Chromogenic Approach for Oxygen Sensing Using Tapered Coreless Optical Fibre Coated with Methylene Blue

Abstract

In this paper, a Methylene Blue (MB)-coated tapered coreless (TCL) optical fibre sensor is proposed and experimentally investigated for oxygen sensing in the near-infrared (NIR) wavelength range of 993.5 nm. The effect of TCL diameter and MB sol–gel coating thickness on the sensitivity of the sensor was also investigated. A maximum sensitivity of 0.19 dB/O₂% in the oxygen concentration range of 0–37.5% was achieved for a TCL fibre sensor with a 2 µm taper waist diameter and a 0.86 µm MB sol–gel coating thickness, with a response time of 4 min. The sensor provides reproducible results even after 7 days and is shown to be highly selective to oxygen compared to argon and ethanol at the same concentration.

1. Introduction

Fresh air generally contains 23.5% oxygen, and levels below 19.5% are considered oxygen-deficient by the Occupational Safety and Health Administration (OSHA) [1]. Measuring oxygen levels is crucial for life support, workplace safety, biochemistry, medicine, and various industries [2]. Consequently, many types of sensors have been developed to accurately measure oxygen.

Optical methods are popular because they are non-invasive and can replace traditional electrodes [2]. These methods include direct spectroscopy, absorption-metric and luminescent probes, and ratio-metric sensing, which are valued for their selectivity and visual sensing [3]. New materials and techniques, such as longwave-excitable fluorophores and silicone rubber embedded with indicators and enzymes, have been created [4]. Other technologies like potentiometric, ampere-metric, paramagnetic, and tuneable diode laser spectroscopy have been examined for their theory, operation, and application limits [5].

Fluorescence quenching is a common method for oxygen sensing. It involves a reduction in fluorescence intensity when a fluorophore interacts with oxygen molecules, allowing for the detection of oxygen levels in various environments [6,7]. This method is highly sensitive, enabling the detection of even trace amounts of oxygen, and supports real-time monitoring [8]. It can also be designed to selectively detect oxygen amidst other gases [8]. However, it can be interfered with by other substances that quench fluorescence, and the response time can vary, impacting the speed of detection [8]. The stability and longevity of the sensor materials also affect reliability [6,9].

Phosphorescence quenching is another method for oxygen sensing. It offers high sensitivity and selectivity, and the quenching follows a linear Stern–Volmer relationship at physiological oxygen levels [10]. Platinum and palladium porphyrins are often used in this method for biological systems [11]. This method is reliable for real-time oxygen measurement in tumours [12] and stable for water-based biosensors [13]. However, it can face issues with excitation and emission crosstalk, and the long lifetimes of some dyes require specific lasers and long pixel dwell times, affecting data acquisition rates [14].

Absorption spectroscopy, especially Tuneable Diode Laser Absorption Spectroscopy (TDLAS), is useful for oxygen sensing across different applications. TDLAS has been used to assess seal quality in the pharmaceutical industry by detecting residual oxygen in vial headspaces [15] and to monitor oxygen in fuel cells [16]. It also enables sensitive dual sensing for oxygen and pressure using deep UV absorption spectroscopy [17]. TDLAS is highly sensitive, making it suitable for trace oxygen detection [18], and it is non-invasive and useful for scenarios where direct contact is not possible [19]. However, TDLAS can face stability issues and narrow measurable bandwidths, and its accuracy can be affected by other gases or environmental conditions [20].

Colourimetric sensors detect oxygen through colour changes, offering a simple and visual analysis method, ideal for point-of-care use [21]. They are cost-effective and provide rapid detection for various gases, including oxygen [22]. These sensors use chemical reactions with compounds like Methylene Blue and Indigo to indicate oxygen levels [23], or systems like haemoglobin–oxyhaemoglobin [24]. This study proposes using Methylene Blue immobilised on a tapered coreless optical fibre sensor for oxygen sensing. Methylene Blue changes colour from its reduced to oxidised state, indicating oxygen presence.

