Flexible and Reusable Ag Coated TiO2 Nanotube Arrays for Highly Sensitive SERS Detection of Formaldehyde

Quantitative analysis of formaldehyde (HCHO, FA), especially at low levels, in various environmental media is of great importance for assessing related environmental and human health risks. A highly efficient and convenient FA detection method based on surface-enhanced Raman spectroscopy (SERS) technology has been developed. This SERS-based method employs a reusable and soft silver-coated TiO2 nanotube array (TNA) material, such as an SERS substrate, which can be used as both a sensing platform and a degradation platform. The Ag-coated TNA exhibits superior detection sensitivity with high reproducibility and stability compared with other SERS substrates. The detection of FA is achieved using the well-known redox reaction of FA with 4-amino-3-hydrazino-5-mercapto-1,2,4-triazole (AHMT) at room temperature. The limit of detection (LOD) for FA is 1.21 × 10−7 M. In addition, the stable catalytic performance of the array allows the degradation and cleaning of the AHMT-FA products adsorbed on the array surface under ultraviolet irradiation, making this material recyclable. This SERS platform displays a real-time monitoring platform that combines the detection and degradation of FA.


Introduction
Formaldehyde (FA), a colorless molecule that is a strong irritant, is a major hazard to human health and has been identified as a carcinogen [1][2][3]. The commercial product is an aqueous solution, and a 35-40% aqueous solution is commonly called formalin [4,5]. At present, many FA detection methods have been studied, including high-performance liquid chromatography (HPLC) [6], gas chromatography-mass spectrometry (GC-MS) [7], and fluorescence analysis [8]. However, these methods have certain disadvantages, such as a long detection time, complicated preprocessing steps, and incomplete detection ability. Therefore, it is important to establish a simple, fast, green, sensitive, and selective trace FA detection method.
Raman spectroscopy is an important analytical technique that provides molecular information by obtaining structural fingerprints of molecular vibrational levels [9,10]. However, Raman spectroscopy has the disadvantages of very weak signals and low reproducibility due to the inelastic scattering of radiation. Surface-enhanced Raman scattering (SERS), a fast, sensitive, and nondestructive spectroscopy tool for identifying and detecting chemical/biological species, has received considerable attention because it significantly amplifies the effect of adsorbed molecules on the target Raman signal [11][12][13][14][15][16][17][18]. Figure 1A shows scanning electron microscopy (SEM) images of the TNA material prepared by surface pretreatment of pure titanium foil through two anodizing methods in a 0.5 wt% NH 4 F ethylene glycol system at a controlled voltage of 60 V. From the top view, it can be seen that the surface of the TNA consists of a nanotube array, and the diameter of the tubes is 100-200 nm. From the side view, it can be seen that the TNA is open at the top and closed at the bottom. At the bottom of the film, there is a thin and dense barrier layer. The barrier layer is uniformly distributed, perpendicular to the substrate nanotube array, and the tube length is 4-7 µm.   Figure S1. Most Ag nanoparticles were distributed on the TNA surface. During the preparation of SEM samples, Ag may aggregate during the drying step. The SERS test can be completed in 1 min; during this time, the Ag nanoparticles are moving, and their distribution is more uniform than that shown in the figure. Figure 2 shows the preliminary preparation scheme in our experiment. The figure illustrates the preparation of the TNA on titanium foil by the anodization method. Silver nanoparticles were applied to the TNA by coating a silver sol to ensure the formation of an SERS substrate with the attached precious metal. Studies have reported that Ag-TiO 2 composites can be used to detect and degrade organic pollutants. We optimized an Ag-TiO 2 material to explore whether the detection of FA was satisfactory.   Figure S1. Most Ag nanoparticles were distributed on the TNA surface. During the preparation of SEM samples, Ag may aggregate during the drying step. The SERS test can be completed in 1 min; during this time, the Ag nanoparticles are moving, and their distribution is more uniform than that shown in the figure. Figure 2 shows the preliminary preparation scheme in our experiment. The figure illustrates the preparation of the TNA on titanium foil by the anodization method. Silver nanoparticles were applied to the TNA by coating a silver sol to ensure the formation of an SERS substrate with the attached precious metal. Studies have reported that Ag-TiO2 composites can be used to detect and degrade organic pollutants. We optimized an Ag-TiO2 material to explore whether the detection of FA was satisfactory. In this work, the SERS method is based on the reaction of excess AHMT