- Email: [email protected]
Optical fiber-based evanescent ammonia sensor
Sensors and Actuators B 110 (2005) 252–259
Optical fiber-based evanescent ammonia sensor Wenqing Cao, Yixiang Duan ∗ C-ACS, Los Alamos National Laboratory, MS K484, Los Alamos, NM 87545, USA Received 27 August 2004; received in revised form 19 January 2005; accepted 3 February 2005 Available online 2 March 2005
Abstract An optical fiber-based evanescent gaseous ammonia sensor is designed and developed. The sensing dye, bromocresol purple (BCP), is immobilized in the substitutional cladding using sol–gel process. The sensing properties of the optical fiber sensor to gaseous ammonia at room temperature are presented. This newly developed ammonia sensor exhibits good reversibility and repeatability. The effect of different carrier gases, argon, nitrogen, and air on sensing properties of the ammonia sensor is investigated. The sensor with air as carrier gas has the best response time and sensitivity. In order to improve the response time of the optical fiber evanescent ammonia sensor, an elevated ambient temperature is applied and thoroughly investigated. A fast response time of 10 s was obtained at 55.5 ◦ C with the carrier gas of air or argon. These experimental results have demonstrated that a fast response optical fiber evanescent gaseous ammonia sensor can be constructed by applying slightly elevated ambient temperature. Published by Elsevier B.V. Keywords: Ammonia; Sensors; Optical fiber; Evanescent field; Sol–gel
1. Introduction Ammonia has been widely used in the production of explosives, fertilizers, and as an industrial coolant [1]. The excess presence of ammonia in the atmosphere may create potential hazards to human being and ecosystems. Inhalation of only a small dose of ammonia vapor may cause acute poisoning to people. Threshold limit of ammonia concentration in air is only 25 ppm for human beings [2]. In addition, the fast, realtime quantification of atmospheric ammonia concentration is particularly important for environmental chemistry because ammonia is the most abundant alkaline component in the atmosphere. A substantial part of the acids generated in the atmosphere through sulfur dioxide (SO2 ) and nitrogen oxide (NOx ) can be neutralized by ammonia vapor (NH3 ). In such a case, ammonium ion (NH4 + ) is formed through this chemical reaction and counts as a major component of atmospheric aerosol that plays major roles in precipitation and nucleation process [3]. Due to an increasing environmental awareness ∗
Corresponding author. Tel.: +1 505 6659219; fax: +1 505 6655982. E-mail address: [email protected] (Y. Duan).
0925-4005/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.snb.2005.02.015
and pollution control, accurate detection and near real-time monitoring of gaseous ammonia have spread interests and wide applications in environmental science and occupational inspection. Traditionally, the detection of gaseous ammonia was performed by potentiometric electrodes within laboratory [4–10]. The advantages of the method are accurate, sensitive, and selective. However, it has significant limitations such as consumptive of the analyte, expensive and static instruments, and requiring the presence of an experienced operator. Another kind of ammonia sensing device, which was used for routine operations, is commercial infrared gas analyzers. The infrared devices, although sensitive, are nevertheless expensive and bulky [11]. Relative simple detectors based on semiconductivity of SnO2 [12] and MoO3 [13] thin films can detect gaseous ammonia, but these devices have some kind of restrictions in reproducibility, stability, sensitivity, selectivity, and a limited active sensor lifetime. In recent years, several new approaches have been reported for preparing optical gaseous ammonia sensors [1,14–32]. These devices utilize the reaction of ammonia vapor with either a pH-dependent dye material or a pH-sensitive film [32,33],
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
which undergoes a suitable color change or an absorption change. In general, these sensing mechanisms are based on monitoring the absorption or fluorescence characteristics of indicator dyes/sensing films entrapped within a membrane deposited onto a waveguiding substrate or an optical fiber as substituted cladding. The targeted ammonia molecules interact with the immobilized indicator, resulting in changes in absorbance or emission spectra, which are monitored using a proper detector module via an optical fiber or planar waveguide. Evanescent wave absorption is an effective technique for performing such analysis [1]. When a beam of light propagates along an optical fiber, the electromagnetic field does not abruptly fall to zero at the core/cladding interface. Instead, the overlap of the incoming beam and the internally reflected beam leads to a field that penetrates into the medium next to the core. This electromagnetic field, which tails but does not propagate into the second medium, is called the evanescent field. Its intensity I(z) decays exponentially with the distance z perpendicular to the interface as follows: z I(z) = I0 exp − (1) dp where I0 is the intensity of the incident radiation. The depth of penetration dp of the evanescent field is related to the angle of incidence θ at the interface, refractive indexes of core n1 and cladding n2 , and the wavelength of the radiation λ as follows [14,15,34]:
dp = 2π
λ n21 sin2 θ − n22
(2)
The evanescent field is able to interact with dye materials contained within the coating. This mechanism can be employed in rather ingenious way to develop fiber-optic evanescent gaseous ammonia sensors. In this kind of sensors, the parameter measured is the light intensity carried by the fiber that is modulated by the change in the absorption spectrum of the chromophore sensitive to NH3 . Optical fiber chemical sensors (OFCSs) have some distinctive characteristics. The small size and flexibility of the sensor design makes them ideal tools for in situ and in vivo analysis. Because optical fiber can easily transmit Chemically encoded information between the spectrometer and a remote sample, optical fiber sensors are particularly suitable for monitoring various environmental hazards including either hostile or not easily accessible events. In addition, optical fibers are relatively insensitive to sources of noise, such as radioactivity and electric fields, and signals acquired with optical fibers are much less prone to environmental interferences than those transmitted through electrical wires. Furthermore, optical fibers are able to channel a high density of information, such as wavelength, polarization, and phase. The ability to analyze each of these parameters enhances both the quality and quantity of chemical information obtained by OFCSs. However, conventional OFCSs based on dye indicators do have some limitations that should be recognized in
253
applications of sensors. First of all, almost all dyes are temperature sensitive in their response characteristic and most dyes have a critical temperature above which they tend to cease response completely and they may get irreversibly dissociated. Secondly, the dye indicators have their own chemical properties and natural lifetimes that usually decrease with increasing temperature, and their behaviors may be largely dependent upon the ambient optical conditions, particularly in the presence of ultraviolet light [35]. Because evanescent field reduces the amount of light impinging on the dye by many orders of magnitude, optical fiber sensors based on evanescent field can overcome this problem. In addition, if the chemical reactions involved in the sensing process are too slow or irreversible, the sensor cannot be used to provide a fast, continuous, and accurate response. These drawbacks are typical challenges and obstacles in optical fiber sensor development. Although a series of OFCSs based on dye indicators have been reported [1,11,16–18,24,26–28,36–41], to the best of our knowledge, so far there is no enough attention paid to overcome these limitations. For example, temperature effects on the sensing properties as well as on sensor response time have not been studied so far. Based on our previous experiences, temperature should be a critical parameter to either sensing films or dye indicator, as both are organic components with limited heat tolerance. However, on the other hand, temperature should also be critical for chemical reaction involved in the sensing process. To better understand sensing mechanism and improve optical sensor performance, in this paper, we have designed and developed an optical fiber-based evanescent wave sensor for gaseous ammonia sensing. Sol–gel film is used to encode bromocresol purple on the surface of a bared fiber core, and evanescent absorption is measured through a spectrometer. Gas sensing properties of the optical fiber evanescent sensor and ammonia responses versus concentration and temperature are reported. Meanwhile, the effects of carrier gases such as argon, nitrogen and air on signal intensity and response time are also thoroughly investigated and compared.
