useful chemical
2
2
feedstocks that include syngas (CO + H ), an intermediate for further Fischer–Tropsch
2
processes, carbon-based simple organic fuels, or utility chemicals [2-14]. But the CO2
reduction reactions (CO RRs) are multi-electron and multi-proton processes with considerable
2
kinetic barriers (overpotentials) requiring development of the carefully designed
electrocatalytic materials. In comparison to electrochemical water splitting and hydrogen
production by electrolysis, feasibility of the efficient electrolytic CO -conversion is far away
2
from being practical [10].
The first step in CO RR involves electrocatalytic activation of the CO molecule to
2
2
.-
form CO2 radical which is believed to be energetically demanding (-1.9 V vs. standard
hydrogen electrode, SHE) due to high stability of the CO structure itself [2-8]. A rational
2
model takes into account strong adsorptive (chemical) interactions of carbon dioxide with the
.-
catalytic surface thus stabilizing the CO2 radical, or the reaction intermediate, and shifting
the CO -electroreduction potential toward less negative values [4,7,10]. Copper-based
2
catalysts have attracted special attention due to their unique properties permitting CO RR to
2
oxo-hydro-carbons (e.g., alcohols, aldehydes) and hydrocarbons with substantial Faradaic
efficiencies [3,4,7,15-20]. Because the CO RR products mentioned above are in the state of
2
further reduced in comparison to carbon monoxide or formate, the actual electroreductions are
expected to involve more than two electrons per CO molecule. On the whole, by choosing an
2
appropriate catalyst, or by proper designing its surface structure (morphology, lattice strain,
4

composition or intentional modification) [2,4,8], the two-electron CO RR to carbon monoxide
2
or formic acid becomes feasible even at low overpotentials.
Palladium-based electrocatalysts exhibit capability to convert selectively CO to CO or
2
formate [21-30], thus permitting clean production of syngas or electrosynthesis of formic acid
which of importance to preparation of fine chemicals, including pharmaceuticals, and to
application in direct formic acid fuel cells. The performance of Pd catalysts is highly
dependent on interfacial phenomena and the nature of binding of CO to Pd surfaces, namely
2
on the formation of intermediates such as CO and COOH [22,26]. Among important issues
*
*
is the structure of Pd active sites which reflects its size, morphology, grain boundary,
composition or even the presence of active ligands. Furthermore, distinctive interactions of
carbon dioxide with palladium affecting both absorbing and adsorbing of hydrogen are known
to influence both activity and selectivity of CO RR with respect to hydrogen evolution.
2
Like platinum, palladium can be considered as one of the most important metals in
catalysis, also for electrochemical reactions, including oxygen reduction, formic acid
oxidation, etc. [31-35]. In a case of the oxidation of HCOOH, Pd shows an excellent
performance, even superior, in comparison to Pt. Despite the ability of palladium to both
adsorb and absorb hydrogen readily, the chemical interactions between CO and the adsorbed
2
monoatomic H are rather weak, at least in comparison to those exiting on platinum surfaces.
Indeed, surface concentration of the CO -reduction products appearing in the hydrogen
2
(UPD-H) region on Pd is more than 50 times lower than that on Pt [36,37]. Also, no
electrolyte-soluble products of the interaction of CO with the Pd-electrode surface are
2
formed in the UPD-H potential range. Obviously, the nature and effectiveness of adsorptive
interactions between CO and the catalytic metal surface play affects the CO RR pathways.
2
2
For example, the interfacial reduction of CO to CO on platinum inhibits further CO -
2
2
reduction in favor of the water reduction and generation of hydrogen [38]. Because the
5

adsorption affinity of the surface ―reduced-CO ‖ is weaker on palladium, in comparison to
2
platinum (or even other group VIIIB metals including Fe and Ni), the self-poisoning of Pd
surfaces becomes less pronounced and the CO-type reduced species tend to act as the reaction
intermediates [39,40]. Furthermore, the formation of substantial amounts of the bulk absorbed
hydrogen on Pd affects the adsorption affinity of the species originating from CO RR while
2
decreasing the competitive hydrogen evolution [40].
The present work follows our initial observation about possibility of electrochemical
0
generation of nanostructured Pd within the coordination architecture of tridentate Schiff-
II
base-ligands of the Pd -complex, [Pd(C H N O )Cl ] ∙MeOH [27]. The Schiff base ligands
14
12
2
3
2 2
with N,O-donors are employed in the metal-ion-directed assemblies of coordination
architectures [41,42]. These ligands are fascinating and versatile complexing agents because
their metal complexes are known to serve as luminescent probes [43], as catalysts in organic
syntheses [44] and for specific DNA cleavage reactions [45], as well as in antibacterial [46]
and anticancer [47] drugs. The well-defined Schiff base ligands containing nitrogen donor
atoms are considered as ‗privileged ligands‘ because of their capability to stabilize different
metals in various oxidation states, and their complexes are extensively studied due to their
synthetic flexibility, selectivity and sensitivity toward a variety of metal ions [48]. The
structural diversity of palladium-driven polynuclear species [49] have permitted self-
assembling supramolecular architectures of different nuclearity [50]. The palladium
complexes with Schiff bases are known to exhibit high activities toward addition
polymerization of norbornene [51], and to function as promising alternatives for
phosphine/carbene species exhibit air and moisture stability and high catalytic activity in
Heck reaction [52,53]. Furthermore, mononuclear complexes of Pd(II) ions with
thiosemicarbazone ligands have been found to act as antitumour agents [54] or they show a
groove binding mode to CT-DNA [55]. Following the successful synthesis of complexes of
6

