1. Introduction
Gas sensors are indispensable devices for monitoring air quality in living spaces, where a high level of safety and comfort is required. This is especially true in modern buildings and vehicles that have become airtight to save energy and be soundproof, and in underground malls where natural ventilation is difficult, there is a potential for hazardous gasses to reach concentrations that are harmful to the human body and that pose a risk of fire or other danger. Optical gas sensors, which provide optical signals when specific gasses are detected, have advantages over electricity-based gas sensors, which produce electrical signals in response to gas detection. These advantages include excellent resistance to electromagnetic noise; high explosion resistance and safety due to the absence of electrical spark generation; non-contact and remote signal readout from the sensor element; and compatibility with optical fibers and waveguides. Because of these advantages, which are not found in electricity-based gas sensors, optical gas sensors are attracting attention as next-generation sensors [
1,
2,
3,
4].
Ozone (O
3) has both strong oxidizing and corrosive properties. On the other hand, it is easy to handle, because it typically decomposes gradually into non-toxic oxygen. This oxidizing property provides applications that include sterilization and deodorizing of air and water, inactivation of viruses (including COVID-19), and treatment of diseases and external wounds [
5,
6,
7]. Recently, small ozone generators for home deodorizing have even become widespread. However, ozone is highly toxic to the human body when its concentration in the air exceeds a few ppm, so ozone sensors are necessary [
8,
9,
10,
11]. Various sensing technologies have been developed and put into practical use for ozone detection and concentration measurement in air, including semiconductor [
12,
13], electrochemical [
14], ultraviolet (UV) absorption [
15], gas detection tube [
16,
17], chemiluminescent [
18,
19], and iodometric [
20] methods [
10,
12]. However, each of these conventional technologies has its own advantages and disadvantages, and there is a continuing need to develop ozone-sensing technologies that can detect ozone more easily and quickly and that reduce the size and weight of sensor devices [
11].
We have recently reported that the ppm level of ozone in the air can be detected at room temperature and atmospheric pressure using the reversible, photoluminescence (PL) intensity change of core–shell quantum dot (QD) films composed of group II-VI semiconductors, such as CdSe/ZnS prepared on a planar glass substrate [
11,
21,
22,
23,
24]. Among the variety of optical ozone sensors, the detection method using the change in PL intensity of the sensor material due to ozone has the advantage of expanding the dynamic range of the output signal compared to the method using the optical absorption change in the sensor material [
25]. Optical ozone sensors using the reversible change in PL intensity of QDs are considered to be safer than semiconductor ozone sensors, because the optical ozone sensors using QDs do not require high-temperature heating, are more resistant to electromagnetic noise, and have the potential to be smaller and lighter than electrochemical ozone sensors, chemiluminescent ozone sensors, and UV-absorption-type ozone densitometers because of the simple and all-solid-state structure of the optical ozone sensors using QDs [
11,
24].
The PL quenching of QDs by ozone is assumed to be due to the oxidizing power of the adsorbed ozone and/or the weakening of the bonds between the inorganic QDs and the organic surfactant molecules that coat inorganic QDs, resulting in a change in the electronic state of the QDs and an increase in surface defects and consequent enhancement of the non-radiative recombination of the charge carriers in the QDs. We also reported that the reversible decrease in PL intensity due to ozone exhibited by QDs can be enhanced by forming composites with noble metal nanoparticles, such as Au and Pt, which have gas adsorption/desorption properties and local electric field enhancement effects, with the QD film [
22]. The QD film prepared on a planar glass substrate or a noble metal/QD composite film showed a relatively rapid response speed (seen as a decrease in PL intensity) when exposed to ozone, but the recovery speed (increase in PL intensity) in ozone-free air after exposure to ozone was still slow. In the case of QD films on planar glass substrates, the thinner the QD layer was, the lower the possible PL intensity level was, and the lower the resulting signal-to-noise ratio of the gas sensor output signal was. Porous glass is known to have good gas distribution due to its continuous pores, as well as a high adsorption capacity for various gasses due to its large specific surface area [
26], so it is expected to be a high-performance gas sensor material if QDs can be integrated into it. In this study, in order to speed up the adsorption and desorption of ozone molecules, we reduced the thickness of the QD layer (compared to the conventional QD film on a planar glass substrate), while at the same time dispersing and immobilizing QDs on the inner walls of the pores of a porous glass substrate to ensure that the sample contains a sufficiently large amount of QDs to produce a high PL intensity and thus maintain a high signal-to-noise ratio. The ozone sensor characteristics of the porous glass with dispersed QDs were investigated.
