Plasma-Enabled Amorphous TiO2Nanotubes as Hydrophobic Support for Molecular Sensing by SERS

Plasma-enabled amorphous TiO2 nanotubes as hydrophobic support for molecular sensing by SERS Nicolas Filippin,1‡ Javier Castillo-Seoane,1,2‡ M. Carmen López-Santos,1,2 Cristina T. Rojas,1 Kostya (Ken) Ostrikov,3,4 Angel Barranco,1 Juan R. Sánchez-Valencia,1,2,* and Ana Borrás1* Dr. Alejandro Nicolas Filippin, Javier Castillo-Seoane, Dr. Juan R. Sánchez-Valencia, Dr. Cristina T. Rojas, Dr. Angel Barranco, Dr. Ana Borrás 1 Nanotechnology on Surfaces and Plasma Group, Materials Science Institute of Seville (Consejo Superior de Investigaciones Científicas – Universidad de Sevilla) c/Américo Vespucio 49, 41092, Sevilla, Spain. Javier Castillo-Seoane, Dr. Juan R. Sánchez-Valencia, Dr. Carmen López-Santos 2 Departamento de Física Atómica, Molecular y Nuclear, Universidad de Sevilla, Avda. Reina Mercedes, E-41012, Seville, Spain. Prof. Kostya (Ken) Ostrikov. E-mail: [email protected] 3 School of Chemistry and Physics, Queensland University of Technology, Brisbane, QLD 4000, Australia. 4 CSIRO-QUT Joint Sustainable Processes and Devices Laboratory, Lindfield NSW 2070, Australia. ‡ These two authors have contributed equally to the work KEYWORDS: Ag plasmon, superhydrophobicity, light scattering, PICT enhancement mechanism, core/shell nanostructures. ABSTRACT: We devise a unique hetero-nanostructure-array to overcome a persistent issue of simultaneously utilizing the Surface-Enhanced Raman scattering, inexpensive, Earth-abundant materials, large surface areas, and multifunctionality, to demonstrate near single-molecule detection. Room-temperature plasma-enhanced chemical vapour deposition and thermal evaporation provide high-density arrays of vertical TiO2 nanotubes decorated with Ag nanoparticles. The role of the TiO2 nanotubes is threefold: (i) providing high surface area for the homogeneous distribution of supported Ag nanoparticles, (ii) increasing the water contact angle to achieve superhydrophobic limits and (iii) enhancing the Raman signal by synergizing the localized electromagnetic field enhancement (Ag plasmons) and charge-transfer chemical enhancement mechanisms (amorphous TiO2) and by increasing the light scattering because of the formation of vertically aligned nanoarchitectures. As a result, we reach Raman enhancement factor up to 9.4 x 107 satisfying the key practical device requirements. The enhancement mechanism is optimized through the interplay of the optimum microstructure, nanotube/shell thickness, Ag nanoparticles size distribution, and density. Vertically aligned amorphous TiO2 nanotubes decorated with Ag nanoparticles with a mean diameter of 10-12 nm provide enough sensitivity for near-instant concentration analysis with an ultra-low few-molecule detection limit of 10-12 M (Rh6G in water), and the possibility to scale up device fabrication. Introduction Chemical sensors have a crucial role in our daily life: the knowledge about the exact chemical composition of the air we breathe or the water we drink is crucial for our health as well as the early diagnosis of diseases. In this regard, Surface Enhanced Raman Spectroscopy (SERS) based sensors1,2 have extraordinary potential due to their outstanding characteristics such as an impressively high sensitivity which can lead even to single-molecule detection,3,4 a very high selectivity provided by a Raman spectrum whose peaks can be taken as a fingerprint determining unambiguously the analyte5 including label-free sensing and nondestructive detection approaches.6 Two mechanisms can be accounted for the enhancement of the Raman scattering intensity, namely the electromagnetic (EM) enhancement associated with the localized surface plasmon resonances (LSPR) on noble metallic nanostructures, and the chemical (charge transfer, CT) enhancement from the photo-induced dynamic electron transfer between the probe molecules and nearby metallic or semiconducting nanostructures. The strongest electromagnetic enhancement factors (EF up to 1014) are usually related to the formation of the so-called hot-spots regions between metallic nanoparticles which can even further enhance the amplification of the Raman signal with respect to isolated metal nanoparticles,2 offering the possibility of developing ultra-sensitive SERS-based biosensors.7–10 It is worth mentioning that the LSPR frequencies of metal nanoparticles such as Ag and Au vary in a broad range depending on the nanoparticles' sizes, shapes, and environment. Therefore, a major challenge to increase the EM enhancement factors is the optimization of the hot-spots density and resonance intensity of the LSPRs to be able to detect the target molecules. The CT enhancement, on the other hand, is a weaker process (EF up to 103) attributed to the charge transfer induced by the molecule–metal interaction. Thus, under favorable energy band alignment conditions, electrons can be transferred between the molecule and the sensor, involving new resonant electronic transitions of the adsorbed analyte.11 