with FA ( Figure 3A). Under basic conditions, FA and AHMT undergo a condensation reaction to form 6-mercapto-5triazolo [4,3-b]-s-tetrazine (MTT). It can be seen from Figure 3B that with a decrease in FA concentration, the characteristic band intensity in the ultraviolet-visible (UV-vis) absorption spectrum has a tendency to first increase and then decrease. An FA concentration of 10 −4 M cannot be detected. Therefore, the sensitivity of UV-vis spectroscopy to FA is not ideal. Thus, we prepared a recyclable SERS substrate with photocatalytic performance to detect FA. We also investigated the reproducibility, stability, and recycling characteristics of the SERS substrate for use as a catalyst. To examine these properties, we performed the following experiments. In this work, the SERS method is based on the reaction of excess AHMT with FA ( Figure 3A). Under basic conditions, FA and AHMT undergo a condensation reaction to form 6-mercapto-5-triazolo [4,3-b]-s-tetrazine (MTT). It can be seen from Figure 3B that with a decrease in FA concentration, the characteristic band intensity in the ultraviolet-visible (UV-vis) absorption spectrum has a tendency to first increase and then decrease. An FA concentration of 10 −4 M cannot be detected. Therefore, the sensitivity of UV-vis spectroscopy to FA is not ideal. Thus, we prepared a recyclable SERS substrate with photocatalytic performance to detect FA. We also investigated the reproducibility, stability, and recycling characteristics of the SERS substrate for use as a catalyst. To examine these properties, we performed the following experiments.   Figure 4A shows the SERS spectra of a mixture of AHMT and AHMT-MTT in the 600-1800 cm −1 region using the Ag-sol substrate. Figure 4B shows the scheme of the SERS sample deposition and measurement. As shown in Table 1, the bands at 710 and 832 cm −1 are due to the S-C-N tensile vibration and N-C-N tensile vibration of the AHMT and MTT molecules, respectively. The six rings of MTT make its N-C-N tensile vibration stronger than that of AHMT. The SERS band at 1473 cm −1 is significantly enhanced, which is attributed to the tensile vibration mode of C-C rings. The other main bands of AHMT are observed at 1217, 1286, and 1391 cm −1 , which are attributed to in-ring breathing vibrations and in-plane deformation. Compared to AHMT, several bands of MTT are blueshifted due to the six rings.   Figure 4A shows the SERS spectra of a mixture of AHMT and AHMT-MTT in the 600-1800 cm −1 region using the Ag-sol substrate. Figure 4B shows the scheme of the SERS sample deposition and measurement. As shown in Table 1, the bands at 710 and 832 cm −1 are due to the S-C-N tensile vibration and N-C-N tensile vibration of the AHMT and MTT molecules, respectively. The six rings of MTT make its N-C-N tensile vibration stronger than that of AHMT. The SERS band at 1473 cm −1 is significantly enhanced, which is attributed to the tensile vibration mode of C-C rings. The other main bands of AHMT are observed at 1217, 1286, and 1391 cm −1 , which are attributed to in-ring breathing vibrations and in-plane deformation. Compared to AHMT, several bands of MTT are blueshifted due to the six rings.  Figure 4A shows the SERS spectra of a mixture of AHMT and AHMT-MTT in the 600-1800 cm −1 region using the Ag-sol substrate. Figure 4B shows the scheme of the SERS sample deposition and measurement. As shown in Table 1, the bands at 710 and 832 cm −1 are due to the S-C-N tensile vibration and N-C-N tensile vibration of the AHMT and MTT molecules, respectively. The six rings of MTT make its N-C-N tensile vibration stronger than that of AHMT. The SERS band at 1473 cm −1 is significantly enhanced, which is attributed to the tensile vibration mode of C-C rings. The other main bands of AHMT are observed at 1217, 1286, and 1391 cm −1 , which are attributed to in-ring breathing vibrations and in-plane deformation. Compared to AHMT, several bands of MTT are blueshifted due to the six rings.     Figure 4C shows the SERS spectra of AHMT-FA with various concentrations of FA (from 1.44 × 10 −2 to 1.44 × 10 −9 M). With decreasing FA concentration, the intensity of the Raman band at 1473 cm −1 greatly decreased, and the band was clearly observed down to an FA solution concentration of 1.44 × 10 −9 M, relative to the band of the probe solution (AHMT). Therefore, this method is reasonable and suitable for the detection of FA with a detection limit of 1.44 × 10 −9 M. Figure 5 shows the Raman intensity of the characteristic band at 1286 cm −1 at different concentrations. The test was repeated six times for each concentration. The results also show the average and standard deviation for each concentration during the test. The accuracy of the SERS method for detecting FA is shown. It can be seen from Figure 5 that as the MTT concentration decreases, the average value and standard deviation also decrease. The results show that the SERS method is sensitive to the detection of FA and further prove the feasibility of this method.

