Tape of the truth: Ta2O5 nanopore array formed under broad potential range and SERS potential after silver sputtering

The deposition of a plasmonic metal layer on a nanostructured oxide surface is one of the important methods of preparing a platform for surface-enhanced Raman scattering (SERS) measurements. In this contribution, we describe the formation of SERS substrates by the deposition of a silver layer on ordered a Ta2O5 nanopore array. The influence of various experimental anodization process parameters on the morphology of a Ta2O5 nanopore array was carefully studied. It was found that the formation of a Ta2O5 nanopore array is possible under a broad potential range (15–50 V) in a highly acidic solution containing F− ions. In some cases, the nanopore array structures were covered by an outer layer rich in F− and SO42− ions, which could easily be removed using adhesive tape or by sonication. The deposition of an Ag layer led to SERS activity. The optimal Ag layer thickness was specified based on SEM and DRS measurements. The SERS substrates formed exhibited high point-to-point, sample-to-sample and time durability.


Introduction
The cross section of Raman scattering is very low [1,2]. Therefore, for decades, Raman scattering was not considered a very sensitive analytical method. Its limitations, though, can be overcome by the application of plasmonic nanostructures. There are two main mechanisms explaining this type of enhancement: electromagnetic [3] and chemical [4] (also known as charge transfer). Both mechanisms assume a contribution by a plasmonic surface, and so the method is called SERS-surface-enhanced Raman scattering. In SERS, the Raman signal is significantly enhanced by the presence of plasmonic nanostructures. A very important parameter in SERS measurements is what is known as 'enhancement factor' (EF). This parameter determines the ratio of the measured SERS signal to the normal Raman signal for the same amount of molecules [5][6][7]. Currently, typical EF for the most active samples range from 10 5 -10 6 [6]. In some cases, even single-molecule detection is possible [8][9][10].
The SERS enhancement factor (EF) is an essential parameter for characterizing the SERS effect. In order to calculate EF, the signal intensities and the number of molecules under SERS and normal Raman conditions have to be compared. In this way, the EF permits a simple, quick comparison between two different SERS platforms. However, a good quality SERS substrate should provide a stable EF over the entire substrate (point-to-point analysis) and from substrate to substrate (sample-to-sample analysis), and should demonstrate high chemical and temporal stability [11].
Nowadays, many research groups from around the world are trying to form new, efficient, stable SERS platforms. SERS spectra can be collected from a sol of plasmonic nanoparticles [12,13], but the resulting measured signal is not homogeneous due to the presence of hot spots [14]. A promising SERS preparation method involves coating highly ordered nanomaterials with a plasmonic layer. The nanostructures can be prepared by many methods: chemical etching [15][16][17], epitaxial growth [18][19][20], solvothermal methods [21][22][23] or anodization [24][25][26][27]. In some reports, such a template was covered by a thin plasmonic film [16,18] in others by plasmonic nanoparticles [26,28]. An especially efficient SERS platform is one based on anodized aluminum oxide (AAO), where the AAO was used as a template for plasmonic metal deposition, and then dissoluted, leading to the formation of densely packed plasmonic nanoparticles [29,30].
In this work, we studied the influence of various parameters on the formation of a Ta 2 O 5 nanopore array by means of the anodization method in a highly acidic electrolyte containing Fions [31]. So far, in the literature, Ta 2 O 5 nanopore arrays have only been reported under a narrow potential range [32]. In this contribution, it was found that a Ta 2 O 5 nanopore array can be formed under a wide (15-50 V) potential range. In some cases, scanning electron microscope (SEM) analysis revealed some surface contamination. X-ray photoelectron spectroscopy (XPS) and energydispersive X-ray spectroscopy (EDS) analyses showed high concentrations of F and S in such a layer. Such contamination can be easily removed using adhesive tape or by. The Ta 2 O 5 nanopore array was then coated with Ag layers of various thicknesses. The structural and optical properties of the obtained materials were investigated. Based on SEM and DRS (diffuse reflectance spectroscopy) measurements, the optimal thickness of the deposited Ag thin film was determined. The SERS platform based on a Ta 2 O 5 nanopore array coated with an Ag layer exhibited a high EF, as well as good point-to-point, sample-to-sample and temporal durability. The significant advantages of the proposed Ta 2 O 5 -based SERS platform are fast anodization-the anodization time is below 3 min-and high repeatability in comparison with other nanomaterials fabricated by anodization [26,28,33] or chemical methods [34].