2. Experimental 2.1. Optical fiber preparation Plastic Clad Silica (PCS) multi-mode optical fiber with core diameter of 600 m was cut into 50 cm length and both ends were polished using 2000-grit and 3 um polishing film. The plastic jacket of the PCS fiber was stripped from the central portion of the fiber in 6 cm length. Since the evanescent tail has only a small fraction of power, our ammonia sensor was designed and constructed using a bent U-shape fiber probe to enhance sensitivity. The diameter of the semicircle of the U-shape structure is about 1.5 cm. The revealed cladding was removed by chemical etching through immersing the fiber in a 50% HF solution for 10 min.
254
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
2.2. Sol–gel film fabrication Sol–gel is an optically transparent glasslike material formed by hydrolysis and polymerization of metal alkoxides or metalorganic compounds at lower temperature. The silica glass formed by this method is a porous matrix that contains interconnected pores formed by a three-dimensional network of SiO2 . In the present investigation, tetraethyl ortho silicate (TEOS, Acros Organics, 99+% (gc)) was used as the precursor for the sol preparation since the refractive index of the porous silica film produced is less than that of the fiber core. The ammonia sensitive dye used in the present case is bromocresol purple (BCP, Acros Organics, Indicator Grade), as it is chemically more stable and is more resistant to oxidation than other indicators [17]. A magnetic stirrer was employed to mix the sol–gel reagents. The procedure of sol–gel film fabrication is as following: (1) 10 ml deionized water was acidified by 37% HCl to pH = 1. (2) 45 ml of TEOS was mixed with 55 ml of ethanol for 20 min. (3) 10 ml of the acidified water was added and mixing continued for 1 h. (4) 10 ml of ethanol containing 10 mg bromocresol purple was added, with stirring being continued for 1 h. (5) The uncladded surface of the PCS fiber was treated with 30% HNO3 for 10 min to activate the OH groups at the core surface. (6) The treated fiber was coated using dip coating procedure. (7) The coated fiber was dried overnight in an oven at 75 ◦ C and stored in ambient conditions away from direct sunlight for 1 month prior to characterization. This storage period ensures that post-fabrication structural evolution of the silica matrix is complete [42]. 2.3. Sensor set up Experimental set up for the sensing system is shown schematically in Fig. 1. A cylindrical chamber with inner diameter of 2.5 cm and thickness of 0.2 cm was customized. Gas inlet and outlet were positioned opposite as shown in
Fig. 1. The upside channel is used as gas inlet and the downside channel as outlet. Because the analyte gas, ammonia is lighter than any carrier gases used, this design will help the analyte gas evenly pass through the test chamber when ammonia gas was introduced from the up to down way. In this system, the mass flow rate (MFR) and concentration of the analyte gas were controlled using two FC-260 Mass Flow Controllers that were electronically controlled via the type 58784-channel power supply/readout unit. A 5000 ppm compressed ammonia gas with argon as balance gas was diluted by carrier gas such as ultra high purity argon or nitrogen. The test chamber was wrapped using an electrical resistance belt heater for temperature influence testing. The inner temperature of the chamber was calibrated and monitored using a digital thermometer. A LS-1 Tungsten Halogen Light Source and USB2000 Miniature Fiber Optic Spectrometer (Ocean Optics Inc.) were employed to characterize the absorption properties of the analyte gas.
3. Results and discussion 3.1. Sensing properties at room temperature The sensing properties of optical fiber-based evanescent gaseous ammonia sensor were investigated. A dark spectrum was first stored with the presence of argon carrier gas only, and then a sample spectrum was collected with the targeted ammonia gas introduced into the chamber. By subtracting the stored dark spectrum from sample spectrum, we got the pure absorption spectrum contributed solely from sample ammonia. Fig. 2 shows the absorption spectra obtained with different concentrations of ammonia from 145 to 3518 ppm. The carrier gas is argon and the mass flow rate was set at 1000 sccm. All absorption peaks for various concentrations of ammonia were recorded together for comparison purpose. It is clear from Fig. 2, the intensity of absorption increases with concentration and a maximum peak at wavelength of 596 nm was identified for all the concentrations tested. Fig. 3 shows the variation of absorption versus concentrations of ammonia gas at a fixed wavelength 596 nm. It is noticed from Fig. 3 that the sensitivity of the sensor system at lower con-
Fig. 1. Schematic of the experimental setup for optic fiber-based evanescent gaseous ammonia sensor.
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
255
Fig. 2. Absorption spectra for different concentration ammonia. Fig. 5. Response time, recovery time, and reversibility.
Fig. 3. Absorption versus concentrations of ammonia gas at 596 nm.
centration of ammonia is higher than that at higher concentrations. The response of the optical fiber evanescent sensor versus different concentrations of ammonia is plotted in Fig. 4 in logarithmic scale. It is obvious that the response of the sensor is logarithmic linear with the concentration of ammonia at least to 1000 ppm. This linear shape bends toward X-axial at concentrations higher than 1000 ppm. With a fixed dye concentration, a lower concentration of ammonia can completely react with the dye in the substitutional cladding. However, when the concentration of ammonia increases, the
Fig. 4. Calibration curve and linear dynamic range in logarithmic scale.
reaction between ammonia and dye would get gradually saturation, which leads to the linear shape bending over 1000 ppm concentration. This is probably the reason that the sensitivity at lower concentration of ammonia is higher than that at higher concentrations and the signal response versus concentration cannot maintain logarithmic linear relationship. The mass flow controllers used in our experiment can dilute gas in maximum of about 34.5 times. Since the concentration of the compressed ammonia gas is 5000 ppm, the lowest obtainable concentration in this experimental set up is about 145 ppm with the limitations by mass flow controllers. For this reason, concentrations lower than 145 ppm were not measured due to the experimental set up limitation. However, based on Fig. 4, we can estimate the detect limit of our ammonia sensor by extending the linear line to X-axial, where the cross point corresponding to the concentration should be the detection limit for the sensor. In this way, a detection limit of about 10 ppm is obtained. Therefore, the linear dynamic range of our sensor is estimated about two orders of magnitude from 10 to 1000 ppm. To investigate the response time, recovery time, and reversibility, the response of optical fiber sensor to 145 ppm ammonia was characterized and shown in Fig. 5 with argon as carrier gas at a flow rate of 1000 sccm. Based on the definition of response time that is determined by the time interval between the 10% and 90% of the stationary value [43], the response time of the optical sensor to analyte is about 5 min. In the same way, the recovery time is identified to be about 20 min. When the ammonia gas was introduced, the ammonia molecules diffuse into the inner layer of the sensing film where the dye complex can absorb the evanescent field. Because of the sensing mechanism, this type of sensors usually have relatively longer response and recovery time. Similar results concerning response and recovery time were also reported in the literature [16,17,25,31]. However, in this research, we found that this limitation can be somehow minimized through either optimization of the thickness of substitutional cladding or control of ambient temperature. When the cladding thickness is close to the depth of penetration dp of the evanescent field expressed in Eq. (2), the sensor has the
256
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
Fig. 6. Comparison of different carrier gases.
