scent molecular electronic devices from gold nanoparticles and terminal alkyne functionalised monolayer fi lms

Departamento de Qúımica F́ısica, Facultad 50009, Spain. E-mail: smartins@unizar.es Instituto de Nanociencia de Aragón (INA) y (LMA), Edicio i+d Campus Rio Ebro, Esquillor, s/n, 50018 Zaragoza, Spain Department of Chemistry, University of Dur Department of Chemistry, University of Live UK Instituto de Microelectrónica de Barcelona Bellaterra, Spain School of Chemistry and Biochemistry, Un Highway, Crawley, Perth, 6009, Australia Instituto de Ciencias de Materiales de Aragó 50009 Zaragoza, Spain Cite this: J. Mater. Chem. C, 2014, 2, 7348


Introduction
2][3] These test-bed junctions have revealed the critical roles played by not only the chemical structure of the backbone of the molecular component but also the nature of the electrode-molecule contact. 4,5Despite the substantial number of fundamental problems that have been faced by this emerging area of science, such as the uctuations in electrodemolecule contacts and hence coupling, quantum interference effects and the nature of the tunnelling to hopping transition, 6 and challenges for the imminent future including progressing device designs beyond the mimicry of conventional circuits, the manipulation of electron spin as well as charge transport designs and gated charge transport, 7 it is now possible to realistically contemplate molecular electronics as a potential, albeit future, technology with which to augment present-day silicon microelectronics technology. 8uilding on the success of measurements of molecular electronic characteristics in test platforms, the next signicant challenge to be addressed before molecular electronics can be considered a truly viable and scalable technology is the construction of robust and reliable metal-molecule contacts.A wide variety of molecular functional groups have been tested as contacting groups for various substrates (principally gold), including thiols, 9,10 amines, 11,12 carboxylic acids, [12][13][14] dithiocarboxylic acids, 15 esters, 16 pyridyl, [17][18][19] cyano, 20,21 isocyano, 22,23 isothiocyanato, 24 nitro, 20 methylselenide, 25 methylthiol, 25 dimethylphosphine, 25 trimethylsilane, [26][27][28][29][30][31] fullerenes, 32,33 etc.][36][37] The second important challenge in molecular electronics is the fabrication of the top contact electrode in two terminal sandwich-based metal-organic monolayer-metal devices.8][39][40] A wide variety of techniques to deposit the top metal electrode have been described in the literature including direct and indirect evaporation, 38,[41][42][43][44][45][46][47][48][49] use of liquid metals, 38,43,45,50,51 ip chip lamination, 45,52 electrodeposition, [53][54][55] and surface-diffusion-mediated deposition. 56We have recently shown that adsorption of aurate ([AuCl 4 ] À ) onto an ammonium ion terminated Langmuir-Blodgett lm followed by photoreduction is also a convenient, "so" method of assembling metal-molecule-metal structures with potential applications in the scaleable fabrication of molecular electronic devices. 57The most signicant problems in the deposition of the second electrode are those related to damage of the functional molecules during the metallization of a monolayer or penetration of the second metal through the organic lm, which results in a short circuit, rendering the device unusable. 37,40In addition, in most of these methods the metalmolecule contact is rather weak, whilst robust metal-organic junctions are required to improve the device performance, reproducibility, and stability.
Prompted by the current landscape, in this contribution we explore the fabrication of a metallic electrode on top of a monolayer of a 'wire-like' bis(phenylene ethynylene) derivative, leading to the formation of robust and conducting metalmolecule-metal (or GNP) junctions contacted via a covalent Au-C bond which results in the easiest electron transfer between the metal and the organic monolayer.This work is inspired by the recent experiments of Maity et al., 58,59 who obtained acetylide-decorated Au clusters from arylacetylene (-C 6 H 4 -C^C-H) terminated organic compounds and Au nanoparticles weakly stabilized by polyvinylpyrrolidone (PVP).These authors have demonstrated that the terminal hydrogen in the alkyne (-C^C-H) group is lost during this process with concomitant formation of an Au-C^C s-bond, accompanied by a weakening of the -C^Cbond of the alkynyl group.The signicant reduction of the pH value of the aqueous phase observed by these authors aer the ligand exchange indicated that the ligation of the alkynyl group proceeds via deprotonation of the alkyne.This method obviates the need for the formation of aryl radicals (e.g. from diazonium compounds) or alkynyl anions (e.g. from abstraction of the alkynyl proton with a strong base) used in other methods for formation of metal-C bonds.
Complementary DFT calculations concerning the adsorption of an ethynylbenzene radical on Au(111) have also shown that a strong covalent bond is formed with the surface upon removal of the terminal hydrogen-atom of the ethynyl group. 60In addition, Au-C covalent bonds have been recently studied in metalmolecule junctions, 58,61,62 exhibiting high contact conductance resulting from the direct metal-carbon coupling, [63][64][65] and prompting further consideration of this direct Au-C bond as a surface contacting group.
In the present work, an arylacetylene terminated molecule is used to fabricate and study metal-monolayer-metal junctions.A monolayer of an oligo (phenylene-ethynylene) (OPE) derivative, namely 4-((4-((4-ethynylphenyl)ethynyl)phenyl)ethynyl) benzoic acid (HOPEA), Fig. 1, assembled as a directionally oriented Langmuir-Blodgett (LB) monolayer 66 has been incubated in a dispersion of unprotected gold nanoparticles (GNPs), resulting in chemisorption of GNPs to the LB lm through Au-C bonds to give robust sandwich-like device structures in an experimentally simple fashion.