2. Materials and Methods
A decane dispersion of red-emitting CdSe/ZnS core–shell QDs (QD concentration: 1 μM) capped with trioctylphosphine and trioctylphosphine oxide (Q21721MP, Invitrogen
TM, Thermo Fisher Scientific Inc. (Waltham, MA, USA)) was diluted 1000-fold with toluene to prepare a mixed decane/toluene dispersion of QDs (QD concentration: 1 nM). This decane/toluene-mixed dispersion of QDs was absorbed by capillary action into a porous glass substrate (20 mm square; 1 mm thick; average pore size: 50 nm; specific surface area: 80 m
2/g; void fraction: 49.1%; bulk density: 1.12; Akagawa Glass Co., Ltd. (Osaka, Japan) [
27]) at room temperature. The porous glass with dispersed QDs was then kept under vacuum at room temperature in a portable aspirator (MDA-015, ULVAC KIKO Inc. (Miyazaki, Japan)) to remove the decane and toluene, and the QDs were dispersed and immobilized on the inner walls of the pores of the porous glass. Since all the pores in the porous glass are contiguous with each other, the gas distribution is good, and rapid gas adsorption/desorption is possible. The porous glass substrate is 1 mm thick, which is advantageous for loading a large amount of QDs and obtaining a high PL intensity from the entire sample.
The fine structures of the porous glass with dispersed QDs and porous glass without QDs were observed using a scanning electron microscope (SEM) (SU-8230, Hitachi (Tokyo, Japan)) operating at an acceleration voltage of 2 kV. SEM observations were made after osmium (Os) deposition on the samples to prevent charge-up.
The optical ozone sensor characteristics due to changes in PL properties of the porous glass with dispersed QDs were measured at room temperature (25 °C) and atmospheric pressure (1 atm) using the experimental apparatus and methods previously described [
11,
21]. Air containing ozone at a defined concentration (in the range of 0.5–200 ppm) was generated from an ozone generator (EG-OG-L-AIST, EcoDesign, Inc. (Saitama, Japan)) and distributed at 100 mL/min onto the optical quartz gas cell (internal volume: 41 mL) (C-type, Sawada Quartz Co., Ltd. (Kyoto, Japan)) containing porous glass with dispersed QDs. The ozone generator used dry (moisture-free) synthetic air (78.5% nitrogen + 21.5% oxygen) (Taiyo Nippon Sanso JFP Co. (Tokyo, Japan)) as the source gas, which was pulsed with UV light (wavelength: 185 nm) in the chamber to convert oxygen to ozone through a photochemical reaction; the ozone concentration was controlled by the intensity and frequency of the UV irradiation. The ozone generator was connected to a UV-absorption-type ozone concentration meter (UV ozone monitor OZM-7000GNW, OKITROTEC Co., Ltd. (Tokyo, Japan)) to generate air containing ozone at a defined concentration by feedback control.
