Color Changes in Ag Nanoparticle Aggregates Placed in Various Environments: Their Application to Air Monitoring

Fresh Ag nanoparticles (NPs) dispersed on a transparent SiO2 exhibit an intense optical extinction band originating in localized surface plasmon resonance (LSPR) in the visible range. The intensity of the LSPR band weakened when the Ag NPs was stored in ambient air for two weeks. The rate of the weakening and the LSPR wavelength shift, corresponding to visual chromatic changes, strongly depended on the environment in which Ag NPs were set. The origin of a chromatic change was discussed along with both compositional and morphological changes. In one case, bluish coloring followed by a prompt discoloring was observed for Ag NPs placed near the ventilation fan in our laboratory, resulted from adsorption of large amounts of S and Cl on Ag NP surfaces as well as particle coarsening. Such color changes deduce the presence of significant amounts of S and Cl in the environment. In another case, a remarkable blue-shift of the LSPR band was observed for the Ag NPs stored in the desiccator made of stainless steel, originated in the formation of CN and/or HCN compounds and surface roughening. Their color changed from maroon to reddish, suggesting that such molecules were present inside the desiccator.


Introduction
Environmental pollution has become an important problem all over the world. For instance, nitrogen oxide (NO x ), which plays a major role in the formation of ozone and acid rain, is one of the most dangerous air pollutants. Continued exposures to NO 2 cause increased incidence of acute respiratory infection [1,2]. In addition, sulphur compounds such as sulphur oxides (SO x ) and hydrogen sulphide (H 2 S) are also well known to air pollutants. Even short-term exposures to SO 2 are linked with respiratory effects including breathing difficulty and asthma symptoms [1,2]. The H 2 S, produced by many industrial processes and decomposition of oil, is a very poisonous, corrosive, flammable and explosive gas [3].
Focusing on industrial materials, e.g., electronics and semiconductor material, copper and silver are extensively used because of their high electrical conductivity, ductility and malleability. They are, however, inevitably corroded by reacting with H 2 S in ambient air to produce their sulfide such as Cu 2 S and Ag 2 S [4][5][6]. The atmospheric corrosion caused by the pollutant gases is becoming a significant factor in the reliability of electrical equipment. As electronics continue to decrease in size, it is important to be aware of the pollutant gases around the electrical equipment to prevent corrosion risk.
In recent years, particulate matter (PM) has also attracted considerable attention due to adverse effects to human health and materials corrosion. It is known to that the PM can be carried deep into the lungs, worsening lung diseases [1]. Moreover, adsorption of PM containing sulphur compounds on copper and its alloys corrodes locally the metals, Teijin, Tokyo, Japan). These samples are referred to as Ag NPs/SiO2, Ag NPs/Si, NPs/HOPG and Ag NPs/PEN, hereafter. A desktop rotary-pumped sputter coater, SC MKII, developed by Sanyu Electron Co. Ltd., Tokyo, Japan was used with a silver s tering target (purity: 99.8%). The pressure of the Ar gas was ~1 Pa, and the current typically 7 mA at a voltage of 1.2 keV during the sputter deposition with a 30 s durat The samples became gradually darker immediately after the deposition. Therefore, samples were aged in vacuum (<7.0 × 10 −5 Pa) for 7 days prior to air exposure, in orde stabilize the Ag NPs, e.g., morphology and crystallinity. The samples were stored at different places; on a table (in ambient air), near a ventilation fan, inside a clean desicc (class 100 of cleanroom classification) made of acrylic resin, inside a conventional de cator (referred to as "metal desiccator") made of stainless steel, in our laboratory w the illumination light was almost turned off. All the Ag NP samples of Ag NPs/SiO2 NPs/Si, Ag NPs/HOPG and Ag NPs/PEN were subjected to the same conditions.
The Ag NPs/SiO2 samples were used for visual confirmation for chromatic chan as well as Ultraviolet-visible (UV-Vis) measurements. The color of the as-prepared NPs/SiO2 sample (aged in vacuum for 7 days) was maroon as shown in Figure 1a. UV extinction spectra were taken in standard transmission geometry in the wavelength ra of 200-800 nm.
