Detection of Alcohol Content in Food Products by Lossy Mode Resonance Technique

To grow the ZnO nanorods on the middle core part of the optical fiber, a well-established technique named "Hydrothermal technique" was used. First of all, 0.878 g of zinc acetate was dissolved in 40 ml of ethanol then stirred at 50 °C for 90 min. On the other hand, 0.160 g of NaOH was added in 40 ml of ethanol then stirred at same temperature for 60 min. With the vigorous stirring, the second solution was added dropwise into first solution to prepare the solution for the growth of seed layer of ZnO. This solution was stirred on magnetic stirrer for 120 min at 70 °C. Optical fiber was dipped into the prepared solution to grow seed layer of ZnO on fiber core. Fiber was dipped for 15 min and then annealed for 15 min at 150 °C in the furnace in our laboratory. This process was repeated three times. For growth of ZnO nanorods, another sample was prepared. 0.0881 g HMTA and 0.1878 g zinc nitrate hexahydrate were mixed in 250 ml of deionized water and stirred at 100 °C for 8 h, 12 h, 16 h and 20 h vigorously for four different probes to approach the best results. Finally, ZnO nanorods grown fibers were rinsed in deionised water at 100 °C for 1 h without stirring. Then ZnO nanorods grown optical fiber probes were sent to MNIT, Jaipur for characterisation.

To deposit the bulk layer of ${{\rm{TiO}}}_{2}$ on the central core part of the optical fiber, a simple procedure was performed. 1.5 g of ${{\rm{TiO}}}_{2}$ powder was dissolved into 50 ml of deionized water. The solution was stirred for 1 h continuously at room temperature on magnetic stirrer. The optical fiber was dipped in the above solution to coat a layer of the solution on uncladded portion of the optical fiber. Coated layer was annealed at 150 °C in the tube furnace for 2 h. Finally, the fiber was rinsed with distilled water and kept it idle for three days in room temperature. A transparent layer of ${{\rm{TiO}}}_{2}$ was deposited on fiber core. For the confirmation of sample grown on the fiber, the probe was characterized using FESEM and EDS spectrometry.

Field Emission Scanning Electron Microscope (FESEM) use electrons for imaging, morphological study as the light microscopes use the visible light. It is a powerful and multipurpose tool for characterization of the materials. Figure 1a shows images of the ZnO nanorods grown on the fiber with nano size and micro size resolution respectively. These nanorods are of different dimensions and randomly oriented therefore they cannot be considered in localised mode. Diameter and length of these nanorods ranges between 200–300 nm and 2.5 μm respectively. Nanorods grown on fiber core are not hollow from inside. The image confirms solid nanorods with almost same length and diameter grown using hydrothermal technique. Similarly, Fig. 1b represents the ${{\rm{TiO}}}_{2}$ bulk layer deposited on the optical fiber core. Thickness of the deposited ${{\rm{TiO}}}_{2}$ layer is of the order of 50 μm. It shows the non-uniformity of thickness of bulk layer deposited and clarifies that no nanostructure is fabricated on the optical fiber core.

Zoom In Zoom Out Reset image size

Energy Dispersive Spectroscopy (EDS/EDX) is a technique used for the confirmation of the elements in the sample. Figure 1c represents the EDS image for ZnO nanorods grown on optical fiber. Presence of both Zn and O can be analysed by the corresponding peaks shown in the image. Si, Zn, O and Au are corresponding to the inner K, K, K and M shells respectively. These shells correspond to the electron emission shells due to incidence of X-rays. Weight % of Si, Zn, O and Au are observed as 39.13, 37.19, 14.10 and 9.58 respectively, with total weight % of 100. Atomic % of Si, O, Zn and Au were observed as 48.17, 19.66, 30.48 and 1.68 respectively, with a total atomic % of 100. On the other hand, Fig. 1d represents the EDS image for ${{\rm{TiO}}}_{2}$ bulk layer grown on uncladded portion of the optical fiber. Presence of both Ti and O can be analysed by the corresponding peaks shown in the image. All the elements Si, Ti, O and C show peaks corresponding to the K-shell. Weight % of Si, Ti, O and C are observed as 11.27, 33.99, 33.19 and 18.24 respectively, with total weight % of 100. Atomic % of Si, O, Zn and Au are observed 8.53, 15.09, 44.10 and 32.28 respectively, with a total atomic % of 100.