Hence, contrary to the EM enhancement, the CT mechanism depends on both the substrate and the molecule. Such enhancement is not only ensured by noble metals but also opens the scope for using semiconductor and transition metal oxides as effective SERS-active supports. However, the bottleneck in such semiconductor-based SERS substrates is their relatively weak enhancement factors, which hinders their practical applications.12 Several approaches have been proposed to enhance the CT mechanism. For example, pre-irradiation of semiconductor nanomaterials with UV light to generate oxygen vacancies provides new resonant states and the associated additional enhancement in Raman signal (EF up to 104), whereas the oxygen plasma treatment can induce additional active sites on the surface.13,14 In general, the strongest CT mechanism of semiconductors can be attributed to the Photon-In- duced Charge Transfer (PICT) process between the substrate and the target molecules.13,15 The PICT enhancement mechanism does not rely on chemisorption of a molecule on the surface of the semiconductor. Indeed this enhancement mechanism is based on the matching of both, between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of the molecule and the energy levels of Valence Band (VB) and Conduction Band (CB) of the substrate, and between the incident photon energy and the bandgap of the semiconductor. Recent reports of SERS enhancement relying on CT mechanism focus on semiconducting nanoparticles and quantum dots, Si and Ge nanowires/nanotubes, ZnO nanosheets, NiO nanofibers, Cu2O nanowires, WO3 hierarchical nanostrutures, -Fe2O3 and TiO2 nanostructured surfaces and nanoparticles, and other nanostructured oxides.16,17,18 Despite the advances in the exploitation of both EM and CT mechanisms, the main remaining challenge for the analytical application of SERS is the lack of substrates with both high and uniform signal enhancement suitable for large-scale production.19 Consequently, one of the main drawbacks of large-area plasmonic SERS sensors with high enhancement factors is the need for homogeneous dispersion of the hot spots on the samples. This dispersion necessitates the sophisticated nanofabrication based on nanolithography,20–22 which requires expensive equipment and time, or chemical methods, with further drawbacks related to the reduced control of the particle dispersion over the surfaces (e.g., due to uncontrolled particle aggregation), reproducibility and high environmental costs.23,24 Several options to address these issues have been proposed, such as the synthesis of nanoparticles with different shapes such as nanorods,25 cubes,26 or stars,23 which enhance the electromagnetic field in the vicinity of the edges,20 or the use of template-assisted methods. The latter methods include the assembly of metal nanoparticles on the template surface via the layer-by-layer (LBL) self-as- sembly,27,28, and the in situ reductive or vacuum deposition of Ag or Au nanocrystals onto templates or nanostructured supports.29–31 The template-assisted methods are promising for fabrication of SERS substrates that can add additional functionalities such as photocatalytic activity and recyclability,31–35 compatibility with microfluidic systems,30 bactericidal properties,36 in situ SERS monitoring of chemical reactions,33,37,38 the further enhancement of the Raman signal through the PICT mechanism13,14 or applying Photon Induced Enhanced Raman Spectroscopy (PIERS).15 Thus, in recent articles in the literature, the coupling between Ag plasmonic nanoparticles and multifunctional TiO2 nanostructures has been the combination of choice.7,12,26,30,31,34,38 For instance, the decoration with Ag nanoparticles of TiO2 nanotubes formed by anodic oxidation of Ti foils.30,31,34 In this context, a recent alternative achieving both high sensitivity and reproducibility is the combination of plasmonic probe sensors with wettability controlled supports, including superwettable interfaces,39 hydrophilic-hydro- phobic combinations40,41 and high water contact angle supports.21 Concretely, hydrophobic surfaces allow concentrating molecules on a reduced area of the sensor, which can be even limited to micro-areas for superhydrophobic surfaces (with WCA > 150º). In this way, small-volume droplets can be deposited onto a hydrophobic surface covered with SERS active probes and analyzed once the droplet is dried and the analyte concentrated.42 This approach has been recently demonstrated for the ultra-sensitive detection of common antibiotic residues in water.43 Very recently, the optimization of a hydrophobic support with hydrophilic nanoprobes has allowed the ultrasensible detection of tumor markers avoiding the coffee-ring effect characteristic of the SERS analysis on hydrophobic surfaces.44 Thus, although the underlying principle is simple, the main limitation of a supported superhydrophobic SERS- based sensor is that the surface should not interfere with the Raman signals. A further challenge is to work out the components of the sensor with synergistic