- Email: [email protected]
Development and evaluation of optical fiber NH3 sensors for application in air quality monitoring
Atmospheric Environment 66 (2013) 1e7
Contents lists available at SciVerse ScienceDirect
Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv
Development and evaluation of optical fiber NH3 sensors for application in air quality monitoring Yu Huang, Lucas Wieck, Shiquan Tao* Department of Mathematics, Chemistry & Physics, West Texas A&M University, Canyon, TX 79016, USA
h i g h l i g h t s < Optical fiber NH3 sensors have been developed. < The feasibility of using the sensors for air quality monitoring has been investigated. < The sensors are reversible, can be used for continuous monitoring NH3 in air. < The sensors are highly sensitive, can detect NH3 in ppbv level. < Cross responses of the sensors to moisture, CO2 and temperature change have been investigated.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 November 2011 Received in revised form 22 June 2012 Accepted 23 June 2012
Ammonia is a major air pollutant emitted from agricultural practices. Sources of ammonia include manure from animal feeding operations and fertilizer from cropping systems. Sensor technologies with capability of continuous real time monitoring of ammonia concentration in air are needed to qualify ammonia emissions from agricultural activities and further evaluate human and animal health effects, study ammonia environmental chemistry, and provide baseline data for air quality standard. We have developed fiber optic ammonia sensors using different sensing reagents and different polymers for immobilizing sensing reagents. The reversible fiber optic sensors have detection limits down to low ppbv levels. The response time of these sensors ranges from seconds to tens minutes depending on transducer design. In this paper, we report our results in the development and evaluation of fiber optic sensor technologies for air quality monitoring. The effect of change of temperature, humidity and carbon dioxide concentration on fiber optic ammonia sensors has been investigated. Carbon dioxide in air was found not interfere the fiber optic sensors for monitoring NH3. However, the change of humidity can cause interferences to some fiber optic NH3 sensors depending on the sensor’s transducer design. The sensitivity of fiber optic NH3 sensors was found depends on temperature. Methods and techniques for eliminating these interferences have been proposed. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Optical fiber chemical sensor Ammonia sensor Air quality monitoring Concentrated animal feeding Manure Feedlot
1. Introduction Ammonia emission from agricultural activities has attracted increased attention in recent years due to the adaptation of large scale concentrated animal feeding practices and the rural development which brought residents closer to the animal feeding operation (AFO) facilities (Rhoades et al., 2010; Wathes et al., 1997; Gates et al., 2008). Ammonia from AFO has been implicated to the formation of fine aerosol particles. US EPA also recognizes that ammonia in the atmosphere is indirectly involved in the formation of ground ozone (http://www.epa.gov/region9/animalwaste/
* Corresponding author. E-mail address: [email protected] (S. Tao). 1352-2310/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atmosenv.2012.06.072
problem.html). Metal oxide aerosol particles can catalyze the oxidation of ammonia in air to form NOx, which are precursors of ground ozone (Renard et al., 2004; Yuan et al., 1994). In order to understand ammonia atmospheric chemistry, to develop ammonia emission control techniques, as well as to provide baseline data for air quality standards, it is necessary to develop technologies for continuous, real-time, long-term monitoring ammonia spatial distribution. Present state-of-the-art techniques for monitoring NH3 concentration in air quality programs depend on field-sampling/labanalysis (Cole et al., 2005; Phillips et al., 2001; McGinn et al., 2003) or the use of expensive instruments (Mukhtar et al., 2009; Maeda and Takenaka, 1993; Roberts et al., 1988). The first method requires frequent field trips and has a high labor-cost. It is too expensive to obtain high time/spatial-resolution distribution of NH3
2
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
concentration in air with this method. Chemiluminescence analyzers are presently available for continuously monitoring trace NH3 concentration in air (Mukhtar et al., 2009; Maeda and Takenaka, 1993). However, these instruments are expensive, and monitoring hundreds of sites using such instruments in air quality monitoring programs is difficult and cost prohibitive. There are several sensor technologies reported for detecting/ monitoring ammonia gas (Malyshev and Pislyakov, 2003; Marquis and Vetelino, 2001; Varghese et al., 2003; Yuan and El-Sherif, 2003; Jin et al., 2001; Morales-Bahnik et al., 1994; Huang and Tao, 2011; Tao et al., 2008, 2006, 2007; Guo and Tao, 2007). Among the reported sensor technologies, fiber optic NH3 sensors have the advantages of high sensitivity, fast response, reversible sensing, robustness, low cost of both fabrication and operation, compatibility with present fiber optic communication techniques for forming sensor network and communicating obtained information of NH3 spatial and time distribution among interested parties. The challenges in developing a sensor for continuous, long-term NH3 monitoring in air include: 1) low analyte concentration (subppmv), 2) continuously changing sample matrix (moisture and CO2), and 3) continuously changing sample temperature. Although many fiber optic NH3 sensors have been reported, there has been no systematic study of the effect of these parameters (relative humidity, CO2, temperature) on the NH3 sensors. We have been developing fiber optic sensor technologies for applications of monitoring trace ammonia (Huang and Tao, 2011; Tao et al., 2008, 2006, 2007; Guo and Tao, 2007). These sensors are sensitive, and can detect NH3 in air to low ppbv level. The sensors also have fast response. This paper reports our continued efforts to evaluate the potential of fiber optic NH3 sensors for applications in air quality monitoring, and the development of approaches to eliminate potential interferences. These include assessing the sensitivity of fiber optic NH3 sensors for air quality monitoring, investigating the effect of air composition change (humidity, CO2) and temperature on sensor’s response to ammonia. The approaches for eliminating potential interferences are also discussed. 2. Experimental 2.1. Instruments An optical fiber compatible ultraviolet/visible (UV/Vis) spectrometer (Redtide620, OceanOptics, Inc., Dunedin, FL) and an optical fiber compatible light source (LS-1, OceanOptics, Inc.) were used for measuring optical absorption response of optical fiber sensor probes. A dynamic gas calibrator (Thermo Scientific Model 146i, Franklin, MA) was used to dilute a standard NH3 gas sample with compressed air to make test air samples. A tube furnace (MTI Model GSL-1500X-50, MTI Corp., Richmond, CA) was used to heat air gas samples during some experiments for testing the effect of varying temperature.
