Detection of Water Vapor by Chemiluminescence

Abstract

We examined the possibility of detecting water vapor by chemiluminescence using the reaction of popular “chemical light” (bis(2,4,5-trichlorophenyl-6-carbopentoxyphenyl)oxalate with H₂O₂). H₂O₂ is released from sodium percarbonate exposed to water molecules as in the oxygen bleach. The release of H₂O₂ by water vapor was confirmed by mass spectrometry in a vacuum. The chemiluminescence from the mixed reagents was observed when exposed to water vapor. This method opens the way to locally detect the faulty points of water barrier films and observe the real-time failure of the barrier films during bending tests of flexible packing materials. A molecular dynamics simulation was performed to study the diffusion of H₂O₂ molecules in polymers.

1. Introduction

Water vapor barrier performance is a very important property for the implementation of various devices, including organic light-emitting diodes (OLEDs) [1], thin film transistors (TFTs), and liquid crystal displays (LCDs) [2], organic or perovskite solar cells [3], and lithium-ion batteries [4]. The criterion for the water vapor barrier is very severe in OLEDs (10⁻⁶ g m⁻² day⁻¹ level) [1], while they are less severe in other thin film devices (TFTs and solar cells) and thick film devices (LCDs and batteries). There are various methods to detect water vapor transmission rate (WVTR) for film substrate materials and coatings: the calcium test method [5,6,7], differential pressure measurement [8], mass spectrometry [9], mass spectroscopy with a cold trap [10,11,12], electrochemical detection [13], highly sensitive optical absorption using cavity ringdown [8,14], etc. We have found that the adsorption of water molecules on the instrumental wall gives the deviating background by temperature fluctuation [12], and a method that uses chemical reactions, such as the calcium test, has an advantage for reducing the instrument sizes and, thus, the effect of wall outgasses.

There are two directions to improve the WVTR measurement. First, it has been practically impossible to detect WVTR locally. Since the major problem is the permeation of water vapor from cracks in the barrier coatings or pinholes formed during manufacturing, it is desirable to develop a position-sensitive technique of WVTR measurement. This local detection has been only achievable by using the calcium test in combination with optical microscopy [15]. However, the ordinary calcium test only sees the change in the appearance of the calcium metal after the reaction with water, in which the metallic luster of calcium becomes white/transparent Ca(OH)₂. It takes a very long time (e.g., weeks and months) to see the change for high-barrier films or coatings [5,6,7,15]. Although there are proposals to use electrical detection of the degradation of calcium to shorten the measurement time [16], it is not possible to locally detect the water vapor transmission in a pixel-wise manner. Second, all of the measurement methods so far are designed for static measurement of the whole sample, in which the sample is fixed to a rigid sample holder and cannot move. It is desirable to measure flexible films and their coatings in real time in motion, for example, during repeating bending tests to reduce the testing time. However, there is a high possibility of making a mistake in the handling, i.e., detaching the sample from the bending test machine and setting it to the WVTR measurement instrument repeatedly.

This paper describes the attempt toward the WVTR measurements with position sensitivity and for an unrigid measurement setup which will enable the local and real-time monitoring of WVTR properties in motion. It involves chemiluminescence initiated by water molecules. Luminescence is used in analytical chemistry and related areas [17,18], but no chemiluminescence was used in WVTR measurement. There are various types of chemiluminescence [19], such as involving firefly luciferin [20], Vargula luciferin [21], aequorin [22], luminol [23], 1,2-dioxcetane [24], etc. Cyalume chemiluminescence [25,26] (Figure 1) is one of them. There are various mechanisms specific to each chemiluminescence, for example, twisted intramolecular charge transfer, chemically induced electron exchange luminescence, etc. The luminol reaction and luciferin chemiluminescence are caused by the reaction with oxygen with a catalyst or an enzyme. The Cyalume chemiluminescence involving peroxyoxalate is initiated by the decomposition of bis(2,4,5-trichlorophenyl-6-carbopentoxyphenyl)oxalate (TCPO) or its derivative by H₂O₂ to yield C₂O₄ (peroxyoxalate) in the excited state. The Cyalume reaction is used as chemical lights with various colors, in which excitation of C₂O₄ is transferred to fluorescent dye molecules with various colors. Some substances release H₂O₂ in the existence of water. One of them is sodium percarbonate (Na₂CO₃·1.5H₂O₂) which is the main ingredient of oxygen bleach [27]. Here, we did experiments to show the possibility of using Cyalume chemiluminescence in combination with sodium percarbonate to detect water molecules. We demonstrated that the detection of water vapor is possible and examined the sensitivity. We discuss whether the expected sensitivity is high enough for the local WVTR detection of high-barrier coatings, and the remaining issues for practical applications. Additionaly, we studied the effect of organic solvents on the diffusion of an H₂O₂ molecule in a polymer matrix which has been encountered in the experiment.