Results and Discussion
1473 cm −1 Breathing vibration 1217, 1286, 1391 cm −1 Figure 4C shows the SERS spectra of AHMT-FA with various concentrations of FA (from 1.44 × 10 −2 to 1.44 × 10 −9 M). With decreasing FA concentration, the intensity of the Raman band at 1473 cm −1 greatly decreased, and the band was clearly observed down to an FA solution concentration of 1.44 × 10 −9 M, relative to the band of the probe solution (AHMT). Therefore, this method is reasonable and suitable for the detection of FA with a detection limit of 1.44 × 10 −9 M. Figure 5 shows the Raman intensity of the characteristic band at 1286 cm −1 at different concentrations. The test was repeated six times for each concentration. The results also show the average and standard deviation for each concentration during the test. The accuracy of the SERS method for detecting FA is shown. It can be seen from Figure 5 that as the MTT concentration decreases, the average value and standard deviation also decrease. The results show that the SERS method is sensitive to the detection of FA and further prove the feasibility of this method. .2701 X, which shows a linear relationship and has a correlation coefficient of 0.9639. The limit of detection (LOD) was determined according to the IUPAC recommendations (the minimally acceptable signal intensity must be three times greater than the standard deviation of the blank signal). As depicted in Figure 5, this intensity approximately corresponds to 10 −6.92 M (i.e., 1.21 × 10 −7 M). Finally, in terms of reproducibility and repeatability, SERS measurements were obtained on 20 spots, randomly selected on the Ag-TNA substrates. (see Figure  S2 in the Supplementary Material).
The degradation process during the experiments was monitored by the changes in the band intensity of AHMT-FA in the SERS spectra. Taking 1.44 × 10 −4 M as an example, as shown in Figure  6A, lines a-h correspond to Ag-TNA substrates used to adsorb AHMT-FA and irradiated under ultraviolet light at 254 nm for 40, 50, 60, 70, 100, 130, 160, and 190 min, respectively. Line i is the Raman spectrum of Ag-TNA without adsorbed probe molecules. The comparison shows that the probe molecules, AHMT and FA, were completely degraded after approximately 3 h of irradiation. This result shows that the Ag-TNA substrate has a self-cleaning function under ultraviolet The calculated limit of detection (LOD) is given by the black line. The standard curve of the AHMA-FA solution obeys the equation Y = 5.1399 + 0.2701 X, which shows a linear relationship and has a correlation coefficient of 0.9639. The limit of detection (LOD) was determined according to the IUPAC recommendations (the minimally acceptable signal intensity must be three times greater than the standard deviation of the blank signal). As depicted in Figure 5, this intensity approximately corresponds to 10 −6.92 M (i.e., 1.21 × 10 −7 M). Finally, in terms of reproducibility and repeatability, SERS measurements were obtained on 20 spots, randomly selected on the Ag-TNA substrates. (see Figure S2 in the Supplementary Materials).
The degradation process during the experiments was monitored by the changes in the band intensity of AHMT-FA in the SERS spectra. Taking 1.44 × 10 −4 M as an example, as shown in Figure 6A, lines a-h correspond to Ag-TNA substrates used to adsorb AHMT-FA and irradiated under ultraviolet light at 254 nm for 40, 50, 60, 70, 100, 130, 160, and 190 min, respectively. Line i is the Raman spectrum of Ag-TNA without adsorbed probe molecules. The comparison shows that the probe molecules, AHMT and FA, were completely degraded after approximately 3 h of irradiation. This result shows that the Ag-TNA substrate has a self-cleaning function under ultraviolet irradiation, and this function can overcome the disadvantage of being able to use SERS substrates only once.
Molecules 2020, 25, x FOR PEER REVIEW 6 of 11 irradiation, and this function can overcome the disadvantage of being able to use SERS substrates only once. As mentioned in the introduction, organic molecules on the surface of TiO2 can be degraded by the generation of free radicals and oxidizing substances on the surface of TiO2 under ultraviolet irradiation. In visible light, the localized surface plasmon resonance (LSPR) of AgNPs is excited and decays to generate hot electrons. The energy of these electrons is higher than the potential barrier between TiO2 and AgNPs, and the electrons can jump to the conduction band of TiO2 and generate holes in AgNPs. The interface state density of the TNA and Ag is relatively large, and most of the hot electrons are captured by the interface state, thus greatly reducing the total number of electrons that can reach the conduction band of TiO2. The remaining electrons that reach the conduction band of TiO2 migrate to the surface of TiO2, combine with molecular oxygen adsorbed on the surface to form · O 2− , or combine with ·OH, resulting from the decomposition of H2O in the oxygen vacancy to form ·OH. Note that ·O 2− and ·OH have strong catalytic activity and are the main active species