The influence of molecular mobility on the properties of networks of gold nanoparticles and organic ligands

Summary We prepare and investigate two-dimensional (2D) single-layer arrays and multilayered networks of gold nanoparticles derivatized with conjugated hetero-aromatic molecules, i.e., S-(4-{[2,6-bipyrazol-1-yl)pyrid-4-yl]ethynyl}phenyl)thiolate (herein S-BPP), as capping ligands. These structures are fabricated by a combination of self-assembly and microcontact printing techniques, and are characterized by electron microscopy, UV–visible spectroscopy and Raman spectroscopy. Selective binding of the S-BPP molecules to the gold nanoparticles through Au–S bonds is found, with no evidence for the formation of N–Au bonds between the pyridine or pyrazole groups of BPP and the gold surface. Subtle, but significant shifts with temperature of specific Raman S-BPP modes are also observed. We attribute these to dynamic changes in the orientation and/or increased mobility of the molecules on the gold nanoparticle facets. As for their conductance, the temperature-dependence for S-BPP networks differs significantly from standard alkanethiol-capped networks, especially above 220 K. Relating the latter two observations, we propose that dynamic changes in the molecular layers effectively lower the molecular tunnel barrier for BPP-based arrays at higher temperatures.


Experimental Section
Part 1: Synthesis of charge stabilized monodispersed gold nanoparticles in water Synthesis of monodispersed gold nanoparticles was made according to the following method. All solid reagents were purchased from Fluka and all organic liquid reagents were mainly purchased from Sigma Aldrich. The reagents were used in the state as they were received and some were diluted to the desired concentration. Reagents used and purchased from other suppliers will be mentioned. Further during the whole synthesis process (preparation of gold nanoparticles and 2D Au-NP-S-BPP arrays) we used Millipore demineralized water.
Monodispersed spherical gold particles with a diameter ~10 nm in water were synthesized through use of a method described by Slot and Geuze [1]. Charge stabilized gold nanoparticles (~10 12 particles/mL) were prepared through reduction of The SAc-BPP molecule was obtained by a cross-coupling reaction of 4-ethynyl-2,6di(pyrazol-1-yl)pyridine with 4-iodo-1-(thioacetyl)benzene using Sonogashira conditions. The procedure to synthesize the SAc-BPP molecule has been published elsewhere [2]. To obtain a higher yield the procedure was adapted and is described below.    In the hole of a Teflon block was pipetted 300 μl Millipore water. Then 30 L of the gold nanoparticles functionalized with SAc-BPP molecules in chloroform were deposited on to the water layer. Immediately the thin 2D single layer of Au-NP-S-BPP S12 array was formed on top of the water after evaporation of the chloroform in a fume hood (see Supporting Information, Figure S5).
The self-assembled 2D Au-NP-S-BPP array was picked up using a PDMS stamp (holding the stamp with a tweezer) from the water layer in the Teflon hole. The PDMS stamp with the 2D Au-NP-S-BPP array layer was then blown dry with nitrogen gas.
Here was used polydimethylsyloxane (PDMS) stamps for microcontact printing that were fabricated through a Sylgard 184 silicone elastomer (purchased from Dow Corning). Through mixing a prepolymer gel with a curing agent (Sylgard 184) the PMDS mixture was formed and poured inside a master model of a rectangular shape.
Through degassing the PDMS mixture at 20 ºC for at least 30 minutes, the bubbles from previous mixing will escape out of the PDMS mixture. The mixture was then baked in an oven at 60 °C for 90 minutes. The obtained silicone PDMS was cooled down and later removed from the master model. The silicone PDMS was cut into small cubic PDMS stamps of desired sizes. The prepared PDMS stamps were placed in a glass beaker containing ethanol (absolute). This beaker of PDMS stamps was S13 washed in ethanol through sonication by an ultrasonic bath for at least 15 minutes.
The prepared PDMS stamps were dried with nitrogen gas and stored in a storage box to be ready for microcontact printing of 2D Au-NP-S-BPP arrays. cooling stage was used for temperature dependent measurements. And the sample was allowed to equilibrate for 20 minutes at each temperature before collection of a S15 spectrum. All spectra were collected n = 3 times. The Raman band of a silicon wafer at 520 cm −1 was used to calibrate the spectrometer along with the Rayleigh line. The spectral data were acquired and analyzed using LabSpec software (Labspec, Horiba/Jobin-Yvon Group).