Optical fibre sensors are valuable for various applications, including oxygen sensing, due to their fast response time, immunity to electromagnetic interference, and non-consumptive measurement of oxygen [25]. Past approaches have used materials like metalloporphyrins for their luminescent properties and stability [26], and polymer matrices for immobilising oxygen-sensitive fluorophores [27]. Studies have optimised these sensors by selecting better solvents and incorporating antenna dyes to enhance sensitivity and accuracy [28]. All-polymer optical fibre sensors have been developed for specific uses like monitoring dissolved oxygen in fish farms [29]. Multi-point sensors based on laser absorption spectroscopy have been created to measure oxygen at different points simultaneously [30], and temperature compensation mechanisms have enabled real-time gas detection [31].

In the past, various designs for optical fibre sensors have been investigated for oxygen sensing. These include sensors based on fluorescence quenching [32], Fabry–Perot interferometers for dissolved oxygen monitoring [33], and large-diameter fibres with fluorescent materials [34]. This paper explores a TCL fibre optic sensor with Methylene Blue for oxygen detection. The sensor’s colour change indicates oxygen concentration, affecting the absorption of the TCL fibre’s evanescent field. Most optical oxygen sensors use dynamic luminescence quenching [35] and optical fibres to transmit fluorescence. The proposed sensor uses NIR wavelengths and a chromogenic approach for oxygen detection in low-interference environments.

This paper will describe (i) the preparation of a Methylene Blue sol–gel to functionalise the TCL fibre optic sensor and (ii) the fabrication of the TCL fibre optic structure used for oxygen sensing. Experimental results investigating the influence of the tapering diameter of the TCL fibre optic sensor and the MB sol–gel coating thickness on sensitivity will be presented, with further experimental observations on response time, repeatability, and selectivity towards oxygen discussed. Finally, conclusions on the suitability and application of this proposed method of oxygen sensing will be given.

2. Experiment

2.1. Methylene Blue (MB) Sol–Gel Preparation

Methylene Blue (C₁₆H₁₈ClN₃S·3H₂O) was entrapped in sol–gel, which is an optically transparent glass-like material formed by the hydrolysis and polymerisation of metal alkoxides or metal–organic compounds at room temperature. A porous sol–gel matrix formed by a three-dimensional network of SiO₂ allows oxygen to permeate the polymer and interact with dissolved MB [36].

The procedure for preparing the MB sol–gel coating is outlined below:

• A 0.006 gm amount of Methylene Blue is dissolved in 3 mL ethanol and sonicated for 20 min.
• A 1 mL volume of deionised water is acidified by 30 µL of HCL (0.04 mol/V).
• A 4.5 mL volume of TEOS is mixed with 5.5 mL of ethanol by stirring them for 20 min.
• The acidified water prepared in Step 2 is added to the above solution in Step 3 and stirred continuously for 20 min.
• A 1 mL volume of the Methylene Blue solution prepared in Step 1 is added to the above solution (Step 4) and stirred for another 20 min.
• The above solution is ready to be used for coating the TNC optical fibre structure.

2.2. Functionalise Tapered Coreless Optical Fibre Sensor

The coreless optical fibre (NCF125, POFC Corp., Zhunan Township, Taiwan) is a silica cylinder of 125 µm diameter, with the same refractive index throughout, unlike conventional optical fibres, which have core and cladding layers with different refractive indexes. In this experiment, the TCL fibre structure was fabricated by heating a section of coreless optical fibre at a temperature high enough to soften the silica fibre (~1100 °C) and stretch it into a biconically tapered shape [37], as shown in Figure 1a. In this paper, two waist diameters of 2 and 4 µm are investigated for sensitivity comparison. The TCL was coated with the MB sol–gel solution prepared following the above procedure using a dip-coating method. To investigate the influence of coating thickness on sensitivity, three different coating thicknesses of MB sol–gel (i.e., 0.86, 0.69 and 0.39 µm) were then applied to the fabricated TCL fibre structures with tapered waist diameters of 2 µm and 4 µm, respectively. Therefore, a total of 6 sensors with two different diameters and three different coating thicknesses are investigated in this paper.