optimized response time and recovery time. However, when the cladding thickness decreases further, the system sensitivity will decrease due to less absorption of evanescent field by the colored dye complex. In addition, we found that the ambient temperature significantly affects the response time for this type of sensor. This issue will be further discussed in Section 3.3. As shown in Fig. 5, our ammonia sensor exhibits good reversibility and repeatability. 3.2. Effect of carrier gas Although different gases have been used as carrier gas for optical fiber ammonia sensor, the effect of different carrier gases on the sensing properties has not been taken into account in the ammonia sensor studies. In order to better understand the effect of carrier gas, the response of the sensor to 145 ppm ammonia vapor at the wavelength of 596 nm and room temperature (22.1 ◦ C) was characterized using various carrier gases, including Ar, N2 , and Air. The response spectra and recovery curves were compared in Fig. 6. It is clear that the response time and relative absorption change of the sensor vary with the different carrier gases. Interestingly, the baselines of the three different carrier gases are also different, in an order of Ar, N2 , and air from high to low. The response times of the sensor are 5, 3.3, and 2.7 min for the
carrier gases of Ar, N2 , and air, respectively. In consideration of baseline change when using different carrier gases, we have compared the sensor response in relatively absorption change rather than absolute amplitude. Amazingly, the sensor using air as carrier gas has the largest relative absorption change and the sensor using Ar as carrier gas has the smallest relative absorption change, although in the later case of Ar, a higher baseline is identified (Fig. 6). The reason for air to have a better response for ammonia measurement is not clear right now. However, a possible explanation in the view of reaction mechanism may help to understand the process. As it is well know that some high electro-affinity elements such as oxygen and nitrogen can form hydrogen bond with O H and N H structure, there is possibility that using nitrogen or air as carrier gas may help to stabilize the targeted ammonia molecules through forming hydrogen bond, which should have better interaction with the dye molecules immobilized on the sensing film. Since the dye molecule, BCP, has aromatic structure with three benzene rings that have double bonds and electrons which is mobile, it is helpful for interactions between the targeted ammonia molecule and the dye molecules when ammonia molecule forms a hydrogen bond with nitrogen or oxygen, since oxygen or nitrogen possesses either double or triple bonds and – interaction is easier than –σ interaction when ammonia molecule alone (without hydrogen bond). Careful analysis of Fig. 6, we also found that using air as carrier gas has better sensor response than using nitrogen. This phenomenon can also be explained based on the above mechanism as air has about twenty percent of oxygen and oxygen has higher electro-affinity than nitrogen does, which means a stronger hydrogen bond is formed with oxygen than with nitrogen. In this point of view, a better interaction between the analyte and the BCP molecules happens and results in a higher sensitivity in the sensor response. 3.3. Effect of temperature on the sensing properties As mentioned in the previous section, the dyes are temperature sensitive in their response characteristic. The effect of temperature on the sensing property of optical fiber sen-
Fig. 7. Temperature effect on sensor response in air: (a) T = 22.1, 35.7, and 55.5 ◦ C; (b) T = 55.5 and 80.0 ◦ C.
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
257
Fig. 8. Temperature effect on sensor response in Ar: (a) T = 22.1, 35.7, and 55.5 ◦ C; (b) T = 55.5 and 80.0 ◦ C.
sors has been investigated with the carrier gases of Ar and air. In the case of air as carrier gas, the response curves of the sensor to 145 ppm ammonia at temperatures of 22.1, 35.7, 55.5, and 80.0 ◦ C, are plotted in Fig. 7(a) and (b). When the temperature increase from 22.1 to 55.5 ◦ C, the response time reduces from 470 s to about 10 s quickly. After 55.5 ◦ C, the response time has no significant change with the temperature further increase up to 80.0 ◦ C. Very similar results shown in Fig. 8(a) and (b) were obtained with the carrier gas of argon. The variations of response times with the ambient temperatures for the carrier gases of argon and air were plotted in Fig. 9. Compared with the carrier gas of argon, the sensor using air as carrier gas has faster response time to ammonia in the temperature range from 22.1 to 55.5 ◦ C. In the higher temperature range from 55.5 to 80.0 ◦ C, both of the carrier gases have almost same response time of about 10 s to 145 ppm ammonia. The response times for the carrier gases of argon and air, respectively, were improved 47 and 33 times from room temperature of 22.1 ◦ C to an elevated temperature of 55.5 ◦ C. This result indicates that slightly increase of ambient temperature for the sensor can significantly improve the system performance, particularly the response time. The response time of the optical fiber evanescent ammonia sensor depends on the gas diffusion time from the surface of cladding to absorption layer of evanescent field and reaction time between the dye and ammonia. Even though the elevated temperature can reduce both of diffusion time and reaction
time, the reaction time dominates the reduction of response time in such a narrow variation of temperature from 22.1 to 55.5 ◦ C.
4. Conclusion We have designed and developed an optical fiber-based evanescent gaseous ammonia sensor. The dye, bromocresol purple (BCP) was immobilized in the substitutional cladding using sol–gel process. The sensing properties of the optical sensor to gaseous ammonia at room temperature were characterized. The response of the sensor is logarithmic linear with the concentration of ammonia under 1000 ppm. This newly developed ammonia sensor exhibits good reversibility and repeatability. The effect of different carrier gases, argon, nitrogen, and air on sensing properties of ammonia sensor was investigated. The sensor using air as carrier gas has the best response time and sensitivity. In order to improve the response time of optical fiber evanescent ammonia sensor, an elevated ambient temperature was applied and thoroughly investigated. A very fast response time of 10 s was obtained at 55.5 ◦ C with the carrier gas of air or argon. The response times for the carrier gases of argon and air was improved 47 and 33 times from room temperature 22.1 to 55.5 ◦ C, respectively. These experimental results have demonstrated that a fast response optical fiber evanescent gaseous ammonia sensor can be constructed by applying slightly elevated ambient temperature.
References
Fig. 9. Response time at different temperatures.
[1] C. Malins, A. Doyle, B.D. MacCraith, F. Kvasnik, M. Landl, P. Simon, L. Kalvoda, R. Lukas, K. Pufler, I. Babusik, Personal ammonia sensor for industrial environments, J. Environ. Monit. 1 (5) (1999) 417–422. [2] http://www.boc.com/capability/gases/Gases.pdf. [3] E.P. Felix, A.A. Cardoso, Colorimetric determination of ammonia in air using a hanging drop, Instrum. Sci. Technol. 31 (3) (2003) 283–294. [4] F. Bailescu, V.V. Cosofret, Applications of Ion-Selective Membrane Electrodes in Organic Analysis, Ellis Horwood, Chichester, 1978.