previous works are: 1) a bent optical fiber probe coated with a bromocreasol purple (BCP) doped solegel silica polymer (Tao et al., 2006), 2) a bent optical fiber probe having a dual layer poly (methylmethacrylate) (PMMA)/chlorophenol red (CPR) coatings (Huang and Tao, 2011). The transducer of one of the new sensors is a bent optical fiber probe having a polyaniline coating. The transducer of the second new sensor is a bent optical fiber probe having a Fe(III)-porphyrin complex doped PMMA coating. The sensing mechanisms (chemical reactions) of these two new sensors are similar to those already reported by other research groups (Yuan and El-Sherif, 2003; Jin et al., 2001; Morales-Bahnik et al., 1994). The BCP-doped solegel silica coated sensor probe and the dual PMMA/CPR coated probe were made by following our previously reported procedures (Huang and Tao, 2011; Tao et al., 2006). In order to make a polyaniline coated bent optical fiber sensor probe, 2.0 mL of an acidic solution (1.0 M HCl) containing 0.10 M aniline was mixed with 2.0 mL of 0.10 M NH4S2O8 solution. A pre-cleaned bent optical fiber probe was inserted into the mixed solution for 12 h. During the oxidization polymerization process, a thin layer of polyaniline was formed on the surface of the bent optical fiber core. This procedure is similar to a reported method for preparing a polyaniline membrane based NH3 sensor (Yuan and El-Sherif, 2003; Jin et al., 2001). Fifty milligram of a Fe(III)-porphyrin complex (Fe(III) meso-tetra-(o-dichlorophenyl) porphine chloride, Frontier Scientific, Inc, Logan, UT) was dissolved together with 0.20 g PMMA resin in 4.0 mL acetone to form a coating solution. A dip-coating method was followed to coat a Fe(III)-porphyrin doped PMMA membrane on a bent optical fiber probe (Huang and Tao, 2011; Guo and Tao, 2007; Tao et al., 2006). 2.3. Laboratory set-up for testing the sensor probes and test method The laboratory set-up for testing the sensor probes is diagrammatically shown in Fig. 1. A bent optical fiber sensor probe to be tested was set inside a laboratory-made climate chamber. Light from the light source was injected into the bent optical fiber sensor probe from one end. The light transmitted through the probe is detected by the optical fiber compatible spectrometer and converted to absorbance signal with a computer program provided by the spectrometer maker. Compressed air was used to dilute a standard NH3 gas sample to make test air samples of different NH3 concentration. The
2.2. Fiber optic NH3 sensors A fiber optic NH3 sensor consists of the optical fiber compatible light source, the optical fiber compatible UV/Vis spectrometer, and a bent optical fiber probe having a specially tailored coating on the surface of the bent fiber core. The two ends of the bent optical fiber probe are connected to the light source and spectrometer, respectively. Two fiber optic sensors investigated in this work were developed from our previous works (Huang and Tao, 2011; Tao et al., 2006). Two additional fiber optic sensor techniques, which have potential in applications of monitoring trace NH3, were also investigated. The transducers of the two fiber optic NH3 sensors developed from our
Fig. 1. A schematic diagram shows the laboratory setup for testing the fiber optic NH3 sensor probes exposed to air samples of different composition at different temperatures.