Preparation of Ta 2 O 5 nanopores array
The tantalum foil was divided into pieces 0.6 cm 9 1.2 cm in size and then cleaned by successive immersion in acetone, ethanol and water, and sonicated in each solution for 5 min. The initial SEM analysis showed that the tantalum foil was flat at the nanoscale and did not require any flattening procedure. The synthesis of the Ta 2 O 5 nanopore array was carried out in a two-electrode cell where the Ta foil was used as the anode and a platinum rod served as the cathode. The anodization was carried out in a solution composed of 7.16 ml H 2 O, 1.83 ml of concentrated HF and 74 ml of concentrated H 2 SO 4 , under a constant potential in a range of 15-50 V for 60-500 s. After the anodization, the foil was rinsed with distilled water and then immersed in water for sonication for 2 min.

Deposition of an Ag layer
The deposition of a silver layer on the Ta 2 O 5 nanopore array substrates was carried out using a Quorum Q150ES sputter coater. In all cases, a constant current of 20 mA was applied. The thickness of the layer deposited was controlled by changing the duration of the deposition, based on a prepared calibration curve.

Experimental techniques
Scanning electron microscopy (SEM) analyses of the nanoparticles formed were carried out using a Merlin field emission scanning electron microscope (SEM) (Zeiss, Germany) equipped with an energy-dispersive X-ray microanalysis (EDS) probe (Bruker).
A Microlab 350 spectrometer (Thermo Electron, East Grinstead, UK) was used for the XPS measurements. AlKa (hm = 1486.6 eV, 300W) was used as the radiation source. Signals were recorded with a hemispherical analyzer at constant pass energies of 100 eV (survey spectra) or 40 eV (high-resolution spectra). The background was corrected using the Shirley model. In order to deconvolute the spectra, a mixed Gaussian/Lorentzian asymmetric function with a constant value of G/L = 0.3 was used. The position of the bands was corrected relative to the position of the C1s carbon band at 284.5 eV. An Advantage Surface Chemical Analysis program was used to analyze and develop the results.
Diffuse reflectance spectra (DRS) were recorded using a UV-2600i Shimadzu spectrometer equipped with an integrating sphere with the internal detectors working in the 1400-200 nm field.
Raman spectra were collected with a Horiba Jobin-Yvon Labram HR800 spectrometer equipped with an Olympus microscope with a 50 9 long distance objective, a holographic grating with 600 grooves/ mm and a Peltier-cooled (1024 9 256 pixel) CCD detector. A diode pumped; frequency doubled Nd:YAG laser (532 nm) provided the excitation radiation.