258
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
[5] W.E. Morf, E. Pungor, S.W. Incsedy (Eds.), The Principles of IonSelective Electrodes and Membrane Transport, Elsevier, Amsterdam, 1981, pp. 402–406. [6] M.E. Meyerhoff, Polymer membrane electrode-based potentiometric ammonia gas sensor, Anal. Chem. 52 (1980) 1532–1534. [7] Y.M. Fraticelli, M.E. Meyerhoff, Automated determination of ammonia with a potentiometric gas sensor and flowing internal electrolyte, Anal. Chem. 53 (1981) 992–997. [8] P. Buhlmann, E. Pretsch, E. Bakker, Carrier-based ion-selective electrodes and bulk optodes. 2. Ionophores for potentiomeric and optical sensors, Chem. Rev. 98 (1998) 1593–1687. [9] S.J. West, S. Ozawa, K. Seiler, S.S.S. Tan, W. Simon, Selective ionophore-based optical sensors for ammonia measurement in air, Anal. Chem. 64 (1992) 533–540. [10] A. Morales-Bahnik, R. Czolk, H.J. Ache, An optochemical ammonia sensor-based on immobilized metalloporphyrins, Sens. Actuators B: Chem. 19 (1994) 493–496. [11] P.S. Kumar, A.V. Scaria, C.P.G. Vallabhan, V.P.N. Nampoori, P. Radhakrishnan, Long-period grating in multimode fiber for ammonia gas detection, in: Proceedings of the SPIE, vol. 5279, No. 1, International Society for Optical Engineering, 2004, pp. 331–335. [12] T. Seiyama, A. Kato, K. Fujiishi, M. Nagatani, A new detector for gaseous components using semiconductive thin films, Anal. Chem. 34 (11) (1962) 1502–1503. [13] A.K. Prasad, P.I. Gouma, D.J. Kubinski, J.H. Visser, R.E. Soltis, P.J. Schmitz, Reactively sputtered MoO3 films for ammonia sensing, Thin Solid Films 436 (1) (2003) 46–51. [14] Y. Zaatar, D. Zaouk, J. Bechara, A. Khoury, C. Llinaress, J.P. Charles, Fabrication and characterization of an evanescent wave fiber optic sensor for air pollution control, Mater. Sci. Eng. B: Solid State Mater. Adv. Technol. 74 (1–3) (2000) 296–298. [15] S.Q. Tao, C.B. Winstead, R. Jindal, J.P. Singh, Optical-fiber sensor using tailored porous sol–gel fiber core, IEEE Sens. J. 4 (3) (2004) 322–328. [16] E. Scorsone, S. Christie, K.C. Persaud, P. Simon, F. Kvasnik, Fibreoptic evanescent sensing of gaseous ammonia with two forms of a new near-infrared dye in comparison to phenol red, Sens. Actuators B: Chem. 90 (1–3) (2003) 37–45. [17] P.S. Kumar, P.G. Vallabhan, V.P.N. Nampoori, P. Radhakrishnan, Fiber optic evanescent wave sensor for ammonia gas, in: Proceeding of the Society of Photo-Optical Instrumentation Engineers (SPIE), Parts 1 and 2, vol. 5280, 2004, pp. 617–621. [18] R.A. Potyrailo, G.M. Hieftje, Distributed fiber-optic chemical sensor with chemically modified plastic cladding, Appl. Spectrosc. 52 (8) (1998) 1092–1095. [19] V. Ruddy, An effective attenuation coefficient for evanescent wave spectroscopy using multimode fiber, Fiber Integrated Opt. 9 (2) (1990) 143–151. [20] S.H. Kim, G.S. Byun, S. Hong, Determination of ammonia in environmental samples using the diode laser/fiber optic colorimetric spectrometer, Opt. Eng. 39 (7) (2000) 1985–1988. [21] N. Matsuoka, S. Yamaguchi, K. Nanri, T. Fujioka, D. Richter, F.K. Tittel, Yb fiber laser pumped mid-IR source based on difference frequency generation and its application to ammonia detection, Jpn. J. Appl. Phys. Part 1 40 (2A) (2001) 625–628. [22] R. Claps, F.V. Englich, D.P. Leleux, D. Richter, F.K. Tittel, R.F. Curl, Ammonia detection by use of near-infrared diode-laserbased overtone spectroscopy, Appl. Opt. 40 (24) (2001) 4387– 4394. [23] I.M. Raimundo, R. Narayanaswamy, Simultaneous determination of relative humidity and ammonia in air employing an optical fibre sensor and artificial neural network, Sens. Actuators B: Chem. 74 (1–3) (2001) 60–68. [24] P. Simon, M. Landl, M. Breza, F. Kvasnik, New NIR dyes for ammonia sensing, Sens. Actuators B: Chem. 90 (1–3) (2003) 9–14. [25] E. Scorsone, S. Christie, K. Persaud, P. Simon, F. Kvasnik, Fibre optic evanescent sensing of gaseous ammonia with a near infrared
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33] [34]
[35] [36]
[37]
[38]
[39] [40]
[41]
[42]
[43]
dye for leak detection, in: Proceeding of The Society of PhotoOptical Instrumentation Engineers (SPIE), Parts 1 and 2, vol. 4829, 2003, pp. 978–979. P.S. Kumar, C.P.G. Vallabhan, V.P.N. Nampoori, P. Radhakrishnan, LED-based fiber optic evanescent wave ammonia sensor, in: Proceedings of the SPIE, vol. 4969, The International Society for Optical Engineering, 2003, pp. 166–173. C. Malins, T.M. Butler, B.D. MacCraith, Influence of the surface polarity of dye-doped sol–gel glass films on optical ammonia sensor response, Thin Solid Films 368 (1) (2000) 105–110. C. Malins, M. Landl, P. Simon, B.D. MacCraith, Fibre optic ammonia sensing employing novel near infrared dyes, Sens. Actuators B: Chem. B51 (1–3) (1998) 359–367. T. Grady, T. Butler, D. MacCraith, D. Diamond, M.A. McKervey, Optical sensor for gaseous ammonia with tuneable sensitivity, Analyst 122 (8) (1997) 803–806. A. Messica, A. Greenstein, A. Katzir, U. Schiessl, M. Tacke, Fiberoptic evanescent wave sensor for gas detection, Opt. Lett. 19 (15) (1994) 1167–1169. S. Christie, E. Scorsone, K. Persaud, F. Kvasnik, Remote detection of gaseous ammonia using the near infrared transmission properties of polyaniline, Sens. Actuators B: Chem. 90 (1–3) (2003) 163– 169. Z. Jin, Y.X. Su, Y.X. Duan, Development of a polyanilinebased optical ammonia sensor, Sens. Actuators B 72 (2001) 75– 79. Z. Jin, Y.X. Su, Y.X. Duan, An improved optical pH sensor based on polyaniline, Sens. Actuators B 71 (12) (2000) 118–122. G. Stewart, W. Johnstone, Evanescently coupled components, in: J. Dakin, B. Culshaw (Eds.), Optical Fiber Sensors Volume Three: Components and Subsystems, Artech House, Boston, 1988–1997, pp. 69–101. B. Culshaw, Optical fiber sensor technologies: opportunities andperhaps-pitfalls, J. Lightwave Technol. 22 (1) (2004) 39–50. T. Yagi, N. Kuboki, Y. Suzuki, N. Uchino, K. Nakamura, K. Yoshida, Fiber-optic ammonia sensors utilizing rectangular-cladding eccentriccore fiber, Opt. Rev. 4 (5) (1997) 596–600. P.K. Choudhury, T. Yoshino, On the pH response of fiber optic evanescent field absorption sensor having a U-shaped probe: an experimental analysis, Optik 114 (1) (2003) 13–18. R.A. Potyrailo, L.A. Mikheenko, P.S. Borsuk, S.P. Golubkov, P.M. Talanchuk, Portable photometric ammonia gas analyser using pHsensitive polymer dye-films: spectral optimization of performance under field operating conditions, Sens. Actuators B: Chem. 21 (1994) 65–70. J. Lin, Recent development and applications of optical and fiber-optic pH sensors, Trends Anal. Chem. 19 (9) (2000) 541–552. P. Simon, S. Sekretar, B.D. MacCraith, F. Kvasnik, Near-infrared reagents for fibre-optic ammonia sensors, Sens. Actuators B: Chem. B39 (1–3) (1997) 252–255. S.K. Spear, S.L. Patterson, M.A. Arnold, Flow-through fiber-optic ammonia sensor for analysis of hippocampus slice perfusates, Anal. Chim. Acta 357 (1–2) (1997) 79–84. C. McDonagh, F. Sheridan, T. Butler, B.D. MacCraith, Characterisation of sol–gel-derived silica films, J. Non-Cryst. Solids 194 (1–2) (1996) 72–77. A. D’Amico, C. Di Natale, A. Taroni, Sensors parameters, sensors for domestic applications, in: Proceedings of the First European School on Sensors (ESS’94), Castro Marina, Leee, Italy, September 12–17, 1994, pp. 3–13.
Biographies Wenqing Cao received the BSc and MSc degrees from Xidian University, China, in 1986 and 1992, respectively, and the PhD from Nanyang
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259 Technological University, Singapore, in 2002. He is currently a postdoc with Los Alamos National Laboratory, USA. His research interests are in optical fiber chemical sensors, semiconductor oxide-based gas sensors, MEMS/MOEMS and their packaging. Yixiang Duan received his BS from Fudan University in 1982 and MS from Chinese Academy of Sciences in 1988. In 1994, he received his
259
PhD jointly from Jilin University/Indiana University. He is currently a staff scientist and principal investigator in Los Alamos National Laboratory. His research interests include optical fiber sensors, chemical and biological sensors, handheld detectors and field portable instruments.