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
compressed air has a 15% relative humidity (RH). The compressed air contains about 350 ppm CO2 and trace NH3. The sensors investigated in this work are not sensitive enough to detect trace NH3 in the compressed air. The compressed air was used in this work without further treatment. In order to make testing air samples of different RH, water vapor was added to the air sample by passing part of the diluted air sample through an acoustic humidifier (Crane EE-5301G Ultrasonic Cool Mist Nursery Humidifier, Crane USA, Bensenville, IL) as shown in Fig. 1. The relative humidity of the air sample was adjusted through changing the flow rate ratio of the two gas lines shown in Fig. 1. A humidity sensor (TJ-USB/RHUSB, Omega Engineering, Inc., Stamford, CT) was connected to the exit port of the climate chamber to monitor RH. Additional CO2 was added to the air sample by mixing compressed air with a 5% CO2 gas (balanced with N2). The flow rate ratio of the compressed air and 5% CO2 gas was controlled through adjusting the flow rate of the two flow-meters as shown in Fig. 1. In order to test the effect of temperature on sensor’s response, a gas sample was passed through a copper tube (i.d. 5 mm, o.d. 8 mm), which was placed inside the tube furnace. The gas sample was heated by the furnace while flowing through the copper tube. A thermocouple was deployed in the climate chamber for continuous monitoring of gas temperature during these tests. 2.4. Sample gases A standard gas sample containing 100 ppmv NH3 balanced with N2 (Airgas South, Inc., Dallas, TX) was used. Testing air gas samples were made by diluting this standard gas sample with compressed air as described above. 3. Result and discussion 3.1. Sensitivity of fiber optic sensors Two fiber optic NH3 sensors developed from our previous work have been tested for detecting NH3 in air samples with detection limit in low ppbv level. Fig. 2 shows the time response of a BCPdoped solegel silica coating based fiber optic NH3 sensor developed from our previous work (Tao et al., 2006). This sensor has fast response (response time < 1 min) and a detection limit of 9.6 ppbv NH3 in air. In considering the fact that the reported NH3 concentration at concentrated cattle feeding facilities is several tens of ppbv to hundreds of ppbv, (Rhoades et al., 2010) and several ppmv to tens of ppmv in poultry houses, (Wathes et al., 1997), this sensor
3
is sensitive enough for monitoring NH3 in air quality monitoring program related to animal feeding practices. The sensitivity of the CPR/PMMA dual coatings based fiber optic NH3 sensor is comparable with the BCPesolegel silica coating based fiber optic NH3 sensor (Huang and Tao, 2011). Two bent optical fiber sensor probes using polyaniline coating and Fe(III)-porphyrin-doped PMMA coating have also been tested for sensing NH3 in air. The sensitivity of these two sensor probes for monitoring NH3 in air is not as high as that of the pH indicator (BCP, CPR) based sensor probes. The detection limit of these two sensors estimated from our preliminary work is in ppmv range, which is comparable with the detection limit of reported fiber optic NH3 sensors using similar reagents (Yuan and El-Sherif, 2003; Jin et al., 2001; Morales-Bahnik et al., 1994). These sensors can also be used for air quality monitoring in some animal feeding facilities. 3.2. Cross-response to moisture change and methods of eliminating the interference Water vapor is a major component of air, and its concentration in outdoor air is continuously changing with time and location. Therefore, a sensor developed for air quality monitoring applications must be free of moisture cross response, or a method can be used to eliminate moisture cross response. Most reported fiber optic NH3 gas sensors, including the four sensors discussed in this paper, have cross response to moisture (Huang and Tao, 2011; Tao et al., 2006; Raimundo and Narayanaswamy, 2001). The mechanism of moisture cross response is different depending on sensing mechanism and transducer structure. For a fiber optic sensor using a sensing reagent immobilized polymer coating as a transducer, polymer coating’s microstructure and hydrophilic property, as well as the reactivity of sensing reagent with water vapor affect the sensor’s response to moisture change. For example, the BCPesolegel silica coating based sensor developed from our previous work has cross response to moisture. Fig. 3 shows calibration curves of the sensor for monitoring NH3 in air of different RH% (Tao et al., 2006). The slope of the calibration curves at different humidity level is almost the same. However, the interception of the calibration curves increases with the increase of RH% in air sample. We have previously described the mechanism of this moisture cross response (Tao et al., 2006; Xu et al., 2004). In brief, the moisture cross response of this sensor is caused by, 1) the porous solegel silica coating scatters evanescent wave (EW) of light guided inside the optical fiber; and 2) porous solegel silica coating absorbs water vapor
0.2
RH = 47%
Abs.= 0.0057x + 0.0327 (RH = 47%)
RH = 60%
0.18
RH = 70%
Abs. = 0.0058x + 0.0444 (RH = 60%) (R = 0.979) Abs = 0.0059x + 0.0637 (RH = 70%) (R = 0.9831) Abs= 0.0058x + 0.0772 (RH = 80% (R = 0.9844)
RH = 80%
(R = 0.9863)
0.16
Absorbance
0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0
5
10
15
20
NH3 concentration (ppmv) Fig. 2. Time response of a bent optical fiber NH3 sensor having a BCP-doped solegel silica coating exposed to air samples of different NH3 concentration.
Fig. 3. Calibration curves of a bent optical fiber NH3 sensor having a BCP-doped solegel silica coating for detecting NH3 in air of different humidity level.
4
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
from air sample it exposed to, which enhances the EW-scattering and causes an interference signal. Two approaches have been investigated to eliminate this sensor’s moisture cross response problem. The first approach involves using a different sensing reagent and polymer combination. A coating on the surface of an optical fiber will scatter EW of light guided inside the optical fiber if it has multiple phases in microscopic scale. This can be a solid/gas phase structure, likes that in solegel silica coating. It can also be solid/solid multi-phases. Fig. 3 shows the example of moisture cross response of a sensor that has a solid/gas multi-phase coating. Fig. 4 shows the spectral response of a bent optical fiber NH3 sensor having a BCP immobilized PMMA coating as a transducer to air samples containing the same NH3 concentration, but different RH%. In addition to optical absorption spectrum caused by the reaction product of NH3 with BCP (peak absorption wavelength at around 600 nm), EWscattering signals in the broad wavelength range are also observed. This is an example of EW-scattering caused by a coating having micro solid (BCP)/solid (PMMA) heterogeneous structure. In this case, the sensing reagent is trapped as particles (solid phase) in the PMMA membrane (another solid phase). With the increase of relative humidity the solid reagent particles absorb more water, which causes an increase of particle size. As a result, EW-scattering signal increased. In order to make a fiber optic NH3 sensor which is insensitive to moisture change, the transducer coating has to be a single phase. In the case of using a polymer to immobilize a sensing reagent, this requires: 1) the polymer will form a continuous membrane coating; and 2) the sensing reagent has to be soluble in the polymer in solid phase. One such example is a BCP-doped polyvinyl alcohol (PVC) coating. PVC is a hydrophilic polymer due to the large number of eOH groups in the polymer chain. This polymer forms continuous membrane coating on fiber surface, which does not scatter light. BCP is hydrophilic, can be distributed as molecules in the PVC polymer to form a solid solution. Fig. 5 shows the response of a bent optical fiber NH3 sensor having a BCP-doped PVC coating to air samples of different RH% and its response to an air sample containing 1.0 ppmv NH3. This test result clearly demonstrates that the increase of NH3 concentration causes an absorption spectrum with peak absorption wavelength at 600 nm. The change of humidity from RH ¼ 34%e68% does not cause EW-scattering related baseline shift. Therefore, this sensor does not have cross response to moisture. The second approach for eliminating moisture caused EWscattering interference is monitoring sensing signals at two wavelengths. The signal at one wavelength is the sum of EW-scattering
Fig. 4. Spectral response of a bent optical fiber NH3 sensor having a BCP-doped PMMA coating exposed to NH3-containing air samples of different relative humidity. The spectral responses have two components: an EW-scattering signal in broad wavelength range and an absorption spectrum with peak absorption wavelength at 615 nm.