for catalyzing the degradation of organic matter. Due to the characteristics of TiO2 nanotubes, they have a higher electron transport efficiency than other compounds, which is conducive to the transfer of hot electrons from AgNPs to the conduction band of TiO2, thus increasing the photocatalytic efficiency. Figure 6C shows the intensity of the characteristic band at 1286 cm −1 as a function of degradation time. The results show that as the degradation time increases, the intensity of the Raman band significantly decreases, and the overall Raman signal intensity decreases. Furthermore, after detecting FA, the reversible SERS active substrate could degrade the FA and AHMT adsorbed on the substrate surface into small inorganic molecules and water molecules by photocatalytic degradation ( Figure 6B). This result further proves the self-cleaning function.
To evaluate the degradation performance and recyclability of this Ag-TNA substrate, we performed three detection-degradation cycle experiments on the Ag-TNA substrate. We used 1.44 × 10 −4 M FA as an example (Figure 7). In the first cycle, line a represents the Raman spectrum of AHMT-  As mentioned in the introduction, organic molecules on the surface of TiO 2 can be degraded by the generation of free radicals and oxidizing substances on the surface of TiO 2 under ultraviolet irradiation. In visible light, the localized surface plasmon resonance (LSPR) of AgNPs is excited and decays to generate hot electrons. The energy of these electrons is higher than the potential barrier between TiO 2 and AgNPs, and the electrons can jump to the conduction band of TiO 2 and generate holes in AgNPs. The interface state density of the TNA and Ag is relatively large, and most of the hot electrons are captured by the interface state, thus greatly reducing the total number of electrons that can reach the conduction band of TiO 2 . The remaining electrons that reach the conduction band of TiO 2 migrate to the surface of TiO 2 , combine with molecular oxygen adsorbed on the surface to form · O 2− , or combine with ·OH, resulting from the decomposition of H 2 O in the oxygen vacancy to form ·OH. Note that ·O 2− and ·OH have strong catalytic activity and are the main active species for catalyzing the degradation of organic matter. Due to the characteristics of TiO 2 nanotubes, they have a higher electron transport efficiency than other compounds, which is conducive to the transfer of hot electrons from AgNPs to the conduction band of TiO 2 , thus increasing the photocatalytic efficiency. Figure 6C shows the intensity of the characteristic band at 1286 cm −1 as a function of degradation time. The results show that as the degradation time increases, the intensity of the Raman band significantly decreases, and the overall Raman signal intensity decreases. Furthermore, after detecting FA, the reversible SERS active substrate could degrade the FA and AHMT adsorbed on the substrate surface into small inorganic molecules and water molecules by photocatalytic degradation ( Figure 6B). This result further proves the self-cleaning function.
To evaluate the degradation performance and recyclability of this Ag-TNA substrate, we performed three detection-degradation cycle experiments on the Ag-TNA substrate. We used 1.44 × 10 −4 M FA as an example (Figure 7). In the first cycle, line a represents the Raman spectrum of AHMT-FA reacting for 16 h (first reaction), and line b represents the Raman spectrum after 3 h of 254 nm irradiation (first ultraviolet irradiation). In the second cycle, line c indicates the Raman spectrum of AHMT-FA reacting for 16 h using the substrate in line b (second reaction), and line d indicates the Raman spectrum after 3 h of 254 nm irradiation (second ultraviolet irradiation). In the third cycle, line e indicates the Raman spectrum of AHMT-FA reacting for 16 h (third reaction), and line f indicates the Raman spectrum after 3 h of 254 nm irradiation (third ultraviolet irradiation). The Ag-TNA SERS substrate used in this experiment can be completely reused. This figure also shows how a reversible SERS-active substrate works; after detecting FA via SERS, the substrate can photocatalytically degrade FA and AHMT adsorbed on the substrate surface, exhibiting a self-cleaning function. This result provides prospects for the on-site detection and degradation of organic pollutants by SERS. At the same time, FA is a major indoor pollutant. The existing commercial FA detection method is mainly based on electrochemical sensing technology, and its shortcomings are low specificity and an insufficient ability to discriminate volatile organic compounds (VOCs). In this work, the preparation method of Ag-TNA is simple, and it is easy to prepare in a large area. At the same time, Ag-TNA is an excellent SERS substrate for FA detection with good SERS activity and photocatalytic degradation performance. In the future, we will try to detect gaseous formaldehyde based on this work. This method has great potential to complement the selectivity of existing formaldehyde gas detection methods.

