High performance multi-purpose nanostructured thin films by inkjet printing: Au micro-electrodes and SERS substrates

Nanostructured thin metal films are exploited in a wide range of applications, spanning from electrical to optical transducers and sensors. Inkjet printing has become a compliant technique for sustainable, solution-processed, and cost-effective thin films fabrication. Inspired by the principles of green chemistry, here we show two novel formulations of Au nanoparticle-based inks for manufacturing nanostructured and conductive thin films by using inkjet printing. This approach showed the feasibility to minimize the use of two limiting factors, namely stabilizers and sintering. The extensive morphological and structural characterization provides pieces of evidence about how the nanotextures lead to high electrical and optical performances. Our conductive films (sheet resistance equal to 10.8 ± 4.1 Ω per square) are a few hundred nanometres thick and feature remarkable optical properties in terms of SERS activity with enhancement factors as high as 107 averaged on the mm2 scale. Our proof-of-concept succeeded in simultaneously combining electrochemistry and SERS by means of real-time tracking of the specific signal of mercaptobenzoic acid cast on our nanostructured electrode.

Ne laser emitting at 632.8 nm (nominal power of 15mW) and a diode laser emitting at 785 nm (nominal power of 30mW).Sheet resistance has been measured with a four-probe system from Ossila, whose probe spacing is equal to 1.27 mm; three runs were recorded and mediated for each sample.Electrochemical measures were recorded using a CHI 660c Electrochemical work station.
LASER ABLATION: Gold nanoparticles were prepared by laser ablation synthesis in solution (LASiS) as previously reported. 2In brief, a plate of pure gold was placed in the bottom of a glass flask filled with 10µM NaCl water solution.A 9ns pulsed Nd:YAG laser (Quantel QSmart 850), operating at the 1064 nm fundamental wavelength and at 20Hz repetition rate, was focused on the target surface.The ablation was periodically monitored recording extinction spectra of the colloid and stopped once the AuNP concentration reached about 1 to 2 nM concentration.The nanoparticles concentration was calculated according to their molar extinction coefficient, obtained by the model presented in reference. 3NP_Cys: a fairly assumed sub-monolayer coverage of L-cysteine on AuNP was obtained by adding a stock solution of L-cysteine to the AuNP solution prepared by LASiS (2.0 nM) reaching a final concentration of L-Cys equal to 0.05 µM, so that about 24 molecules of L-cysteine per AuNP.The pH was adjusted to 10 by adding some drops of NaOH 1 M.The solution was stirred overnight at room temperature.Afterwards, centrifugation at 3000 rcf for 30 minutes was carried out.The supernatant was removed achieving a final concentration of AuNPs equal to 1.6•10 -8 M.This ink formulation was used without further processing.
AuNP_PVA: 10 mg/mL PVA in water was added to the AuNP solution prepared by LASiS (1.2 nM), achieving a final concentration of PVA equal to 1 mg/mL.The colloid was centrifuged at 3000 rcf for 30 minutes.The supernatant was removed.The obtained concentration of AuNPs was equal to 2.4•10 -8 M.This ink formulation was used without further processing.
Electrochemical-SERS measures: a brand-new electrode was inkjet-printed following the procedure already described for the pAuNP_PVAHT series, but designed as in Figure S15a.It was incubated overnight in MBA 2.6mM in NaOH 0.1M and rinsed with NaOH 0.1M before the experiments.A dedicated electrochemical cell was 3D-printed as well (Figure S15b), in which the electrode may be placed at the bottom and crewed in between the cell and a steal plate.An Ag/AgCl reference electrode and a Pt wire, as counter electrode, were used.The potentiometer was set to run multiple cycles in between -0.5 and -1.2V, at a scan rate or 0.01V/s.NaOH 0.1M was used as electrolyte.All of these were placed on the sample plate of the Raman instrument and continuous spectra acquisition (633 nm laser excitation at 6mW, 3s acquisition time, 10x objective magnification) were run along the cyclic voltammetry measurements.Raman spectra were baseline corrected and the band at about 1580 cm -1 was integrated for further analysis.that induce an increase of the local environmental refractive index. 4n Figure.S1 the UV-vis spectra of the pristine AuNPs, AuNP_Cys and AuNP_PVA are reported together with their TEM images and their corresponding size distributions.From the UV-vis spectra, it is possible to locate the plasmonic peak of the AuNPs at 520 nm for the as-synthesised AuNP, and at 522 nm and 523 nm for the AuNP_Cys and AuNP_PVA inks respectively.TEM analysis reveals a slightly higher and less polydisperse distribution, at 15.6(±6.1)nm and 16.1(±8.8)nm for AuNP_Cys and AuNP_PVA, respectively, compared to the laser ablated AuNPs spheres, which have a diameters distribution centred at 9.8(±7.9)nm.This dimension difference is a consequence of the sorting obtained by centrifugation, which is well-known to separate particles by size, with larger particles sedimenting faster than the smaller ones, indeed providing the narrower size distribution.The spectra were acquired with the 633 nm excitation source using a 50x objective.