Optical fiber-based evanescent ammonia sensor Wenqing Cao, Yixiang Duan ∗ C-ACS, Los Alamos National Laboratory, MS K484, Los Alamos, NM 87545, USA Received 27 August 2004; received in revised form 19 January 2005; accepted 3 February 2005 Available online 2 March 2005
Abstract An optical fiber-based evanescent gaseous ammonia sensor is designed and developed. The sensing dye, bromocresol purple (BCP), is immobilized in the substitutional cladding using sol–gel process. The sensing properties of the optical fiber sensor to gaseous ammonia at room temperature are presented. This newly developed ammonia sensor exhibits good reversibility and repeatability. The effect of different carrier gases, argon, nitrogen, and air on sensing properties of the ammonia sensor is investigated. The sensor with air as carrier gas has the best response time and sensitivity. In order to improve the response time of the optical fiber evanescent ammonia sensor, an elevated ambient temperature is applied and thoroughly investigated. A fast response time of 10 s was obtained at 55.5 ◦ C with the carrier gas of air or argon. These experimental results have demonstrated that a fast response optical fiber evanescent gaseous ammonia sensor can be constructed by applying slightly elevated ambient temperature. Published by Elsevier B.V. Keywords: Ammonia; Sensors; Optical fiber; Evanescent field; Sol–gel
1. Introduction Ammonia has been widely used in the production of explosives, fertilizers, and as an industrial coolant [1]. The excess presence of ammonia in the atmosphere may create potential hazards to human being and ecosystems. Inhalation of only a small dose of ammonia vapor may cause acute poisoning to people. Threshold limit of ammonia concentration in air is only 25 ppm for human beings [2]. In addition, the fast, realtime quantification of atmospheric ammonia concentration is particularly important for environmental chemistry because ammonia is the most abundant alkaline component in the atmosphere. A substantial part of the acids generated in the atmosphere through sulfur dioxide (SO2 ) and nitrogen oxide (NOx ) can be neutralized by ammonia vapor (NH3 ). In such a case, ammonium ion (NH4 + ) is formed through this chemical reaction and counts as a major component of atmospheric aerosol that plays major roles in precipitation and nucleation process [3]. Due to an increasing environmental awareness ∗
Corresponding author. Tel.: +1 505 6659219; fax: +1 505 6655982. E-mail address: [email protected] (Y. Duan).
0925-4005/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.snb.2005.02.015
and pollution control, accurate detection and near real-time monitoring of gaseous ammonia have spread interests and wide applications in environmental science and occupational inspection. Traditionally, the detection of gaseous ammonia was performed by potentiometric electrodes within laboratory [4–10]. The advantages of the method are accurate, sensitive, and selective. However, it has significant limitations such as consumptive of the analyte, expensive and static instruments, and requiring the presence of an experienced operator. Another kind of ammonia sensing device, which was used for routine operations, is commercial infrared gas analyzers. The infrared devices, although sensitive, are nevertheless expensive and bulky [11]. Relative simple detectors based on semiconductivity of SnO2 [12] and MoO3 [13] thin films can detect gaseous ammonia, but these devices have some kind of restrictions in reproducibility, stability, sensitivity, selectivity, and a limited active sensor lifetime. In recent years, several new approaches have been reported for preparing optical gaseous ammonia sensors [1,14–32]. These devices utilize the reaction of ammonia vapor with either a pH-dependent dye material or a pH-sensitive film [32,33],
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
which undergoes a suitable color change or an absorption change. In general, these sensing mechanisms are based on monitoring the absorption or fluorescence characteristics of indicator dyes/sensing films entrapped within a membrane deposited onto a waveguiding substrate or an optical fiber as substituted cladding. The targeted ammonia molecules interact with the immobilized indicator, resulting in changes in absorbance or emission spectra, which are monitored using a proper detector module via an optical fiber or planar waveguide. Evanescent wave absorption is an effective technique for performing such analysis [1]. When a beam of light propagates along an optical fiber, the electromagnetic field does not abruptly fall to zero at the core/cladding interface. Instead, the overlap of the incoming beam and the internally reflected beam leads to a field that penetrates into the medium next to the core. This electromagnetic field, which tails but does not propagate into the second medium, is called the evanescent field. Its intensity I(z) decays exponentially with the distance z perpendicular to the interface as follows: z I(z) = I0 exp − (1) dp where I0 is the intensity of the incident radiation. The depth of penetration dp of the evanescent field is related to the angle of incidence θ at the interface, refractive indexes of core n1 and cladding n2 , and the wavelength of the radiation λ as follows [14,15,34]:
dp = 2π
λ n21 sin2 θ − n22
(2)
The evanescent field is able to interact with dye materials contained within the coating. This mechanism can be employed in rather ingenious way to develop fiber-optic evanescent gaseous ammonia sensors. In this kind of sensors, the parameter measured is the light intensity carried by the fiber that is modulated by the change in the absorption spectrum of the chromophore sensitive to NH3 . Optical fiber chemical sensors (OFCSs) have some distinctive characteristics. The small size and flexibility of the sensor design makes them ideal tools for in situ and in vivo analysis. Because optical fiber can easily transmit Chemically encoded information between the spectrometer and a remote sample, optical fiber sensors are particularly suitable for monitoring various environmental hazards including either hostile or not easily accessible events. In addition, optical fibers are relatively insensitive to sources of noise, such as radioactivity and electric fields, and signals acquired with optical fibers are much less prone to environmental interferences than those transmitted through electrical wires. Furthermore, optical fibers are able to channel a high density of information, such as wavelength, polarization, and phase. The ability to analyze each of these parameters enhances both the quality and quantity of chemical information obtained by OFCSs. However, conventional OFCSs based on dye indicators do have some limitations that should be recognized in
253
applications of sensors. First of all, almost all dyes are temperature sensitive in their response characteristic and most dyes have a critical temperature above which they tend to cease response completely and they may get irreversibly dissociated. Secondly, the dye indicators have their own chemical properties and natural lifetimes that usually decrease with increasing temperature, and their behaviors may be largely dependent upon the ambient optical conditions, particularly in the presence of ultraviolet light [35]. Because evanescent field reduces the amount of light impinging on the dye by many orders of magnitude, optical fiber sensors based on evanescent field can overcome this problem. In addition, if the chemical reactions involved in the sensing process are too slow or irreversible, the sensor cannot be used to provide a fast, continuous, and accurate response. These drawbacks are typical challenges and obstacles in optical fiber sensor development. Although a series of OFCSs based on dye indicators have been reported [1,11,16–18,24,26–28,36–41], to the best of our knowledge, so far there is no enough attention paid to overcome these limitations. For example, temperature effects on the sensing properties as well as on sensor response time have not been studied so far. Based on our previous experiences, temperature should be a critical parameter to either sensing films or dye indicator, as both are organic components with limited heat tolerance. However, on the other hand, temperature should also be critical for chemical reaction involved in the sensing process. To better understand sensing mechanism and improve optical sensor performance, in this paper, we have designed and developed an optical fiber-based evanescent wave sensor for gaseous ammonia sensing. Sol–gel film is used to encode bromocresol purple on the surface of a bared fiber core, and evanescent absorption is measured through a spectrometer. Gas sensing properties of the optical fiber evanescent sensor and ammonia responses versus concentration and temperature are reported. Meanwhile, the effects of carrier gases such as argon, nitrogen and air on signal intensity and response time are also thoroughly investigated and compared.
2. Experimental 2.1. Optical fiber preparation Plastic Clad Silica (PCS) multi-mode optical fiber with core diameter of 600 m was cut into 50 cm length and both ends were polished using 2000-grit and 3 um polishing film. The plastic jacket of the PCS fiber was stripped from the central portion of the fiber in 6 cm length. Since the evanescent tail has only a small fraction of power, our ammonia sensor was designed and constructed using a bent U-shape fiber probe to enhance sensitivity. The diameter of the semicircle of the U-shape structure is about 1.5 cm. The revealed cladding was removed by chemical etching through immersing the fiber in a 50% HF solution for 10 min.