. Figure 1. (a) A photograph of as-prepared Ag/SiO2; (b) and an atomic force microscopy (AFM) image of as-prepared Ag/Si sample that was aged in vacuum for 7 days.
The Ag NPs/Si samples were fabricated to observe morphology of Ag NP aggreg with atomic force microscopy (AFM). In the AFM observation, the dynamic force m corresponding to the tapping mode, of the Nanonavi Station manufactured by SII Na technology Inc., Chiba, Japan was employed with a Si cantilever, SI-DF20, with a tip dius of curvature less than 10 nm. The lateral sizes of individual Ag NPs in the aspared Ag NPs/Si sample (aged in vacuum for 7 days) were 20-30 nm as shown in Fig  1b. The root mean square (RMS) surface roughness of the as-prepared Ag NPs/Si wa timated to be 2.1 nm.
The Ag NPs/HOPG samples were used to analyze chemical compositions and bo at Ag NP surfaces using Rutherford backscattering spectrometry (RBS) and X-ray ph electron spectroscopy (XPS). RBS was conducted with 2 MeV-4 He ions produced fro single-ended Van de Graaff (VdG) accelerator of Hiroshima University. Backscattered ions were detected with a surface barrier detector placed at an angle of 165 with res to an incident beam. RBS analysis also can provide the areal density of Ag atoms in samples. A typical areal density was estimated to be 1.6 × 10 16 Ag atoms/cm 2 . XPS u Mg Kα radiation (hν = 1253.6 eV) was performed with a JEOL 9010 X-ray spectrom (JEOL Ltd., Akishima, Japan). The Ag NPs/Si samples were fabricated to observe morphology of Ag NP aggregates with atomic force microscopy (AFM). In the AFM observation, the dynamic force mode, corresponding to the tapping mode, of the Nanonavi Station manufactured by SII Nanotechnology Inc., Chiba, Japan was employed with a Si cantilever, SI-DF20, with a tip radius of curvature less than 10 nm. The lateral sizes of individual Ag NPs in the as-prepared Ag NPs/Si sample (aged in vacuum for 7 days) were 20-30 nm as shown in Figure 1b. The root mean square (RMS) surface roughness of the as-prepared Ag NPs/Si was estimated to be 2.1 nm.
The Ag NPs/HOPG samples were used to analyze chemical compositions and bonds at Ag NP surfaces using Rutherford backscattering spectrometry (RBS) and X-ray photoelectron spectroscopy (XPS). RBS was conducted with 2 MeV-4 He ions produced from a single-ended Van de Graaff (VdG) accelerator of Hiroshima University. Backscattered 4 He ions were detected with a surface barrier detector placed at an angle of 165 • with respect to an incident beam. RBS analysis also can provide the areal density of Ag atoms in the samples. A typical areal density was estimated to be 1.6 × 10 16 Ag atoms/cm 2 . XPS using Mg Kα radiation (hν = 1253.6 eV) was performed with a JEOL 9010 X-ray spectrometer (JEOL Ltd., Akishima, Japan). The Ag NPs/PEN samples were prepared to detect heavy elements adsorbed on Ag NP surfaces using particle-induced X-ray emission (PIXE). The beams of 2 MeV-protons produced from the VdG accelerator and a Si (Li) detector with an appropriate filter made of polyethylene terephthalate was used for the PIXE analysis. Figure 2 shows the photographs of Ag NPs/SiO 2 samples stored in various environments. All the Ag NPs/SiO 2 samples were aged in vacuum for 7 days prior to air exposure. The color of the as-prepared Ag NPs/SiO 2 sample was maroon as shown in Figure 1a and insets in each photograph in Figure 2. The color faded gradually for the sample stored in ambient air. The chromatic change from maroon to burgundy were clearly recognizable after 14 days. For the Ag NPs/SiO 2 stored into the clean desiccator, its color turned still light violet even after 56 days. For the Ag NPs/SiO 2 stored into the metal desiccator, the color became terracotta after 14 days and almost transparent after 56 days. The color of the Ag NPs/SiO 2 stored near the ventilation fan faded quickly even after 14 days. This sample looked almost transparent after storing for 56 days. produced from the VdG accelerator and a Si (Li) detector with an appropriate