Results and discussion
Ta 2 O 5 nanopore array characterization

Structural characterization
The Ta 2 O 5 nanopore array surfaces were formed by a standard anodization process in which the nanopores grew perpendicular to the tantalum substrate. During the anodization, the Ta foil acted as an anode, and hence the metallic tantalum was oxidized to the Ta 5? ions. Subsequently, the Ta cations were transferred into the Ta 2 O 5 . The structure of the formed Ta 2 O 5 nanopore array depended on the balance between the rate of electrochemical formation of the oxide film at the metal foil surface and the rate of oxide dissolution in acids. The platinum mesh, which served as a counter electrode, acted as the cathode on which hydrogen is formed. The whole process can be described by the following equations: It was assumed that the rate of oxide dissolution was much higher at the pore base than on the walls due to the field-focusing effect [35]. The formation of tantalum oxide with hydrogen occurs mainly at the pore base, which led to an effect of self-induction acidification resulting in a pH gradient [36]. The high viscosity of the electrolyte limited the diffusion rate and thereby helped maintain the pH gradient. This effect permitted the nanopore array to form.
The mixing of water with concentrated acid is a highly exothermal reaction, and so just after being prepared the electrolyte is hot. Before the anodization, the electrolyte solution was cooled to 22°C in an ice bath. Many experimental parameters can effect the growth of a Ta 2 O 5 nanopore array [37]. One of these is the distance between the anode and the cathode. We found that the optimal distance was 1 cm. The distance between the electrode affects the current density, where increasing the distance led to a decrease in the current density [38]. A small electrode distance resulted in a significant increase in the electric field, which in turn promoted the electrochemical dissolution process and led to the formation of nanotubes having a well-defined shape. It was found that, with increasing electrode distance, the number of defective nanotubes increased. Even small changes in this parameter lead to a drop in the reaction yield. The other important factor was the ratio of the surfaces of the two electrodes. Samples with well-defined Ta 2 O 5 nanopores were only formed when the Pt surface was at least as large as the surface of the Ta foil (Fig. 1).
Many literature reports have looked at the formation of nanotubes or nanopore arrays in a two-step anodization process. Such a procedure is used when the layer formed using a one-step anodization is not highly ordered, and some tubes/pores are significantly deformed. In such a situation, the formed layer can be removed by sonication. For this purpose, the analyzed sample was immersed in a plastic cup filled with Millipore water placed in the center of the sonication bath. The sonication process was conducted for 1 min. During the sonication, cavitation bubbles induced by the high-frequency pressure (sound) waves appear across the sample surface. The sonication bath we used employs a 40 kHz ultrasound wave. The accumulative effect of the imploding bubbles leads to the cleaning of the surface. The anodization is then repeated in conditions the same or similar conditions to those of the first step of the process. By removing of the original layer, a surface having regular wells and holes is obtained. These surface structures are places for the preferential growth of highly ordered nanotubes or nanopores. However, it was found that, in the case of Ta 2 O 5 , twostep anodization did not lead to any significant structural changes in the nanopore array formed. Additionally, in the method used, the samples were sonicated after anodization to remove the overgrowth layer rich in Fand SO 4 2ions, but no detachment of the Ta 2 O 5 nanopore array was observed. This effect could be explained by the extremely compact and thin layer that was formed.

Surface contamination
To date, many literature reports about the anodic method of fabricating Ta 2 O 5 nanopore arrays have noted that such a layer can be formed over a wide potential range, typically at 15 V [32,39]. Some reports do not describe the anodization effect at higher or lower potentials, while others only report the limits of potential within which efficient Ta 2 O 5 nanopores can be formed. In this study, the impact of the applied potential on the formation of the nanopore array was carefully studied. It was found that tantalum foil can be successfully anodized into a nanopore array over a broad potential range of from 15 up to 50 V. However, SEM analysis showed that, even under the same anodization conditions, some samples contains an organized Ta 2 O 5 nanopore array, while others exhibited different structures. An in-depth analysis showed that contaminated layer cover the well-organized nanopores array. such other structures is organized as nanopores array, which can be seen in Fig. 2. The SEM analysis showed that the duration of the anodization process does not have a significant impact on the size of Ta 2 O 5 nanopores. The diameter of the pores can be tuned by tuning the applied potential. It was found that the average diameter of the nanotubes increased with increasing anodization voltage. The average size of the pores formed at 15 V was 39 ± 4 nm, while at 30 V it increased to 66 ± 6 nm. This effect is in line with a previous literature report showing that tube spacing is directly proportional to the applied voltage [40]. It is worth noting that neither the time nor the applied potential affected the thickness of the nanopore array formed. The lower inset of Fig. 1 shows an SEM-FIB cross section of a Ta 2 O 5 nanopore array. The average thickness of the layer formed was estimated as 30-50 nm, depending on the voltage. To understand the formation process of both layers, EDS measurements were made on both types of surface. The analysis showed that, in case of the organized Ta 2 O 5 nanopore array, the Ta concentration was about 57%, while the oxygen concentration was 8.33%. This formula is close to that of Ta 2 O 5 structures. Importantly, the only other element present in the samples was carbon (5%). The situation was completely different in the samples having a different morphology, where the tantalum was only 26.8%, the oxygen concentration was 6%, and the sulfur and fluoride signals could be found in high concentrations. The sulfur concentration even reached a value close to 6.1%, and fluoride as high as 9%. The results showed that, depending on some unknown factor, in some processes a pure Ta 2 O 5 nanopore array was formed, while in other part of the anodization (even under the same conditions) different structures rich in fluoride and sulfur formed. Also, the XPS analysis confirmed a high level of contamination of the surfaces by S and F. Figure 3a and b show XPS Survey spectra of non-contaminated and contaminated surfaces. In the case of the polluted sample, the S concentrations is estimated as 11%, and that of F as about 6%, while for the treated samples the S concentration is below 2%, and the F concentration is below the detection limit. The deconvolution procedure for the Ta 4f high-resolution spectra of the Ta 2 O 5 nanopore array gives two well-resolved spin-orbit components: Ta 4f 7 and Ta 4f 5 , respectively, with a 1:1.3 ratio of the band areas, full width at half maximum (FWHM) equal to 2.5 ± 0.1 eV, and a band separation of 2.1 eV. Also, the Ta4d and 4p3 could be found as doublets at 233 and 243 eV, and singlet at 403 eV, respectively. The maximum at 532 eV ± 0.1 eV related to the O 1 s spectrum assigned to the oxygen lattice confirmed the presence of Ta 2 O 5 .
The presence of fluoride and sulfur in the layer could point to some surface contamination coming from the electrolyte solution composed of concentrated H 2 SO 4 and HF acid. Another possible explanation is to assume that such a layer is a highly compacted and deformed layer of Ta 2 O 5 nanotubes highly contaminated by Fand SO 4 2ions. The presence of SO 4 2-ions was assigned by the XPS peak located at 168-170 eV, while the typical value for S 2is lower, in a range of 158-162 eV. Due to the high growth rate, the nanotubes stick together and deform during the anodization process. Some attempts were made to remove the contaminated layer from the surface. It was found that sonication in water led to a remove a surface contamination layer and exposition of the nanopore Ta 2 O 5 surface.. The same result can be achieved by tearing off the layer using adhesive tape. To confirm our suspicion, we examine the adhesive tape under SEM; it was found that structures rich in sulfide and fluoride were present. Figure 2b and c shows SEM micrographs of the removed contaminated layer on the adhesive tape. As can be seen, the contaminated layer consists of long (longer than 1 lm) wires, which suggests that the layer comes from deformed, long Ta 2 O 5 nanotubes.