Figure 1a shows a schematic of the coated TCL fibre optic sensor structure detailing both the initial thickness and fabricated waist region. For this experiment, two waist diameters were produced to allow an experimental comparison of the diameter’s influence to be undertaken. A non-tapered (125 µm) coreless fibre optic sensor was also coated with 0.86 µm MB sol–gel, and its response was evaluated against that of the fabricated structures. A scanning electron microscope (SEM) image of the coated fibre is shown in Figure 1b, showing the dimensions of coating thickness and waist diameter for one of the fabricated TCL fibre optic sensors. The cross-section samples of the coated TCL fibre optic sensor were prepared by cutting the fibre around the waist region, using a ruby fibre cleaver. Figure 1c shows the full-length SEM image of the fabricated TCL fibre optical sensor with a waist diameter of 4 µm. The waist length of the tapered section was around 10 mm, followed by a transition length of 10 mm on both sides.

2.3. Experimental Setup

The experimental apparatus used to characterise the proposed sensor is shown in Figure 2. An airtight box of 1-L volume is used as a gas chamber, with two inlets and one outlet for the ingress and egress of oxygen (test gas) and nitrogen (carrier gas). The oxygen concentration in the gas chamber is controlled by varying the oxygen flow rate using a flow controller. To eliminate the effect of gas flow, the inlets and outlet were closed after allowing the gases to flow through the chamber for 2 min, thus achieving a steady-state environment with desired concentrations of oxygen while taking the readings. Measurements were taken every minute for a 10-min duration, using a Yokogawa AQ6370D optical spectrum analyser (OSA). A broadband light source with a stable output in the NIR region of 800–1150 nm is used to launch light into the fibre optic sensor.

3. Results and Discussion

NIR spectra (800–1150 nm) were observed for all seven fabricated sensors, with 993.5 nm chosen for the experiment for two reasons. First, MB has good absorption at short wavelengths. Second, there is a peak or dip point available around 993.5 nm, making it easier to monitor intensity changes more accurately compared to flat spectra. For the TCL fibre optic sensor with a taper diameter of 2 µm and an MB sol–gel coating thickness of 0.86 µm, a decrease in output intensity with increasing oxygen concentration was observed, with data shown in Figure 3.

The oxygen concentration was gradually increased from 0% to 37.5% by volume. The OSA was programmed to automatically record the light intensity at a wavelength of 935.5 nm (i.e., peak wavelength) every minute for the entire duration of 10 min (including 2 min of gas flow and 8 min of steady-state analysis) as shown in Figure 3a. A decrease in light intensity with increasing oxygen concentration can be clearly seen, and is explainable by the absorption of the evanescent field by the MB sol–gel layer. The spectral response of the TCL fibre optic sensor with a 2 µm taper diameter and 0.86 µm coating thickness of MB sol–gel, over a wavelength range of 2 nm (central wavelength of 993.5 nm), to the above oxygen concentrations is shown in Figure 3b. Again, a reduction in light intensity as a result of increasing oxygen concentration is recorded. The observed sensor response to changing oxygen concentrations can also be observed in Figure 3a. A somewhat stochastic intensity variation can be noted during the first 2 min of any oxygen concentration change as a result of the dynamic environment inside the gas chamber due to gas flow. However, once the inlet and outlet valves are closed, the sensor shows a stable response. The maximum response time observed for this sensor is 4 min, excluding the 2 min of dynamic state in the gas chamber. Similar response times were observed for the other six sensors.

Figure 4 shows the change in intensity as a function of oxygen concentration (from 0% to 37.5% oxygen concentration by volume) for all seven sensor structures investigated. It can be observed that all sensors display a sensitivity to oxygen concentration within the range investigated, and do so by a linear profile. Furthermore, the intensity change for the fabricated tapered waist sensors is shown to be most sensitive to coating thickness, with intensity changes for thicker coatings greater than for thinner coatings exposed to comparable oxygen concentrations. A significant improvement in oxygen sensitivity can be observed for the MB sol–gel coating thickness of 0.86 µm, when the TCL sensor diameter is reduced from 125 µm to 2 µm. This is because a smaller taper diameter allows for a stronger evanescent field in the waist region, increasing the absorption by the MB sol–gel layer. However, TCL fibre optic sensors with taper waist diameters of 2 µm and 4 µm exhibit similar sensitivity to those with the same coating thickness, suggesting that a taper diameter below 4 µm yields little further improvement in oxygen sensitivity. This is explainable by considering that MB sol–gel reaches absorption saturation at this level of oxygen concentration. By contrast, and as mentioned, the absorption of the evanescent field increases with an increasing number of coating layers and thus coating thickness, thereby improving the oxygen sensitivity. The sensitivity of all six sensors was calculated by finding the best fit to the experimental data, as shown in Figure 4 and determining slope. The layer thickness of MB sol–gel is measured using SEM. Figure 5 shows the predicted sensitivity values as a function of coating thickness for a range of fibre sensor diameters.