PRINTING FEATURES:
The electrodes were patterned on glass substrates employing the Dimatix Materials Printer (DMP-2800, Fujifilm) equipped with 10 pL droplet cartridges.The glass slides were cleaned by immersion in acid piranha for 30 minutes.Afterwards, they were abundantly rinsed with distilled water and dried by N2 flow.The temperature of the platen was set to 40°C and the cartridge was let at ambient temperature.The waveform was optimized and it is illustrated in

ENHANCEMENT FACTOR CALCULATION PROCEDURE
A sample of MBA powder was used for the Raman measure (Figure S11), whereas for SERS experiment, a drop of 4-MBA solution (about 1 mM) was cast over the substrates and left to react for three hours without letting the drop dry.Afterwards, they were rinsed with water and dried gently under N2 flow.At this point, a μ-Raman map was carried out to investigate the efficiency of the samples as SERS substrates.
For calculating the EF the following equation has been followed: where ISERS is the surface-enhanced signal intensity of the probe at a specific band, Ibulk is the corresponding Raman intensity of its powder for the same band, NSERS and Nbulk are the numbers of molecules probed, respectively for the SERS and Raman measurements. 18ulk and ISERS were extrapolated from the experimental Raman and SERS spectra, respectively.
To determine Nbulk, an accurate approach was followedIt was assumed that the entire focal volume contributes to the measured signal.First, it is necessary to measure the effective sample volume irradiated by the laser.The confocal volume is given by: 19 where w0 is the probe volume radius and z0 is the half-height.Z0 can be measured experimentally and w0 can be derived by the Rayleigh limit given by: where n is the refractive index of the sample.
Z0 was experimentally measured by acquiring Raman spectra along the z-direction.From a gaussian fit of the curve obtained by plotting the intensity of the 4-MBA powder signal vs the position along the z-direction, the full width at half maximum, FWHM, was extrapolated.It is known that this value corresponds to 2z0. 19 is also known that the density of the 4-MBA is equal to 1.5 g•cm -3 , and thus, the number of molecules of probed by the Raman measure on the powder sample was accordingly calculated. 20different approach was used for the calculation of NSERS.In this case, not the confocal volume was used, but the sample surface under the laser spot.Again, z0 was measured experimentally by acquiring Raman spectra of the printed sample along the z-direction and, consequently, w0 was calculated.The area of the laser spot (A) was then given by: Ideally, it has been considered that a monolayer of 4-MBA was formed on top of the printed sample.
The geometric area under laser irradiation does not rigorously correspond to the real area of the sample accessible to the 4-MBA molecules, due to the roughness and nanostructured surfaces of the samples.Thus, to have an estimation as reliable as possible, the roughness measured by AFM measurements were employed.From the AFM, the real surface area was calculated, and thus, from the surface hindrance of the 4-MBA molecules, which is known to be equal to 0.38 nm 2 , 21 NSERS was estimated.
The measurements and relative calculations were performed for both the 633 and 785 nm excitations.Raman peak of MBA, where the displacement vectors are represented in red and the dipole derivative unit vector in orange.

Figure S1 .
Figure S1.UV-vis spectra, TEM images and nanoparticles size distribution of a) AuNPs prepared

Figure S2 .
Figure S2.Morphological characterization of the different substrates: a) dark-field optical

Figure S3 .
Figure S3.Morphological characterization of the different substrates: a) dark-field optical

Figure S6 :
Figure S6: Morphological characterization of the different substrates using dark-field optical

Figure S7 :
Figure S7: AFM topography images for the PVA rich regions in pAuNP_PVALT samples, the

Figure S8 :
Figure S8: Distribution of the elevations measured for the AFM images of Figure 2 and S2, S3

Figure S9 .
Figure S9.(a) Raman SERS spectrum of pAuNP_PVALT mediated over 6400 spectra collected in

Figure S12 :
Figure S12: Raman spectra of the MBA powder acquired using a 50x objective and 633 nm

Figure S13 :
Figure S13: SERS maps relative to the inkjet-printed substrates functionalized with MBA.The

Figure S14 :
Figure S14: SERS maps relative to the inkjet-printed substrates functionalized with MBA.The

Figure S14 .
Figure S14.The firing voltage was adjusted in the range of 20-25 V, depending on the nozzle used.

Figure S15 :
Figure S15: Waveform used and corresponding parameters for the jetting and non-jetting profile.