The Conductance and Thermopower Behavior of Pendent Trans-Coordinated Palladium(II) Complexes in Single-Molecule Junctions

The present work provides insight into the effect of connectivity within isomeric 1,2-bis(2-pyridylethynyl)benzene (bpb) palladium complexes on their electron transmission properties within gold|single-molecule|gold junctions. The ligands 2,2′-((4,5-bis(hexyloxy)-1,2-phenylene)bis(ethyne-2,1-diyl))bis(4-(methylthio)pyridine) (Lm) and 6,6′-((4,5-bis(hexyloxy)-1,2-phenylene)bis(ethyne-2,1-diyl))bis(3-(methylthio)pyridine) (Lp) were synthesized and coordinated with PdCl2 to give the trans-Pd(Lm or p)Cl2 complexes. X-ray photoelectron spectroscopy (XPS) measurements shed light on the contacting modes of the molecules in the junctions. A combination of scanning tunneling microscopy–break junction (STM–BJ) measurements and density functional theory (DFT) calculations demonstrate that the typical lower conductance of meta- compared with para-connected isomers in a molecular junction was suppressed upon metal coordination. Simultaneously there was a modest increase in both conductance and Seebeck coefficient due to the contraction of the HOMO–LUMO gap upon metal coordination. It is shown that the low Seebeck coefficient is primarily a consequence of how the resonances shift relative to the Fermi energy.

DCM was added, the organic layer was collected and dried over MgSO4 and filtered before the solvent was removed from the filtrate.The residue was purified using Kugelrohr distillation (120 °C, 0.1 mbar) to give a colorless oil.Yield: 3.91 (91%).

X-ray Crystallography
X-ray single crystal data for Pd p •CH2Cl2 were collected on a Bruker D8 Venture diffractometer with a Photon III C14 MM CPAD detector, using λMo-Kα radiation (λ = 0.71073Å) from a IμS microsource with focusing mirrors and corrected for absorption by numerical integration. 5The crystal was cooled with a Cryostream 700+ (Oxford Cryosystems) open-flow N2 cryostat.The structure was solved by dual-space intrinsic phasing method 5 and refined by full-matrix least squares against F 2 of all data using SHELXL program 5 on Olex2-1.5 platform. 6Crystal data and experimental details are listed in Table S1.Crystallographic data for the structure have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication CCDC 2345694.
The Pd p molecule (Figure S19) lies astride a crystallographic twofold axis, sharing its site with small admixtures of certain by-products.Thus, the substituent at C( 9) is an overlay of

Conductance peaks exploration
Similar conductance values should be related to similar junction geometries.In solution, we would expect a reduced interaction of the molecules with the gold surface and between each other.Indeed, the conductance measurements in solution show that one of the peaks, the yellow one, became predominant, therefore it is assigned to the MeS…SMe contacted junction.The reference compounds served as a test to corroborate that the yellow peak is the connection with the SMe groups, where the connection to the N-groups was found to describe a lower conductance.

d) XPS results
Binding energies were calibrated according to the C1s peak at 284.6 eV.The C 1s peak was chosen, rather than for example Au 4f, for consistency with all the samples (powders and SAMs) although possible inaccuracies of using the C 1s binding energy have been reviewed. 7S peak fitting was carried out using the CASA software with the Shirley BG type. 8 where clearly a peak attributed to Pd(0), Pd3p3/2, is observed at 340.0 eV.Note that the Pd3p1/2 peak is obscured by the Au4d5/2 peak due to the gold substrate.