Ag(I), Cu(II), Co(II) and Zn(II) [56] with ligand L (Figure 1), we have concentrated on the
preparation of an analogous complex with palladium(II), initially with the metal-to-ligand (L)
stoichiometry of 1:2 but, later, of the system containing only one molecule of ligand and two
chloride ions [57-59]. Herein we explore a unique palladium complex with the characteristic
N,N coordination mode:
To utilize the system in electrocatalysis, the Pd-complex has been deposited (by
introducing the appropriate amount of its ink) onto the electrode surface and, subsequently,
subjected to electroreduction through voltammetric potential cycling. In our preliminary
0
communication [27], thus electrochemically-generated Pd nanocenters were demonstrated to
exhibit reasonably high activity toward CO RR in near-neutral medium of CO -saturated
2
2
KHCO . In this work, special attention is paid to physicochemical identity of the palladium
3
complex (with its detailed crystallographic, IR and NMR features), electrochemical
0
generation and characterization of Pd nanocenters, feasibility of stabilizing and activating
interactions between Pd and N,N-coordination ligands, as well as to identification of the
reaction products with the ultimate goal of explaining of high intrinsic catalytic activity of
palladium supported within tridentate Schiff-base-ligand network. It is also reasonable to
expect that, by formation and immobilization of Pd nanoparticles with supramolecular
architectures of Schiff-base-ligands containing nitrogen donor atoms, the distribution and
resistance to agglomeration of catalytic palladium should be improved. In comparison to
conventional Pd nanoparticles, the performance of the Schiff-base-ligand-supported Pd
nanoparticles is much higher during CO -reduction both under voltammetric and
2
chronoamperometric conditions.
2
. Experimental
2
.1. Chemicals
7

All chemicals were commercial materials of analytical grade purity, except the
palladium complex, which was synthesized according to the procedure described below. Pd
black and 5wt% Nafion® perfluorinated resin solution were purchased from Sigma–Aldrich.
Sulphuric acid, potassium bicarbonate and potassium formate were obtained from POCh
(Gliwice, Poland) and Sigma–Aldrich.
Solutions were prepared using doubly distilled and subsequently de-ionized (Millipore
Milli-Q) water. They were deoxygenated by bubbling with ultrahigh purified argon for 20 min
prior to the electrochemical experiment. Measurements were made at room temperature
(22±2°C).
2
.2. Synthesis of palladium complex
The metal salts were used without further purification as supplied from Aldrich. The ligand L
was prepared as described before [56]. NMR spectra were run on Varian Gemini 400 MHz
-4
spectrometer using TMS as an internal reference. Mass spectra for acetonitrile solutions ~10
-3
mol dm were determined using a Waters Micromass ZQ spectrometer. Sample solutions
were introduced into the mass spectrometer source with a syringe pump at a flow rate of 40
3
-1
o
µdm min with a capillary voltage of +3 kV and a desolvation temperature of 300 C. Source
o
temperature was 120 C. The cone voltage (V ) was set to 30 V to allow transmission of ions
c
without fragmentation processes. Scanning was performed from m/z = 200–1000 for 6 s, and
1
0 scans were summed to obtain the final spectrum. Microanalyses were obtained using a
Perkin–Elmer 2400 CHN microanalyzer.
PdLCl ] ∙MeOH. To the Schiff base ligand L (0.032 g, 0.125 mmol), first dissolved
[
2
2
in 15 ml of MeOH, PdCl was added (0.022 g, 0.125 mmol). Yellowish solution immediately
2
turned red and solid of the same color precipitated. After 24 hours of stirring under ambient
conditions, the precipitate was filtered, washed well with consecutively cold MeOH and Et O.
2
8

Air drying gave 0.040 g of red complex with 73% yield. Single crystals suitable for X-Ray
characterization were collected from the filtrate, upon slow evaporation of saturated
+
+
methanolic/ether solution of the complex. ESI-MS(+): m/z (%) = 421 (100) {[PdLCl]∙Na } ,
+
+
-
3
61 (80) [PdL] , 833 (30) [Pd L Cl ] . ESI-MS(-): m/z = 431 (100) [PdLCl ] : IR (KBr):
2 2 3 2
ν(O-H) = 3362, ν(C-H) = 3086, 3057, 3014, ν (CH ) = 2954, ν (CH ) = 2840, ν (C=O) =
ar
as
3
s
3
s
1
1
702, ν(C=C) = 1624, 1596, 1584, 1562, 1542, ν(C=N) = 1491, 1485, 1432, ρ(C-H) =
py py py
–1 1
160, 1114, 1108, 1095, γ(C-H) = 926, 906, 883, 828, 819, 765, 639, 548, 513 cm ; H
py
1
2
NMR (400 MHz, d -DMSO): δ = 8.80 (s, 1H, H ), 8.39 (dd, 1H, H , J = 5.50Hz, J = 1.20Hz,
6
a
e
3
1
2
1
J = 1.05Hz), 8.21 (dt, 1H, H , J = 7.83Hz, J = 1.51Hz), 8.08 (d, 1H, H , J = 1.96Hz), 7.82
c
h
1
2
3
1
2
(
dd, 1H, H , J = 7.81Hz, J = 1.05Hz, J = 0.90Hz), 7.69 (td, 1H, H , J = 5.58Hz, J =
b
d
3
1
2
2
3
.25Hz, J = 1.51Hz), 7.64 (dd, 1H, H , J = 9.03Hz, J = 1.96Hz), 6.58 (d, 1H, H , J=9.03Hz),
f g
.80 (s, 3H, OMe) ppm. Analytical calculations: stoichiometric (%) for
[
Pd(C H N O )Cl ] ∙MeOH (899.15): C, 38.74; H, 3.14; N, 6.23; found: C, 38.66; H, 3.25;
14 12 2 3 2 2
N, 6.20.
X-Ray crystallography. Very thin red needles of the complex were obtained from the
methanol/ether solution; X-ray diffraction data were collected at 130(1) K by the ω-scan
technique on Agilent Technologies four-circle SuperNova diffractometer (Atlas detector) with
mirror-monochromatized CuK radiation (λ=1.54178Å). The data were corrected for Lorentz-
α
polarization and absorption effects [60]. Accurate unit-cell parameters were determined by a
least-squares fit of 1068 reflections of highest intensity, chosen from the whole experiment.
The structures were solved with SIR92 [61] and refined with the full-matrix least-squares
2
procedure on F by SHELXL97 [62]. Scattering factors incorporated in SHELXL97 were
used. All non-hydrogen atoms were refined anisotropically, hydrogen atoms from O10-H
groups were found in the difference Fourier maps and freely refined, all other hydrogen atoms
were placed in the calculated positions, and refined as ‗riding model‘ with the isotropic
9