filter made of polyethylene terephthalate was used for the PIXE analysis. Figure 2 shows the photographs of Ag NPs/SiO2 samples stored in various environments. All the Ag NPs/SiO2 samples were aged in vacuum for 7 days prior to air exposure. The color of the as-prepared Ag NPs/SiO2 sample was maroon as shown in Figure 1a and insets in each photograph in Figure 2. The color faded gradually for the sample stored in ambient air. The chromatic change from maroon to burgundy were clearly recognizable after 14 days. For the Ag NPs/SiO2 stored into the clean desiccator, its color turned still light violet even after 56 days. For the Ag NPs/SiO2 stored into the metal desiccator, the color became terracotta after 14 days and almost transparent after 56 days. The color of the Ag NPs/SiO2 stored near the ventilation fan faded quickly even after 14 days. This sample looked almost transparent after storing for 56 days. The chromatic changes can be examined in a quantitative way using optical extinction spectroscopy. Figure 3 shows the optical extinction spectra in the wavelength range 200-800 nm obtained from the four kinds of Ag NPs/SiO2 samples presented in Figure 2. All the Ag NPs/SiO2 samples were aged in a vacuum for 7 days prior to air exposure. After the aging in a vacuum, the LSPR band intensity and wavelength of Ag NPs/SiO2 stored in ambient air were 505 ± 6 nm and 0.38 ± 0.02, respectively. The intensity of the LSPR band located at 400-700 nm gradually weakened with ambient air exposure. For the Ag/SiO2 stored in the clean desiccator, the rate of LSPR intensity decrease was lower than that for the sample stored in ambient air. The intensity of the LSPR band maintained at 0.16 even after 56 days. For the Ag NPs/SiO2 stored into the metal desiccator, the LSPR band shifted toward shorter wavelengths drastically. The most remarkable decrease in the LSPR band intensity was observed for the sample stored near the ventilation fan. The reproducibility of chromatic changes and optical extinction spectra of each Ag NPs/SiO2 sample was confirmed. The chromatic changes can be examined in a quantitative way using optical extinction spectroscopy. Figure 3 shows the optical extinction spectra in the wavelength range 200-800 nm obtained from the four kinds of Ag NPs/SiO 2 samples presented in Figure 2. All the Ag NPs/SiO 2 samples were aged in a vacuum for 7 days prior to air exposure. After the aging in a vacuum, the LSPR band intensity and wavelength of Ag NPs/SiO 2 stored in ambient air were 505 ± 6 nm and 0.38 ± 0.02, respectively. The intensity of the LSPR band located at 400-700 nm gradually weakened with ambient air exposure. For the Ag/SiO 2 stored in the clean desiccator, the rate of LSPR intensity decrease was lower than that for the sample stored in ambient air. The intensity of the LSPR band maintained at 0.16 even after 56 days. For the Ag NPs/SiO 2 stored into the metal desiccator, the LSPR band shifted toward shorter wavelengths drastically. The most remarkable decrease in the LSPR band intensity was observed for the sample stored near the ventilation fan. The reproducibility of chromatic changes and optical extinction spectra of each Ag NPs/SiO 2 sample was confirmed.   For the Ag NPs/HOPG stored in ambient air, into the clean desiccator the metal desiccator, N, O and S were detected. In addition to these elements, amount of Cl was detected on Ag NPs/HOPG stored in ambient air for 44 days. In spectra of Ag NPs/HOPG stored near the ventilation fan, peaks corresponding to inant elements such as N, O, S, Cl were found. In addition to assignable peaks contaminants, there are some peaks, for example, at energies of 1.64 MeV (in Fi and 1.85 MeV (in all the spectra), that cannot be determined by RBS because of mass resolution for heavy elements. Therefore, we tried to detect trace amounts o element impurities adsorbed on Ag NPs by PIXE analysis using proton beams.   