Fig. 5. Spectral responses of a bent optical fiber NH3 sensor having a BCP-doped PVA coating exposed to air samples of different humidity and an air sample containing 1.0 ppmv NH3.
signal and the absorbance of NH3/reagent reaction product. The sensing signal at a second wavelength is caused by EW-scattering only. Since the EW-scattering signal is independent of wavelength in broad wavelength range (Xu et al., 2004), a computer program can be developed to substrate the EW-scattering signal, and give out a sensing signal which is only related to trace NH3 in the air sample. Some sensing reagent immobilized in a polymer itself reacts with water vapor. The reaction product absorbs light and causes spectral response. It will be difficult to eliminate this kind of interference if the absorption wavelength of the waterereagent reaction product is close to that of the NH3ereagent reaction product. Fig. 6 shows such an example. Fe(III)-porphyrin reacts with both water and NH3 to form complexes, and the peak absorption wavelengths of the formed complexes are very close to each other. This phenomenon could happen to many fiber optic sensors using the complex formation of metal ions with NH3 as a sensing mechanism. It is inappropriate to use these sensors in air quality monitoring.
Fig. 6. Spectral response of a bent optical fiber NH3 sensor having an Fe(III)-porphyrin complex doped PMMA coating exposed to air samples of different humidity and NH3 concentration. These test results demonstrate the potential spectral interference of relative humidity change on the sensor for monitoring NH3 in air.
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
3.3. Potential interference of CO2 to fiber optic NH3 sensors In ambient air, CO2 concentration is usually hundreds to thousands of times higher than that of NH3. In addition, CO2 concentration in cattle feeding facilities also continuously changes with time and location. Therefore, in order to develop a fiber optic sensor for monitoring NH3 in air quality program it is necessary to investigate potential interference from CO2 in air. Many fiber optic NH3 sensors use NH3’s basic property to generate a sensing signal. For example, pH indicators have been used in several reported fiber optic sensors as sensing reagents (Huang and Tao, 2011; Tao et al., 2006; McMurray and Albadran, 1999). The reported fiber optic sensors using polyaniline are also based on NH3’s basic property (Yuan and El-Sherif, 2003; Jin et al., 2001). Because CO2 is an acidic oxide it will react with pH indicators and polyaniline in the existence of water. Therefore, it is possible that the change of CO2 concentration in air causes a change of sensing signal and interferes such fiber optic sensor’s function of detecting NH3 in air. On the other hand, CO2 is a very weak coordinating ligand. Optical fiber NH3 sensors based on the formation of coordinating compounds of NH3 with metal compounds are usually free of CO2 interference. Fig. 7 is the spectral response of the BCP-doped solegel silica coating based fiber optic NH3 sensor exposed to NH3-containing air samples of different CO2 concentration. In this case, the change of CO2 concentration from 350 ppm to 1350 ppm does not cause significant interference to the sensor’s function of sensing NH3 in air. Although this sensor uses a pH indicator as a sensing reagent, the pH indicator is distributed on the surface of solegel silica particles surface. Silica particle surface is covered with SieOH, which is a weak acid. The pH indicator molecules on solegel silica particle surface exist as acidic form. The change of trace CO2 concentration in air samples does not change the form of pH indicator molecule on solegel silica particle surface, and thus, does not interfere the sensor’s function of monitoring NH3 in air. The response of PCR/PMMA dual layers coated bent optical fiber sensor probe to CO2 in air was also investigated. It was found that the change of CO2 concentration in air will cause interference of a freshly-made PCR/PMMA probe, but does not cause interference once the probe was exposed to air over-night.It is possible that during the long-time air exposure the PCR in the coating reacted with CO2 (350 ppm) in air, and changed to acidic form. Therefore,
further change of CO2 concentration in air does not cause interference to the sensor’s function of sensing NH3 in air. It was reported that the emission rate of CO2 in cattle feeding facilities is about 200 times that of NH3 (Leytem et al., 2011). Carbon dioxide concentration in cattle feeding facilities changes in the range from 400 ppmv to 700 ppmv. In considering these facts and the test results of this work, it can be conclude that the change of CO2 concentration in AFOs will not interfere the fiber optic sensors for monitoring NH3.