Structural and optical characterization of Ta 2 O 5 nanopore array surface after Ag layer deposition
It is obvious that Ta 2 O 5 nanopore array structures will change when an Ag layer is deposited. Figure 4 shows SEM images of the samples coated with an Ag layer (under a constant current of 20 mA) for (A) 15, (B) 30, (C) 45 and (D) 60 s. As can be seen, the deposition of a layer of silver for 15 s made almost no change to the structure of the Ta 2 O 5 nanopore array. A longer deposition time led to a decrease in the size of the nanopores, and the pores become more plumpy. A longer deposition time also led to a flattening of the Ta 2 O 5 nanopore array structure, meaning that the deposited Ag layer completely covered it. As can be seen, the PVD technique applied led to a surface deposition of Ag on the nanopore array, where most of the deposited material was located on the surface and did not penetrate the interior of the nanopores. The optical properties of plasmonic nanomaterials can be modeled by various quantum calculations [29], but here diffuse reflectance spectra (DRS) were measured for the Ta 2 O 5 nanopore array coated by an Ag layer in a range of 800-350 nm. Those DRS spectra are presented in Fig. 5. It was found that a broad peak with a maximum located at 610 nm dominated in all the recorded spectra. The peak intensity reached a maximum for the sample coated by an Ag layer for 30 s, whereas the absorbance of the samples coated with a thicker layer was lower, and they were quite similar to each other. This effect is related to the smoothing and flattening of the Ta 2 O 5 nanopore array after the deposition of such a thick layer, which is supported by the SEM image. The lowest absorbance was exhibited by the samples coated with an Ag layer deposited for 15 s. Such a low absorbance can explain the low SERS activity of this sample. The number of free electrons from the SPR was very low, much lower than for any of the other samples, and so the SERS activity was also lower.

SERS measurements
The Ag layer was deposited on the Ta 2 O 5 nanopore arrays using the sputtering technique. The thickness of the deposited Ag layer was controlled by changing the deposition time, and all of the samples were covered using the same current (20 mA). It was found that, in order to use a Ta 2 O 5 nanopore array as a SERS platform after Ag nanolayer deposition, no annealing process was needed. Therefore, the samples were not annealed and Ag deposition was done right after the samples were dried after sonication. The samples covered with a layer of silver were then immersed for 24 h in a saturated aqueous solution of p-MBA to chemisorb this Raman reporter. After chemisorption, the samples were rinsed with water to remove excess p-MBA from the surface, and then dried at room temperature for 1 h.