Results and discussion
Monomolecular LB lms incorporating HOPEA were deposited onto quartz, mica, glass and gold substrates initially immersed in the water subphase that were withdrawn resulting in organized molecular lms in which the carboxylate group is in contact with the substrate and the acetylenic (-C^C-H) moiety is oriented towards the air. 66These lms were incubated in a dispersion of freshly prepared GNPs for 0.5-4 hours.Immediately aer the removal of the substrates from the incubation solution the lms were thoroughly rinsed with Milli-Q® water to eliminate physisorbed GNPs from the lm surface.Finally, the lms were allowed to dry.The incubation time was optimized using quartz crystal microbalance (QCM) experiments (Fig. 2).The variation in the resonator frequency before and aer the incubation process is indirectly related to the mass of GNPs incorporated on the QCM substrate by means of the Sauerbrey equation which establishes that: 67 where f 0 is the fundamental resonant frequency of 5 MHz, Dm (g) is the mass change, A is the electrode area, r q is the density of quartz (2.65 g cm À3 ), and m q is the shear modulus (2.95 Â 10 11 dyn cm À2 ).A decrease in the frequency with the time during the incubation process reveals the incorporation of GNPs onto the LB lm until an incubation time of 2.5 hours, aer which the frequency remains constant indicating that further incorporation of GNPs onto the LB lm does not occur.
UV-vis spectra were recorded before and aer the incubation process in order to verify the assembly of GNPs onto the organic layer (Fig. 3).For the purpose of comparison, the inset of Fig. 3 shows the UV-vis spectrum of the aqueous dispersion of GNPs.The UV-vis spectrum of the pristine LB lm features one absorption band centered at 280 nm, attributable to p-p* electronic transitions associated with the OPE backbone of HOPEA. 66The spectrum of the LB lm aer the incubation process features two absorption bands at 259 nm and at ca. 540 nm, [68][69][70][71] revealing the incorporation of the GNPs onto the LB lm.A signicant red-shi of the surface plasmon resonance (SPR) band from 515 nm in the dispersion to 545 nm for the immobilized GNPs on the LB lm surface is indicative of a change in the dielectric constant surrounding the nanoparticle.In addition, this shi in the maximum absorption position of the SPR band is consistent with an interaction of the GNPs and the organic lm. 61he assembly of GNPs onto the lm has also been demonstrated by XPS.Fig. 4 shows the XPS spectrum of a HOPEA LB lm on a glass substrate aer the incubation process in a dispersion of GNPs.][74][75] However, neither QCM, UV-vis spectroscopy nor XPS results provide any information about the distribution, shape, or size of the GNPs on the surface of the lm.To investigate these issues, Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM) images of LB lms before and aer incubation in a dispersion of GNPs were obtained (Fig. 5).The pristine HOPEA lm shows a very homogeneous surface.In contrast, aer incubation in the dispersion of GNPs both AFM and SEM images show the appearance of spots distributed all over the lm revealing the presence of GNPs.The diameter of the GNPs immobilized on the monolayer surface determined by SEM was in the 15-30 nm range whilst the GNP height is in the 6-25 nm range as determined from AFM images.These compare with the 7-28 nm range of particle sizes determined for the initial dispersions by Dynamic Light Scattering (DLS) methods, see Experimental section.
In addition, incubation of HOPEA monomolecular LB lms caused signicant modication to the appearance, physical characteristics, and composition of the lm surface.Surface modication is veried by differences in contact angle measurements from pristine and incubated lms of HOPEA in a dispersion of GNPs.The contact angle of water onto a pristine HOPEA LB lm deposited onto a glass substrate is 60 , consistent with a well-formed hydrophobic LB lm.Aer incubation of the HOPEA monomolecular LB lm in a dispersion of GNPs for 2.5 h the contact angle is reduced to 43 indicating a more hydrophilic surface, with a contact angle value surprisingly close to the value associated with a bare gold substrate (39 ).