Figure 4b-e show the RBS spectra of Ag NPs/HOPG after air exposure for 14 days and 44 days. For the Ag NPs/HOPG stored in ambient air, into the clean desiccator and into the metal desiccator, N, O and S were detected. In addition to these elements, a small amount of Cl was detected on Ag NPs/HOPG stored in ambient air for 44 days. In the RBS spectra of Ag NPs/HOPG stored near the ventilation fan, peaks corresponding to contaminant elements such as N, O, S, Cl were found. In addition to assignable peaks for such contaminants, there are some peaks, for example, at energies of 1.64 MeV (in Figure 4c) and 1.85 MeV (in all the spectra), that cannot be determined by RBS because of its poor mass resolution for heavy elements. Therefore, we tried to detect trace amounts of heavy element impurities adsorbed on Ag NPs by PIXE analysis using proton beams. Nanomaterials 2021, 11, x FOR PEER REVIEW 6 of 13  Figure 5 shows the PIXE spectra of the Ag/PEN samples stored in various environments. Very weak peaks at 3.61 and 3.84 keV overlapped with background signals corresponding to Sb Lα1 and Lβ1 X-rays, respectively, which appear even for a bare PEN, indicating that Sb comes from the foil. Two intense peaks of 22.2 and 24.9 keV correspond to Ag Kα1 and Ag Kβ1 X-ray, respectively. Unfortunately, the impurity elements heavier than Ag, except for Sb, were not detected by the proton-PIXE. In Figure 5c, several peaks, e.g., Cr Kα1 (5.41 keV), Fe Kα1 (6.40 keV), Ni Kα1 (7.47 keV) X-rays, originated in extrinsic impurities adsorbed on Ag NPs. These impurities are coincident with the constituent elements of the metal desiccator. Furthermore, in Figures 5c,d, the Br Kα1 X-ray was observed at an energy of 11.92 keV, indicative of adsorption of Br on Ag NPs.  Figure 5 shows the PIXE spectra of the Ag/PEN samples stored in various environments. Very weak peaks at 3.61 and 3.84 keV overlapped with background signals corresponding to Sb Lα 1 and Lβ 1 X-rays, respectively, which appear even for a bare PEN, indicating that Sb comes from the foil. Two intense peaks of 22.2 and 24.9 keV correspond to Ag Kα 1 and Ag Kβ 1 X-ray, respectively. Unfortunately, the impurity elements heavier than Ag, except for Sb, were not detected by the proton-PIXE. In Figure 5c, several peaks, e.g., Cr Kα 1 (5.41 keV), Fe Kα 1 (6.40 keV), Ni Kα 1 (7.47 keV) X-rays, originated in extrinsic impurities adsorbed on Ag NPs. These impurities are coincident with the constituent elements of the metal desiccator. Furthermore, in Figure 5c,d, the Br Kα 1 X-ray was observed at an energy of 11.92 keV, indicative of adsorption of Br on Ag NPs.  Figure 6a displays the narrow-scan XPS C 1s spectra of Ag NPs/HOPG after air posure for 14 days. The C 1s binding energy (BE) for the Ag NPs/HOPG stored into metal desiccator was higher than that for other samples, as can be seen in Figure 6a. the sample stored into the metal desiccator, an intense N 1s peak was clearly observe shown in Figure 6b. The higher C 1s BE and the intense N 1s peak indicate that CN an HCN compound forms on Ag NP surfaces [27]. Figure 6c shows the narrow-scan A spectra of Ag NPs/HOPG. As demonstrated by the previous XPS study [28], the Ag BE of 368.6 eV is assigned to Ag2S. Since the Ag 3d binding energies for a metallic Ag Ag compounds such as AgCl, Ag2S, AgO and Ag2O are 367.9 ± 0.4 eV [29][30][31], it is diff to assign each peak. In XPS analysis, the nature of Ag can be clearly seen in 4d vale band spectra as shown in Figure 6d. The valence band spectrum of the Ag NPs/HO stored into the metal desiccator became very sharp compared to that of the as-prepa Ag NPs/ HOPG. The band width, originating from spin-orbit splitting, for the as-prepa sample and Ag NPs/HOPG stored into the metal desiccator were 3.3 eV and 1.9 eV spectively. In contrast, the spectral changes for the samples placed in ambient air, in the clean desiccator and near the ventilation fan were rather small, but the band wi for these three samples were slightly narrower than that for the as-prepared sample. observed band narrowing is evidence that Ag reacts with outermost layer of contamin to partly form compounds such as AgCl, Ag2S and AgCN. The band width of the NPs/HOPG stored in ambient air, into the clean desiccator and near the ventilation were 3.1 eV, 3.2 eV and 2.8 eV, respectively, after 14 days. Further band narrowing been recognized in valence band spectra of Ag NPs samples stored for 56 days (not sho here).  