Initially, the experiment was performed for 10% concentration of ethanol and absorption spectrum was recorded to observe the absorption peak and the variation of absorbance with resonance wavelength. Maximum absorbance was observed at 491.042 nm corresponding to this sample as a sensing medium. After that, all the samples were used as sensing medium and inserted in the flow cell one by one with concentration difference of 10%. Absorption spectra for varying concentration of ethanol are shown in Fig. 4a. Shift in resonance clearly indicates the sensitivity of layer deposited on fiber with change in refractive index. For 10%, 20%, 30%, 40% and 50% concentrated ethanol, resonance wavelength was observed at 491.042 nm, 543.586 nm, 577.573 nm, 588.543 nm and 606.797 nm respectively. As the concentration of ethanol is increased, the red shift in resonance wavelength can be clearly observed. For 10% change in concentration, there is approximately 52.544 nm shift in resonance wavelength. From the sensitivity curve in Fig. 4c, sensitivity of ZnO NR grown optical fiber was calculated and it was found to be 28898.46 nm RIU⁻¹.

Zoom In Zoom Out Reset image size

Similar experiment was performed with different concentration of methanol. Absorption spectra for varying concentration of methanol are shown in Fig. 4b. Shift in resonance wavelength towards increasing wavelength can be observed with increase in concentration of methanol in DI water. Resonance wavelengths were observed as 551.927 nm, 593.525 nm, 607.460 nm, 628.323 nm and 645.836 nm corresponding to the 10%, 20%, 30%, 40% and 50% concentrated methanol samples respectively. For 10% change in concentration of methanol, there is nearly, 41.598 nm of maximum change in wavelength. From Fig. 4d, taking the slope of the sensitivity curve, sensitivity value of 23481.52 nm RIU⁻¹ was observed.

To check the durability and repeatability of the sensor, the experiment was repeated five times on the same set of the optical fiber probe at different times. A minor change in resonance wavelength was observed within the approximate limit of 0.2%. These changes in resonance wavelengths might be due to the environmental effects and coupling of light in optical fiber. Error bars in Figs. 4c, 4d depicts the small variation in the resonance shift and confirms the sustainability of the fiber probe.

Initially, the experiment for transmission spectra was performed with ${{\rm{TiO}}}_{2}$ deposited optical fiber probe and ethanol as a sensing region. Already prepared samples with varying concentration of ethanol are used as sensing region. Change in concentration of ethanol changes the refractive index of the sample. Transmission spectra were affected with the concentration of ethanol which can be observed from Fig. 5a with response time of 4 sec. Shift in resonance wavelength towards red end of the spectrum can be observed with increase in the concentration of ethanol in DI water. Resonance wavelengths were observed as 403.782 nm, 409.869 nm, 412.911 nm, 419.669 nm and 427.772 nm corresponding to 10%, 20%, 30%, 40% and 50% concentrated ethanol samples respectively. With 10% change in ethanol concentration, change in resonance wavelength of 8.103 nm can be observed and the sensitivity curve showing variation of resonance wavelength with varying concentration of ethanol is shown in Fig. 5c. From this curve, sensitivity of the sensor obtained was 6177.85 nm RIU⁻¹.

Zoom In Zoom Out Reset image size

For methanol as a sensing medium, similar experiment was performed to observe the transmission spectra for methanol. The prepared methanol concentrated samples were used as sensing region. Effect of change in concentration of methanol can be observed from the transmission spectra from Fig. 5b. Resonance wavelength shift towards increasing wavelength can be observed with increase in concentration of methanol in DI water. Resonance wavelengths can be observed as 415.615 nm, 428.785 nm, 430.135 nm, 434.183 nm and 450.026 nm corresponding to the 10%, 20%, 30%, 40% and 50% concentrated methanol samples respectively. Sensitivity curve is shown in Fig. 5d depicting the change in resonance wavelength of 13.17 nm with 10% change in methanol concentration. From this curve, sensitivity of the sensor was obtained to be 7962.88 nm RIU⁻¹ with response time of 4 s. This sensor becomes more effective due to its fast response as compared with the sensor with ${{\rm{TiO}}}_{2}$ -Hydroxyapatite nano-composites. ⁷⁶ Larger shift in resonance wavelength with concentration provides a ground for better detection capability of fiber probe. For comparison of the performance parameters of our sensor, experiment was performed with deionised water, 50% concentrated ethanol and 50% concentrated methanol as sensing region with refractive indices 1.3242, 1.3581 and 1.3352 at room temperature, respectively. Resonance wavelengths were observed at 407.164 nm,427.772 nm and 450.026 nm for DI water, ethanol and methanol respectively. This red shift in resonance wavelength can be seen from Fig. 5e, showing high response of probe 2 towards methanol as compared to ethanol and DI water. The experiment was repeated five times to check the repeatability of the optical fiber probe and the variation in resonance wavelength was observed approximately less than 0.3%. These minor changes in resonance wavelength may be due to environmental effects and coupling of light in optical fiber. Error bars in Figs. 5c and 5d depicts the small variation in resonance shift and confirms the sustainability of the fiber probe. Therefore, it reflects the high durability of the sensor which is a crucial requirement of any sensing probe.