The previous set of experiments demonstrates the assembly of GNPs onto the monomolecular lm.Nevertheless, two   further important questions concerning the strategy proposed in this paper to fabricate the top contact electrode in metalorganic monolayer-GNP sandwich structures have still to be addressed: (i) what is the nature of the gold-organic layer interaction?and (ii) does the fabrication of individual GNP electrodes induce the formation of short-circuits due to metallic contact between the bottom and top-contact electrodes or, in contrast, do the GNPs serve as well-behaved top contact electrodes?
As mentioned in the Introduction of this paper, it has been demonstrated 58,59,75,76 that terminal alkynes react with GNPs with simultaneous abstraction of acetylenic hydrogen to give Au-C^C stabilized structures. 51,68,69With a view to exploring related processes between the terminal alkyne moiety in surface bound HOPEA lms and the deposited GNPs vibrational spectroscopy is a particularly useful tool, with several Raman and SERS studies 61,63,65,77 having shown evidence for the formation of a carbon-gold covalent bond in related systems.Fig. 6 shows the Raman spectra of HOPEA in the solid state as a powder, as well as the SERS spectrum of a HOPEA pristine LB lm deposited on silver mirrors and the spectrum aer the incubation procedure in a dispersion of GNPs.All spectra show three major vibrational bands at 2216, 1598, and 1131 cm À1 .The highest wavenumber band is associated with the localized vibrational motion of the internal alkyne moiety; the band at 1598 cm À1 is assigned to the symmetric stretch of the three aromatic rings along the long axis of the molecule, and the one at 1131 cm À1 arises from symmetric C-H bending of the phenyl C-H bonds (there are 12 phenyl C-H bonds in HOPEA). 78,79This band at ca. 1131 cm À1 is also present in an analogous compound to HOPEA without the terminal acetylene group, i.e., (4-[4-(phenylethynyl)phenylethynyl benzoic acid). 79The spectrum of the HOPEA LB lm aer the incubation process exhibits a low intensity band at 399 cm À1 indicative of the presence of a covalent Au-C s-bond. 56Such a band is not present in the Raman spectrum of HOPEA powder, nor in the spectra of pristine HOPEA LB lms.In addition, the small band at 2110 cm À1 , due to the terminal C^C group, in the powder Raman spectrum and the SERS spectrum of the HOPEA pristine lm is broadened and down-shied to 1990 cm À1 aer incubation of the lms in a dispersion of GNPs, also indicative of the formation of a -C^C-Au bond. 80he electrical characteristics of the structures fabricated as described above were determined from collection and analysis of current-voltage (I-V) curves arising from charge ow across the organic lms.Analysis of the I-V curves not only gives information concerning the conductance in the metal-monolayer-GNP assemblies but can also be used to rule out the formation of electrical short-circuits by incursion of the GNPs through the organic lm that eventually lead to direct metal-tometal contacts.As described in a previous paper, I-V curves were recorded with a conductive-AFM (Bruker ICON) in the Peak Force Tunneling AFM (PF-TUNA™) mode. 57This operation mode of the AFM is a valuable method for conductivity mapping of delicate samples since it avoids lateral forces that may otherwise damage the tip coating and the so sample surface.This mode also facilitates the use of cantilevers with low spring constant.The peak force tunneling AFM used here combines a conducting AFM tip and associated low-noise current amplier with a "tapping" mode AFM to probe current ow through these  metal-molecule-GNP junctions.Nevertheless a compromise has to be reached with respect to the set-point force before recording the I-V curves.This compromise involves not applying a too high peak force during the measurements, which would result in large deformation of the monolayer underlying