3.4. Effect of temperature on fiber optic NH3 sensor’s response Temperature is another factor that needs to be taken into consideration when designing a fiber optic NH3 sensor for applications in air quality monitoring because: 1) air temperature is continuously changing with time and location; and 2) temperature affects chemical equilibrium constant. Almost all the reported fiber optic NH3 sensors are based on NH3 reaction with a sensing reagent inside a coating on fiber surface. Such a reaction can be described by following equation:
NH3 þ SR5SRNH3
(1)
Where, SR is sensing reagent (metal compound or pH indicating compound) in the coating and SRNH3 is reaction product formed inside the coating. The sensing signal is proportional to the concentration of reaction product, SRNH3. The reaction is reversible and is in equilibrium with gas phase NH3 concentration. The equilibrium constant of this reaction can be expressed as:
Kc ¼
½SRNH3 ½NH3 ½SR
(2)
Because the sensing reagent is usually in large excess compared to trace NH3 in gas phase, the concentration (or quantity) of reaction product, SRNH3, and thus the sensing signal in absorbance, can be expressed as:
Abs: ¼ K½SRNH3 ¼ Kc0 ½NH3
(3)
Where Kc0 is a combined constant including reaction equilibrium constant, absorption coefficient, interaction path-length of EW with the reaction product on fiber’s surface and sensing reagent concentration. Equation (3) indicates that a linear calibration curve (absorbance against NH3 concentration) should be obtained for such a fiber optic NH3 sensor at a decided temperature (constant Kc0 value), and the sensing signal is proportional to NH3 concentration in gas phase. From thermochemistry, the equilibrium constant of a chemical reaction is related to temperature as described by following equation: o 1 DHrxn
lnKc ¼
Fig. 7. Spectral response of a bent optical fiber NH3 sensor having a BCP-doped solegel silica coating as a transducer exposed to air samples of different CO2 concentration. The change of CO2 concentration in air does not cause interference to this sensor’s function of monitoring NH3 in air.
5
R
T
þ
DSorxn R
(4)
o is the standard enthalpy change of the reaction, and Where, DHrxn DSorxn is the standard entropy change of the reaction. Therefore, if a fiber optic NH3 sensor is calibrated at different temperatures, the slope of the calibration curves, which is proportional to Kc, should have a relationship with temperature similar to that described by Equation (4). The effect of temperature on fiber optic NH3 sensor was investigated using a bent optical fiber sensor probe having a BCP-doped solegel silica coating as an example. This sensor has been calibrated for sensing NH3 in air at different temperatures. Fig. 8 shows the obtained calibration curves. A plot of ln Kc0 against reciprocal of
6
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
function of monitoring trace NH3 in air. This interference can be eliminated through calibrating the sensor at different temperatures to find out the relationship of slope of calibration curves with temperature, and use a temperature sensor together with the fiber optic NH3 sensor in air quality monitoring. From this work it can be concluded that among the four tested fiber optic NH3 sensors developed from our works, the fiber optic NH3 sensor using a bent optical fiber probe coated with BCP-doped solegel silica is most appropriate for monitoring NH3 in animal feeding facilities. In further work we will build prototype sensors and test the longterm stability of the sensors in a real cattle feedlot. Fig. 8. Calibration curves of a bent optical fiber NH3 sensor for detecting NH3 in air of different temperature. Temperature change affects the slope of the calibration curves.
This work is supported by USDA National Institute of Food and Agriculture through award# 2010-65112-20521.
1/T (K -1) 0 0.00326 -1
0.00328
0.0033
0.00332
0.00334
0.00336
0.00338
References
ln Kc'
-2 -3 -4 ln Kc' = 29027 (1/T) - 99.711
2
(R = 0.8955)
-5 -6 Fig. 9. A plot of ln Kc0 (Kc0 is the slope of calibration curves in Fig. 8) against the reciprocal of temperature (1/T). This linear relationship agrees with Equations (3) and (4), which provides a method for calibrating the sensor for detecting NH3 in air of changing temperature.
temperature (1/T) is showing in Fig. 9. A linear relationship of ln Kc0 with (1/T) can be established as:
lnKc0 ¼ 29027
Acknowledgment
1 99:71 R2 ¼ 0:8955 T
(5)
This relationship agrees with Equations (3) and (4). This gives a method to correct temperature effect on fiber optic NH3 sensors. A fiber optic NH3 sensor can first be calibrated in the laboratory at different temperatures to find out the relationship of calibration curve’s slope with temperature before deploying the sensor to for field applications. A temperature sensor can be deployed together with the fiber optic NH3 sensor in the field. The NH3 concentration in air can be calculated from the sensor’s absorbance signal and temperature using Equations (5) and (3) with a computer program.