254
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
2.2. Sol–gel film fabrication Sol–gel is an optically transparent glasslike material formed by hydrolysis and polymerization of metal alkoxides or metalorganic compounds at lower temperature. The silica glass formed by this method is a porous matrix that contains interconnected pores formed by a three-dimensional network of SiO2 . In the present investigation, tetraethyl ortho silicate (TEOS, Acros Organics, 99+% (gc)) was used as the precursor for the sol preparation since the refractive index of the porous silica film produced is less than that of the fiber core. The ammonia sensitive dye used in the present case is bromocresol purple (BCP, Acros Organics, Indicator Grade), as it is chemically more stable and is more resistant to oxidation than other indicators [17]. A magnetic stirrer was employed to mix the sol–gel reagents. The procedure of sol–gel film fabrication is as following: (1) 10 ml deionized water was acidified by 37% HCl to pH = 1. (2) 45 ml of TEOS was mixed with 55 ml of ethanol for 20 min. (3) 10 ml of the acidified water was added and mixing continued for 1 h. (4) 10 ml of ethanol containing 10 mg bromocresol purple was added, with stirring being continued for 1 h. (5) The uncladded surface of the PCS fiber was treated with 30% HNO3 for 10 min to activate the OH groups at the core surface. (6) The treated fiber was coated using dip coating procedure. (7) The coated fiber was dried overnight in an oven at 75 ◦ C and stored in ambient conditions away from direct sunlight for 1 month prior to characterization. This storage period ensures that post-fabrication structural evolution of the silica matrix is complete [42]. 2.3. Sensor set up Experimental set up for the sensing system is shown schematically in Fig. 1. A cylindrical chamber with inner diameter of 2.5 cm and thickness of 0.2 cm was customized. Gas inlet and outlet were positioned opposite as shown in
Fig. 1. The upside channel is used as gas inlet and the downside channel as outlet. Because the analyte gas, ammonia is lighter than any carrier gases used, this design will help the analyte gas evenly pass through the test chamber when ammonia gas was introduced from the up to down way. In this system, the mass flow rate (MFR) and concentration of the analyte gas were controlled using two FC-260 Mass Flow Controllers that were electronically controlled via the type 58784-channel power supply/readout unit. A 5000 ppm compressed ammonia gas with argon as balance gas was diluted by carrier gas such as ultra high purity argon or nitrogen. The test chamber was wrapped using an electrical resistance belt heater for temperature influence testing. The inner temperature of the chamber was calibrated and monitored using a digital thermometer. A LS-1 Tungsten Halogen Light Source and USB2000 Miniature Fiber Optic Spectrometer (Ocean Optics Inc.) were employed to characterize the absorption properties of the analyte gas.
3. Results and discussion 3.1. Sensing properties at room temperature The sensing properties of optical fiber-based evanescent gaseous ammonia sensor were investigated. A dark spectrum was first stored with the presence of argon carrier gas only, and then a sample spectrum was collected with the targeted ammonia gas introduced into the chamber. By subtracting the stored dark spectrum from sample spectrum, we got the pure absorption spectrum contributed solely from sample ammonia. Fig. 2 shows the absorption spectra obtained with different concentrations of ammonia from 145 to 3518 ppm. The carrier gas is argon and the mass flow rate was set at 1000 sccm. All absorption peaks for various concentrations of ammonia were recorded together for comparison purpose. It is clear from Fig. 2, the intensity of absorption increases with concentration and a maximum peak at wavelength of 596 nm was identified for all the concentrations tested. Fig. 3 shows the variation of absorption versus concentrations of ammonia gas at a fixed wavelength 596 nm. It is noticed from Fig. 3 that the sensitivity of the sensor system at lower con-
Fig. 1. Schematic of the experimental setup for optic fiber-based evanescent gaseous ammonia sensor.
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
255
Fig. 2. Absorption spectra for different concentration ammonia. Fig. 5. Response time, recovery time, and reversibility.
Fig. 3. Absorption versus concentrations of ammonia gas at 596 nm.
centration of ammonia is higher than that at higher concentrations. The response of the optical fiber evanescent sensor versus different concentrations of ammonia is plotted in Fig. 4 in logarithmic scale. It is obvious that the response of the sensor is logarithmic linear with the concentration of ammonia at least to 1000 ppm. This linear shape bends toward X-axial at concentrations higher than 1000 ppm. With a fixed dye concentration, a lower concentration of ammonia can completely react with the dye in the substitutional cladding. However, when the concentration of ammonia increases, the
Fig. 4. Calibration curve and linear dynamic range in logarithmic scale.
reaction between ammonia and dye would get gradually saturation, which leads to the linear shape bending over 1000 ppm concentration. This is probably the reason that the sensitivity at lower concentration of ammonia is higher than that at higher concentrations and the signal response versus concentration cannot maintain logarithmic linear relationship. The mass flow controllers used in our experiment can dilute gas in maximum of about 34.5 times. Since the concentration of the compressed ammonia gas is 5000 ppm, the lowest obtainable concentration in this experimental set up is about 145 ppm with the limitations by mass flow controllers. For this reason, concentrations lower than 145 ppm were not measured due to the experimental set up limitation. However, based on Fig. 4, we can estimate the detect limit of our ammonia sensor by extending the linear line to X-axial, where the cross point corresponding to the concentration should be the detection limit for the sensor. In this way, a detection limit of about 10 ppm is obtained. Therefore, the linear dynamic range of our sensor is estimated about two orders of magnitude from 10 to 1000 ppm. To investigate the response time, recovery time, and reversibility, the response of optical fiber sensor to 145 ppm ammonia was characterized and shown in Fig. 5 with argon as carrier gas at a flow rate of 1000 sccm. Based on the definition of response time that is determined by the time interval between the 10% and 90% of the stationary value [43], the response time of the optical sensor to analyte is about 5 min. In the same way, the recovery time is identified to be about 20 min. When the ammonia gas was introduced, the ammonia molecules diffuse into the inner layer of the sensing film where the dye complex can absorb the evanescent field. Because of the sensing mechanism, this type of sensors usually have relatively longer response and recovery time. Similar results concerning response and recovery time were also reported in the literature [16,17,25,31]. However, in this research, we found that this limitation can be somehow minimized through either optimization of the thickness of substitutional cladding or control of ambient temperature. When the cladding thickness is close to the depth of penetration dp of the evanescent field expressed in Eq. (2), the sensor has the
256
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
Fig. 6. Comparison of different carrier gases.