the GNPs, or too little force which would result in inadequate contact between the tip and the surface leading to ineffective electrical probing.The deformation or damage of the monolayer has been investigated as a function of the tip loading force (set-point force) to determine the most suitable set-point force compromise.Fig. 7 shows a c-AFM image of a HOPEA LB lm aer the incubation process in a GNP dispersion using a setpoint force of 3.35 nN.Four gold nanoparticles labeled as GNP1, GNP2, GNP3, and GNP4 are clearly visible in this image.In the set-point force range between 3.35 and 10 nN, the analysis shows practically constant heights of 25, 19.5, 13.5, and 11.5 nm for GNP1, GNP2, GNP3, and GNP4, respectively.The practically constant GNP height in this set-point force range indicates that no signicant deformation of the monolayer occurs and that the GNPs are not substantially pushed into the monolayer by the AFM tip.If the set-point force is increased up to 16.75 nN, the section analysis of the GNPs gives heights of 21, 16.2, 10 and 9 nm for GNP1, GNP2, GNP3, and GNP4 respectively.These reduced height values reveal deformation of the monolayer for this applied set-point force.A further increase in the deformation of the monolayer occurs when a set-point force of 25.2 nN is applied since the section analysis shows heights of 18.8, 14.3, 8.8, and 7.4 nm for GNP1, GNP2, GNP3 and GNP4, respectively.Moreover, if the set-point force is tuned down to low values, e.g. to 6.7 nN, the section analysis of these GNPs shows heights of 24.2, 19.1, 13 and 11.3 nm.This restoration of height values indicates that the deformation produced in the monolayer aer applying a high set-point force (25.2 nN) is elastic and does not induce permanent damage of the organic layer.
Having evaluated the inuence of the applied set-point force on the monolayer, the I-V curves were recorded.To record these current-voltage curves, the c-AFM tip was located on the nanoparticle and a bias voltage was applied between the sample and the tip and tip voltage was swept (AE1.1 V) with the LB-coated Au substrate set to ground.When a set-point force of 3.35 nN was used any resulting current which might have owed is below the detection limit (see conductance versus force data in Fig. 7), whilst for a set-point force between 3.35 and 10 nN the I-V curves show low conductance.These results suggest that when low set-point forces are used to record the I-V curves, the electrical contact between the tip and the GNP is not sufficient to support measureable electrical current ow.Nevertheless, when a higher set-point force (between 16.7 and 25.1 nN) was applied, the I-V curves show a signicant conductance (Fig. 7), revealing that for these set-point forces there is a good electrical contact between the tip and the GNP, while no damage to the monolayer occurs.Fig. 8 shows a representative I-V curve of all the curves (ca.550) recorded on different GNPs when a set-point force of 16.7 nm was applied.The bottom inset of Fig. 8 shows the conductance histogram built from all the experimental data (10 I-V curves on each of different GNPs) in the À0.45 to 0.45 V ohmic region for each of the 550 I-V curves obtained experimentally by placing the AFM tip on top of different GNPs applying a set-point force of 16.7 nN.Different GNPs yield different I-V curves whose slope (and thereby conductance of the device) increases with the GNP diameter.All the curves measured exhibit the typical shape observed for metal-molecule-metal junctions, with a linear section only at relatively low bias voltages and increasing curve gradients at higher bias.Importantly, only curves with this behavior were observed, both  over GNPs and on the organic monolayer not covered by GNPs (top inset of Fig. 8), and no low resistance traces characteristic of metallic short circuits were obtained over a wide range of setpoint forces which rules out the presence of short-circuits conrming that robust and reliable top-contacts have been constructed without damaging the underlying organic monolayer lm.