Another dry synthetic air cylinder was connected to the gas distribution system via a three-way stopcock in order to quickly switch the atmosphere in the optical quartz gas cell from ozone-free air to ozone-containing air, alternately. The porous glass with dispersed QDs in the optical quartz gas cell emitted PL when excited by UV light (wavelength: 365 nm), which was irradiated from above through an optical band-pass filter for 365 nm (MX0365, Asahi Spectra Co., Ltd. (Tokyo, Japan)). The UV light source for excitation was a Spot Cure SP-11 (USHIO LIGHTING Inc. (Tokyo, Japan)) with an optical quartz fiber. The PL was transmitted by the optical quartz fiber to a spectro-multichannel photodetector (MCPD-7000, MCPD System Version 4.0.2.8, Otsuka Electronics Co., Ltd. (Osaka, Japan)), where the spectrum and intensity of the PL were measured simultaneously at all visible wavelengths by a photodiode array [
11,
21,
22].
3. Results and Discussion
Figure 1 and
Figure 2 show SEM images of the porous glass without QDs and with dispersed QDs, respectively. The bright area in the SEM images corresponds to the silica glass skeleton, while the dark area corresponds to pores between the silica glass skeleton. The contrast between bright and dark areas in the SEM images shows that these two samples have similar porous structures. As shown in
Figure 1, the inner pores of the porous glass without QDs were continuous, with an average pore size of ca. 50 nm.
Figure 2 shows that the continuous pore structure was maintained after the QDs were dispersed in the porous glass, and the average pore size was ca. 50 nm, the same as in the porous glass without QDs. Regarding the structural differences between the two samples, the inner pore walls of the porous glass without QDs are very smooth, and no surface roughness exceeding ca. 5 nm in height is observed (
Figure 1), whereas the inner pore walls of the porous glass with dispersed QDs are less smooth and show a surface roughness of ca. 5–10 nm in height (
Figure 2). This surface roughness is assumed to be the result of the deposition of QDs on the inner pore walls. No surface roughness exceeding ca. 10 nm in height was observed on the inner pore walls of the porous glass with dispersed QDs. These SEM observation results indicate that the QDs were immobilized on the inner pore walls of the porous glass in a highly dispersed state, without forming large aggregates or thick layers with diameters or thicknesses exceeding ca. 10 nm.
Figure 3 shows a schematic diagram depicting the structure of the porous glass without QDs, the structure of the porous glass containing CdSe/ZnS QDs in a highly dispersed state, and the process by which the bright PL of the QDs is reversibly quenched by ozone exposure.
Figure 4a–d show photographs of the porous glass with dispersed QDs under room light and UV irradiation (wavelength: 365 nm), as well as the porous glass with dispersed QDs in an optical quartz gas cell. The porous glass with dispersed QDs was translucent under room light (
Figure 4a,c) but emitted bright red PL when UV-irradiated (wavelength: 365 nm) (
Figure 4b,d). The uniform PL of the sample indicated that the QDs were uniformly dispersed and immobilized throughout the porous glass.
Figure 5 shows the PL spectra of the porous glass with dispersed QDs in air and in air containing ozone at concentrations of 0.5, 20, and 200 ppm. The peak PL wavelength in air was 646 nm, which was 6 nm blue-shifted from the peak PL wavelength of 652 nm exhibited by a film of the same QDs that was prepared on a planar glass substrate [
22,
23]. Due to the quantum size effect, the effective size of the QDs decreases/increases as the degree of aggregation of the QDs decreases/increases, and the PL wavelength becomes shorter/longer [
28,
29], suggesting that the QDs are highly dispersed without aggregation in the porous glass. The PL intensity of the porous glass with dispersed QDs in ozone-free air rapidly decreased as it was replaced by ozone-containing air, and then recovered quickly and reversibly when the atmosphere was returned to ozone-free air. The degree of PL quenching (decreased ratio of PL intensity) increased with each increase in ozone concentration. The PL intensity change due to ozone was always reversible in the ozone concentration range tested (0–200 ppm), and the PL peak wavelength and spectral shape were identical and independent of the ozone concentration, indicating that QDs in porous glass are not irreversibly degraded or decomposed by ozone exposure up to a concentration of 200 ppm. This suggests that the porous glass with dispersed QDs has the potential to be a reliable ozone sensor material with good long-term stability.