optimized response time and recovery time. However, when the cladding thickness decreases further, the system sensitivity will decrease due to less absorption of evanescent field by the colored dye complex. In addition, we found that the ambient temperature significantly affects the response time for this type of sensor. This issue will be further discussed in Section 3.3. As shown in Fig. 5, our ammonia sensor exhibits good reversibility and repeatability. 3.2. Effect of carrier gas Although different gases have been used as carrier gas for optical fiber ammonia sensor, the effect of different carrier gases on the sensing properties has not been taken into account in the ammonia sensor studies. In order to better understand the effect of carrier gas, the response of the sensor to 145 ppm ammonia vapor at the wavelength of 596 nm and room temperature (22.1 ◦ C) was characterized using various carrier gases, including Ar, N2 , and Air. The response spectra and recovery curves were compared in Fig. 6. It is clear that the response time and relative absorption change of the sensor vary with the different carrier gases. Interestingly, the baselines of the three different carrier gases are also different, in an order of Ar, N2 , and air from high to low. The response times of the sensor are 5, 3.3, and 2.7 min for the
carrier gases of Ar, N2 , and air, respectively. In consideration of baseline change when using different carrier gases, we have compared the sensor response in relatively absorption change rather than absolute amplitude. Amazingly, the sensor using air as carrier gas has the largest relative absorption change and the sensor using Ar as carrier gas has the smallest relative absorption change, although in the later case of Ar, a higher baseline is identified (Fig. 6). The reason for air to have a better response for ammonia measurement is not clear right now. However, a possible explanation in the view of reaction mechanism may help to understand the process. As it is well know that some high electro-affinity elements such as oxygen and nitrogen can form hydrogen bond with O H and N H structure, there is possibility that using nitrogen or air as carrier gas may help to stabilize the targeted ammonia molecules through forming hydrogen bond, which should have better interaction with the dye molecules immobilized on the sensing film. Since the dye molecule, BCP, has aromatic structure with three benzene rings that have double bonds and electrons which is mobile, it is helpful for interactions between the targeted ammonia molecule and the dye molecules when ammonia molecule forms a hydrogen bond with nitrogen or oxygen, since oxygen or nitrogen possesses either double or triple bonds and – interaction is easier than –σ interaction when ammonia molecule alone (without hydrogen bond). Careful analysis of Fig. 6, we also found that using air as carrier gas has better sensor response than using nitrogen. This phenomenon can also be explained based on the above mechanism as air has about twenty percent of oxygen and oxygen has higher electro-affinity than nitrogen does, which means a stronger hydrogen bond is formed with oxygen than with nitrogen. In this point of view, a better interaction between the analyte and the BCP molecules happens and results in a higher sensitivity in the sensor response. 3.3. Effect of temperature on the sensing properties As mentioned in the previous section, the dyes are temperature sensitive in their response characteristic. The effect of temperature on the sensing property of optical fiber sen-
Fig. 7. Temperature effect on sensor response in air: (a) T = 22.1, 35.7, and 55.5 ◦ C; (b) T = 55.5 and 80.0 ◦ C.
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
257
Fig. 8. Temperature effect on sensor response in Ar: (a) T = 22.1, 35.7, and 55.5 ◦ C; (b) T = 55.5 and 80.0 ◦ C.
sors has been investigated with the carrier gases of Ar and air. In the case of air as carrier gas, the response curves of the sensor to 145 ppm ammonia at temperatures of 22.1, 35.7, 55.5, and 80.0 ◦ C, are plotted in Fig. 7(a) and (b). When the temperature increase from 22.1 to 55.5 ◦ C, the response time reduces from 470 s to about 10 s quickly. After 55.5 ◦ C, the response time has no significant change with the temperature further increase up to 80.0 ◦ C. Very similar results shown in Fig. 8(a) and (b) were obtained with the carrier gas of argon. The variations of response times with the ambient temperatures for the carrier gases of argon and air were plotted in Fig. 9. Compared with the carrier gas of argon, the sensor using air as carrier gas has faster response time to ammonia in the temperature range from 22.1 to 55.5 ◦ C. In the higher temperature range from 55.5 to 80.0 ◦ C, both of the carrier gases have almost same response time of about 10 s to 145 ppm ammonia. The response times for the carrier gases of argon and air, respectively, were improved 47 and 33 times from room temperature of 22.1 ◦ C to an elevated temperature of 55.5 ◦ C. This result indicates that slightly increase of ambient temperature for the sensor can significantly improve the system performance, particularly the response time. The response time of the optical fiber evanescent ammonia sensor depends on the gas diffusion time from the surface of cladding to absorption layer of evanescent field and reaction time between the dye and ammonia. Even though the elevated temperature can reduce both of diffusion time and reaction
time, the reaction time dominates the reduction of response time in such a narrow variation of temperature from 22.1 to 55.5 ◦ C.
4. Conclusion We have designed and developed an optical fiber-based evanescent gaseous ammonia sensor. The dye, bromocresol purple (BCP) was immobilized in the substitutional cladding using sol–gel process. The sensing properties of the optical sensor to gaseous ammonia at room temperature were characterized. The response of the sensor is logarithmic linear with the concentration of ammonia under 1000 ppm. This newly developed ammonia sensor exhibits good reversibility and repeatability. The effect of different carrier gases, argon, nitrogen, and air on sensing properties of ammonia sensor was investigated. The sensor using air as carrier gas has the best response time and sensitivity. In order to improve the response time of optical fiber evanescent ammonia sensor, an elevated ambient temperature was applied and thoroughly investigated. A very fast response time of 10 s was obtained at 55.5 ◦ C with the carrier gas of air or argon. The response times for the carrier gases of argon and air was improved 47 and 33 times from room temperature 22.1 to 55.5 ◦ C, respectively. These experimental results have demonstrated that a fast response optical fiber evanescent gaseous ammonia sensor can be constructed by applying slightly elevated ambient temperature.
References
Fig. 9. Response time at different temperatures.
[1] C. Malins, A. Doyle, B.D. MacCraith, F. Kvasnik, M. Landl, P. Simon, L. Kalvoda, R. Lukas, K. Pufler, I. Babusik, Personal ammonia sensor for industrial environments, J. Environ. Monit. 1 (5) (1999) 417–422. [2] http://www.boc.com/capability/gases/Gases.pdf. [3] E.P. Felix, A.A. Cardoso, Colorimetric determination of ammonia in air using a hanging drop, Instrum. Sci. Technol. 31 (3) (2003) 283–294. [4] F. Bailescu, V.V. Cosofret, Applications of Ion-Selective Membrane Electrodes in Organic Analysis, Ellis Horwood, Chichester, 1978.