2. Materials and Methods

2.1. Materials

TCPO, sodium percarbonate, polymethylmethacrylate (PMMA), and rubrene were purchased from Sigma-Aldrich Japan and used as purchased. Methyl acetate and chloroform (Sigma-Aldrich Japan, Tokyo, Japan) were completely dried by immersing vacuum-heated (300 °C) zeolite 4 A before use. The water level was checked by the Karl Fischer method (Kyoto Electronics Manufacturing Co. Ltd., Kyoto, Japan) to be undetectable (<5 ppm).

2.2. Emission of H₂O₂ from Sodium Percarbonate

Gas emission analysis was performed using a lab-made vacuum system (Figure 2) equipped with a turbo-molecular pump (base pressure < 1 × 10⁻⁵ Pa without baking) and quadrupole mass spectrometer (QMS; QIG-066, Canon Anelva, Kawasaki, Japan) to detect the emission of H₂O₂ (34 AMU). Water vapor was supplied through a variable leak valve and a stop valve. Distilled water was put in a glass tube welded to the ICF-34 flange and the stop valve was purified by the repeated suck–thaw procedure. A total of 200 mg of sodium percarbonate powder was put in the vacuum chamber.

2.3. Luminescence Initiated by Water Vapor in the Ambient

The chemiluminescent sheet was prepared from TCPO and rubrene mixed in a PMMA matrix coated with sodium percarbonate. The procedure was as follows: (1) TCPO 10 mg and rubrene 5 mg were dissolved in methyl acetate 5 mL. (2) PMMA 30 mg was dissolved in chloroform 10 mL. (3) Solutions (1) and (2) were thoroughly mixed and cast on a glass plate. It was dried by heating at 70 °C in an oven until the PMMA-based film started showing surface adhesion. (4) Finely grounded sodium percarbonate powder was bar coated on the sticky surface of the film (3). Solutions (5) and (4) were vacuum-dried by rotary pumping.

Light emission was measured by a spectrometer equipped with 1800/cm grating with a cooled CCD detector (DV-401, ANDOR Technology, Belfast, Great Britain). Green laser (semiconductor laser excited Nd: YVO₄, 532nm, 5 mW) was used to excite rubrene for reference. The measurements were performed at room temperature.

2.4. Demonstration of Local Detection

We demonstrated local detection by making a sensor sheet with a transparent barrier and a star-shaped opening. TCPO 10 mg, rubrene 5 mg, and sodium percarbonate 30 mg were dispersed in dry methyl acetate 5 mL by sonication. It was coated on filter paper (ADVANTEC No. 2, 260 μm-thick) and completely dried in a vacuum desiccator. The filter paper was laminated with a polyolefine sheet (200 μm-thick) using a non-water-base glue from both sides. One side has an opening with a star shape of about 8 mm. The optical image in the dark was recorded with a color CMOS camera (Sony, Tokyo, Japan).

2.5. Molecular Dynamics Simulations

It is necessary to make H₂O₂ molecules diffuse into the materials containing TCPO and rubrene to initiate chemiluminescence. We performed molecular dynamics simulations to examine the diffusion of an H₂O₂ molecule in PMMA to see the effect of solvent methyl acetate that is observed experimentally. For the calculation, we used LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator, version 2022) [28] using 63 parallel processors. A total of 1.3–1.6 × 10⁵ atoms were treated in the simulations. Sixteen PMMA 100-mer chains and an H₂O₂ molecule were mixed in a cubic unit cell and the cell volume was diminished under NVT (constant number-volume-temperature) conditions to achieve 5 atm pressure. For the “with-the-solvent” simulation, 1920 methyl acetate molecules were premixed with the polymer. The simulation employed a generalized AMBER force field (GAFF) [29]. The LAMMPS data file was prepared by moltemplate [30] and antechamber [31] with the atomic charges calculated by Gaussian 16 W [32] at the B3LYP/6-31G+ level of theory. The simulation step was 0.1 fs, and the temperature was set to 1000 K to mix the system for 100 ps and then kept at 300 K for 100 ps using a Nosé–Hoover thermostat with 1 atm barostat. The trajectory was recorded after thermal equilibrium was achieved. The trajectory of the H₂O₂ molecule was visualized by OVITO visualization software [33].