displacement parameters set at 1.2 (1.5 for methyl groups) times the U value for appropriate
eq
non-hydrogen atom.
Crystal data: 2(C H Cl N O Pd)·CH OH, M =899.15, monoclinic, P2 , a=
14
12
2
2
3
3
r
1
3
1
4.1468(13)Å, b=7.7338(3)Å, c=15.818(2)Å, β=108.712(13)°, V=1639.1(3)Å , Z=2,
-3
-1
F(000)=892, d =1.82g·cm , μ (CuK )=12.30mm , 6802 reflections collected, of which 4856
x
α
symmetry-independent (R =0.019), 4816 with I>2σ(I). Final R=0.045, wR2=0.127, S=1.06;
int
-3
max/min residual electron density in the final Δρ map: 1.25/-1.55 e·Å .
CCDC-940669 contains supplementary crystallographic data for the complex. These
data can be obtained free of charge via www.ccdc.cam.ac.uk/data request/cif, by e-mailing
data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, UK.
2
.3. Electrochemical Measurements
All electrochemical measurements were performed with the use of CH Instruments
60E workstations in a three-electrode configuration. The reference electrode was a
7
mercury/mercury sulfate electrode (Hg/Hg SO ). All potentials are reported vs. Reversible
2
4
Hydrogen Electrode (RHE). The gold rod was used as a counter electrode. The glassy carbon
electrode served as the working electrode substrate. Before each modification, glassy carbon
electrode was polishing with aqueous alumina slurries (grain sizes from 0.5 to 0.05 μm) on a
polishing cloth.
-3
All electrochemical measurements were carried out in 0.1 mol dm KHCO solution
3
under an atmosphere of Ar (pure, Air Products) or carbon dioxide (CO , 99.99 % Air
2
Products).
2
.4. Procedures involving Pd Nanoparticles
1
0

The suspension (ink) of the palladium complex was prepared as follows: a known
3
amount (1.85 mg) of the palladium complex was dispersed in 100 μdm of acetonitrile. Later
3
1
,5 μdm of the appropriate suspension was dropped onto the surface of a glassy carbon
electrode by a micropipet. The resulting Pd-complex-layer was subjected to electroreduction
-1
through voltammetric potential cycling (25 full cycles at the scan rate of 50 mV s ) from 1.10
II
0
to -0.68 V (the latter potential value was sufficient to cause Pd -reduction and generate Pd
-3
sites) in the deoxygenated 0.1 mol dm KHCO .
3
Comparative measurements were based on commercially available Pd nanoparticles.
The respective aqueous suspension was prepared by dispersing (0.071 g) of palladium black
3
in 2 cm of distilled water, followed by its sonication for 2 h. Later, the appropriate sample (2
3
μdm ) was distributed over the glassy carbon electrode surface with a micropipet.
As a rule, the Pd-based catalytic films described above were covered with ultra-thin
3
layers of Nafion polyelectrolyte by depositing 1 μdm of the Nafion solution (prepared by
introducing 5mass% of the commercial Nafion solution into ethanol at 1 to 10 volume ratio).
The loading of Pd in the catalytic layers (on glassy carbon) originating from both
-2
commercial Pd-black and Pd in Pd-complex was ca. 100 μg cm . Before measurements, the
catalytic electrodes were conditioned by application of 25 voltammetric potential cycles at the
-1
-3
5
0 mV s scan rate between -0.68 V and 1.10 V in 0.1 mol dm KHCO . Diagnostic cyclic
3
-1
voltammograms have been recorded at the scan rate of 10 mV s .
The electrochemically active surface area (ECSA) of metallic palladium nanoparticles
was estimated upon measuring the surface oxide reduction charge (from cyclic voltammetry).
The data was normalized with respect to the active surface area. Knowing that the charge for
II
0
-3
the Pd-surface-oxide reduction (the Pd to Pd cathodic peak) in 1 mol dm H SO was
2
4
-2
reported to be equal to 424 C cm [36,37], the following ESCA values characteristic of
tridentate-Schiff-base-ligand supported Pd nanostructures and conventional Pd nanoparticles
1
1

2
(
deposited on glassy carbon) were estimated: 0.7 and 1.2 cm , respectively. Based on eight
measurements, each performed following independent preparation of the respective Pd-
containing systems, the reproducibility of the above determinations was on the level of 10%
(standard deviation). Nevertheless, due to uncertainty of the assumed stoichiometry of Pd-
oxide and unpredictable contribution from double layer charging, the above method is less
reliable than that based on H-adsorption [36,37]; but the latter one is not applicable to metals
(such as Pd) absorbing hydrogen.
2
.5. IR measurements
Infrared spectra were recorded using the Nicolet iN10-MX-FTIR instrument (Thermo
Scientific). The experiments were carried out in the reflectance mode and with use of MCT/A
2
type detector. The sampled area was equal to 100 μm . The spectral resolution was equal to 4
−
1
cm and typically 16 scans were averaged for a single spectrum.
Thin layers of tridentate Schiff-base ligand, the ligand supported Pd nanostructures,
and of the complex with palladium(II) were deposited on surface of the glassy carbon plate.
The spectrum of the bare glassy carbon surface was considered as the background.
2
.6. XPS measurements
XPS experiments were carried out using SPECS Surface Nano Analysis GmbH
Berlin, Germany) instrument equipped with XR 50 MF X-ray source and m-Focus 600
(
monochromator (600 mm Rowland Circle), using monochromatized X-ray Al Ka emission
line, photon energy 1486.6 eV, operating at 120 W, and Phoibos 150 hemispherical analyzer
with 150 mm radius, NAP version, equipped with 2D-DLD detector. The system base
-10
pressure was in the 10 mbar range. Spectra were fitted using CasaXPS software, version
1
2