coupling for the Raman enhancement. Amorphous TiO2 is an excellent candidate to fulfill these requirements as support of plasmonic nanoparticles. On the one hand, it has been recently reported for the chemical enhancement of the Raman signal, in the form of nanoparticles and core@shell nanostructures.43 On the other hand, TiO2 nanostructured surfaces provide tunable hydrophobicity since their water contact angle depends strongly on the morphology, microstructure, porosity, and surface chemistry.45,46 Thus, TiO2 one-dimensional fibers or tubes can be used to significantly increase hydrophobicity.46,47 Moreover, these surfaces show reversible conversion from superhydrophobic to superhydrophilic states that are easily achievable by UV or visible illumination and thermal treatments.48,49 An additional striking property of one-dimensional supported TiO2 nanostructures for the enhancement of the SERS signal relies on the light-cou- pling resonance and multiple light scattering effects. Scheme 1. Our innovative approach for the fabrication of low-cost, ultra-sensible SERS sensors and prospective applications. The light scattering effect makes the photons have a longer optical path and longer time to interact with the plasmon nanoparticles and also enhance the charge transfer mechanism by boosting the concentration of photogenerated electron-hole pairs. In this context, the scope and key innovation of this work are represented in Scheme 1. Here we aim to demonstrate an advanced procedure for the production of SERS sensors with high efficiency and homogeneity, low cost, and low analyte consumption. The proprietary vacuum-plasma soft-template assisted method is applied to fabricate vertically aligned supported one-dimensional TiO2 nanotubes (NTs) decorated with silver nanoparticles (Ag NPs) as a SERS active element. The control in the synthesis of these nanostructures allows us to generate a highly hydrophobic architecture optimized for SERS with a high concentration of exposed active sites for the support of plasmonic Ag NPs. The method provides an effective route for the optimization of the density of hot-spots, nanoparticle and nanotubes diameters, and control of the water contact angle. Besides, it relays on widely established protocols working at a large scale in microelectronics and deposition of optical coatings. We optimize the SERS platform to test the Rhodamine 6G (Rh6G) as a reference analyte, showing an enhancement factor as high as 9.4 x 107 owing to the synergistic SERS coupling of the EM and CT mechanisms, light scattering, and intrinsic ability to concentrate the analyte at the surface of the tips of the nanoparticle-decorated nanotubes. EXPERIMENTAL SECTION Substrates. Samples were fabricated on n-type Si(100) and fused silica. The substrates were cut in different sizes, up to 5 x 5 cm2, rinsed in acetone, and dried with a flux of dry nitrogen. Materials. Metal-free phthalocyanine (H2Pc) was purchased from Fluka and used as received. Titanium tetraisopropoxide (TTIP) and Rhodamine 6G were acquired from Sigma-Aldrich and used as received. Silver with a purity of 99.99% purity was acquired from Kurt J. Lesker. The Rh6G solutions were prepared with ultrapure Milli-Q water immediately before use. Sample fabrication. The fabrication procedure consists of a multistep vacuum / plasma combination process as shown in Scheme 2. It can be divided into four stages: I) Formation of a thin layer (200 nm) of mesoporous and amorphous TiO2 by Plasma Enhanced Chemical Vapor Deposition (PECVD) which provides the necessary roughness for the subsequent growth of phthalocyanine Nanowires (NWs).50 The deposition conditions were: plasma source operating at 650 W, substrates at room temperature, the total pressure in the chamber was  8.5 x 10-3 mbar, and oxygen was employed as a reactive gas. II) Phthalocyanine NWs were grown by vapor transport in mild vacuum, denoted in the scheme as Organic Physical Vapor Deposition (OPVD) by means of a Knudsen effusion EM CT Synergy X ✓ X MOx Ag NPs Conventional Ag/MOxstructures Ultra-low detection EM CT Synergy TiO2 Ag NPs Vertically Aligned TiO2-NTs@Ag-NPs ✓ ✓ Innovation Tunable plasmons Multifunctional TiO2 Plasma &Vacuum Deposition Supported Low-cost Templates ↑Vertical Alignment ↑High density ↑Wettability control ↑Light management ↑Nano-chemistry on the tip ✓ Applications Ultra-low detection Scale-up Re-use Low-cost Ultra-low detection Multifunctional Self-cleaning Cell under the following conditions: Ar pressure of 2 x 10-2 mbar, substrate temperature of 180 ºC and growth rate of 0.4 Å/s (measured by QCM). The final thickness was set to 1.2 kÅ in QCM, which roughly corresponds to 2-3 μm long NWs. These ONWs will be used as the 1D-scaffold for the deposition of the outer TiO2 shell. III) Growth of the TiO2 shell by PECVD to form core@shell (H2Pc@TiO2) hybrid NWs. The TiO2 shell was deposited by PECVD under the following conditions: plasma source operating at 650 W, Ar and O2 were used in a relation of 9:1, while keeping the total pressure close to 6 x 10-3 