3. Results and Discussion

3.1. Emission of H₂O₂ from Sodium Percarbonate When Exposed to Water Vapor

It is established in the literature that (1) the chemiluminescence occurs from the TCPO-fluorescent dye system when H₂O₂ is present, and (2) H₂O₂ is released by the reaction between sodium percarbonate and a bulk liquid H₂O. For water vapor detection by combining (1) and (2), the critical question is whether a small amount of H₂O molecules can release H₂O₂ molecules quantitatively. To examine this point, we set 200 mg of sodium percarbonate in a vacuum chamber and exposed it to a controlled amount of H₂O gas, while the released H₂O₂ was detected by QMS. Figure 3 shows the result. Figure 3a indicates the detection of supplied water (18 AMU) along with the total pressure when the variable leak valve V1 was set to introduce 2 × 10⁻³ Pa as the total pressure. The V2 valve was open and closed at the timing as seen in the change of the total pressure. Figure 3b shows the emission of H₂O₂ (34 AMU) using the same opening of the V1. It is noted that the detected signal intensity in Figure 3b is very stable, and it is confirmed that the H₂O₂ is emitted from sodium percarbonate quantitatively as the introduction of H₂O₂ vapors. The instability of the QMS signal of H₂O (18 AMU) is due to the adsorption of water molecules on or around the components of the ionization chamber of QMS [12,34].

No 34 AMU signal was at background level (1 step fluctuation in Figure 3b) when sodium percarbonate powder was not placed in the vacuum chamber.

3.2. Chemiluminescence from a Polymer Composite Induced by Water in the Air

Since it is now established that the water vapor can release H₂O₂ molecules in a quantitative and repeatable manner from sodium percarbonate powder in a vacuum, we fabricated a chemiluminescence detection sheet composed of TCPO and rubrene mixed in PMMA and sodium percarbonate.

Figure 4 shows the optical emission spectrum from the sheet. Figure 4a shows the photoluminescence of rubrene in the sheet as irradiated by the green laser (532 nm), showing the luminescence of rubrene. Figure 4b shows the sheet exposed to the ambient air (23 °C, RH 35%) from which no signal was detected. Figure 4c shows the emission spectrum from the sheet when it was exposed to the vapor of dry methyl carbonate in the ambient air. The luminescence of rubrene was observed. This result indicates that the chemiluminescence does not occur only with exposing the sheet to the water in the ambient atmosphere but becomes observable when the solvent of the polymer coexisted. We consider that the migration of H₂O₂ to TCPO or excited C₂O₄ species to rubrene is necessary to induce chemiluminescence. Thus, it is desirable to develop a polymer matrix material that allows the migration of those species for practical measurement by this technique. In Section 3.4, the molecular dynamics simulation results are discussed whether the methyl acetate molecule really enhances the migration of H₂O₂.

3.3. Demonstration of Position-Sensitive Detection of Water Vapor

The advantage of using optical imaging of chemiluminescence is the ability to position-sensitive detection of failure points/areas of WVTR coatings. We demonstrated it by making a sensor sheet with a transparent barrier and a star-shaped opening. The sensor sheet with the opening was prepared as described in the Section 2.4. The sample was exposed to moisture (40 °C, RH 90%) in the dark while the optical image was recorded with a CMOS camera. Figure 5a shows the image of chemiluminescence caused by moisture. The intensity of the chemiluminescence was strong enough to be observable by the naked eye. The spatial resolution of the water-induced chemiluminescence is ~300 μm as shown in Figure 5b,c. The broadening of the signal probably comes from water molecules diffusing through the thickness of the filter paper substrate (260 μm-thick) which can be improved by designing thinner detection layers.

The difference from the previous section is the use of filter paper and high-water vapor pressure in this experiment. The high concentration of water molecules adsorbed on the filter paper probably mediated the migration of the chemical species involved so that chemiluminescence was easily observed.

3.4. Molecular Dynamics Simulation

Figure 6 shows the result of molecule dynamics simulations to see the diffusion of an H₂O₂ molecule in the polymer matrix at 300 K and 1 atm. Figure 6a shows the atomistic particle image of the simulation cell. The size of the cell is 115 Å and 134 Å for the PMMA without and with methyl acetate, respectively. Figure 6b shows the distribution of methyl acetate molecules after equilibration. It indicates that the methyl acetate molecules distribute uniformly. Figure 6c,d show the trajectories of the H₂O₂ molecule for 30 ps. It is easily seen that the volume containing the trajectory is different. From the calculation, they are 3.0 × 10² ų and 9.9 × 10² ų for Figure 6c,d, respectively. It indicates that diffusion is enhanced when methyl acetate is mixed. The estimation of the diffusion constant will be given in the Section 3.