2
.3.17PR1.1 and the quantitative analysis were performed using CasaXPS built-in Scofield
relative sensitivity factors.
2
.7. Gas chromatographic analysis
Gas chromatograph (Agilent Technologies 7890A GC System) equipped with a silica-
based Stationary Phase GS-GASPRO column, thermal conductivity detector (TCD) and flame
ionization detector (FID) was used for analysis of gaseous products. Helium was used as the
carrier gas; and sulfur hexafluoride (SF ) as the internal standard.
6
3
. Results and Discussion
3
.1. Synthesis and characterization of Pd(II) Complex
Tridentate Schiff base ligand L was synthesized via typical condensation reaction
between methyl 3-amino-4-hydroxybenzoate and picolinaldehyde as reported earlier [58]. The
synthetic protocol applied to palladium(II) complex has intentionally omitted the use of
-
external bases such as Et N, N or AcO [57] thus aiming at exerting exclusive coordination of
3
3
ligands nitrogen atoms and leaving aromatic hydroxyl group in its native form. Reaction of
PdCl with L ligand in equimolar ratio has resulted in formation of the complex in which
2
N O-donor Schiff base coordinates the Pd(II) ion in bidentate mode to give the following
2
formula: [PdLCl ]. Electrospray mass spectrometry revealed the presence of peaks that
2
-
+
+
correspond to [Pd(L)Cl ] , [Pd(L)] and [Pd(L)Cl]∙Na species, which under ESI-MS
2
+
conditions may associate to form [Pd (L) Cl ] . Nonetheless, isotopic distribution and
2
2
3
spectrum as a whole (Fig. 2) does not give rise to other species than those M:L stoichiometry
1
of 1:1. H NMR studies of the complex in solution confirm that strictly mononuclear system
is formed of the following formula [PdLCl2].
1
3

The mononuclear form was retained in the solid state, as indicated by single crystal X-
Ray studies, with additional MeOH molecule in the crystal lattice. It is interesting to note that
no spontaneous deprotonation of phenolic OH group occurred, as for instance in complexes
with the same ligand L [56] or similar ligands [58,59] where, without the need of exogenous
base, formation of pseudo-grid dimers was observed. This fact was further corroborated by
-1
the IR spectrum (for simplicity not shown here), where broad band was found at 3362 cm
which was ascribed to ν(O-H) vibration. It should also be emphasized that aromatic ring with
non-coordinating hydroxyl group was free to rotate, what is not out of significance when
considering the surroundings of the palladium(II) ion and, indirectly, electrochemical
behavior of palladium within [PdLCl ] complex.
2
It is apparent from crystallographic data (available in Supplementary Material) that
the structural unit in the crystal structure of the palladium complex consists of the fairly-
strongly hydrogen-bonded two complex molecules and one solvent – methanol molecule (Figs
S1 and S2). Geometrical parameters are summarized in Table S1. In the crystal structure (Fig.
3
), there are some relatively short C-H···O hydrogen bonds joining the structural units. The
hydrogen bond data are listed in Table S2. It should be emphasized that Pd(II) ions are four-
coordinated (by two ligand nitrogen and two chlorine atoms) in the square-planar
environment which of importance to catalysis, including CO -reduction catalysis by tunable
2
square-planar transition-metal complexes [2,4,6,8,10].
3
.2. Electrochemical generation of Pd catalytic sites
To generate catalytic nanocenters within the palladium complex, a glassy carbon electrode
II
modified with [Pd (C H N O )Cl ] ∙MeOH (deposited as ink from suspension, according to
1
4
12
2
3
2 2
the procedure described in Experimental section) has been subjected to 25 full voltammetric
-1
cycles (at 50 mV s ) in the potential range starting from 1.10 down to -0.68 V in the
1
4

-3
deoxygenated 0.1 mol dm KHCO (Fig. 4). The lower potential limit has been chosen to be
3
II
0
negative enough (-0.68 V) to assure reduction of Pd -sites (to Pd ) within the complex layer.
The increasing voltammetric peak currents, which have appeared during continuous
voltammetric potential cycling (Fig. 4), should be correlated with hydrogen sorption
(adsorption, absorption) and desorption phenomena (overlapping with hydrogen evolution at
the negative potential limits) characteristic of the behavior of metallic palladium. are
0
consistent with the formation of Pd catalytic centers. It is noteworthy that the stationary
current responses have been obtained upon finalizing the voltammetric potential cycling.
II
0
Although the latter observation does not mean complete transformation of Pd to Pd but it
implies no futher electroreduction processes within the complex layer under experimental
conditions of Fig. 4.
Based on structural information, dense crystal packing and existence of hydrogen
bonds and multiple coordination planes (Figs S1, S2, S3 and S4; Tables S1 and S2), the
0
electrochemically generated Pd centers should be distributed three-dimensionally within the
rigid hydrogen-bonded network of supramolecular ligands. Having in mind the NMR
spectrum of the original robust square-planar coordination architecture (Fig. 2), it is likely that
the palladium sites are in the close vicinity of N-atoms originating from the nitrogen-
containing ligands (Fig. 1).
3
.3. Infrared characterization
The IR spectrum of the free ligand is shown in Fig. 5A (Curve a). The very broad band
-1
at 3369 cm originates from the O-H stretching motion of the phenyl group. The intense band
-1
at 1703 cm corresponds to the C=O stretching mode of the COOH group. The band at 1603
-1
cm originates from the mixed contribution of the benzoic acid, pyridine ring deformation
-1
and the C=N stretching mode of the N=C bridge. The bands at 1521 and 1439 cm are
1
5