4. Conclusions From the results of this work, it can be concluded that: 1) Fiber optic NH3 sensors using bent optical fiber probes having pH indicators as sensing reagents can achieve low-ppbv level detection limit, which is sensitive enough for monitoring trace NH3 in air of animal feeding facilities. 2) The change of air CO2 concentration does not affect pH-indicator based fiber optic sensor’s function for monitoring NH3. Carbon dioxide in air also does not affect the function of metal-complex formation based fiber optic NH3 sensors. 3) Humidity level change in air affects some fiber optic sensor’s function for monitoring NH3 in air. This interference can be eliminated by using appropriate sensing reagent/polymer combinations to form solid solution coatings on optical fiber surface, or by measuring optical sensing signals at two wavelengths and using a computer program to obtain a sensing signal caused by NH3 only. 4) Air temperature change affects all fiber optic NH3 sensor’s
Cole, N.A., Clark, R.N., Todd, R.W., Richardson, C.R., Gueye, A., Greene, L.W., McBride, K., 2005. Influence of dietary crude protein concentration and source on potential ammonia emissions from beef cattle manure. Journal of Animal Science 83, 722e731. Gates, R.S., Casey, K.D., Wheeler, E.F., Xin, H., Pescatore, A.J., 2008. U.S. broiler housing ammonia emissions inventory. Atmospheric Environment 42, 3342e3350. Guo, H., Tao, S., 2007. “Silver nanoparticles doped silicate nanocomposites coated on an optical fiber for ammonia sensing”. Sensors & Actuators 123, 578e582. Huang, Y., Tao, S., 2011. An optical fiber sensor probe using a PMMA/CPR coated bent optical fiber as a transducer for monitoring trace ammonia. Journal of Sensor Technogies 1, 29e35. Jin, Z., Su, Y., Duan, Y., 2001. Development of a polyaniline-based optical ammonia sensor. Sensors & Actuators B72, 75e79. Leytem, A.B., Dungan, R.S., Bjorneberg, D.L., Koehn, A.C., 2011. Emissions of ammonia, methane, carbon dioxide, and nitrous oxide from dairy cattle housing and manure management systems. Journal of Environmental Quality 40, 1383e1394. Maeda, Y., Takenaka, N., 1993. Chemiluminescence determination of trace amounts of ammonia and halogen species in the environment. Proceedings of SPIE e The International Society for Optical Engineering 1715 (Optical Methods in Atmospheric Chemistry), pp. 185e193. Malyshev, V.V., Pislyakov, A.V., 2003. Dynamic properties and sensitivity of semiconductor metal-oxide thick-film sensors to various gases in air gaseous medium. Sensors & Actuators B96, 413e434. Marquis, B.T., Vetelino, J.F., 2001. A semiconducting metal oxide sensor array for the detection of NOx and NH3. Sensors & Actuators B77, 100e110. McGinn, S.M., Janzen, H.H., Coates, T., 2003. Atmospheric pollutants and trace gases: atmospheric ammonia, volatile fatty acids, and other odorants near beef feedlots. Journal of Environmental Quality 32, 1173e1182. McMurray, H.N., Albadran, J., 1999. Colorimetric and fluorimetric polymer membrane gas-sensing materials. MRS Bulletin 24, 55e59. Morales-Bahnik, A., Czolk, R., Ache, H.J., 1994. An optochemical ammonia sensor based in immobilized metalloporphyrins. Sensors & Actuators B18-19, 493e496. Mukhtar, S., Mutlu, A., Lacey, E.E., Parnell, C.B., 2009. Seasonal ammonia emissions from a free-stall dairy in central Texas. Journal of the Air & Waste Management Association 59, 613e618. Phillips, V.R., Lee, D.S., Scholtens, R., Garland, J.A., Sneath, R.W., 2001. A review of methods for measuring emission rates of ammonia from livestock buildings and slurry or manure stores, Part 2: monitoring flux rates, concentrations and air flow rates. Journal of Agricultural Engineering Research 78, 1e14. Raimundo, I.M., Narayanaswamy, R., 2001. Simultaneous determination of relative humidity and ammonia in air employing an optical fiber sensor and artificial neural network. Sensors & Actuators B74, 60e68. Renard, J.J., Calidonna, S.E., Henley, M.V., 2004. Fate of ammonia in the atmosphere e a review for applicability to hazardous releases. Journal of Hazardous Materials 108, 29e60. Rhoades, M.B., Parker, D.B., Cole, N.A., Todd, R.W., Caraway, E.A., Auvermann, B.W., Topliff, D.R., Schuster, G.L., 2010. Continuous ammonia emission measurements from a commercial beef feedyard in Texas. Transactions of the ASABE 53, 1823e1831. Roberts, J.M., Langford, A.O., Goldan, P.D., Fehsenfeld, F.C., 1988. Ammonia measurements at Niwot Ridge, Colorado and Point Arena, California using the tungsten oxide denuder tube technique. Journal of Atmospheric Chemistry 7, 137e152. Tao, S., Fanguy, J.C., Xu, L., 2006. Optical fiber ammonia sensing probes using reagent immobilized porous silica coating as transducers. Sensors & Actuators B115, 158e163.
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7 Tao, S., Gong, S., Fanguy, J.C., Hu, X., 2007. The application of a light guiding flexible tubular waveguide in evanescent wave absorption optical sensing. Sensors & Actuators B120, 724e731. Tao, S., Fanguy, J.C., Sarma, T.V.S., 2008. A fiber optic sensor for monitoring trace ammonia in high temperature gas samples with a CuCl2-doped porous silica optical fiber as a transducer. IEEE Sensors Journal 8, 2000e2007. Varghese, O.K., Gong, D., Dreschel, W.R., Ong, K.G., Keat, G., Grimes, C.A., 2003. Ammonia detection using nanoporous alumina resistive and surface acoustic wave sensors. Sensors & Actuators B94, 27e35.
7
Wathes, C.M., Holden, M.R., Sneath, R.W., White, R.P., Phillips, V.R., 1997. Concentrations and emission rates of aerial ammonia, nitrous oxide, methane, carbon dioxide, dust and endotoxin in UK broiler and layer houses. British Poultry Science 38, 14e28. Xu, L., Fanguy, J.C., Soni, K., Tao, S., 2004. Optical fiber humidity sensor based on evanescent wave scattering. Optics Letters 29, 1191e1193. Yuan, J., El-Sherif, M.A., 2003. Fiber-optic chemical sensor using polyaniline as modified cladding material. IEEE Sensors Journal 3, 5e12. Yuan, J., Feng, B., Lu, J., Liu, H., Liu, D., 1994. Effects of NH3 on N2O formation and destruction in fluidized bed coal combustion. Journal of Thermal Science 3, 278e282.