Determination of the optimal thickness of the deposited Ag layer
The typical SERS spectrum of p-MBA is dominated by two strong bands, at 1077 cm -1 and 1587 cm -1 , which are due to the v 12 and v 8a vibrations of the aromatic ring, respectively [41,42]. The weaker band in the SERS spectrum of p-MBA, at 1185 cm -1 , is due to the d(C-H) vibration [43]. In some cases, and additional band at 1400 cm -1 appears, indicating the ionization of the carboxylic groups of p-MBA [42]. For each sample, 50 SERS spectra at various points were measured, and then the averaged spectra were plotted. The highest intensity of the p-MBA band located at 1580 cm -1 was found for the samples coated with silver for 30 s. For the sample coated with a thinner layer (15 s), the SERS signal was significantly lower. For longer evaporation times, it was found that the signal decreased. The SERS spectra recorded, and the dependence of the obtained SERS intensity on the thickness of the deposited silver layer, are presented in Fig. 6. It is probably the case that, when the deposited silver layer is above a certain threshold, the initial surface roughness (induced by the Ta 2 O 5 nanopores) is lost due to the flattening effect of the deposited metal layer. We observed the same effect for ZrO 2 NTs, where the maximum EF was obtained for an Au film thickness of about 20 nm [33]. The same behavior was found in case of silver decorated ZnO hexagonal nanoplates [34]. A short sputtering time led to the formation of a rough silver nanofilm, which may be beneficial for SERS enhancement, while a longer deposition time led to a relatively smooth Ag film. Therefore, all further experiments were carried out using Ta 2 O 5 nanopore array substrates covered with an Ag layer deposited for 30 s.

Time stability, point-to-point and sample-to-sample analysis
Of course, EF is a very important parameter of the SERS platform, but a good-quality SERS substrate should provide stable EF over the entire substrate, from substrate to substrate, and should exhibit high chemical and temporal stability. For this purpose, we investigated the temporal stability of the formed SERS platform, and carried out point-to-point and sample-to-sample analyses. All these analyses were performed on the most SERS-active platform-the Ta 2 O 5 nanopore array coated with silver for 30 s.  For the point-to-point analysis, SERS spectra of the chemisorbed p-MBA were collected from 50 different spots. It was found that the RSD was 8.66%. For the sample-to-sample analysis, SERS spectra from five samples prepared under the same conditions were collected. Here, the RSD (relative standard deviation) was 4.72%. Our results showed that the sample surface exhibited good homogeneity over the whole sample surface, and that the samples preparation method used is highly reproducible and repeatable.
The point-to-point and sample-to-sample analyses are presented in Fig. 7.
For the sample that showed the greatest enhancement of the SERS spectrum of the p-MBA molecule, the measurement was repeated after 30 days in order to check the substrate's temporal stability. The signal declined by only 22%, still allowing a very clear spectrum of the measured molecule to be obtained. The percentage was calculated relative to the intensity obtained the band at a frequency of 1587 cm -1 .  b sample-to-sample analysis of SERS spectra recorded from 5 different samples formed under the same conditions.

Conclusions
In summary, an efficient anodization method for synthesizing Ta 2 O 5 nanopore arrays under a wide range of potential in a highly acidic solution containing Fions was developed. The influence of electrolyte concentration, anodization time and applied voltage on the size and structure of the resulting Ta 2 O 5 nanopore arrays was determined. It was demonstrated that Ta 2 O 5 having a well-defined nanopore structures can be formed in a fast, repeatable process over a wide potential range. Surface contamination, rich in Fand SO 4 2ions, can be easily removed using adhesive tape or by sonication. The nanostructures formed were then covered with various Ag layers using the sputtering deposition method. Optical and structural characterizations were made to determine the optimal deposited Ag film thickness. The SERS samples formed exhibited high SERS activity, as well as high point-to-point, sample-to-sample and temporal stability. We believe that, with their plasmonic properties and good temporal stability, such Ta 2 O 5 -Ag nanopore arrays could be widely used-not only as SERS platforms, but also in other plasmonic fields such as in photocatalysis.