Tuning the thermoelectrical properties of anthracene-based self-assembled monolayers

It is known that the electrical conductance of single molecules can be controlled in a deterministic manner by chemically varying their anchor groups to external electrodes.


Materials and Methods
All reactions were performed with the use of standard air-sensitive chemistry and Schlenk line techniques, under an atmosphere of nitrogen. No special precautions were taken to exclude air during any work-ups. All commercially available reagents were used as received from suppliers, without further purification. 4-Ethynylthioanisole and 4-(ethynyl)phenyl-tert-butylthioether were synthesised through adapted literature procedures. 1,2 Solvents used in reactions were collected from solvent towers sparged with nitrogen and dried with 3 Å molecular sieves, apart from DIPA, which was distilled onto activated 3 Å molecular sieves.

Instrumentation
1 H and 13 C{ 1 H} NMR spectra were recorded with a Bruker Avance 400 MHz spectrometer and referenced to the residual solvent peaks of either CDCl3 at 7.26 and 77.2 ppm, respectively. Coupling constants are measured in Hz. Mass spectrometry analyses were conducted by Dr. Lisa Haigh of the Mass Spectrometry Service, Imperial College London. Infrared spectra were recorded on a PerkinElmer Spectrum FT-IR spectrometer

Synthesis
General procedure for the coupling of terminal alkynes to bromoanthracenes -specific details regarding molar equivalents and column conditions are reported below.

Au preparation
The ultra-flat gold was prepared by modifying the template stripped (TS) method of Whitesides and Pinkhassik on mica (Agar scientific). 11 The Si wafer (in 5 mm x 5 mm) was ultra-sonicated in acetone, methanol and isopropanol in succession and then cleaned with oxygen plasma for 5 minutes. The cleaned wafer was glued onto the gold deposited Mica substrate with Epotek 353nd epoxy adhesive to form Si/Glue/Au/Mica sandwich structure. After 40 minutes curing of glue at 150oC, the ultra-flat TS gold was obtained by eliminating the mica with a knife. The prepared gold was scanned by AFM for 3-5 random spots as a quality test. For all cases, the average roughness of the gold was about 0.1 nm.

SAM growth
0.1 mM solution with target SAM molecules dissolved in toluene was prepared, with 10 minute deoxygenating by nitrogen bubbling. The freshly cleaved TS-Au without any further treatment was immersed into the solution, and incubated for 24 hours in vacuum. After SAM growth, the sample was rinsed with toluene, ethanol and isopropanol for several times to remove physisorped molecules. The sample after rinsing was blown with nitrogen for drying, and incubated in vacuum oven (10-2 mbar) overnight at 35oC for solvent evaporation.

QCM measurement
New QCM crystal (5mm diameter, f0 = 10MHz, from icryst) was cleaned by oxygen plasma for 10 minutes, immersed in hot DMF (100oC) for 2 hours, and in room temperature DMF overnight, washed with ethanol and isopropanol, and dried in vacuum oven for 20 hours at 35oC. The QCM measurement is done by an openQCM system. The initial resonance frequency of the cleaned substrate was recorded, and the substrate was used for SAMs growth. The procedure for SAMs growth was the same as for TS gold. The frequency of QCM crystal after SAMs growth was then recorded. The difference between the frequency before and after SAMs growth, Δf, can be used to determine the occupation area of a single molecule on the substrate via Sauerbrey equation: Where A' the electrode area, A the crystal area, Δf the frequency change, Mw the molecular weight, NA the Avogadro's number, µ the shear modulus of quartz, ρ the density of quartz, f0 the initial frequency.

2.4Seebeck measurement:
The Seebeck in this work is measured by thermoelectric atomic force microscopy (THEAFM). An Pt coated blunt probe was used as the top electrode, and the temperature of the bottom electrode (Au substrate) was controlled by a peltier stage. A Cu-CuNi thermocouple was used to calibrate the sample temperature after the measurement. The probe temperature was estimated by using a KMT thermistor probe (Pd-NiCr coating) with similar radius curvature to the Pt coated probe, and the temperature of the apex part of the thermistor probe was recorded at different sample temperature by a Maxwell bridge. Since the radius curvature of the thermistor probe was similar with the probe we use, we assume the apex temperature was similar in these two cases. The temperature difference between sample and probe was defined as ΔT. The voltage between sample and probe, VTherm, was extracted by a voltage pre-amplifier (SRS560) with a low pass filter, and recorded by the computer. The laser deflection of probe shifted a bit during temperature change due to thermal expansion, and we set it to a constant value manually by regulating the setpoint. The Seebeck coefficient was obtained from equation:

cAFM characterisation
The electrical characterization in this work was obtained by conductive AFM. The Pt coated probe was used as the top contact. A sweeping bias voltage from a function generator was applied between the gold and probe for current flow. The current signal between sample and probe was amplified by an IV converter (DLPCA200, Femto) and recorded by a computer. Figure S15: The distribution of differential conductance of SAMs 1-4.
The force between sample and probe was controlled by the laser deflection and fixed at 2nN. The contact area between sample and probe was estimated via Johnson-Kendall-Roberts (JKR) contact model. The contact radius, r, was calculated via equation: r the contact radius, F the loading force from probe to sample (2nN), R the radius of the probe, v1 and v2 the Poisson ratio of the material, E1 and E2 the Young's Modulus for probe and SAMs.
The radius of the probe was obtained from SEM image, and estimated to be 25 nm. The Young's modulus was obtained from AFM in peakforce mode, which was about 2 GPa for all the SAMs. Other parameters were obtained from literature working on similar systems.

SAM quality
The SAMs sample on TS gold was measured by AFM (multi-mode 8, Brucker) in peak force mode.
The roughness of the sample surface was obtained by nano-scope 9.0 software. The SAMs thickness was obtained by nano-scratching. A small region of molecular thin film was scratched by a stiff AFM probe at high force (50 nN), and the peak force mode was used to scan a large region. The scratched window can be observed at the large scale scan, and the height difference indicates thickness of the film as shown in Figure s16. The measured height and the corresponding tilting angle was listed in Table S2. Figure S16: The surface topography with nanoscratching and the height distribution of SAMs 1-4 characterized by AFM. Figure S17: FWHM of conductance (circles) and height (squares) distributions, obtained from Gaussian fits to the histograms in Figure 3, for each anchor group.

Optimised DFT Structures of Isolated Molecules
Using the density functional code SIESTA, 7,8 the optimum geometries of the isolated molecules 1-4 were obtained by relaxing the molecules until all forces on the atoms were less than 0.01 eV / Å as shown in Figure S18. A double-zeta plus polarization orbital basis set, norm-conserving pseudopotentials, an energy cut-off of 250 Rydbergs defined the real space grid were used and the local density approximation (LDA) was chosen to be the exchange correlation functional. We also computed results using GGA and found that the resulting transmission functions were comparable with those obtained using LDA. 9

Frontier orbitals of the molecules
The plots below show isosurfaces of the HOMO, LUMO, HOMO-1 and LUMO+1 of isolated molecules 1-4.

Product rule
Wave function plots for isolated molecules with their optimised geometries (Figures S14-S17) show iso-surfaces of the HOMO, LUMO, HOMO-1 and LUMO+1 of isolated molecules of the studied molecules. The orbital product rule [12][13][14] predicts CQI in the HOMO-LUMO gap for the molecules of study, because from Figures S19-S22, the product of the HOMO (LUMO) amplitudes at opposite ends of the molecules are negative (positive). Table S2 summarises the signs of these orbital products.

Binding energy of molecules on Au
To calculate the optimum binding distance between pyridyl/methyl sulphide anchor groups and Au(111) surfaces, we used DFT and the counterpoise method, which removes basis set superposition errors (BSSE). The binding distance d is defined as the distance between the gold surface and the N/SMe terminus of the pyridyl/methyl sulphide group. Here, compound 4 is defined as entity A and the gold electrode as entity B. The ground state energy of the total system is calculated using SIESTA and is denoted E . The energy of each entity is then calculated in a fixed basis, which is achieved using ghost atoms in SIESTA. Hence, the energy of the individual 4 in the presence of the fixed basis is defined as E and for the gold as E . The binding energy is then calculated using the following equation: We then considered the nature of the binding depending on the gold surface structure. We calculated the binding to a Au pyramid on a (111) surface with the nitrogen/methyl sulphide atom binding at a 'top' site and then varied the binding distance d. Figure S23 (left) shows that a value of d = 2.3 Å gives the optimum distance, at approximately 0.5 eV. As expected, the pyridyl anchor group binds favourably to under-coordinated gold atoms. For SMe d = 2.7 Å gives the optimum distance, at approximately 0.4 eV.