258
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259
[5] W.E. Morf, E. Pungor, S.W. Incsedy (Eds.), The Principles of IonSelective Electrodes and Membrane Transport, Elsevier, Amsterdam, 1981, pp. 402–406. [6] M.E. Meyerhoff, Polymer membrane electrode-based potentiometric ammonia gas sensor, Anal. Chem. 52 (1980) 1532–1534. [7] Y.M. Fraticelli, M.E. Meyerhoff, Automated determination of ammonia with a potentiometric gas sensor and flowing internal electrolyte, Anal. Chem. 53 (1981) 992–997. [8] P. Buhlmann, E. Pretsch, E. Bakker, Carrier-based ion-selective electrodes and bulk optodes. 2. Ionophores for potentiomeric and optical sensors, Chem. Rev. 98 (1998) 1593–1687. [9] S.J. West, S. Ozawa, K. Seiler, S.S.S. Tan, W. Simon, Selective ionophore-based optical sensors for ammonia measurement in air, Anal. Chem. 64 (1992) 533–540. [10] A. Morales-Bahnik, R. Czolk, H.J. Ache, An optochemical ammonia sensor-based on immobilized metalloporphyrins, Sens. Actuators B: Chem. 19 (1994) 493–496. [11] P.S. Kumar, A.V. Scaria, C.P.G. Vallabhan, V.P.N. Nampoori, P. Radhakrishnan, Long-period grating in multimode fiber for ammonia gas detection, in: Proceedings of the SPIE, vol. 5279, No. 1, International Society for Optical Engineering, 2004, pp. 331–335. [12] T. Seiyama, A. Kato, K. Fujiishi, M. Nagatani, A new detector for gaseous components using semiconductive thin films, Anal. Chem. 34 (11) (1962) 1502–1503. [13] A.K. Prasad, P.I. Gouma, D.J. Kubinski, J.H. Visser, R.E. Soltis, P.J. Schmitz, Reactively sputtered MoO3 films for ammonia sensing, Thin Solid Films 436 (1) (2003) 46–51. [14] Y. Zaatar, D. Zaouk, J. Bechara, A. Khoury, C. Llinaress, J.P. Charles, Fabrication and characterization of an evanescent wave fiber optic sensor for air pollution control, Mater. Sci. Eng. B: Solid State Mater. Adv. Technol. 74 (1–3) (2000) 296–298. [15] S.Q. Tao, C.B. Winstead, R. Jindal, J.P. Singh, Optical-fiber sensor using tailored porous sol–gel fiber core, IEEE Sens. J. 4 (3) (2004) 322–328. [16] E. Scorsone, S. Christie, K.C. Persaud, P. Simon, F. Kvasnik, Fibreoptic evanescent sensing of gaseous ammonia with two forms of a new near-infrared dye in comparison to phenol red, Sens. Actuators B: Chem. 90 (1–3) (2003) 37–45. [17] P.S. Kumar, P.G. Vallabhan, V.P.N. Nampoori, P. Radhakrishnan, Fiber optic evanescent wave sensor for ammonia gas, in: Proceeding of the Society of Photo-Optical Instrumentation Engineers (SPIE), Parts 1 and 2, vol. 5280, 2004, pp. 617–621. [18] R.A. Potyrailo, G.M. Hieftje, Distributed fiber-optic chemical sensor with chemically modified plastic cladding, Appl. Spectrosc. 52 (8) (1998) 1092–1095. [19] V. Ruddy, An effective attenuation coefficient for evanescent wave spectroscopy using multimode fiber, Fiber Integrated Opt. 9 (2) (1990) 143–151. [20] S.H. Kim, G.S. Byun, S. Hong, Determination of ammonia in environmental samples using the diode laser/fiber optic colorimetric spectrometer, Opt. Eng. 39 (7) (2000) 1985–1988. [21] N. Matsuoka, S. Yamaguchi, K. Nanri, T. Fujioka, D. Richter, F.K. Tittel, Yb fiber laser pumped mid-IR source based on difference frequency generation and its application to ammonia detection, Jpn. J. Appl. Phys. Part 1 40 (2A) (2001) 625–628. [22] R. Claps, F.V. Englich, D.P. Leleux, D. Richter, F.K. Tittel, R.F. Curl, Ammonia detection by use of near-infrared diode-laserbased overtone spectroscopy, Appl. Opt. 40 (24) (2001) 4387– 4394. [23] I.M. Raimundo, R. Narayanaswamy, Simultaneous determination of relative humidity and ammonia in air employing an optical fibre sensor and artificial neural network, Sens. Actuators B: Chem. 74 (1–3) (2001) 60–68. [24] P. Simon, M. Landl, M. Breza, F. Kvasnik, New NIR dyes for ammonia sensing, Sens. Actuators B: Chem. 90 (1–3) (2003) 9–14. [25] E. Scorsone, S. Christie, K. Persaud, P. Simon, F. Kvasnik, Fibre optic evanescent sensing of gaseous ammonia with a near infrared
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33] [34]
[35] [36]
[37]
[38]
[39] [40]
[41]
[42]
[43]
dye for leak detection, in: Proceeding of The Society of PhotoOptical Instrumentation Engineers (SPIE), Parts 1 and 2, vol. 4829, 2003, pp. 978–979. P.S. Kumar, C.P.G. Vallabhan, V.P.N. Nampoori, P. Radhakrishnan, LED-based fiber optic evanescent wave ammonia sensor, in: Proceedings of the SPIE, vol. 4969, The International Society for Optical Engineering, 2003, pp. 166–173. C. Malins, T.M. Butler, B.D. MacCraith, Influence of the surface polarity of dye-doped sol–gel glass films on optical ammonia sensor response, Thin Solid Films 368 (1) (2000) 105–110. C. Malins, M. Landl, P. Simon, B.D. MacCraith, Fibre optic ammonia sensing employing novel near infrared dyes, Sens. Actuators B: Chem. B51 (1–3) (1998) 359–367. T. Grady, T. Butler, D. MacCraith, D. Diamond, M.A. McKervey, Optical sensor for gaseous ammonia with tuneable sensitivity, Analyst 122 (8) (1997) 803–806. A. Messica, A. Greenstein, A. Katzir, U. Schiessl, M. Tacke, Fiberoptic evanescent wave sensor for gas detection, Opt. Lett. 19 (15) (1994) 1167–1169. S. Christie, E. Scorsone, K. Persaud, F. Kvasnik, Remote detection of gaseous ammonia using the near infrared transmission properties of polyaniline, Sens. Actuators B: Chem. 90 (1–3) (2003) 163– 169. Z. Jin, Y.X. Su, Y.X. Duan, Development of a polyanilinebased optical ammonia sensor, Sens. Actuators B 72 (2001) 75– 79. Z. Jin, Y.X. Su, Y.X. Duan, An improved optical pH sensor based on polyaniline, Sens. Actuators B 71 (12) (2000) 118–122. G. Stewart, W. Johnstone, Evanescently coupled components, in: J. Dakin, B. Culshaw (Eds.), Optical Fiber Sensors Volume Three: Components and Subsystems, Artech House, Boston, 1988–1997, pp. 69–101. B. Culshaw, Optical fiber sensor technologies: opportunities andperhaps-pitfalls, J. Lightwave Technol. 22 (1) (2004) 39–50. T. Yagi, N. Kuboki, Y. Suzuki, N. Uchino, K. Nakamura, K. Yoshida, Fiber-optic ammonia sensors utilizing rectangular-cladding eccentriccore fiber, Opt. Rev. 4 (5) (1997) 596–600. P.K. Choudhury, T. Yoshino, On the pH response of fiber optic evanescent field absorption sensor having a U-shaped probe: an experimental analysis, Optik 114 (1) (2003) 13–18. R.A. Potyrailo, L.A. Mikheenko, P.S. Borsuk, S.P. Golubkov, P.M. Talanchuk, Portable photometric ammonia gas analyser using pHsensitive polymer dye-films: spectral optimization of performance under field operating conditions, Sens. Actuators B: Chem. 21 (1994) 65–70. J. Lin, Recent development and applications of optical and fiber-optic pH sensors, Trends Anal. Chem. 19 (9) (2000) 541–552. P. Simon, S. Sekretar, B.D. MacCraith, F. Kvasnik, Near-infrared reagents for fibre-optic ammonia sensors, Sens. Actuators B: Chem. B39 (1–3) (1997) 252–255. S.K. Spear, S.L. Patterson, M.A. Arnold, Flow-through fiber-optic ammonia sensor for analysis of hippocampus slice perfusates, Anal. Chim. Acta 357 (1–2) (1997) 79–84. C. McDonagh, F. Sheridan, T. Butler, B.D. MacCraith, Characterisation of sol–gel-derived silica films, J. Non-Cryst. Solids 194 (1–2) (1996) 72–77. A. D’Amico, C. Di Natale, A. Taroni, Sensors parameters, sensors for domestic applications, in: Proceedings of the First European School on Sensors (ESS’94), Castro Marina, Leee, Italy, September 12–17, 1994, pp. 3–13.
Biographies Wenqing Cao received the BSc and MSc degrees from Xidian University, China, in 1986 and 1992, respectively, and the PhD from Nanyang
W. Cao, Y. Duan / Sensors and Actuators B 110 (2005) 252–259 Technological University, Singapore, in 2002. He is currently a postdoc with Los Alamos National Laboratory, USA. His research interests are in optical fiber chemical sensors, semiconductor oxide-based gas sensors, MEMS/MOEMS and their packaging. Yixiang Duan received his BS from Fudan University in 1982 and MS from Chinese Academy of Sciences in 1988. In 1994, he received his
259
PhD jointly from Jilin University/Indiana University. He is currently a staff scientist and principal investigator in Los Alamos National Laboratory. His research interests include optical fiber sensors, chemical and biological sensors, handheld detectors and field portable instruments.