-1
characteristic of the pyridine ring. The two bands at 1290 and 1208 cm are related to the C-N
-1
and C-C stretching of the imine group. The band at 1105 cm is related to the C-O stretching
of the phenyl group.
Upon examination of the IR spectrum of the Pd(II)-complex (Fig. 5A, Curve b), it is
apparent that the most noticeable change concerns the spectral range between 1300 and 1100
-1
cm . Obviously, upon formation of the Schiff base Pd(II)-complexes, both nitrogen atoms are
involved in the coordination of the palladium ion. Furthermore, the O-H stretching band
-1
observed at 3369 cm in the ligand spectrum (Fig. 5A, Curve a) tends to disappear after the
-1
complexation (Curve b). The group of bands around 3000 cm , which are visible in the
spectrum of the complex, should be attributed to CH stretching modes. Most likely, these
bands have been effectively covered by the broad O-H band in the free-ligand spectrum. The
absence of the O-H in the Pd(II)-complex spectrum suggests that the phenol group is
dissociated, and it presumably interacts with the positively charged palladium ion. The
infrared spectra are consistent with the structure of the complex proposed on the basis of
NMR results (Fig. 2).
0
The IR spectrum recorded following electrochemical generation of Pd centers within
the complex (Fig. 5A, Curve c) shows that, in comparison to the untreated complex layer
(Curve b), the main bands are not affected by the electrochemical treatment, thus indicating
that the molecular structure of the supramolecular network of ligands is largely unchanged.
-1
There are, however, small differences in the spectral range below 1100 cm . This range is
dominated by the out-of-plane modes, which are sensitive to the conformation. The observed
dissimilarity implies some alterations in the ligand conformation upon the change of the redox
state of the palladium ion. On the whole, the geometry of the ligand seems to be unaffected
upon electroreduction of the complex.
1
6

To get better insight into differences between the spectra of the complex before and
after the reduction, we have enlarged the portions recorded in the 800 - 1050 (Fig. 5B) and
-1
-1
2
800 - 3200 cm (Fig. 5C) ranges, respectively. First, the band at 980 cm , which is
II
characteristic of the Pd -containing complex, undergoes splitting after the electrochemical
reduction. It can be postulated on the basis of studies with 4-hydroxybanzoic acid and
-1
pyridine [63,64] that the band at 980 cm (Fig. 5B, Curve b) comprises contributions from the
out-of-plane C-H deformation modes of both aromatic rings of the ligand molecule. The
splitting of this band implies changes in surroundings of the aromatic C-H groups upon the
complex reduction. Another difference refers to the overlapping bands at 930, 914 and 905
-1
-1
cm (Fig. 5B, Curve b) which are transferred to two well separated bands at 926 and 905 cm
after the complex reduction (Fig. 5B, Curve c). These bands are likely to stand for the out-of-
plane deformation of the carbonyl (COOMe) group of the ligand (Fig. 1) and the out-of-plane
-1
modes of C-H [63,64]. But the most intense band of COOMe appears at 1703 cm (Fig. 5A,
Curve a), and it does not change significantly upon the complex reduction (Fig. 5A, Curve c).
Thus, it seems that the C-H groups are mostly affected by the reduction. Finally, while the
-1
band appearing at 2994 cm and corresponding to the C-H stretching modes have been
II
observed for the Pd -complex (Fig. 5C, Curve b), they do not exit after the electroreduction
(Fig. 5C, Curve c). These bands are likely to shift to higher frequencies and overlap with the
-1
neighboring band at 3120 cm (Fig. 5C, Curve c).
1
7

In conclusion, some electroreduction-induced reorganizations in the vicinity of
aromatic C-H groups are apparent from the infrared data. They may reflect changes of
dihedral angles between two aromatic rings of the ligand molecule as well as alterations of the
C-H…O hydrogen bonding network. The presence of the C-H…O hydrogen bonds has
already postulated on the basis of crystallographic data (Supplementary Material; Figs S1
and S2). But the original robust square-planar coordination architecture seems to be retained
following electroreduction of the complex.
3
.4. Microscopic examination
In our further studies aiming at physicochemical characterization and evaluation of
0
performance of Pd nanocenters generated within tridentate-Schiff-base-ligand network, we
have also considered conventional palladium black nanoparticles as the reference system.
Morphologies (Fig. 6) of (a) pristine Pd(II) complex with tridentate-Schiff-base-ligands, (b)
Pd nanostructures generated within the ligand network and (c) typical palladium nanoparticles
have been examined using transmission electron microscopy (TEM). The data of Fig. 6b are
consistent with the generation of Pd nanoparticles within the complex and formation of
ligand-supported Pd nanoparticles (ligand-nPds). The sizes of ligand-nPds have been ranging
from 5 to 10 nm. For comparison, the sizes of the commercial particles of palladium (nPds)
have been on the level 15 nm. Both nanostructured catalytic materials tend to form
agglomerates.
To get more information about morphology and composition of the films of Pd
complex with tridentate-Schiff-base-ligands before and after the electrochemical generation of
0
Pd catalytic sites, we have performed additional scanning electron microscopic (SEM)
experiments. Fig. 7 illustrates SEM images of the films (deposited on glassy carbon) of (a) the
1
8

initial Pd(II) complex with tridentate-Schiff-base-ligands and (b) the system after
electroreduction (by voltammetric potential cycling). The presence of palladium has been
confirmed with surface microanalysis (EDS), and the existence of Pd is particularly evident
from Pd-mapping images (white spots in images on the right in Figs 7a and 7b). The images
of Fig. 7b are consistent with the formation (electrogeneration) of Pd nanoparticles or
0
dendritic Pd -nanostructures (ligand nPds) on surfaces or within the films initially deposited
as elongated crystals of the Pd(II)-complex (Fig. 7a). The distribution of elements (carbon,
nitrogen, oxygen, palladium, chlorine and palladium) in the complex layer
[
Pd(C H N O )Cl ) ∙MeOH] has been before and after electrochemical reduction (Fig. 4).
14 12 2 3 2 2
On the basis of the EDS analysis, it can stated that the amounts of palladium: in the reduced
system and in the initial Pd(II)-complex are on the levels, 10.6 atomic% and 9.7 atomic%. In
other words, the relative amounts of palladium before and after the electrochemical generation
0
of Pd sites are comparable, and the observed difference may reflect the error of determination
and local agglomeration of Pd centers upon reduction. The analytical data also imply the
presence of carbon, oxygen, nitrogen and chlorine, as well as some potassium which
originates from the supporting electrolyte (used for generation of ligand-nPds).
3
.5. Electrochemical characterization
First, the electrochemical identity of palladium nanostructures generated within
tridentate-Schiff-base-ligand network, was addressed the in such a model acid medium as
-3
deoxygenated 0.5 mol dm H SO . Fig. 8 illustrates cyclic voltammograms recorded for
2
4
ligand-nPds (red lines) in comparison to conventional nPds (black lines). As it is commonly
accepted for the behavior of palladium nanoparticles, hydrogen sorption and desorption peaks
(
below 0.25 V) as well as responses characteristic of the Pd oxide formation and its reduction
above 0.55 V) appear in both cases (black and red lines) during voltammetric potential
(
cycling in acid medium. Consequently, it is reasonable to postulate that Pd nanostructures
1
9