Contents lists available at SciVerse ScienceDirect
Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv
Development and evaluation of optical fiber NH3 sensors for application in air quality monitoring Yu Huang, Lucas Wieck, Shiquan Tao* Department of Mathematics, Chemistry & Physics, West Texas A&M University, Canyon, TX 79016, USA
h i g h l i g h t s < Optical fiber NH3 sensors have been developed. < The feasibility of using the sensors for air quality monitoring has been investigated. < The sensors are reversible, can be used for continuous monitoring NH3 in air. < The sensors are highly sensitive, can detect NH3 in ppbv level. < Cross responses of the sensors to moisture, CO2 and temperature change have been investigated.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 November 2011 Received in revised form 22 June 2012 Accepted 23 June 2012
Ammonia is a major air pollutant emitted from agricultural practices. Sources of ammonia include manure from animal feeding operations and fertilizer from cropping systems. Sensor technologies with capability of continuous real time monitoring of ammonia concentration in air are needed to qualify ammonia emissions from agricultural activities and further evaluate human and animal health effects, study ammonia environmental chemistry, and provide baseline data for air quality standard. We have developed fiber optic ammonia sensors using different sensing reagents and different polymers for immobilizing sensing reagents. The reversible fiber optic sensors have detection limits down to low ppbv levels. The response time of these sensors ranges from seconds to tens minutes depending on transducer design. In this paper, we report our results in the development and evaluation of fiber optic sensor technologies for air quality monitoring. The effect of change of temperature, humidity and carbon dioxide concentration on fiber optic ammonia sensors has been investigated. Carbon dioxide in air was found not interfere the fiber optic sensors for monitoring NH3. However, the change of humidity can cause interferences to some fiber optic NH3 sensors depending on the sensor’s transducer design. The sensitivity of fiber optic NH3 sensors was found depends on temperature. Methods and techniques for eliminating these interferences have been proposed. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Optical fiber chemical sensor Ammonia sensor Air quality monitoring Concentrated animal feeding Manure Feedlot
1. Introduction Ammonia emission from agricultural activities has attracted increased attention in recent years due to the adaptation of large scale concentrated animal feeding practices and the rural development which brought residents closer to the animal feeding operation (AFO) facilities (Rhoades et al., 2010; Wathes et al., 1997; Gates et al., 2008). Ammonia from AFO has been implicated to the formation of fine aerosol particles. US EPA also recognizes that ammonia in the atmosphere is indirectly involved in the formation of ground ozone (http://www.epa.gov/region9/animalwaste/
* Corresponding author. E-mail address: [email protected] (S. Tao). 1352-2310/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atmosenv.2012.06.072
problem.html). Metal oxide aerosol particles can catalyze the oxidation of ammonia in air to form NOx, which are precursors of ground ozone (Renard et al., 2004; Yuan et al., 1994). In order to understand ammonia atmospheric chemistry, to develop ammonia emission control techniques, as well as to provide baseline data for air quality standards, it is necessary to develop technologies for continuous, real-time, long-term monitoring ammonia spatial distribution. Present state-of-the-art techniques for monitoring NH3 concentration in air quality programs depend on field-sampling/labanalysis (Cole et al., 2005; Phillips et al., 2001; McGinn et al., 2003) or the use of expensive instruments (Mukhtar et al., 2009; Maeda and Takenaka, 1993; Roberts et al., 1988). The first method requires frequent field trips and has a high labor-cost. It is too expensive to obtain high time/spatial-resolution distribution of NH3
2
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
concentration in air with this method. Chemiluminescence analyzers are presently available for continuously monitoring trace NH3 concentration in air (Mukhtar et al., 2009; Maeda and Takenaka, 1993). However, these instruments are expensive, and monitoring hundreds of sites using such instruments in air quality monitoring programs is difficult and cost prohibitive. There are several sensor technologies reported for detecting/ monitoring ammonia gas (Malyshev and Pislyakov, 2003; Marquis and Vetelino, 2001; Varghese et al., 2003; Yuan and El-Sherif, 2003; Jin et al., 2001; Morales-Bahnik et al., 1994; Huang and Tao, 2011; Tao et al., 2008, 2006, 2007; Guo and Tao, 2007). Among the reported sensor technologies, fiber optic NH3 sensors have the advantages of high sensitivity, fast response, reversible sensing, robustness, low cost of both fabrication and operation, compatibility with present fiber optic communication techniques for forming sensor network and communicating obtained information of NH3 spatial and time distribution among interested parties. The challenges in developing a sensor for continuous, long-term NH3 monitoring in air include: 1) low analyte concentration (subppmv), 2) continuously changing sample matrix (moisture and CO2), and 3) continuously changing sample temperature. Although many fiber optic NH3 sensors have been reported, there has been no systematic study of the effect of these parameters (relative humidity, CO2, temperature) on the NH3 sensors. We have been developing fiber optic sensor technologies for applications of monitoring trace ammonia (Huang and Tao, 2011; Tao et al., 2008, 2006, 2007; Guo and Tao, 2007). These sensors are sensitive, and can detect NH3 in air to low ppbv level. The sensors also have fast response. This paper reports our continued efforts to evaluate the potential of fiber optic NH3 sensors for applications in air quality monitoring, and the development of approaches to eliminate potential interferences. These include assessing the sensitivity of fiber optic NH3 sensors for air quality monitoring, investigating the effect of air composition change (humidity, CO2) and temperature on sensor’s response to ammonia. The approaches for eliminating potential interferences are also discussed. 2. Experimental 2.1. Instruments An optical fiber compatible ultraviolet/visible (UV/Vis) spectrometer (Redtide620, OceanOptics, Inc., Dunedin, FL) and an optical fiber compatible light source (LS-1, OceanOptics, Inc.) were used for measuring optical absorption response of optical fiber sensor probes. A dynamic gas calibrator (Thermo Scientific Model 146i, Franklin, MA) was used to dilute a standard NH3 gas sample with compressed air to make test air samples. A tube furnace (MTI Model GSL-1500X-50, MTI Corp., Richmond, CA) was used to heat air gas samples during some experiments for testing the effect of varying temperature.