Chemicals
AHMT was purchased from Sigma-Aldrich Co., Ltd. (St. Louis, MO, USA) and was used without further purification. Titanium foil and silver nitrate (AgNO 3 ) were also purchased from Sigma-Aldrich Co., Ltd. (St. Louis, MO, USA). FA (40%) was purchased from Liaoning Quanrui Reagent Co., Ltd. Hydrofluoric acid, ammonium fluoride, and ethylene glycol (analytical reagents) were purchased from Aladdin Company and were used without further purification. The water used in the experiment was ultrapure water, and the ethanol used was anhydrous ethanol.

Instruments
SEM characterization was performed by a JEOL 7610 p thermal field emission scanning electron microscope. UV-vis absorption spectra were recorded on a Cary 5000 ultraviolet-visible-near infrared (UV-VIS-NIR) spectrometer (Agilent Technologies, Inc., Santa Clara, CA, USA). XRD testing was conducted by a Rigaku Smartlab X-ray diffractometer using CuKα radiation (λ = 1.5418 Å) at 45 kV and a 200 mA excitation light source. Raman spectroscopy was performed using a Jobin Yvon/Horiba LabRAM HR Evolution confocal micro-Raman spectrometer equipped with a multichannel air-cooled charge-coupled device (CCD) detector. A 532 nm laser was used as the excitation light source. A monocrystalline silicon wafer was used to calibrate the Raman spectrometer. All Raman spectra were measured with a 600 g/mm grating using a 50×, 0.50 NA (OLYMPUS LMPlanFLN, Tokyo, Japan) long-working-distance (LWD) microscope objective. The laser power reaching the sample surface was approximately 5 mW. The acquisition time of Raman spectra was 30 s for each window. For ultraviolet light degradation, a ZF-7 portable ultraviolet analyzer (including 254 nm and 365 nm ultraviolet radiation) produced by Shanghai Daluo Scientific Instrument Co., Ltd. (Shanghai, China) was used. The UV degradation experiment was performed under irradiation with a 254 nm UV lamp with a power of 8 W.