We have calculated the resonance wavelength shift and sensitivity for performance evaluation of our sensor. Here, sensitivity is the ratio of change in resonance wavelength to the change in concentration of the sample used and is defined as,

$$\begin{eqnarray*}&&{\rm{Sensitivity}},S=\displaystyle \frac{{\rm{\Delta }}\lambda }{{\rm{\Delta }}C}\end{eqnarray*}$$

Here, ${\rm{\Delta }}C={C}_{2}-{C}_{1},$ ${C}_{1}$ and ${C}_{2}$represents the concentrations corresponding to the two samples and since the concentration is the function of refractive index as observed from calibration curves in Figs. 3a and 3b, we have used estimated sensitivity in terms of nm/RIU (refractive index unit), ${\rm{\Delta }}\lambda$ represents the change in resonance wavelength or the wavelengths corresponding to the position of LMR dips. Here, resonance shift reached to a value of 52.544 nm for 10% change in concentration for ZnO nanorod grown fiber probe and the sensitivity value is obtained 28898.46 nm RIU⁻¹. Similarly, for ${{\rm{TiO}}}_{2}$ bulk layer deposited fiber probe, resonance shift and sensitivity were obtained 13.17 nm and 7962.88 nm RIU⁻¹ for methanol as sensing region. The sensor response is compared with the ZnO nanowire grown gas sensor for ethanol detection. Sensitivity of ZnO nanorods grown fiber is approximately four time the sensitivity of the ZnO nanowire grown gas sensor as reported in Ref 77. For further evaluation of the quality of our LMR based sensor, the other parameters such as Quality factor (Q) and figure of Merit (FOM) have also been estimated.

Quality factor (Q) is the reciprocal of the central width of the LMR dip along the wavelength axis. This central width is termed as the Full Width at Half Maximum (FWHM). Numerically, higher value of Q- factor corresponds to the lower value of FWHM which gives a narrower curve. Quality factor is expressed as, ⁷⁸

$$\begin{eqnarray*}&&{\rm{Q}}=\displaystyle \frac{1}{{\rm{FWHM}}}\end{eqnarray*}$$

FWHM was measured from the LMR spectra by Gaussian fitting. Estimated value of FWHM for ZnO NRs grown probe (P1) is 168.365 nm for ethanol and 370.309 nm for methanol and for ${{\rm{TiO}}}_{2}$ coated probe (P2) is 115.176 nm and 100.484 nm for ethanol and methanol, respectively. Q factor is calculated by dividing the unit value by corresponding FWHM as expressed above and we got 5.939 μm⁻¹ 2.700 μm⁻¹ for probe P1 and 8.682 μm⁻¹ 9.951 μm⁻¹ for probe P₂ corresponding to ethanol and methanol respectively.

Ratio of sensitivity to FWHM is termed as figure of Merit (FOM) and is expressed as,

$$\begin{eqnarray*}&&{\rm{FOM}}=\displaystyle \frac{{\rm{S}}}{{\rm{FWHM}}}=S\times Q\end{eqnarray*}$$

Figure of Merit is the combined effect of sensitivity (S) and the quality factor (Q). Higher value of FOM corresponds to the lager sensitivity of the sensor and low value of FWHM. A high FOM means sensor will be more sensitive and precise in wavelength location. For probe 1, spectral analysis and calculations give the numerical FOM values as 171.64 RIU⁻¹ and 63.41 RIU⁻¹ for ethanol and methanol, respectively. On the other hand, FOM values for probe 2 were calculated as 53.63 RIU⁻¹ and 79.24 RIU⁻¹ for ethanol and methanol, respectively.