4. Discussion

4.1. Discussion on the Experimental Results

In the previous section, we demonstrated the following things. (1) Sodium percarbonate releases H₂O₂ molecules in a repeated and constant manner when H₂O molecules are supplied from the gas phase with a very small pressure of 2 × 10⁻³ Pa that is 1/50,000,000 of atmospheric pressure. This result is promising for highly sensitive detection of WVTR of the gas barrier films. (2) Chemiluminescence can be initiated by water vapor in the environment when the TCPO-rubrene is fixed in PMMA film and sodium percarbonate is attached. The completely cured polymer film does not emit chemiluminescence which indicates the migration of the involved species is necessary. (3) Position-sensitive chemiluminescence induced by exposure to water vapor is possible. Although there are still many hurdles before the practical use, including the shelf life of the prepared detection films, the use of chemiluminescence for water vapor detection will open a way to enable local failure analysis of the gas barrier films, and, in addition, during motion.

4.2. Discussion on the Molecular Dynamics Simulation

As shown in Section 3.4, molecular dynamics simulation has revealed that the addition of the methyl acetate molecules to PMMA enhances the diffusion of H₂O₂ in the matrix. This is by the experimental observation in Figure 4, in which the chemiluminescence was only observed under the exposure of the sample to the methyl acetate in addition to the moisture in the air. Here, we quantitatively evaluate the effect of methyl acetate in the simulation. From Figure 6c,d, the migration volumes of H₂O₂ in 30 ps (=t) are calculated to be 3.0 × 10² ų (Figure 6c = V_c) and 9.9 × 10² ų (Figure 6d = V_d), respectively. If we regard these volumes (V) can be approximated by spheres, the radius (r) of the spheres can be calculated from V = 4πr³/3. The radius corresponding to V_c and V_d are r_c = 4.15 Å and r_d = 6.18Å, respectively. The concentration of the diffusing particles at the radius of r at time t (C(r,t)) is

$$C\left(r,t\right)=\frac{1}{{\left(4\pi Dt\right)}^{3/2}}\exp \left(-\frac{r^2}{4Dt}\right).$$

If we take $C\left(r,t\right)=1/2$ as a contour of the diffusing particle distribution, D can be determined by using r = r_c or r_d and t = 30 ps mentioned above. The results are D_c = 2.05 × 10⁻¹¹ m²s⁻¹ and D_d = 4.62 × 10⁻¹¹ m² s⁻¹ for without and with methyl acetate, respectively. These are rather reasonable values as compared with the self-diffusion constants of liquid (D~10⁻⁹ m² s⁻¹ [35]). The question is whether the difference in the diffusion constants is not as large as appearing in the experiment, in which chemiluminescence was not observed without methyl acetate. We consider this point can be explained by considering that our simulation used only 100-mer of PMMA, while the actual PMMA polymer is much longer (>10³) which hinders the motion of the molecule. Here, we have confirmed that the molecular dynamics simulation is useful to evaluate the diffusion of H₂O₂ in polymers and will be effective to search the polymer matrix for making the detection sheet.

4.3. Expected Sensitivity

We have so far confirmed that the chemiluminescence can be initiated by water molecules using the TCPO-rubrene system. It is necessary to examine how high the sensitivity can be using this method and whether it is enough to detect water vapor transmission locally. We make some assumptions: (1) A water molecule initiates the reaction that emits one photon. (2) A pinhole or crack will make a 1 × 10⁻⁵ level increase in WVTR in a sample. (3) The sample size is 63.6 cm² (9 cm diameter circle). With these assumptions, the photons emitted from the pinhole or crack will be 2.5 × 10¹⁰ photons s⁻¹. There are various types of highly sensitive imaging photodetectors developed for biological [36] or astronomical [37] research. For example, various types of image intensifiers including amorphous semiconductors [38] and microchannel plates (MCP) [39] have been developed to amplify the image signal with high spatial resolution. Photodetectors have been improved to suppress the dark counts down to 0.05 counts/pixel/s [40] and with high sensitivity to other advantages [41,42,43]. The noise has been suppressed by cooling the system containing detectors [37] and used a statistical algorithm [44]. Even their combination in a single device has been developed [45]. The overall sensitivity of a state-of-the-art detector is, for example, 50% quantum yield, conversion of one photon to one electric pulse (photon counting), with 0.05–3 counts/pixel/s noises with 30 mm resolutions or very wide apertures. Therefore, it is rather easy to detect 2.5 × 10¹⁰ photons s⁻¹ of chemiluminescence from a pinhole/crack area with mm or μm scales with a state-of-the-art camera even without an image intensifier. It is, thus, concluded that the present technique will be useful to locally detect the water permeation of a gas barrier film or coatings. If the detection sheet is prepared as a flexible film and closely attached to the sample film, it will be possible to detect water vapor permeation during the bending test. In this work, we have not examined the proportionality of the chemiluminescence intensity and the WVTR values because it is strongly influenced by the distribution of the sodium percarbonate that was used as powders. However, it is safely said that this technique has the advantage of locally detecting the water permeation with high sensitivity which is very important in the practical application of barrier materials.

