An optical sensor for monitoring of dissolved oxygen based on phase detection

A LCQ Fleet LC-MS System (Thermo Fisher Scientific) was used to collect mass spectra. Elemental analyses were obtained with a Vario MACRO cube element analyzer. ¹H NMR, ¹³C NMR spectra were taken on a Bruker spectrometer.

Dichloro(p-cymene)ruthenium(II)dimer, 4,7-diphenyl-1,10-phenanthroline, tetraethoxysilane (TEOS) and methyltriethoxysilane (MTEOS) were purchased from Aldrich. Ru(Ph₂phen)₃Cl₂ was synthesized and purified as described in the literature [12]. All other reagents and chemicals were from commercial sources and of analytical reagent grade, and used without further purification. Ultrapure water (Millipore, Bedford, MA, USA) was used throughout the experiment.

Sol–gel-derived silica films were fabricated from silicon alkoxide precursors, which undergo hydrolysis and poly-condensation reactions. Temperature controls the densification process [13]. If the temperature is less than 200 ° C, a non-densified porous glass matrix is formed. This microporous glass acts as a support matrix for the analyte-sensitive dyes containing ruthenium compounds that are later added to the silicon alkoxide solution. Dye molecules are entrapped into nanometer-scale cages formed by the cross-linking silicon and oxygen units. Smaller analyte molecules can permeate the matrix and access the dye complex within the pores. By selecting fabrication parameters appropriate to the size of the dopant, molecule leaching is negligible [14].

The process of film fabrication includes mixing a silicon alkoxide precursor with ethanol, water and hydrochloric acid. The precursors used for the sol–gel preparation in this work were TEOS and MTEOS. The ruthenium complex is added to the precursor solution, which is then stirred for 6 h at room temperature. The concentration of the ruthenium complex was 2–3 mg ml⁻¹ with respect to the precursor solution. After stirring, the sol was stored at 70 ° C for 12 h to promote hydrolysis and condensation polymerization. Then, the sol was dip-coated into a draft-free environment on these glass plates using a dip-coating apparatus. Before coating, these glass plates were soaked for 24 h in concentrated nitric acid, and then cleaned sequentially using de-ionized water, methanol and further de-ionized water. After dip-coating, the films were then finally dried for 18 h at 45 ° C [15].

2.3.1. Synthesis of the ruthenium complex.

2.3.2. Film surface modification and sensor performance.

For a dissolved oxygen sensor to be of practical value, a high over-all quenching response is necessary in order to attain good sensitivity at low oxygen concentrations. Quenching response is rooted in the combination of the low concentration oxygen in water and the hydrophilic nature of the sol–gel film surface. TEOS-based films have a high surface coverage of silanol (SiOH) groups [14], which facilitate water adsorption on the surface of the film, making it highly hydrophilic. However, a hydrophobic film surface is considered to be the best reinforcing DO quenching process. Fortunately, this can be achieved with the use of modified silica precursors. We used modified precursor, MTEOS, to replace the majority of the surface SiOH groups with SiCH₃ groups in the sol–gel film fabrication process. In order to optimize the ratio of MTEOS and TEOS, a systematic study was carried out using MTEOS in environments of 100% oxygen and 100% nitrogen. A number of sols were fabricated with increasing ratios of MTEOS to TEOS, ranging from sols with 100% TEOS to sols with 100% MTEOS. Films were fabricated as described earlier and their DO quenching responses in the gas phase were measured. As can be seen from table 1, increasing the ratio of MTEOS could cause the quenching response, Q_DO, to increase as expected.

Table 1.  Quenching response of TEOS and MTEOS films in the gas phase.

Sol–gel film	QDO (±2%) %
TEOS	27
MTEOS–TEOS (1:1)	48
MTEOS–TEOS (2:1)	79
MTEOS–TEOS (3:1)	81
MTEOS	81

Optical oxygen sensors based on both fluorescence intensity and phase fluorimetric measurements of Ru(Ph₂phen)₃Cl₂ have been reported previously [15]. In order to establish a baseline for the oxygen response within a real water sample, preliminary measurements were made within fully oxygenated and fully de-oxygenated water, respectively. The films used for these initial measurements were MTEOS–TEOS-based at pH 1.5 and R = 2. The response was characterized by repeatability, high signal-to-noise ratio and good reversibility. As shown in figure 2, the response time of the sensor was very short (<10 s), the time from the maximum value to half of the maximum value is about 8 s. The quenching response, Q_DO, of the sensing was 87% with a standard deviation of 2%, where the film response was averaged over four readings. The over-all quenching response of the film may be characterized by the following equation:

$$\begin{eqnarray*} &&{Q}_{\mathrm{DO}}=\frac{{I}_{\mathrm{deoxy}}-{I}_{\mathrm{oxy}}}{{I}_{\mathrm{deoxy}}} \end{eqnarray*}$$

where I_oxy and I_deoxy are the fluorescence signal from the film in fully oxygenated and fully de-oxygenated water, respectively. These results clearly indicated that the sensor material had a high sensitivity to dissolved oxygen in water.

A schematic of the system is shown in figure 1. The system contains three modules: the control module, the modulation module and the data processing module. The system uses a low power consumption MSP43f149 controller within the control module. The control module is used to control the emitting source and the reference source emission time sequence, and the modulation module is used to control the frequency of two emitting light sources. The typical frequency is modulated to 5 kHz in use. The data processing module is connected to the input measurement signal. The target signal and the reference signal are processed with a DFT algorithm and the phase of the two signals can be obtained. As the concentration of the oxygen is proportional to the phase difference, combined with the saved calibration slope/curve, the oxygen concentration can then be calculated using this module.

Zoom In Zoom Out Reset image size

Figure 2 is the schematic of the sensor probe itself. A compact probe configuration was used to be compatible with the requirements of different applications. The main part of the probe contains the following: control center, temperature sensor, optical detectors, filter, electrical cables, and two LEDs (blue LED and red LED). All the parts (including the three modules in figure 1) are integrated into a stainless steel enclosure. The length of the probe is about 12 cm and the volume of the control unit is about 15 cm × 10 cm × 8 cm (figure 3). The control center is the most important item, because it is integrated in the probe. All the active systems, including the modulation module and the data processing module are controlled by this unit. The emitting system comprises two LEDs, which are modulated at the same frequency. The only difference between the two LEDs is the central wavelength: the central wavelengths of the red and blue LEDs (both obtained from Production of Shenzhen Weirui, China) are 620 nm and 450 nm, respectively. The current of the two pulsed LEDs is about 10 mA.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

To ensure emission and detection efficiency, the divergence angle of the LEDs is limited to 5 mrad. To avoid the influence of environmental stray light and ensure a high detection accuracy, two techniques were used. First, the lights (two LEDs) were modulated at 5 kHz. Second, a narrow band filter (Shenyang Huibo, with the central wavelength 620 nm) was placed before the detector.

Figure 4 is the schematic of the electronics system. The two LEDs are modulated at 5 kHz square-wave modulation. The blue and red LEDs illuminate the sensor membrane at a fixed time interval. The detector receives the signal filtered by a narrow width filter effective at 620 nm. The controller is the center of the system, which is responsible for emission signal modulation, collection of temperature data and the calculation of the data. Figure 5 is the flow chart schematic for the concentration calculation. First, the blue LED is turned on (the state for the red LED is off) for 500 ms, while the controller carries out the calculation of the DFT algorithm. Then the blue LED is turned off and the red LED (the reference light source) is turned on for 500 ms, while the same DFT algorithm is executed in the controller. After data collection is complete, the phase difference can be calculated. At the same time, temperature data is collected for use as calibration data.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The solubility of oxygen in water and fluorescence films are both easily affected by the water temperature. For example, the dissolved oxygen in 0 ° C water (at the standard atmosphere pressure) is about 14.62 mg l⁻¹ and the dissolved oxygen in 22 ° C water is about 8.74 mg l⁻¹.⁴ The influence of temperature is characterized by recording the response of the sensor as a function of temperature. Then the temperature corrected calibration function was saved in the controller. When the sensor is inserted into a water sample, the effect of temperature variation can be eliminated by using the saved calibration information.

When the system is in an actual application, in situ calibration is necessary. The calibration process is shown as follows: first, the phase value in air is collected as the minimum value for calibration. Then, Na₂SO₃ is gradually added to the water sample until the solution is saturated. After waiting 20 min, the phase value in the saturated water is measured as the maximum value. The true value of the dissolved oxygen in air and in the saturated water is measured using a chemical iodometer in the laboratory. The value of the phase is a one-to-one correspondence to the true value of the linear relation as measured by the iodometric method. A two-point calibration method is used to get the linear relation between the phase value and the linear relation value.