Experimental
In a previous paper we reported the synthesis of HOPEA as well as the optimum conditions needed to fabricate well-ordered, defect-free and directionally oriented LB lms of this compound on mica, gold and quartz substrates. 66The same protocols have been used in this contribution to fabricate Langmuir-Blodgett (LB) lms incorporating HOPEA.A Nima Teon trough with dimensions (720 Â 100) mm 2 , which was housed in a constant temperature (20 AE 1 C) clean room, was used to fabricate the lms.A Wilhelmy paper plate pressure sensor was used to measure the surface pressure (p) of the monolayers.The subphase was a pH 9 NaOH aqueous solution (Milli-Q water, resistivity 18.2 MU A 1 Â 10 À5 M solution of HOPEA in hexane-ethanol (2 : 1) (both solvents purchased from Aldrich and used as received; purity HPLC grade 99% and >99.5%, respectively) was spread onto the aqueous surface.The spreading solvent was allowed to completely evaporate from the surface of the subphase over a period of at least 20 minutes before compression of the monolayer commenced at a constant sweeping speed of 0.015 nm 2 per molecule per min.The solid substrates used for the transfer were cleaned carefully as described elsewhere. 81,82The monolayers were deposited by the vertical dipping method onto several substrates at a constant surface pressure of 18 mN m À1 and a speed of 3 mm min À1 .
The dispersion of GNPs was prepared by adding rapidly 0.5 mL of a 1.0 Â 10 À3 M NaBH 4 aqueous solution to 30 mL of a 1.0 Â 10 À5 M HAuCl 4 aqueous solution with vigorous stirring at 2 C using an ice-water bath.The hydrodynamic diameter of these GNPs was found to be in the 7-28 nm range, as determined by Dynamic Light Scattering (DLS).DLS measurements were carried out using a NanoZS ZEN3600 instrument from Malvern.Incubation of the LB lms in the dispersion of GNPs (at 2 C) took place immediately aer mixing the reactants.
UV-vis spectra were acquired on a Varian Cary 50 spectrophotometer and recorded using a normal incident angle with respect to the lm plane.Atomic force microscopy (AFM) experiments to study the topography of the monolayers were performed by means of a Multimode 8 AFM system from Veeco, in tapping mode.The data were collected with a scan rate of 1 Hz and under ambient air conditions by using a silicon cantilever provided by Bruker, with a force constant of 40 mN m À1 and operating at a resonant frequency of 300 kHz.Scanning electron microscopy (SEM) images were obtained with a JEOL JSM 6400 microscope.X-ray photoelectron spectroscopy (XPS) spectra were acquired on a Kratos AXIS ultra DLD spectrometer with a monochromatic Al Ka X-ray source (1486.6 eV) using a pass energy of 20 eV.The photoelectron take off angle was 90 with respect to the sample plane.To provide a precise energy calibration, the XPS binding energies were referenced to the C1s peak at 284.6 eV.Quartz crystal microbalance (QCM) measurements were carried out using a Stanford Research System instrument and employing AT-cut, a-quartz crystals with a resonant frequency of 5 MHz having circular gold electrodes patterned on both sides.Contact angle experiments were performed with a commercial optical tensiometer Theta Lite from Attension.Raman and surface-enhanced Raman scattering (SERS) spectra were collected using a Confocal Raman Imaging from Witec, model Alpha300M+ with an excitation wavelength of 633 nm.Silver islands (thickness 9.1 nm) were prepared in an Edwards model 306 vacuum coater from a resistively heated tungsten boat.The substrates were Zuzi glass microscope slides cleaned in piranha solution for 30 min (3 : 1 97% H 2 SO 4 : 30% H 2 O 2 ), rinsed with deionized water, and dried in a stream of N 2 .Care: piranha solutions are exceptionally corrosive and highly oxidizing.Contact between piranha solutions and organic materials is considered extremely hazardous and must be avoided.During lm deposition, the background pressure was maintained at 5 Â 10 À7 Torr, and the deposition rate (0.02 nm 3 s À1 ) was monitored on an Electron Beam Evaporator Auto 500 from BOC Edwards.Aer deposition, annealing was performed at 200 C for 60 minutes.The conducting-AFM (c-AFM) measurements were performed with a Bruker ICON microscope under humidity control, ca.30%, with a N 2 ow, in the Peak Force Tunneling AFM (PF-TUNA™) mode, and employing a PF-TUNA™ cantilever from Bruker (coated with Pt/Ir 20 nm, ca. 25 nm radius, 0.4 N m À1 spring constant and 70 kHz resonance frequency).