4.4. Comparison with Currently Available WVTR Measurement Techniques

There are various techniques available for WVTR measurement as explained in the introduction section, and some of them have come to use in industry. The sensitivity is 10⁻⁵–10⁻⁶ g m⁻² day⁻¹ and the acquisition time for one data with enough signal-to-noise ratio is c.a. 1 h which is far from the real-time measurement. It should be noted that this is even after the improvement from classical calcium test or electrochemical detection which required much more time for the measurement or was absolutely impossible for this high-sensitivity measurement.

In the previous section, we have made many assumptions and reached the conclusion that it is possible to locally detect a crack or a pinhole corresponding to 1 × 10⁻⁶ g m⁻² day permeation of the water molecules which emits 2.5 × 10⁹ photons s⁻¹. It should be emphasized that the detection can be real time for 2.5 × 10⁹ photons s⁻¹ with a state-of-the-art camera (noise level 0.5 photons pixel⁻¹ s⁻¹). The present chemiluminescence approach surpasses the sensitivity of the techniques currently available by orders of magnitudes. If the water vapor permeation is uniform over the sample, the detection becomes more difficult than in the case of cracks or pinholes. Still, it is possible to evaluate the WVTR with the great difference between the signal and noise (2.5 × 10⁹ and 0.05 photons pixel⁻¹ s⁻¹, respectively) when photodetection is only involved.

With the experience obtained during the current experiments, we have learned that the intensity of chemiluminescence strongly depends on the history of the raw material, especially sodium percarbonate. If the sodium percarbonate has been exposed to ambient conditions for a long time (e.g., one day), water vapor in the air reacts with sodium percarbonate, and H₂O₂ emission continues for a while. This will make a large background of chemiluminescence. Moreover, the temperature can be a problem because sodium percarbonate releases H₂O₂ at high temperatures. If we can avoid these problems, the sensitivity of the present technique is much better than other techniques currently available.

4.5. Problems and Challenges

We have found several problems in the application of this method to practical water vapor sensors. One is the need for migration of involved chemical species in the detector film. The migration PMMA polymer-based system without methyl acetate in the ambient condition was not enough (Section 3.2). This problem can be solved by using gel-type matrix material to make the detection sheet which can be screened out by computational search by the molecular dynamics simulations as shown in Section 3.4. Another problem is the short shelf life. The filter paper-type detector (Section 3.3) loses its function in several days even when it is kept in a nitrogen glove box (dew point −20 °C). We consider this to be due to the trace amount of water in the environment, causing the spontaneous chemiluminescence and degradation of the material. Additionally, since it is reported that sodium percarbonate starts decomposition by heating [46,47], spontaneous thermal decomposition to release background H₂O₂ may be the problem. For the water, tightly sealing from the environment is one way to solve this problem, but it is rather difficult to prevent the effect of the thermal release of H₂O₂. There are various peroxysolvates [46,47] which contain H₂O₂ as a crystal solvent. The solution for the latter will be using other peroxosolvates that have higher thermal stability. It is noted that this method uses irreversible chemical reactions and, thus, the device is not reusable. The running costs may become problems.

5. Conclusions

We examined the possibility of detecting water vapor by chemiluminescence. The luminescence from the mixed reagents was observed when exposed to water vapor. We have demonstrated that this technique can be used for the evaluation of water permeation with a spatial resolution, and the sensitivity can be high enough to detect cracks or pinholes causing 10⁻⁶ g m⁻¹ day⁻¹ level WVTR. This method opens the way to locally detect the faulty points of water barrier films and observe the failure of barrier films during bending tests.

6. Patents

We have applied for a patent for the present technique.