Figure 6a displays the narrow-scan XPS C 1s spectra of Ag NPs/HOPG after air exposure for 14 days. The C 1s binding energy (BE) for the Ag NPs/HOPG stored into the metal desiccator was higher than that for other samples, as can be seen in Figure 6a. For the sample stored into the metal desiccator, an intense N 1s peak was clearly observed as shown in Figure 6b. The higher C 1s BE and the intense N 1s peak indicate that CN and/or HCN compound forms on Ag NP surfaces [27]. Figure 6c shows the narrow-scan Ag 3d spectra of Ag NPs/HOPG. As demonstrated by the previous XPS study [28], the Ag 3d 5/2 BE of 368.6 eV is assigned to Ag 2 S. Since the Ag 3d binding energies for a metallic Ag and Ag compounds such as AgCl, Ag 2 S, AgO and Ag 2 O are 367.9 ± 0.4 eV [29][30][31], it is difficult to assign each peak. In XPS analysis, the nature of Ag can be clearly seen in 4d valence band spectra as shown in Figure 6d. The valence band spectrum of the Ag NPs/HOPG stored into the metal desiccator became very sharp compared to that of the as-prepared Ag NPs/ HOPG. The band width, originating from spin-orbit splitting, for the as-prepared sample and Ag NPs/HOPG stored into the metal desiccator were 3.3 eV and 1.9 eV, respectively. In contrast, the spectral changes for the samples placed in ambient air, inside the clean desiccator and near the ventilation fan were rather small, but the band widths for these three samples were slightly narrower than that for the as-prepared sample. The observed band narrowing is evidence that Ag reacts with outermost layer of contaminants to partly form compounds such as AgCl, Ag 2 S and AgCN. The band width of the Ag NPs/HOPG stored in ambient air, into the clean desiccator and near the ventilation fan were 3.1 eV, 3.2 eV and 2.8 eV, respectively, after 14 days. Further band narrowing has been recognized in valence band spectra of Ag NPs samples stored for 56 days (not shown here).  Figure 7 shows the AFM images of the Ag NPs/Si sample stored in various environ ments for 14 days. Surface morphology of the Ag NPs samples after air exposure wa different in roughness from that of the as-prepared Ag NPs. The particle coarsening as result of coalescence was observed in all the samples. The RMS surface roughness of th Ag NPs/Si stored in ambient air, into the clean desiccator, into the metal desiccator an near the ventilation fan, were 3.6 nm, 3.5 nm, 6.2 nm and 4.4 nm, respectively. The mos roughened surface was produced on the sample stored inside the metal desiccator, an for the sample stored in the clean desiccator the change in roughness was found to b smallest. In addition, worm-like aggregates of ~150 nm in length and ~50 nm in widt were observed at the surface of the sample stored near the ventilation fan.  Figure 7 shows the AFM images of the Ag NPs/Si sample stored in various environments for 14 days. Surface morphology of the Ag NPs samples after air exposure was different in roughness from that of the as-prepared Ag NPs. The particle coarsening as a result of coalescence was observed in all the samples. The RMS surface roughness of the Ag NPs/Si stored in ambient air, into the clean desiccator, into the metal desiccator and near the ventilation fan, were 3.6 nm, 3.5 nm, 6.2 nm and 4.4 nm, respectively. The most roughened surface was produced on the sample stored inside the metal desiccator, and for the sample stored in the clean desiccator the change in roughness was found to be smallest. In addition, worm-like aggregates of~150 nm in length and~50 nm in width were observed at the surface of the sample stored near the ventilation fan.