Optimised DFT structures of compounds within junctions
Using the optimised structures and geometries for the compounds obtained as described in section 2.1 (above), we again employed the SIESTA code to calculate self-consistent optimised geometries, ground state Hamiltonians and overlap matrix elements for each metal-molecule-metal junction. Leads were modelled as 625 atom slabs, terminated with 11-atom Au (111)

The tilt angle (θ)
In this section, we determine the tilt angle of each compound on a gold substrate, which corresponds to the experimentally measured most-probable break-off distance. Table S3 shows

HOMO-LUMO gaps
The calculated and optically measured HOMO-LUMO gaps are listed in Table S4. Theoretical gaps were calculated for isolated molecules and when the compounds are placed in the junctions, the gap between their HOMO and LUMO transmission resonances are quoted. As shown by the third and fourth columns in Table S4, isolated gaps for compounds 1, 2, 3 and 4 are larger than the gaps between the transmission resonances. This is because the latter are shifted by the real part of the self-energy of the contact to the leads, reflecting the fact that the system is more open when contacted to electrodes. In general, theoretical gaps are smaller than the measured gaps, which is consistent with the fact that DFT is known to underestimate its value. [15][16] log T(E) log T(E)

Transport calculations
The transmission coefficients T(E), obtained from using the Gollum transport code, were calculated for compounds 1-4 based on the tilt angle ranges in Table S3 (different curves of the same colour correspond to different title angles and the yellow line is the average). The LUMO resonance is predicted to be pinned near the Fermi Level of the electrodes for the four molecules, however, we choose Fermi Level to be in the vicinity of mid gap at ±0.5 eV (black-dashed line), as shown in Figure S29. Similarly, the conductance calculated for the same compounds as shown in Figure S29, yellow-lines indicate the average of each compound. S( V/K) S( V/K)

Seebeck coefficient
After computing the transmission coefficients of the four molecules, a study of their Seebeck coefficients was made.
To calculate the Seebeck coefficient of these molecular junctions, it is useful to introduce the nonnormalised probability distribution ( ) defined by where ( ) is the Fermi-Dirac function and ( ) are the transmission coefficients and whose moments are denoted as follows where is the Fermi energy. The thermopower, , is then given by where is the electronic charge.
Supplementary Figure S30 shows the Seebeck coefficients evaluated at room temperature for different energy range − .  3.10 Transport calculations for 1 and 2 (2Py and PySMe) at different anchor-electrode distances.
Figures S29 shows results at the optimum distances (d =0. 23 nm pyridyl and for SMe d =0. 27 nm, for 1, 2, 3 and 4), between each anchor group and the electrodes. However, for 1 and 2 (2Py and PySMe) based SAMs, the theoretical conductances computed using the optimum distance (of 0.23 nm) between the anchor groups and electrodes are significantly higher than the measured values. This occurs, because as shown in Figures S16 and S17, the film quality of these SAMs is poorer than that of 3 and 4 (2SMe-, 2SAc-)-based SAMs. Consequently, the actual anchorelectrode distances in these SAMs is greater than the optimum value and measured to be of the order of 0.50 nm. Hence, for 1 and 2, we show results as d varies from the optimum distance up to 0.50 nm. Figure S31 shows how the conductance decreases with increasing d, whereas the Seebeck coefficient is barely changed. Figure S31: Zero bias transmission coefficient T(E) of molecules 1-2 against electron energy E, for different electrode-anchor distance d (Top panel). Seebeck coefficient of the same molecules 1-2 (middle panel). Room-temperature Seebeck coefficients of 1, 2, 3 and 4 evaluated for a range of the tilt angles (see Table S2), (Lower panel) Figure S32: Electrical and thermoelectric properties of 1-4. For 3 and 4 the optimum distance is 2.3 Å. For comparison, results are shown as this increases up to 5.0 Å.