e) Theoretical Calculations
The optimum geometry of each of the molecules was calculated using the density functional code SIESTA. 9These used a double-zeta polarized (DZP) basis set defined by a confining cut-off of 0.008 Rydbergs, norm conserving pseudopotentials, an energy cut-off of 150 Rydbergs and the generalized gradient approximation (GGA) method to describe the exchange correlation functional.All forces on the atoms were relaxed to a force tolerance of 0.01 eV/Å.The molecule was then contacted to gold electrodes to form the molecular junction.The gold electrodes were modelled as 6 layers of (111) gold each containing 25 atoms which are terminated by a pyramid of 13 gold atoms to mimic an STM tip.A doublezeta basis was used to describe the gold atoms, and a Hamiltonian describing this extended molecule was extracted using SIESTA.The zero bias transmission coefficient (), conductance G and the Seebeck coefficient S were calculated using the quantum transport code Gollum. 10During the optimization apical Au atoms were also relaxed.In each simulation, a pyramid of gold atoms was used as a tip, to model the last few atoms resulting from the breaking of the junction, as discussed, for example, in ref. 11 .
In the case of the ligands L p and L m the chemical structure makes it possible for the geometry to change as the nitrogen atom of the pyridine unit can either point inwards to the center of the molecule or point outwards.This is important as pyridine can act as an anchor unit with gold attaching to the nitrogen atom.Three possible conformations of L p are identified due to the orientation of the pyridine: 1both pyridines pointing inwards, 2both pyridines pointing outwards and 3one pyridine pointing in and one pointing out.The second important orientation is that of the methyl group, which has its minimum energy when it is aligned with the plane of the pyridine ring.Therefore, for molecule L p the methyl group can point left or right, and this leads to 3 possible orientations; Aleft and right, Bright and right and C right and left.We also consider the case where the methyl groups are rotated out of the plane by 90 o which is shown as geometry D. The corresponding relaxed optimum geometry of each of the 12 isomers can be seen in Figure S45, along with the ground state energy Eg relative to the minimum energy 0 eV.In this series, the geometry 3C has the lowest energy, which corresponds to the pyridines pointing in opposite directions and may be due to an interaction between the nitrogen and the hydrogen on the opposite ring.Note that the energy difference to 3A and 3B is very small (0.01 eV), which indicates that the orientation of the methyl groups is less important.As described in ref. 9 , the ionization potential (IP) was calculated using the formula IP=E(N-1)-E(N) and the electron affinity using the formula EA=E(N)-E(N+1) where E is the ground state energy of the molecule and N is the number of electrons

Figure S7. 1 H
Figure S7. 1 H NMR spectrum of L p recorded in CD2Cl2.

Figure S9. 1 H
Figure S9. 1 H NMR spectrum of L m recorded in CD2Cl2.

Figure S11. 1 H
Figure S11. 1 H NMR spectrum of L py recorded in CD2Cl2.

Figure S13. 1 H
Figure S13. 1 H NMR spectrum of SMe p recorded in CDCl3.

Figure S15. 1 H
Figure S15. 1 H NMR spectrum of Pd p recorded in CD2Cl2.

Figure S17. 1 H
Figure S17. 1 H NMR spectrum of Pd m recorded in CD2Cl2.
SMe (96.2(1)%) and Br (3.8(1)%), probably due to incomplete conversion of 2,4dibromopyridine into 2-bromo-4-(methylthio)pyridine (see above).The halogeno ligand is an overlay of Cl(1) (96.3(1)%) and iodine (3.7(1)%) atoms, the latter possibly originating from CuI used in the synthesis of L p .NMR techniques and elemental analysis gave no evidence of the halogeno species which suggests the amount present in the bulk to be negligible and that it is over-represented in the single crystal that was found to diffract to a suitable level.The nhexyl chain is intensely disordered: atoms C(1) to C(4) are disordered between positions A, B and C with the occupancies of 0.45, 0.40 and 0.15, respectively, C(5) is disordered between positions A and B with occupancies 0.6 and 0.4.The dichloromethane molecule of crystallization is also disordered, Cl(2) lying on a twofold axis, Cl(3) and the CH2 group being disordered between two positions related by this axis.Atoms C(1) to C(3) (all S21 positions), C(4A) and C(4C) were refined in isotropic approximation, other non-hydrogen atoms in anisotropic approximation, hydrogen atoms in riding mode.

R1 = 0
.0382, wR2 = 0.0904 Final R indexes [all data] R1 = 0.0453, wR2 = 0.0940b) Photophysical MeasurementsA comparison of the free ligands (L p and L m ) shows that L p has a significantly lower H-L energy gap than L m attributed to the thiomethyl group in the para position extending the system's conjugation.Similar behavior was observed for the metal complexes except the H-L gaps were significantly smaller due to the metal coordination.

Figure S20 .
Figure S20.Electronic absorbance spectra of compounds a) L p , Pd p , SMe p ; b) L m , Pd m and

Figure S21 .
Figure S21.1D conductance histograms of compounds (a) L p , (b) L m , (c) Pd p and (d) Pd m respectively, measured in air conditions; separated in conductance classes by colors (the highest conductance peak in blue, then the second highest in red, the main conductance peak in yellow and the lowest conductance peak in purple), each of them fitted to a Gaussian distribution displayed by dashed lines.The black histogram is the total considering all the classes.The legend shows the mean G value for each class.

Figure S22 .
Figure S22.Comparison of 1D conductance histograms of compound Pd p measured in air from an (a) ~ 1 mM and (b) ~ 0.1 mM DCM solution; separated in conductance classes by colors (the highest conductance peak in blue, then the second highest in red, the main conductance peak in yellow and the lowest conductance peak in purple), each of them fitted to a Gaussian distribution displayed by dashed lines.The black histogram is the total considering all the classes.The legend shows the mean G value for each class.