Preparation of Silver Sol
AgNO 3 (169 mg) was added to a three-necked flask, 100 mL of deionized water and a stir bar were added, and a thermometer and a condenser tube were connected. Sodium citrate (0.01 g) was added to 1 mL of deionized water. Good silver sol was obtained by heating the AgNO 3 and deionized water until slightly boiling, dropwise adding sodium citrate (discoloration within 3 min), and heating for 30 min.

Preparation of the TNA
Titanium foil with a size of 4 × 5 cm was used as the raw material. The surface of the titanium foil was cleaned with acetone, isopropanol, methanol, and ultrasonication. Then, the surface was washed with deionized water and blown with nitrogen before use. Graphite flakes (also ultrasonically cleaned with distilled water) were used as the cathode, and titanium foil was used as the anode. Anode oxidation was performed at room temperature, and 0.5 wt% NH 4 F glycol was used as the electrolyte. After the electrolytic titanium dioxide was washed with distilled water, it was soaked in water; the titanium dioxide film formed on the surface, was peeled off in the water and dried under strong nitrogen flow, and a second round of electrolysis was performed for 2 h. The electrolyzed titanium foil was washed with absolute ethanol and ultrapure water, soaked in absolute ethanol overnight, removed under strong nitrogen flow, blown dry, and calcined in a muffle furnace at 450 • C for 2 h. The TNA was stored in a box protected from light.

Preparation of the Test Solution
At room temperature, 0.0124 g AHMT was added into a 50 mL centrifuge tube. Then

SERS Detection of FA
The calcined TNA was placed into the test liquid for 16 h. Then, the TNA was removed, 150 µL silver sol was added, and the confocal micro-Raman spectrometer with a 532 nm excitation laser was used to examine the test liquids with different concentrations of FA. The Raman signal of the probe (AHMT concentration of 3.34 × 10 −3 M) was detected by the same method.

UV Degradability of AHMT-FA with the Ag-TNA
A test liquid with a concentration of 1.44 × 10 −4 M was selected to evaluate the ultraviolet photodegradability of the SERS substrate. First, the TNA was immersed in the 1.44 × 10 −4 M test solution for 16 h, and then the TNA was removed. Silver sol was added to measure its initial Raman signal; then, it was directly put into a portable ultraviolet analyzer and irradiated with 254 nm ultraviolet light for 3 h. The Raman signal was detected every 10 or 30 min until the degradation was complete.

Recyclable SERS Substrate
(1) As the degradation time increased, the Raman intensity continued to decrease until the Raman signal basically disappeared; this time corresponded to the time required to obtain complete degradation of AHMT-FA. (2) The same TNA was put into the test solution again for 16 h. Then, the silver sol was removed, and the Raman signal of AHMT-FA was detected. (3) The solution was placed under ultraviolet light with a wavelength of 254 nm and irradiated to complete degradation. (4) The experiment was repeated one more time at the end for a total of three replicates to check the substrate recyclability.

Conclusions
In this study, we show a new method for FA detection based on SERS. The combination of a surface-cleaned TNA array material prepared by electrochemical anodization and Ag-sol constitutes a dual-function Ag-TNA FA-sensing material that has a SERS effect and photocatalytic degradation. The results show that the SERS intensity of the FA reaction product has a strong FA concentration dependence, and that the lowest detection concentration of FA in solution can reach 10 −9 M. Furthermore, it has been proven that the composite array is a highly sensitive, stable, self-cleaning and recyclable SERS-active substrate. In addition, the excellent photocatalytic degradation performance and SERS activity of the substrate show great potential for the on-site detection and degradation of organic pollutants by the SERS method.