mbar. The deposition was carried out at room temperature. These conditions lead to the formation of nanoporous and homogeneous amorphous shells.51–53 Two different TiO2 nominal thicknesses were compared, 200 nm and 400 nm. The evacuation of the NWs core is carried out by annealing the samples in air at a temperature of 350 ºC with a ramp of 2 ºC/min to evacuate the H2Pc (not shown in the Schematic). These conditions result in the formation of domed nanotubes (NTs). Please note that this annealing temperature can be reduced using organic molecules with lower sublimation temperature, as many porphyrins and perylenes (see ref.50,54–56). Also, the results were compared with a thin film reference (nano-TiO2 thin film), deposited under the same conditions without the organic template, and with polycrystalline anatase nanotubes fabricated by the same methodology but carrying out additional annealing at 450 ºC during more than 3 hours. IV) Deposition of Ag NPs by thermal evaporation in high vacuum conditions. Silver nanoparticles (Ag NPs) were deposited by thermal vacuum evaporation of Silver at a pressure of 10-6 mbar and a deposition rate of around 0.1 Å/s using an alumina crucible on a tungsten evaporator supplied by Kurt J. Lesker. Three different equivalent thicknesses of Ag were deposited: 1 nm, 2 nm, and 3 nm. Please consider that these numbers correspond to the thickness estimated using a QCM and calibrated using the deposition on Si(100) as reference. The actual diameter of the nanoparticles will be revealed below after statistical analysis of SEM images. V) The last step consists of the deposition of the R6G drops on the surface of the samples to carry out the SERS tests. Microstructural and chemical characterization. SEM micrographs were acquired in a Hitachi S4800 working at 2 kV. Energy Dispersive X-Ray Spectrometry (EDS) analysis (EDX X-Max 80T, Oxford Instruments) was carried out in the same facility working at 20 kV and 15 mm of working distance. The samples were dispersed onto Holey carbon films on Cu or Ni grids from Agar scientific for TEM characterization. HAADF-STEM and HREM images were acquired with a FEI Tecnai F30 S-Twin STEM operated at 200 kV and a FEG S/TEM FEI Tecnai G2F30 at 300 kV, correspondingly. Micro-Raman spectroscopy. All measurements were carried out in LabRAM Horiba Jobin Yvon spectrometer equipped with a confocal microscope with a 100x objective and a green laser of 532 nm wavelength. A 10x objective was utilized to locate the coffe-ring mark produced after the evaporation of the analyte solvent. Water contact angle. The water contact angle (WCA) was measured by the static and dynamic sessile drop method in a contact angle meter OCA 20 from DataPhysics Instruments GmbH using Milli-Q water. In all cases, drops of 5 µL were dispensed. The accuracy of the WCA value given for a droplet is 0.01º and it is obtained after least squared fitting of the droplet profile to a semisphere. UV irradiation for photocatalytic removal of the analyte. Samples were illuminated in air under ambient conditions for 20 min with the full spectrum of a 125 W Xe lamp. Calibration of the light intensity at the position of the samples yielded a photon intensity of 2 W cm-2. RESULTS AND DISCUSSION The Experimental Section gathers the synthetic route for the fabrication of the TiO2 nanotubes decorated with Ag nanoparticles, from now on labeled as TiO2 NTs@Ag NPs, addressing the inner and outer position of the TiO2 and Ag in the radial distribution of the materials, respectively (see Scheme 2). The synthesis methodology consists of a multistep combination of vacuum and plasma-enhanced chemical vapor deposition (PECVD) and has been exploited previously for the development of one-dimensional nanoscale waveguides,47 superhydrophobic surfaces,46,57 excitonic solar cells, 51,52 and piezoelectric nanogenerators.58 One of the great advantages of this method is the straightforward control of the nanotubes density and length as well as on the shells composition, microstructure, and thickness.51,52,56,58,59 The PECVD conditions selected herein for the formation of the nanotubes shell provide the growth of amorphous, continuous, and smooth TiO2 layers as detailed elsewhere56,59 (see also Experimental Section) with pore sizes in the nanometer range (i.e., below 2 nm). Two shell thicknesses were selected for this study, labeled from now on as TiO2 (200 nm) and TiO2 (400 nm), which refers to the nominal thickness measured by SEM on the equivalent thin films, i.e. the film deposited during the same experiment on a flat substrate, without the organic nanowires used as 1D template (see Experimental). Scheme 2. Scalable and low-cost fabrication of a SERS-based sensor device: I) Plasma enhanced chemical vapor deposition of the seed-layer, II) formation of the singlecrystalline ONWs, III) growth of the nanoporous-TiO2 shell, vertical alignment and evacuation of the organic core leaving TiO2 NTs, IV) thermal evaporation of Ag to form the Ag NPs. Figure 1 presents characteristic SEM images of the TiO2 (400 