generated within tridentate-Schiff-base-ligand network exist in a form of metal nanoparticles,
0
exhibit the interfacial redox processes characteristic of Pd rather than the ionic-Pd-centered
II
redox processes. In other words, the reduction of Pd -sites (within the complex) is in a sense
0
irreversible: once Pd centers are formed they behave like metallic palladium. Upon oxidation
0
of Pd sites, palladium oxides are generated, indeed, on their surfaces (Fig. 8) and, apparently,
they are robust enough not to undergo complexation with the surrounding ligands. The latter
statement is in agreement with the inorganic chemical properties of palladium [65].
II
To support our view that the electroreduction of Pd sites within the complex leads to
0
the formation of Pd nanostructures, we provide an indirect proof by performing additional
diagnostic experiments involving the electrocatalytic reduction of oxygen (the results are
presented in Supplementary Materials in Fig. S5). It is well established that only metallic
palladium centers are capable to effectively inducing reduction of oxygen (mostly to water) in
acid medium [66]. Indeed, when the potential has been scanned negatively from 1.18 to 0.04
V, the characteristic O -reduction peak [66] has been observed at ca. 0.7 V for palladium
2
electrogenerated within tridentate-Schiff-base-ligand network (Curve a in Fig. S5). If the
ligand-supported-Pd-sites were ionic, the well-defined O -reduction voltammetric peak would
2
II
0
be observed at more negative (than 0.7 V) potentials, i.e. until reduction of Pd -to-Pd sites
occurs. Indeed, when the freshly prepared (not pre-reduced by voltammetric potential cycling)
II
complex with Pd sites has been deposited on glassy carbon and subjected to the reductive
voltammetric polarization in the presence of oxygen, the O -reduction peak has been shifted
2
0
now to almost 0.4 V (Curbe b in Fig. S5) where catalytic Pd sites are electrogenerated.
Another important observation coming from the data of Fig. 8a is that (in the standard
-3
acid medium, 0.5 mol dm H SO ) the voltammetric responses of both ligand-nPds and
2
4
conventional nPds are almost identical upon normalization of the voltammetric currents
against the electrochemically active surface area (determined as described in Experimental
2
0

section). Thus, no significant structural differences are expected to exist between
polycrystalline commercially available nPds and Pd nanostructures generated within
supramolecular network. The close resemblance of responses characteristic of conventional
nPds and ligand-nPds confirms once more that the electrogenerated palladium is
0
predominantly in a form of Pd (rather than ionic Pd) sites. The fact that the current densities
observed below 0.15 V are somewhat higher in the case ligand-nPds (Fig. 8a) may originate
from more pronounced hydrogen adsorption capabilities of Pd nanostructures existing within
the supramolecular network. The latter observation may reflect morphological differences
between spherical-type conventional nPds and dendritic ligand nPds formed on surfaces or
within eleongated crystals of the complex (Figs 6 and 7).
When the current densities are expressed against the geometric area of the glassy
carbon electrode, the voltammetric responses would be quantitatively different for ligand-
nPds (red lines) and nPds (black lines) (Fig. 8b). This result is consistent with the effectively
2
lower electrochemically active surface area (ESCA) of ligand-nPds (ca. 0.7 cm ), relative to
2
that characteristic of conventional nPds (ca. 1.2 cm ), despite the fact that the initial loading
-2
of 100 µg cm was the same for Pd in both cases. It is reasonable to expect that, due to
interactions with N,N coordination sites of supramolecular ligands, some portions of
palladium surfaces in ligand-nPds have not been electroactive. Furthermore, although
palladium nanostructures are characterized by the well-developed regions for the oxide
monolayer formation and reduction, uncertainty of the present approach concerns mostly
possibility of the formation of multilayer oxide films of undefined stoichiometry, in addition
to unpredictability of double layer corrections which are likely to be dependent on
morphology and degree of dispersion of Pd nanostructures. Thus, is spite of application of the
analogous experimental conditions (same electrolyte and voltammetric polarization
procedures), the obtained ESCA values should be considered as approximates.
2
1

3
.6. XPS measurements
The analysis of surface layers of nPds, ligand-nPds, and the palladium complex was
carried out by XPS experiment in order to obtain information about chemical nature and
structure of the catalytic systems. The data of Figs 9a, 9b and 9c refer to the high-resolution
XPS spectra of Pd 3d core level.
Based on the data obtained from XPS measurements for ligand-nPds and
conventional-nPds (Figs 9a and 9b), two doublets are clearly visible in the spectra recorded in
the Pd 3d region. First doublet (with the BE of Pd 3d5/2 component at 335.1 eV) corresponds
to metallic palladium (about 55% for conventional nPds and 67% for ligand-nPds). But for
conventional Pd nanoparticles, the second doublet (with the BE of Pd 3d5/2 at 336.8 eV) is
located at the BE value characteristic of palladium(II) oxide (Fig. 9a). Interestingly, the Pd
3
d5/2 signal of the second doublet (nonmetallic Pd) is located in the case of ligand-nPds at
even higher BE of 338.2 eV (Fig. 9b); namely, at the value corresponding to the Pd 3d5/2 BE
characteristic of the Pd(II) complex with tridentate Schiff-base ligand (Fig. 9c). Thus
coordinative interactions with Pd(II) surface atoms can be envisioned. At the same time, the
absence of the signal corresponding to palladium(II) oxides (compare Figs 9a and 9b)
suggests that all oxidized Pd atoms are complexed by tridentate Schiff-base ligands, whereas
0
the nanoparticle core retains its metallic (Pd ) character. The observed difference between the
electronic properties of Pd(II) in the conventional nanoparticles and in the ligand nPds can be
rationalized in terms of the contribution coming from the interactions of palladium with N,N
coordination sites of supramolecular tridentate Schiff-base ligands.
3
.7. Electrocatalytic CO -reduction
2
2
2