Discussion
The present work has shown that the adsorbed impurities and the color depends o the environments where Ag NPs/SiO2 were stored for a couple of weeks. This sugges that environmental substances have an effect on the change in color of Ag NPs/SiO2. Co versely, the chromatic change of Ag NPs/SiO2 is indicative of environmental substance enabling one to check the quality or cleanliness of the environment. In this section, chr matic changes of Ag NPs/SiO2 are discussed for each sample in terms of composition and morphological changes, and then the quality of environment in which Ag NPs/Si were stored is also dealt with. For this purpose, chromatic changes measured by LSP characteristics including resonant wavelength and intensity are represented in Figure  The

Discussion
The present work has shown that the adsorbed impurities and the color depends on the environments where Ag NPs/SiO 2 were stored for a couple of weeks. This suggests that environmental substances have an effect on the change in color of Ag NPs/SiO 2 . Conversely, the chromatic change of Ag NPs/SiO 2 is indicative of environmental substances, enabling one to check the quality or cleanliness of the environment. In this section, chromatic changes of Ag NPs/SiO 2 are discussed for each sample in terms of compositional and morphological changes, and then the quality of environment in which Ag NPs/SiO 2 were stored is also dealt with. For this purpose, chromatic changes measured by LSPR characteristics including resonant wavelength and intensity are represented in Figure 8. The LSPR characteristics of the four samples largely change in 14 days. In particular, the normalized LSPR intensity for the sample placed near the ventilation fan decreased down to 0.4 at elapsed time of 14 days as shown in Figure 8a. For this sample, the most drastic positive shift (+55 nm) was observed at that time as shown in Figure 8b. Visual compositional changes in Ag NPs within 14 days, obtained from Table 1, are also shown in Figure 9. The largest increase in impurity elements O (0.11 to 0.27), S (0.04 to 0.14) and Cl (0 to 0.06) are found in the atomic ratio for the sample placed near the ventilation fan. Next, chromatic changes shown in Figure 8 will be discussed along with compositional ( Figure 9) and morphological (Figure 7) features for each Ag NPs/SiO 2 sample, below.        Atomic ratios of M (*) to Ag Firstly, chromatic changes in the sample placed near the ventilation fan are discussed in comparison with those in the samples placed in ambient air and in the clean desiccator. The normalized LSPR band intensity of this sample decreased drastically compared to other samples as shown in Figure 8a. In addition, the LSPR band shifts quickly toward a longer wavelength as can be seen in Figure 8b. In this sample, the large worm-like aggregates due to particle coarsening were observed, and relatively large amounts of S and Cl atoms were detected. In the previous study [32], the Cl adsorption on clean Ag surfaces led to drastic morphological changes due to the formation of silver chloride (AgCl) islands. Similar situations would occur on the Ag NPs stored near the ventilation fan. In our case, the formation of AgCl caused the particle coarsening. Large amount of extrinsic impurities S and Cl as well as particle coarsening resulted in drastic changes in the LSPR intensity and wavelength. Thus, in an atmospheric environment containing S and Cl, the color of the Ag NPs turned bluish and then faded quickly as shown in Figure 2. Conversely, bluish coloring followed by a quick discoloring suggests significant amount of S and Cl in the environment. Of these elements, Cl would be mainly carried by dust such as particle matter (PM) because Cl was not detected in the samples in the clean and metal desiccators. In fact, ammonium salt, NH 4 Cl, is a common component of PM, and Qu et al., [33] investigated that deposition of NH 4 Cl leads to the atmospheric corrosion of zinc. The atmospheric environment of this sample is essentially the same with ambient air. It is probable that a one-way flow of air through the ventilation fan stimulates the adsorption of such extrinsic impurity elements on Ag NPs, resulting in the fastest discoloring. Thus, relatively quick monitoring of environment qualities can be achieved by one-way flow of the air using a ventilation fan.
Secondly, chromatic changes in the Ag NPs stored into the metal desiccator are discussed. For this sample, a remarkably blue-shift of the LSPR band was observed following the slightly red-shift. The color of this sample became reddish after 14 days passed, different from that of the other samples. The atomic ratio N/Ag for this sample increased to 0.66 at 14 days, remarkably higher than that for the other samples. As indicated by XPS and RBS, the Ag-CN and/or Ag-HCN compounds formed on the Ag NPs. It is probable that the origin of the chromatic change may be such compound formation, followed by surface roughening as revealed by AFM. More importantly, reddish coloring indicates the presence of CN and/or HCN inside the metal desiccator. Considering that the cyanide is used in stainless steel manufacturing [34], there is a possibility that it is released from the components of the metal desiccator.