nm) NTs fabricated by the soft-template approach using TiO2 PECVD layers as both “seed” interface with the substrate and as the NTs shell. As clearly seen in panels ab), the protocol provides large areas of high-density vertically aligned TiO2 arrays. The plasma deposition is critical for both the vertical alignment of the TiO2 tips and the conformal deposition of the shell. The increase of the shell thickness from 200 to 400 nm improves the vertical alignment of the nanotubes as shown in the comparison in Figure S1. Such characteristics, along with the high density of 1D supported nanostructures, are expected to yield large water contact angles (WCAs) with high homogeneity over large areas.46,57 The tips surface of the NTs presents a low roughness which may be interpreted as small pores produced as a consequence of the annealing treatment carried out for the elimination of the organic core (see high magnification pictures in Fig. 2 a) and HAADF-STEM images in Fig. 2 b-c). The internal pressure generated as a consequence of H2Pc sublimation is very likely released through these pores.51 Such homogeneously distributed pores may act as nucleation sites for the subsequent formation of the silver nanoparticles, as will be discussed below. This domed or closed morphology differs from previous SERS supports reported in the literature based on open TiO2 nanotubes generated by anodic oxidation of Ti foils.30,31,34 Figure 1. TiO2 nanotubes are vertically aligned, can be produced at a large scale, and uniquely offer dense nucleation sites for Ag nanoparticle decoration on their tips. SEM images of TiO2 NTs deposited on Si(100) for an equivalent shell thickness of 400 nm, a) tilted cross-view at 17º, b-c) top views at different magnifications. The hollow microstructure of the TiO2 NTs after the evacuation of the organic core is revealed by HAADF-STEM in Figure 2 b-c). These micrographs also demonstrates the highly conformal and homogeneous shells formed by the TiO2 deposition conditions. In this case, the TiO2 shells are amorphous, however, the method allows deposition of polycrystalline anatase shells as well as core@shell systems as detailed elsewhere.51,52,56,58,59 We have also prepared anatase TiO2 by increasing the substrate temperature (see Experimental Section) to compare the performance of amorphous and crystalline metal oxide supports. The XRD patterns comparing the amorphous and anatase nanotubes are presented in the Figure S2 of the SI section. Thermal evaporation of silver in high vacuum conditions on flat substrates provides the formation of Ag NPs with size distribution controlled mainly by two experimental parameters, namely, the deposition rate and time. Figure S3 in the Supporting Information presents the Ag NPs deposition on Si (100) for a thickness of 2 nm determined using a Quartz Crystal Monitor (QCM). The nanoparticles are rounded with an average NP area of 78.4 nm2 calculated by the imaging analysis software. Such size distribution results in a strong surface plasmon resonance absorption situated at 430 nm in the corresponding UV-Vis spectrum (see Figure S3). It is a common practice to carry out additional annealing treatments on the metal nanoparticles up to 500-700 ºC to tune the metal nanoparticle plasmons.60– 62 As introduced in Scheme 1, we aim to keep the energy consumption as low as possible in the fabrication protocol. Consequently, our samples are tested as-grown, without further annealing treatment. 5 5 um a) Substrate “seed” layer (TiO2thin film) TiO2nanotubes array b) c) 200 nm3 um Figure 2. Formation of the nanoporous shell. SEM (a) and HAADF-STEM (b-c) images revealing the pores formed on the shell and the empty core after the removal of the organic template. Hence, we propose as an alternative approach, the tuning of the silver plasmon resonances by controlling the deposited silver thicknesses. Figure 3 presents representative SEM images of the equivalent deposition on the vertical TiO2 nanotubes for three different silver nominal thicknesses, 1, 2, and 3 nm, see Figure S3 for the comparison of the UV-Vis transmittance spectra. The comparison between panels a), b) and c) reveals that both the density and size of the nanoparticles depend on the nominal thickness (see statistical analyses in Figure S4). The NTs decorated with 1 and 2 nm of silver present a higher number of nanoparticles than the TiO2 NT@Ag NPs (3 nm) sample (dNPs 4.2, 2.9, and 1.65 x 103 NPs/m2 for 1, 2 and 3 nm correspondingly). The nanoparticles size distributions follow the opposite trend with medium diameters estimated as 7.6  2.2 nm, 11  3 nm, 19  4 nm for the NTs represented in panels (a-c). Besides, it is worth stressing the high homogeneity of the distribution of the Ag nanoparticles on the nanotubes (see also Figure S5) which can be straightforwardly linked to the distribution of pores on the TiO2 nanotubes after removal of the organic template (Figure 2). Thus, the silver nuclei would preferentially form on such pores, working as high surface tension sites, yielding tiny