Conclusions
In this contribution, GNPs have been assembled onto an organic LB lm to give arrays of metal-molecule-GNP devices.Formation of covalent Au-C s-bonds through a simple wetchemical procedure has been demonstrated by SERS whilst I-V curves ruled out the presence of short-circuits.The terminal alkyne moiety therefore appears to be a suitable functional group for the subsequent deposition of metal nanoparticles which can serve as the top contact electrode.The electrical measurements averaged over multiple gold-molecule-GNP assemblies show the suitability of this approach of forming C-Au bonds to the top GNP contact.This could be a useful approach for future nano-electronic assemblies which exploit more robust C-Au junctions rather than the typically more uxional chemisorption chemical bonds used in conventional self-assembly.If needed these GNPs could be employed as seeds for deposition of a contiguous metal lm using complementary methods such as electroless metal deposition subsequently leading to a controlled preparation of the top contact electrode in metal-monolayer-metal devices with a signicant inhibition in the formation of short-circuits.

Fig. 2
Fig. 2 Amount of gold deposited onto a HOPEA monomolecular film incubated in a dispersion of GNPs as a function of the incubation time.

Fig. 3
Fig.3UV-vis spectrum of a pristine HOPEA monomolecular LB film and a spectrum of the same film after incubation in a dispersion of GNPs for 2.5 hours.The inset figure shows the UV-vis spectrum of the aqueous dispersion of GNPs.

Fig. 4
Fig.4XPS spectrum of Au4f photoelectrons of a HOPEA monomolecular LB film on a glass substrate after incubation in a dispersion of GNPs for 2.5 hours.

Fig. 5
Fig. 5 AFM (top) and SEM (bottom) images of a pristine HOPEA monomolecular LB film (left) deposited onto a mica substrate and the same LB film after incubation for 2.5 hours in a dispersion of GNPs (right).Images are 2 Â 2 mm 2 in size and the Z range of the AFM images is 30 nm.For a better view of the size and shape of the GNPs both AFM and SEM images have been magnified (500 Â 500 nm).

Fig. 6
Fig. 6 Raman spectrum of HOPEA powder, SERS spectrum of a pristine HOPEA monomolecular LB film, and SERS spectrum of a HOPEA monomolecular LB film after incubation in a dispersion of GNPs.

Fig. 7
Fig. 7 Height of GNPs determined with the c-AFM at the indicated set-point forces together with the average conductance values measured by locating the tip of the c-AFM on the indicated GNPs.The inset top image shows a representative example of a 350 Â 350 nm 2 image where GNPs can be clearly distinguished and was used to position the c-AFM tip onto the GNPs; the Z range is 50 nm.

Fig. 8
Fig.8Representative I-V curve obtained experimentally by positioning the c-AFM tip on top of a GNP when a set-point force of 16.7 nN was applied.Inset bottom: conductance histogram built from all the experimental data from À0.45 to 0.45 V for each I-V curve obtained (ca.550 curves).Inset top: representative I-V curve obtained by positioning the c-AFM tip on the organic monolayer not covered by GNPs when a set point-force of 6.7 nN was applied; if higher set-point forces are applied damage on the LB film is produced.