Finally, the mechanism of a change in optical properties of Ag NPs/SiO 2 is discussed briefly to understand the behavior shown in Figure 8. As described above, metallic Ag NPs exhibit strong LSPR in the visible region. The formation of compound, e.g., Ag 2 S, AgCl and AgCN, layers on Ag NP surfaces exposed to air reduces the volume of metallic Ag NPs. The reduction in volume of metallic Ag NPs enlarges the gap between adjacent NPs. The volume reduction and the gap widening result in the weakening of LSPR intensity (referred to as "effect 1", hereafter) and the blue-shift of LSPR wavelength (effect 2) [35]. For a single metallic NP covered with a thin layer with a high refractive index, the LSPR position shifts toward longer wavelength (effect 3) together with enhancement in LSPR intensity (effect 4), depending on its refractive index as well as thickness [36]. In fact, McMahon et al. [23] calculated Mie scattering efficiency of a single Ag NP with various Ag 2 S thicknesses, and obtained results concerning the effect 1 and effect 3, i.e., reduced intensity and red-shift of LSPR with increasing Ag 2 S thickness. The calculated changes are identical to the LSPR changes observed for the sample placed near the ventilation fan. The LSPR changes observed for the sample stored in the metal desiccator were unique in the LSPR wavelength, i.e., blue-shift, which could be explained by the effect 1 and effect 2. In this sample, less influence of effect 3 would be due to the lower refractive index of AgCN (1.7 [37]) compared with those of AgCl (2.0 [38]) and Ag 2 S (2.2 [39]). The LSPR wavelength shifts as a function of refractive index were experimentally obtained by Sugawa et al. [40] for Ag NPs. It can be, therefore, considered that the refractive index of the overlayer formed on Ag NP surfaces determines the LSPR wavelength changes, red-shift or blue-shift.
Apart from the chromatic change, it should be noted that Br was detected on Ag NPs inside the metal desiccator, indicating that Ag NPs possess an ability to adsorb very toxic Br vapor. Ag NPs can, therefore, be used as adsorbent to detect trace amounts of toxic Br as well as cyanide because of their high reactivity.

Conclusions
The origin of chromatic changes is examined from the viewpoints of compositional and morphological changes in Ag NPs/SiO 2 stored in various environments to develop a small and cheap device for monitoring of the cleanliness of the atmosphere. It was demonstrated that bluish coloring followed by quick discoloring of Ag NPs/SiO 2 was induced by adsorption of a significant amount of S and Cl as well as particle coarsening due to the formation of AgCl, as was observed for the sample stored near the ventilation fan. Of these elements, Cl was probably carried by the particle matter containing Cl compounds, e.g., NH 4 Cl, considering that Cl was only detected in unenclosed samples. A remarkable blue-shift of the LSPR band, i.e., reddish coloring, was observed for the sample stored into the metal desiccator, resulting from the formation of Ag-CN and/or Ag-HCN compounds as well as surface roughening. The chromatic change into reddish color shows the inclusion of CN and/or HCN molecules inside the environment. Furthermore, the discoloration rate, defined as time required to become transparent, would be sensitive to the concentration of material to be adsorbed, meaning that the discoloration rate is a measure of the cleanliness of the air.
As expected from the present work, Ag NPs/SiO 2 will be a promising material to check the quality of the environment. Long-term monitoring is more important than quick sensing for substances that accumulate in the human body and electronic materials, such as gaseous pollutants and PM. Our findings provide a pathway to develop a device that can easily check the cleanliness of the air by monitoring chromatic changes in Ag NPs/SiO 2 with the naked eye. In addition, a fiber optic sensing system based on optical absorption changes, similar to the chemical gas sensor proposed by Chen et al. [41], can be developed using Ag NPs/SiO 2 , if quick sensing rather than monitoring is required.