nanoparticles growing with narrow size distributions (see Figure S4). The crystal structure of the silver particles was resolved by employing TEM and HRTEM (Figure 4), the latter showing an atomically resolved image of the Ag NP at the tip. The Ag NP digital diffractogram pattern (DDP) in the inset corresponds to the 011 zone axis of the cubic phase of Ag. Figure 3. Silver nanoparticles densely and uniformly anchored over the tips of TiO2 nanotubes enable ultra-sen- sitive molecular detection by enhancing photoresponses from multiple plasmonic hot-spots between the nanoparticles. SEM images of Ag NPs on 400 nm TiO2 NTs for three different silver nominal thicknesses, 1 nm (a), 2 nm (b, d), and 3 nm (c). As mentioned in the introduction, to further improve the sensitivity of a SERS sensor, it is highly desirable to concentrate the analyte. The water contact angle is crucial for the improved sensitivity of the sensor since hydrophobic samples exhibit a reduced liquid-solid contact area that concentrates the analyte molecules during the droplet evaporation (see also Figure S6). In this series of experiments, the water contact angle of the studied samples depended on the equivalent thickness of both the TiO2 shell and Ag NPs. All surfaces under analysis show a strong hydrophobic character with WCAs higher than 120º. Figure 4. Nanocrystalline Ag nanoparticles formed by thermal evaporation on the porous shell. High-resolution TEM micrograph of 3 nm (nominal) Ag NPs on 400 nm TiO2 NTs (a) and a magnified view of one Ag NP (b). The DDP of the selected NP is presented in the inset of b). The WCA is higher for the TiO2 (200 nm) NTs, which present a superhydrophobic behavior with WCA > 160º and roll-off-angles below 10º. Such a difference is in good agreement with the Cassie-Baxter mechanism.63 Thus, the TiO2 NTs (200 nm) surface presents a smaller solid fraction in comparison with the corresponding to TiO2 NTs (400 nm), leaving larger space for air gaps. The combination of a low solid fraction and the hydrophobic character of the PECVD TiO2 surface leads to the development of superhydrophobicity. In regards to the application of these surfaces, Figure 5 a) presents a series of photographs showing the evaporation of a 5 µL drop of water on top of TiO2 NTs (400 nm) decorated with the three silver equivalent thicknesses, 1 (left), 2 (middle) and 3 nm (right). One important characteristic of the evaporation process is that the liquid-solid area is not significantly altered during the evaporation process. To visualize this effect, higher magnification images have been taken at different evaporation times for TiO2 NT decorated with Ag NPs (1 nm) (see Figure 5 b). The initial contact diameter (top) is approximately the same as the one taken at 35 min (bottom). This measured diameter is around 1.9 mm, meaning a solid-liquid contact area of 2.75 mm2, very close to the theoretical value calculated in Figure S6, 2.6 mm2 for an initial WCA of 125º. This is a very important aspect of the evaporation process because it validates the subsequent calculations of the enhancement factors (EFs). To reveal the process dynamics, Figure 5 c) shows the values for the the fitted WCA at different times of the evaporation process, showing how the initial WCA affects the evaporation time. Thus, the higher the initial WCA, the lower the evaporation time with differences in the order of minutes for WCAs variations in the range between 126º and 143º. One of the most important aspects of our methodology is its scalability. Indeed, several steps of the fabrication gathered in Scheme 1 (see Experimental Section) have already been successfully tested at wafer-scale in industrial processes. Therefore, the development of these sensors is scalable to large areas with very high homogeneity in the formation of the nanotubes and nanoparticles, while maintaining the low cost and substrate compatibility features. Figure 5 d) shows photographs of the TiO2 NTs decorated with Ag nanoparticles on a 2 x 3 cm2 silicon substrate. Figure S7 in the SI section gathers the EDS characterization of a sample with a diagonal larger than 5.5 cm. The media and maximum variance in the Ag/Ti ratio (in atomic percentage) is: (2.50  0.27) x 10-3 at%. We have obtained these values after processing the EDS spectra from 5 different points as schematized in the figure. We additionally did the EDS experiment on a 1.5 x 1.5 cm samples with a large amount of silver (equivalent thickness on a flat substrate of 6 nm) and found also a positive result, with media and maximum variance in the Ag/Ti ratio of (9.16  0.56) x 10-3 at%. Please note that samples larger than 10 x 10 cm2 can be reproducibly fabricated by the same procedure using an ample variety of substrates, from glass slides to cellulose, including metal foils and polymers.60,64 Figure 5. High Water Contact Angle surfaces allow the concentration of analytes on millimetric areas and multiple testing on the same sensor. a) Pictures of a 5 μL water drop deposited on top of TiO2 NTs (400 nm) samples with different