Figure S23 .
Figure S23.Comparison of 1D conductance histograms of compounds (a,b) SMe p and (c,d) L py , measured in air and solution respectively; separated in conductance classes by colors (the highest conductance peak in blue, then the second highest in red, the main conductance peak in yellow and the lowest conductance peak in purple), each of them fitted to a Gaussian distribution displayed by dashed lines.The black histogram is the total considering all the classes.The legend shows the mean G value for each class.

Figure S24 .
Figure S24.2D conductance-distance histograms of compound L p for all the measurements with a molecular response recorded in air conditions, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S25 .
Figure S25.2D conductance-distance histograms of compound L p for all the measurements with a molecular response recorded in solution, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S26 .
Figure S26.2D conductance-distance histograms of compound L m for all the measurements with a molecular response recorded in air conditions, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S27 .
Figure S27.2D conductance-distance histograms of compound L m for all the measurements with a molecular response recorded in solution, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S28 .
Figure S28.2D conductance-distance histograms of compound Pd p for all the measurements with a molecular response recorded in air conditions, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S29 .
Figure S29.2D conductance-distance histograms of compound Pd p for all the measurements with a molecular response recorded in solution, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S30 .S30Figure S31 .
Figure S30.2D conductance-distance histograms of compound Pd m for all the measurements with a molecular response recorded in air conditions, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S32 .
Figure S32.2D conductance-distance histograms of compound SMe p for all the measurements with a molecular response recorded in air conditions, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S33 .
Figure S33.2D conductance-distance histograms of compound SMe p for all the measurements with a molecular response recorded in solution, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S34 .
Figure S34.2D conductance-distance histograms of compound L py for all the measurements with a molecular response recorded in air conditions, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S35 . 3 .
Figure S35.2D conductance-distance histograms of compound L py for all the measurements with a molecular response recorded in solution, separated by conductance classes.Inset: relative percentage of the total selected traces contributing to each conductance class.

Figure S37 .
Figure S37.Length histogram obtained with the counts as the points with higher distance value within Gm±2σ for each trace of the Class III of compounds (a) Pd p and (b) Pd m measured in solution.Lorentzian fitting as a solid line.

Figure S38 .Figure S39 .
Figure S38.Examples of individual traces representative of each conductance class of the four compounds measured in air.

Figure S40 .
Figure S40.Examples of individual traces extracted from measurements of L m in air

Figure S41 .
Figure S41.Examples of individual traces extracted from measurements of Pd p in air

Figure S42 .
Figure S42.Examples of individual traces extracted from measurements of Pd m in air conditions exhibiting the conductance switching behavior between the different conductance values represented as dashed lines with the color of each conductance class.The conductance changes are marked with lines indicating the approximated orders of magnitude as ΔG.

Figure S43 .
Figure S43.XPS spectra in the N1s region corresponding to L p , L m , Pd p and Pd m powders and SAMs of L p , L m , Pd p , Pd m and L py .The same XPS spectra were obtained for L p and L m powders; Pd p and Pd m powders, and for their respective SAMs.Therefore, only one spectrum is shown in each panel.

Figure S44 .
Figure S44.XPS spectra in the Au4d/Pd3p region corresponding to Pd p and Pd m SAMs

Figure S45 .Figure S46 .
Figure S45.Possible rotamers of the L p molecule including their ground state energy relative to the global minimum 0 eV.

Figure S48 .
Figure S48.HOMO and LUMO orbitals of (a) Pd p , (b) L p , (c) Pd m and (d) L m

Figure S50 .
Figure S50.T(E) against energy E for rotation angle θ about the equilibrium position 0° in molecule Pd p .

Figure S51 .
Figure S51.T(E) against energy E for rotation angle θ about the equilibrium position 0° in molecule Pd m .

Figure S52 .
Figure S52.Binding energies of Pd p and Pd m molecule to a gold electrode.

Figure S53 .
Figure S53.Molecular junction geometry for (a) Pd p , (b) L p , (c) Pd m and (d) L m .

Table S1 .
Crystal data and structure refinement for structures Pd p .

Table S3 .
The mean length of the plateaus for all L p , L m , Pd p and Pd m recorded in air

Table S4 .
HOMO and LUMO energy levels and Ionization Potential and Electron Affinity of the L p molecules shown in FigureS45

Table S5 .
HOMO and LUMO energy levels and Ionization Potential and Electron Affinity of the L m molecules shown in FigureS46

Table S6 .
HOMO and LUMO energy levels and Ionization Potential and Electron Affinity of the Pd p molecules shown in FigureS47

Table S7 .
HOMO and LUMO energy levels and Ionization Potential and Electron Affinity of the Pd m molecules shown in FigureS47