Figure 6a shows the ozone concentration dependence of the PL intensity of the porous glass with dispersed QDs, and
Figure 6b shows a Stern–Volmer plot using the same data; the Stern–Volmer plot shows the relationship between the inverse of the relative PL intensity and the ozone concentration in air. No saturation of response to ozone was observed when the ozone concentration was increased to 200 ppm, indicating that the porous glass with dispersed QDs can detect ozone over a wide concentration range. The slope of the Stern–Volmer plot decreased with increasing ozone concentrations in a concentration range of 0–50 ppm and remained nearly constant in the ozone concentration range of 50–200 ppm. The shape of the Stern–Volmer plot is similar to the shape of the Stern–Volmer plot for the ozone response of QD films on planar glass substrates reported previously [
21,
22]. This suggests that the mechanism of ozone-induced PL quenching of the porous glass with dispersed QDs is mainly due to collisional or dynamic quenching, and the contribution of static quenching is small, as in the case of QD films on planar glass substrates.
Figure 7 shows the time response of the change in PL intensity of the porous glass with dispersed QDs due to ozone at concentrations of 0.5–200 ppm. Upon exposure to ozone, the PL intensity of the porous glass with dispersed QDs decreased rapidly, and at almost the same speed at all ozone concentrations examined. Subsequently, after switching the atmosphere to ozone-free air, the PL intensity recovered quickly, again at almost the same speed regardless of the ozone concentration tested. Such rapid response and recovery over a wide range of ozone concentrations could be a major advantage of the porous glass with dispersed QDs as an ozone gas sensor material.
Table 1 compares the response times for 0.5 ppm ozone or 200 ppm ozone, as well as the recovery time in air, for the porous glass with dispersed QDs compared to the previously reported responses of a QD film on a planar glass substrate (using the same red-emitting CdSe/ZnS QD (Q21721MP)). The decrease in the PL intensity ratio of each sample after 12 min of exposure to ozone-containing air is used as the reference (100%), and the times taken to reach 50% and 90% of that ratio decrease are defined as the 50% response time and 90% response time, respectively. For the porous glass with dispersed QDs, the 50% response time, 90% response time, 50% recovery time, 90% recovery time, and 100% recovery time were 1, 2, 4, 10, and 14 min for 0.5 ppm ozone, respectively, and 1, 2, 4, 10, and 18 min for 200 ppm ozone, respectively. In contrast, the 50% response time, 90% response time, 50% recovery time, 90% recovery time, and 100% recovery time of the previously reported QD film on a planar glass substrate were 3, 9, 7, 17, and 21 min for 0.5 ppm ozone and 4, 10, 20, 110, and 180 min for 200 ppm ozone, respectively. Thus, the porous glass with dispersed QDs showed faster response and recovery times than conventional QD films on planar glass substrates at both high and low ozone concentrations. In particular, the faster PL recovery after exposure to a high concentration of ozone is very advantageous as a gas sensor material property. In addition, the recovery time after exposure to 200 ppm ozone was reduced to about 10% of that of the QD film on a planar glass substrate. A further advantage is that both the response and recovery speeds of the porous glass with dispersed QDs were almost constant, independent of ozone concentration, which is in contrast to the QD film on a planar glass substrate, where the recovery in air after ozone exposure is significantly slower as the ozone concentration increases [
21,
22]. Since the QDs were immobilized on the inner pore walls of the porous glass in a highly dispersed state and the average thickness of the QD layers is less than ca. 10 nm, which is significantly thinner than the QD film on a planar glass substrate (estimated thickness: several tens nm to several hundred nm by microscopic observation) [
22,
23], we hypothesize that the very thin QD layer on the inner pore walls of the porous glass allows for the fast adsorption and desorption of ozone, resulting in the very rapid response to ozone and concomitantly rapid recovery in air observed in this study.