To get further insight into electrochemical behavior of tridentate-Schiff-base-ligand
supported Pd nanoparticles (ligand-nPds; red lines), relative to the characteristics of
conventional Pd nanoparticles (nPds; black lines), we have performed diagnostic cyclic
-3
voltammetric experiments in 0.1 mol dm KHCO electrolyte, i.e. in the supporting
3
electrolyte used for electrocatalytic measurements (Fig. 10). The actual voltammetric
responses have been recorded in the following potential ranges: (a) from 1.18 to -0.70 V, and
(b) from 1.18 to -0.05 V. While the latter potential range (b) covers surface phenomena
occurring on palladium at positive potentials (>0 V), the previous potential range (a) includes
excursion toward more negative potential values (<0.2 V) where absorption of hydrogen is
accompanied by hydrogen evolution. Here desorption of absorbed hydrogen takes place at
potentials somewhat higher (ca. 0 V) that the onset potential characteristic of hydrogen
evolution (ca. -0.3 V). In both cases (Figs 10a and 10b), hydrogen oxidation and desorption
processes are evident during scanning potential from the negative limit toward positive
values.
Upon medium transfer from H SO (Fig. 8) to the near-neutral KHCO solution (Fig.
2
4
3
5
, Fig. 10a and 10b recorded in different potential ranges), such processes as the sorption of
hydrogen (water reduction, or proton discharge, followed by hydrogen evolution) are
kinetically less facile and the respective peaks are drawn out (Fig. 10a). Among important
issues is the appearance of much higher hydrogen- sorption/hydrogen-evolution currents
(
<0.1 V) for the ligand-nPds (Fig. 10a, red line), in comparison to conventional nPds (Fig.
0a, black line). The coupled phenomenon of hydrogen oxidation and desorption appears in a
form of the broad surface-type peaks in both cases (Fig. 10a) but, in the case of ligand-nPds
red line), at potentials more than 200 mV less positive, relative to conventional nPds (black
1
(
line). Relatively low anodic currents starting at ca. 0.7 V (red line) and 0.85 V (black line) are
consistent with the formation of thin oxide layers on surfaces of Pd nanoparticles. The
2
3

appearance of small cathodic peak at about 0.6 V (Fig. 10b) should be correlated with the
reduction of previously generated (in anodic scans) Pd oxides. Finally, the hydrogen evolution
starts in both cases at potentials below -0.45 V (Fig. 10).
As reported earlier, palladium exhibits electrocatalytic activity toward reduction of
CO [21-27]. The results of present studies (Fig. 11) indicate some differences in the
2
performance of (a) conventional nPds, and (b) ligand-nPds. It is apparent upon comparison of
responses in the presence (solid lines) and absence (dashed lines) of carbon dioxide that the
onset potentials for CO RR can be correlated with the hydrogen evolution reaction regions
2
characteristic of both types of Pd nanoparticles. The maximum current CO RR is attained at
2
ca. -0.3 V on nPds (Fig. 11a) , and at -0.5 V on ligand-nPds (Fig. 11b). On the other hand, the
onset potential for CO RR appears in the case of ligand-nPds at the less negative potential
2
value of ca. -0.05 V (Fig. 11b), in comparison to -0.15 V characteristic of conventional nPds
(
Fig. 11a). What is even more important, is that the CO RR current densities (expressed
2
against the electrochemically active surface area) are approximately five times higher for
ligand-nPds (Fig. 11b), relative to the performance of conventional nPds (Fig. 11a). It is
noteworthy that Pd-free tridentate-Schiff-base-ligands do not exhibit elctrocatalytic activity
toward CO RR in the investigated range of potentials of Fig. 11.
2
Development of broad CO -reduction peak (Fig. 11) reflects competition between the
2
actual CO RR and evolution of hydrogen or absorption of hydrogen. It is reasonable to expect
2
formation of CO-type poisoning species as the reaction products or intermediates. Judging
from the observed much higher CO RR-current-denities, the degree of poisoning of surfaces
2
of ligand-nPds is likely to be less pronounced in comparison to conventional nPds. It should
2
4

be remembered that hydrogen sorption/desorption phenomena seen to be more reversible
ligand-nPds in comparison to conventional nPds (compare dashed lines in Figs 11a and 11b).
During positive potential scans recorded upon CO RR (solid lines in Fig. 11), the
2
observed hydrogen oxidation currents are much less pronounced than those recorded in the
absence of CO (dashed lines in Fig. 11). The appearance of two anodic peaks at potentials
2
higher than 0.5 V (solid lines in Fig. 11) should be correlated with the following phenomena:
hydrogen desorption and oxidation of the CO -reduction products (such as CO). The presence
2
of the adsorbed CO RR products on Pd surface must have been responsible for the partial
2
blocking of hydrogen desorption. The hydrogen oxidation peak is significantly shifted
towards more positive values, and it seems to appear at potentials close to 1 V (i.e. following
oxidative desorption of the poisoning CO-type reduction products).
3
.7. Identification of CO RR products with gas chromatography
2
Gas chromatography was used for determination of volatile (gaseous) CO reduction
2
products obtained during CO RR performed at electrodes modified with conventional nPds
2
and ligand-nPds. After 8 minutes of electrolysis under chronoamperometric conditions and
application of different reduction potentials to the working electrode (the counter electrode
was separated with Luggin capillary), the resulting 30 μl gas sample (taken with a GC injector
from the gas collector) was mixed with 10 μl SF and, then, injected into the gas
6
chromatograph inlet. The gas collector was in the form of a funnel with a septum immersed in
the electrolyte and placed directly above the GC electrode.
The results of gas chromatographic analysis are summarized in Table 1. Out of the
selected electrolysis potentials applied to both catalytic electrodes, the relatively largest
amounts of carbon monoxide (relative to hydrogen) have been obtained at the least negative
potential of -0.82 V. In this respect (CO-generation), the system utilizing ligand-nPds seems
2
5

to be the most efficient, at the relative molar amounts of CO to H are approximately 1:1. In
2
other words, the electrochemical conversion of CO into syngas becomes feasible under the
2
above mentioned conditions.
2
6