Table 1. Experimentally calculated sensitivity for all the TCL optical fibre sensors with different coating layer thicknesses of MB sol–gel.

Sensor No.	Layer Thickness (µm)	Dia. of TNC Optical Fibre Sensor (µm)	Sensitivity (dB/O2%)
1	0.86	2	0.189
2	0.69		0.112
3	0.39		0.034
4	0.86	4	0.171
5	0.69		0.092
6	0.39		0.014
7	0.86	125	0.013

shows the experimentally calculated sensitivity values in dB/O₂% for all six TCL optical fibre sensors. It can be observed from Table 1 that, with an increase in coating thickness, the sensitivity of the proposed sensor improves. This is because a greater coating thickness enhances the interaction between the evanescent field and the surrounding MB sol–gel medium in the tapered waist section. Furthermore, a thicker coating leads to more pronounced absorption when oxygen concentrations are changed. It further increases the surface area for interaction and reduces optical loss, hence improving the sensitivity of the MB sol–gel-coated TCL optical fibre sensor. As shown in Figure 5, for the 2 µm diameter fibre sensor, the sensitivity increases linearly with coating thickness, demonstrating the potential of MB sol–gel coating thickness in oxygen measurements. In our experiments, a TCL fibre optic sensor with a taper diameter of 2 µm and an MB sol–gel coating thickness of 0.86 µm exhibits a maximum sensitivity of 0.189 dB/O₂% in the oxygen concentration range from 0% to 37.5% by volume.

As shown in Figure 6, introducing a 9% by volume of argon and ethanol into the gas chamber does not produce a significant change in intensity compared to the intensity change introduced by the same oxygen concentration. Therefore, the proposed sensor can be considered to have good selectivity for oxygen.

The TCL fibre optic sensor with a 2 µm taper waist diameter and 0.86 µm MB sol–gel coating thickness was exposed to ambient conditions for 7 days and re-tested in the same range of oxygen concentration. Figure 7 shows the good reproducibility of the proposed TCL fibre optic oxygen sensor, suggesting good stability and insignificant performance degradation over these time periods.

4. Conclusions

In this paper, an MB sol–gel-coated tapered coreless (TCL) optical fibre is proposed and experimentally studied for oxygen sensing. The taper waist diameter of the TCL fibre and the coating thickness of MB sol–gel are optimised to improve absorbance at 993.5 nm and increase sensitivity to oxygen concentration. The proposed fibre optic sensor showed no significant improvement in sensitivity when the taper waist diameter was reduced to less than 4 µm, but by increasing the coating thickness, a significant improvement in sensitivity was observed. The maximum sensitivity of 0.189 dB/O₂% was achieved using a TCL fibre with a diameter of 2 µm and an MB sol–gel coating thickness of 0.86 µm. The sensor has a response time of 4 min in a steady-state exposure with good selectivity for oxygen and provides reproducible results even after a period of 7 days in ambient conditions. Compared to fluorescence-based fibre optic sensors, where the optical fibre acts as a waveguide to carry fluorescence generated by the coating material, the proposed sensor uses a chromogenic approach and operates in the near-infrared (NIR) wavelength range. This is useful for detecting and quantifying oxygen in the environment with minimal effects of ambient light, especially in locations where visible light is attenuated. Silica-based optical fibres and sol–gel have shown high temperature resistance of ~1000 °C [38] and radiation resistance [39], making the proposed sensor suitable for monitoring oxygen levels in thermal and nuclear power plants. However, the lifetime of the proposed sensor in such high-temperature and radiative environments was outside the scope of this paper; therefore, periodic maintenance and replacement of the proposed sensor is recommended for such applications.