thicknesses of silver at different times. b) Higher magnification image showing the liquid-solid diameter contact for the surface of the sample TiO2 NTs (400nm)@Ag NPs (1nm). c) Water contact angle evolution following the evaporation of a 5 μL water drop deposited on TiO2 NTs (400nm) samples with different thicknesses of silver. d) Photographs of a TiO2 NTs (400 nm)@Ag NPs (2nm) sample with 5 μL drops of Rh6G/water at different concentrations (left) and after droplets evaporation (middle). The right side picture is contrast-enhanced to highlight the evaporated drops. The photograph in Figure 5 d) shows the ordered drop patterns before (left) and after (middle) their evaporation. The right-hand side of Figure 5 d) presents an enhanced contrast photograph revealing three circles left by the drops, which is a typical “coffee ring” residue due to the concentration of the analyte around the external area of the liquid-solid area.65 This effect is very favorable for a SERS sensor since i) it concentrates even further the analyte along this ring and ii) the dried drops are easily spotted by the naked eye, which makes the subsequent Raman alignment and measurement much easier. The SERS mapping in Figure S8 gathers the maxima for the aromatic C-C stretching band at 1350 cm-1 vibration in the surrounding of the coffering mark. Importantly, the process of concentrating solute in small areas after evaporation does not modify or damage the substrate. Previous references have demonstrated the advantage of the TiO2 as photocatalytic support to allow the removal of the analyte and reutilization of the surface.31-35 In this study, the samples could be easily reused by removing the medium analyte with a 50% ethanol-water mixture and then dried or by photocatalytic removal upon UV light irradiation taking advantage of the TiO2 multifunctional properties. Figure S9 shows how the illumination with UV light of a TiO2 NTs (400 nm) @ Ag NPs (1 nm) completely removes the Rh6G for a solution of the 10- 6 M in water. Thus, along with the aforementioned advantages of this sensing platform, the possibility of reusing surfaces after the analysis is another important aspect of the potential industrial applicability of these structures. Stamplecoskie et al. reported a maximum enhancement factor of the Ag nanoparticles enabled Surface Enhanced Raman effect for an optimal Ag NPs diameter between 30 and 50 nm. Variations in the order of tens of nanometers led to differences in orders of magnitude in Raman intensities.66 Accordingly, in our case, the optimal NP diameter would correspond to the Ag NPs with 3 nm of nominal thickness which presents nanoparticles sizes up to 30 nm (See Figure S4). However, the development of an efficient SERS-based sensor also requires a compromise between the size, shape, density and chemical identity of NPs.2,67–69 In this study, the sample decorated with a silver nominal thickness of 2 nm exhibits smaller particles but with a higher density on the NTs tips (up to 230 particles at the tips) which would eventually increase the formation of hot-spots. Our data indicate that the main contribution to the SERS signal enhancement arises from the tips of the nanotubes. This possibility is supported by the previous reports regarding the nanocarpet effect on 1D nanomaterials as carbon nanotubes and ZnO nanowires.57,70 This effect mitigates the common issue of clustering and reorganization of vertical 1D arrays by capillarity forces after spreading and evaporation of water on their surface. Specifically, the nanocarpet effect can be used as a fingerprint of the droplet profile and depth on vertical nanorods depending on the WCAs.57 Thus, for apparent WCAs > 130º the droplet 0 min 2 min 10 min 20 min 25 min 30 min 35 min 40 min 42 min 43 min 0 min 2 min 10 min 20 min 25 min 30 min 35 min 37 min 39 min 40 min 0 min 2 min 10 min 20 min 25 min 30 min 35 min 40 min 41 min 42 min Ag 1 nm Ag 2 nm Ag 3 nm 0.5 mm 1.87 mm a) b) c) d) 0 5 10 15 20 25 30 35 40 45 0 20 40 60 80 100 120 140 1 nm Ag 2 nm Ag 3 nm Ag Water Contact Angle (º) Time (min) Dried drops 5 mm 5 mm the Development of Metal Nanostructured Layers. Adv. Mater. Interfaces 2017, 4 (14), 1601233. https://doi.org/10.1002/admi.201601233. (55) Alcaire, M.; Sanchez-Valencia, J. R.; Aparicio, F. J.; Saghi, Z.; Gonzalez-Gonzalez, J. C.; Barranco, A.; Zian, Y. O.; Gonzalez-Elipe, A. R.; Midgley, P.; Espinos, J. P.; Groening, P.; Borras, A. Soft Plasma Processing of Organic Nanowires: A Route for the Fabrication of 1D Organic Heterostructures and the Template Synthesis of Inorganic 1D Nanostructures. Nanoscale 2011, 3 (11), 4554–4559. https://doi.org/10.1039/C1NR11001B. (56) Borras, A.; Cotrino, J.; Gonzalez-Elipe, A. R. Type of Plasmas and Microstructures of TiO2 Thin Films Prepared by Plasma Enhanced Chemical Vapor Deposition. J. Electrochem. Soc. 2007, 154 (12), P152. https://doi.org/10.1149/1.2794289. (57) Macias-Montero, M.; Borras, A.; Alvarez, R.; Gonzalez- Elipe, A. R. Following the Wetting of One-Dimensional Photoactive Surfaces. Langmuir 2012, 28 (42), 15047–15055. https://doi.org/10.1021/la3028918. (58) Filippin, A. N.; Sanchez-Valencia, J. R.; Garcia-Casas, X.; Lopez-Flores, V.; Macias-Montero, M.; Frutos, F.; Barranco, A.; Borras, A. 3D Core-Multishell Piezoelectric Nanogenerators. Nano Energy 2019, 58, 476–483. https://doi.org/10.1016/j.nanoen.2019.01.047. (59) Borrás, A.; Barranco, A.; González-Elipe, A. R. Design and Control of Porosity in Oxide Thin Films Grown by PECVD. In Journal of Materials Science; Springer, 2006; Vol. 41, pp 5220– 5226. https://doi.org/10.1007/s10853-006-0431-y. (60) Fang, S.