At potentials lower than -1.3 V, hydrogen evolution predominates over CO RR. The
2
latter effect is even more sound at conventional nPds: at the potentials of -1.52 V and -1.72 V,
no CO but only H has been detected. Thus immobilization of palladium nanostructures
2
within supramolecular network of tridentate-Schiff-base-ligands seems to affect selectivity of
palladium and favors production of CO (over H2).
3
.8. Electrochemical characterization of CO RR products
2
Some information about the reaction products can be obtained from the stripping type
electrochemical experiments [67-69] in which the oxidation voltammetric responses are
recorded following application of different electroreduction potentials, E ‗s (at which
ad
adsorption of the reaction products takes place). Before the linear scan (stripping)
voltammograms have been recorded (Fig. 12), the Pd-containing catalytic electrodes have
been for 1 s at the selected value of E (from 0.18 V to -0.88 V). The reduction products have
ad
been formed and adsorbed on the electrode surface (in addition to simultaneous absorption of
hydrogen on Pd). It was reported earlier that, by holding the catalytic (Pd) electrode at the
potential of 0.50 V (Scheme 1), absorbed hydrogen was removed from Pd electrodes without
oxidizing the adsorbed CO RR reduction. Following the suggested procedure [70], the
2
potential has been maintained at 0.50 V for 3 s. Afterwards, the potential has been scanned
-1
(
10 mV s ) toward positive values. This operation yields charge characteristic of the anodic
products
oxidation of the adsorbed CO RR products (QECSA
).
2
Figure 12 illustrates cyclic voltammetric responses measured for (a) conventional nPds
and (b) tridentate-Schiff-base-ligand supported or ligand-nPds recorded following CO2-
reduction (at different potentials under chronoamperometric conditions, as described above)
-3
in CO -saturated 0.1 mol dm KHCO . The fact the voltammetric stripping peaks for the
2
3
oxidation of the ―reduced-CO ‖ products (generated at different potentials) appear at
2
2
7

potentials ca. 0.87-0.89 V is consistent with the surface oxidation of carbon monoxide as the
predominant CO RR product.
2
The charge characteristic of stripping (oxidation) of ―reduced-CO ‖ (Fig. 12c)
2
products provides important information about catalytic reactivity of conventional nPds and
ligand-nPds. The relationship between the charge determined during oxidation of the
adsorbed products generated during CO RR upon application of different E potentials within
2
ad
the range from 0.60 to 1.20 V (Figs 12a and 12b) is shown in Fig. 12c. Upon decreasing Ead
potentials toward more negative values, magnitudes of charges related to the anodic oxidation
of the adsorbed CO RR products (and normalized against the electrochemically active surface
2
products
area, ECSA), QECSA
, which are initially (i.e., in the E potential range from 0.2 to 0 V)
ad
equal to 0) tend to increase until the Ead potential decreases to -0.6 V and, then, reaches
plateau. An important issue is that, the determined values of charges are significantly higher
for the Pd-catalyst immobilized within the supramolecular framework, in comparison to
conventional Pd nanoparticles. Thus it is reasonable to expect that significantly larger
amounts of CO-type products are generated when palladium exists within tridentate-Schiff-
base-ligand network.
When CO RR is carried out on palladium catalyst in deoxygenated near-neutral
2
aqueous solutions, carbon monoxide together with hydrogen are expected as reaction products
but formation of formate is also feasible [21-26,71]. The appearance of carbon monoxide as
the main CO RR product (at all reduction potentials considered in the present work) is evident
2
form both gas chromatographic analysis and stripping type voltammetry (Fig. 12). Palladium
is a very good catalyst for the oxidation of formic acid and formate, and its appearance as the
CO RR product would contribute to the stripping-type voltammetric oxidation pattern. The
2
blue curve in Fig. 13 illustrates a typical response for the electrooxidation of formate (in
KHCO ) recorded at ligand-nPds, whereas the black curve stands for the stripping peak
3
2
8

characteristic of the oxidation of the CO RR products (as in Fig. 12). The appearance of the
2
single peak (stripping experiment) at about 0.87 V (black curve), namely without any
shoulder at less positive potentials (as in the case of blue curve) implies that carbon monoxide
has been the main reaction product. It is well established [33] that the oxidation of formic acid
or formate proceeds at palladium largely without carbon monoxide intermediate, i.e. the onset
for the formate oxidation is at less positive potentials as it is demonstrated in Fig. 13 (black
curve). Judging from the data of Fig. 13, it can be concluded that, even if some formate is
formed, the amounts are negligible relative to carbon monoxide appearing as the main
reaction product at ligand-nPds.
3
.8. Durability of electrocatalytic performance
The electrocatalytic stability (durability) of conventional nPds and ligand-nPds was
-3
addressed by performing chronoamperometric in CO saturated 0.1 mol dm KHCO3
2
-
solutions upon application potential of the constant potential of -0.51 V for 1 h (Fig. 14). The
initially-observed high currents reflect contributions originating from both the double-layer
charging/discharging and the CO -reduction. The observed currents have decayed readily due
2
to poisoning with CO-type reaction products. Finally, the currents reached the steady-state
response after about 800 s. The existence of steady state currents implies that the system is
well-behaved enough to be applicable to generate carbon monoxide and hydrogen (Table 1)
through electrolysis. Furthermore, the steady-state currents are significantly (approximately
two times) higher in the case of ligand-nPds, relative to the performance of conventional
nPds.
Conclusions
2
9