Part 5: Preparation of devices and setups for conduction experiments
A layer of self-assembled gold nanoparticles, formed on water, is microcontact printed several times on a HAR nanotrench electrodes device [5] via microcontact printing [4] (see Supporting Information, Figure S8). The HAR nanotrench device (gap length ≈100 nm, gap width ≈20 μm) is fabricated using a combination of opticaland E-beam lithography. We used photolithography to make larger contact pads (by first deposition of gold contact pads on a Si-SiO 2 substrate), followed by E-beam lithography to write and link T or bow-shaped nanotrench electrodes to the already gold deposited contacts on the substrate. After e-beam assisted gold deposition and metal lift-off a gold HAR nanotrench (containing electrical interconnects electrodes) device is ready for stamping [5]. Low temperature electrical measurements were performed on the nanotrench devices in a liquid helium bath cryostat. The DC electrical properties of these networks are measured by Agilent semiconductor parametric E5270B analyzer. S16  The resulting 'nanotrenches' of around 120 nm distance between the Au electrodes ensure a favourable aspect ratio for conductance measurements. These Au-NP-S-BPP networks have been obtained via three microcontact printing sequences of Au-NP-S-BPP arrays onto the nanotrench devices. In general, the networks are reasonably well ordered. However, single defects and larger voids (as seen in the image) will be present too. Although the latter may lead to more complicated percolation paths, conductance properties are expected to be dominated by the more ordered regions. We also note that ordering on the SiO 2 and on the Au nanotrench S19 electrodes is similar. The main reason for this is that the Au-NP-S-BPP arrays were self-assembled on water, before they were actually transferred to a substrate.

Modelling the charging energy of 2D Au-NP-S-BPP arrays as function of the dielectric constant
To get a feeling for the expected values of E C for these multilayered Au-NP-S-BPP networks, we plot E C in Figure S13 as a function of the dielectric constant based on three models to approximate the E C for a 2D Au-NP-S-BPP array. The following models are used in Figure S13 to determine the E C of a 2D Au-NP-S-BPP array. The first model is the simple spherical model to calculate the maximum limit of E C for a 2D Au-NP-S-BPP array (see equation S1). (S1) Where r is the radius "r" of the gold nanoparticles in a 2D Au-NP-S-BPP array (earlier determined by TEM of value 0.5*8.5 nm). The dielectric constant ε of a 2D Au-NP-S-BPP array (based on UV-Vis analyses) is set on 2.8.

S25
The second model is called the "sphere in sphere" model which estimates a minimum limit of E C. In the model "sphere in sphere" the total capacitance of a gold nanoparticle is calculated assuming it is fully surrounded by other gold nanoparticles. This is approximated by assuming that the metallic sphere is in a second concentric metal shell [6] (see equation S2).
( ) Where for "d" (the length of two entangled S-BPP molecules) is used 2 nm. The radius "r" remains set 4.25 nm.
The third model is called the "nearest neighbour" model [7] (see equation S3). From Figure S13 we determine that for 2D Au-NP-S-BPP array with a dielectric constant of 2.8, the E C of 2D Au-NP-S-BPP array ranges 0.019-0.060 eV.
Unfortunately, the models are rather crude so, a good quantitative comparison to the data is not possible. Note however that the experimental value of E C is expected to S26 be significantly larger for alkanethiol arrays than for Au-NP-S-BPP arrays, due to the lower dielectric constant in the alkanethiol arrays (see UV-Vis section in article).