-U.; Hsu, C.-L.; Hsu, T.-C.; Juang, M.-Y.; Liu, Y.-C. Surface Roughness-Correlated SERS Effect on Au Island- Deposited Substrate. J. Electroanal. Chem. 2015, 741, 127–133. https://doi.org/10.1016/j.jelechem.2015.01.028. (61) Quan, J.; Zhang, J.; Qi, X.; Li, J.; Wang, N.; Zhu, Y. A Study on the Correlation between the Dewetting Temperature of Ag Film and SERS Intensity. Sci. Rep. 2017, 7 (1), 1–12. https://doi.org/10.1038/s41598-017-15372-y. (62) Pisarek, M.; Holdynski, M.; Roguska, A.; Kudelski, A.; Janik-Czachor, M. TiO2and Al2O3nanoporous Oxide Layers Decorated with Silver Nanoparticles—Active Substrates for SERS Measurements. J. Solid State Electrochem. 2014, 18 (11), 3099– 3109. https://doi.org/10.1007/s10008-013-2375-x. (63) Cassie, A. B. D.; Baxter, S. Wettability of Porous Surfaces. Trans Faraday Soc 1944, 40 (0), 546–551. https://doi.org/10.1039/TF9444000546. (64) Liehr, M.; Wieder, S.; Dieguez-Campo, M. Large Area Microwave Coating Technology. Thin Solid Films 2006, 502 (1), 9–14. https://doi.org/10.1016/j.tsf.2005.07.226. (65) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Contact Line Deposits in an Evaporating Drop. Phys. Rev. E - Stat. Phys. Plasmas Fluids Relat. Interdiscip. Top. 2000, 62 (1 B), 756–765. https://doi.org/10.1103/PhysRevE.62.756. (66) Stamplecoskie, K. G.; Scaiano, J. C.; Tiwari, V. S.; Anis, H. Optimal Size of Silver Nanoparticles for Surface-Enhanced Raman Spectroscopy. J. Phys. Chem. C 2011, 115 (5), 1403–1409. https://doi.org/10.1021/jp106666t. (67) Fan, M.; Lai, F.-J.; Chou, H.-L.; Lu, W.-T.; Hwang, B.- J.; Brolo, A. G. Surface-Enhanced Raman Scattering (SERS) from Au:Ag Bimetallic Nanoparticles: The Effect of the Molecular Probe. Chem Sci 2013, 4 (1), 509–515. https://doi.org/10.1039/C2SC21191B. (68) Le Ru, E.; Etchegoin, P. Principles of Surface-Enhanced Raman Spectroscopy: And Related Plasmonic Effects, 1st ed.; Elsevier Science: Amsterdam, 2009. (69) Yockell-Lelièvre, H.; Lussier, F.; Masson, J. F. Influence of the Particle Shape and Density of Self-Assembled Gold Nanoparticle Sensors on LSPR and SERS. J. Phys. Chem. C 2015, 119 (51), 28577–28585. https://doi.org/10.1021/acs.jpcc.5b09570. (70) Hildebrandt, P.; Stockburger, M. Surface-Enhanced Resonance Raman Spectroscopy of Rhodamine 6G Adsorbed on Colloidal Silver. J. Phys. Chem. 1984, 88 (24), 5935–5944. https://doi.org/10.1021/j150668a038. (71) Macias-Montero, M.; Peláez, R. J.; Rico, V. J.; Saghi, Z.; Midgley, P.; Afonso, C. N.; González-Elipe, A. R.; Borras, A. Laser Treatment of Ag@ZnO Nanorods as Long-Life-Span SERS Surfaces. ACS Appl. Mater. Interfaces 2015, 7 (4), 2331–2339. https://doi.org/10.1021/am506622x. (72) Borras, A.; Alvarez, R.; Sanchez-Valencia, J. R.; Ferrer, J.; Gonzalez-Elipe, A. R. Critical Thickness and Nanoporosity of TiO2 Optical Thin Films. Microporous Mesoporous Mater. 2012, 160, 1–9. https://doi.org/10.1016/j.micromeso.2012.04.035. (73) Wang, X.; Shi, W.; Jin, Z.; Huang, W.; Lin, J.; Ma, G.; Li, S.; Guo, L. Remarkable SERS Activity Observed from Amorphous ZnO Nanocages. Angew. Chem. Int. Ed. 2017, 56 (33), 9851–9855. https://doi.org/10.1002/anie.201705187. (74) Ohsaka, T.; Izumi, F.; Fujiki, Y. Raman Spectrum of Anatase, TiO2. J. Raman Spectrosc. 1978, 7 (6), 321–324. https://doi.org/10.1002/jrs.1250070606. (75) Ling, Y.; Zhuo, Y.; Huang, L.; Mao, D. Using Ag-Em- bedded TiO2 Nanotubes Array as Recyclable SERS Substrate. Appl. Surf. Sci. 2016, 388, 169–173. https://doi.org/10.1016/j.apsusc.2016.01.257. (76) Yang, L.; Jiang, X.; Ruan, W.; Zhao, B.; Xu, W.; Lombardi, J. R. Observation of Enhanced Raman Scattering for Molecules Adsorbed on TiO2 Nanoparticles: Charge-Transfer Contribution. J. Phys. Chem. C 2008, 112 (50), 20095–20098. https://doi.org/10.1021/jp8074145. (77) Ma, Y.; Du, Y.; Chen, Y.; Gu, C.; Jiang, T.; Wei, G.; Zhou, J. Intrinsic Raman Signal of Polymer Matrix Induced Quantitative Multiphase SERS Analysis Based on Stretched PDMS Film with Anchored Ag Nanoparticles/Au Nanowires. Chem. Eng. J. 2020, 381, 122710. https://doi.org/10.1016/j.cej.2019.122710. (78) Le Ru, E. C.; Blackie, E.; Meyer, M.; Etchegoin, P. G. Surface Enhanced Raman Scattering Enhancement Factors: A Comprehensive Study. J. Phys. Chem. C 2007, 111 (37), 13794– 13803. https://doi.org/10.1021/jp0687908. (79) Petryayeva, E.; Krull, U. J. Localized Surface Plasmon Resonance: Nanostructures, Bioassays and Biosensing—A Review. Anal. Chim. Acta 2011, 706 (1), 8–24. https://doi.org/10.1016/j.aca.2011.08.020. 18 Table of Contents