previous works are: 1) a bent optical fiber probe coated with a bromocreasol purple (BCP) doped solegel silica polymer (Tao et al., 2006), 2) a bent optical fiber probe having a dual layer poly (methylmethacrylate) (PMMA)/chlorophenol red (CPR) coatings (Huang and Tao, 2011). The transducer of one of the new sensors is a bent optical fiber probe having a polyaniline coating. The transducer of the second new sensor is a bent optical fiber probe having a Fe(III)-porphyrin complex doped PMMA coating. The sensing mechanisms (chemical reactions) of these two new sensors are similar to those already reported by other research groups (Yuan and El-Sherif, 2003; Jin et al., 2001; Morales-Bahnik et al., 1994). The BCP-doped solegel silica coated sensor probe and the dual PMMA/CPR coated probe were made by following our previously reported procedures (Huang and Tao, 2011; Tao et al., 2006). In order to make a polyaniline coated bent optical fiber sensor probe, 2.0 mL of an acidic solution (1.0 M HCl) containing 0.10 M aniline was mixed with 2.0 mL of 0.10 M NH4S2O8 solution. A pre-cleaned bent optical fiber probe was inserted into the mixed solution for 12 h. During the oxidization polymerization process, a thin layer of polyaniline was formed on the surface of the bent optical fiber core. This procedure is similar to a reported method for preparing a polyaniline membrane based NH3 sensor (Yuan and El-Sherif, 2003; Jin et al., 2001). Fifty milligram of a Fe(III)-porphyrin complex (Fe(III) meso-tetra-(o-dichlorophenyl) porphine chloride, Frontier Scientific, Inc, Logan, UT) was dissolved together with 0.20 g PMMA resin in 4.0 mL acetone to form a coating solution. A dip-coating method was followed to coat a Fe(III)-porphyrin doped PMMA membrane on a bent optical fiber probe (Huang and Tao, 2011; Guo and Tao, 2007; Tao et al., 2006). 2.3. Laboratory set-up for testing the sensor probes and test method The laboratory set-up for testing the sensor probes is diagrammatically shown in Fig. 1. A bent optical fiber sensor probe to be tested was set inside a laboratory-made climate chamber. Light from the light source was injected into the bent optical fiber sensor probe from one end. The light transmitted through the probe is detected by the optical fiber compatible spectrometer and converted to absorbance signal with a computer program provided by the spectrometer maker. Compressed air was used to dilute a standard NH3 gas sample to make test air samples of different NH3 concentration. The
2.2. Fiber optic NH3 sensors A fiber optic NH3 sensor consists of the optical fiber compatible light source, the optical fiber compatible UV/Vis spectrometer, and a bent optical fiber probe having a specially tailored coating on the surface of the bent fiber core. The two ends of the bent optical fiber probe are connected to the light source and spectrometer, respectively. Two fiber optic sensors investigated in this work were developed from our previous works (Huang and Tao, 2011; Tao et al., 2006). Two additional fiber optic sensor techniques, which have potential in applications of monitoring trace NH3, were also investigated. The transducers of the two fiber optic NH3 sensors developed from our
Fig. 1. A schematic diagram shows the laboratory setup for testing the fiber optic NH3 sensor probes exposed to air samples of different composition at different temperatures.
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
compressed air has a 15% relative humidity (RH). The compressed air contains about 350 ppm CO2 and trace NH3. The sensors investigated in this work are not sensitive enough to detect trace NH3 in the compressed air. The compressed air was used in this work without further treatment. In order to make testing air samples of different RH, water vapor was added to the air sample by passing part of the diluted air sample through an acoustic humidifier (Crane EE-5301G Ultrasonic Cool Mist Nursery Humidifier, Crane USA, Bensenville, IL) as shown in Fig. 1. The relative humidity of the air sample was adjusted through changing the flow rate ratio of the two gas lines shown in Fig. 1. A humidity sensor (TJ-USB/RHUSB, Omega Engineering, Inc., Stamford, CT) was connected to the exit port of the climate chamber to monitor RH. Additional CO2 was added to the air sample by mixing compressed air with a 5% CO2 gas (balanced with N2). The flow rate ratio of the compressed air and 5% CO2 gas was controlled through adjusting the flow rate of the two flow-meters as shown in Fig. 1. In order to test the effect of temperature on sensor’s response, a gas sample was passed through a copper tube (i.d. 5 mm, o.d. 8 mm), which was placed inside the tube furnace. The gas sample was heated by the furnace while flowing through the copper tube. A thermocouple was deployed in the climate chamber for continuous monitoring of gas temperature during these tests. 2.4. Sample gases A standard gas sample containing 100 ppmv NH3 balanced with N2 (Airgas South, Inc., Dallas, TX) was used. Testing air gas samples were made by diluting this standard gas sample with compressed air as described above. 3. Result and discussion 3.1. Sensitivity of fiber optic sensors Two fiber optic NH3 sensors developed from our previous work have been tested for detecting NH3 in air samples with detection limit in low ppbv level. Fig. 2 shows the time response of a BCPdoped solegel silica coating based fiber optic NH3 sensor developed from our previous work (Tao et al., 2006). This sensor has fast response (response time < 1 min) and a detection limit of 9.6 ppbv NH3 in air. In considering the fact that the reported NH3 concentration at concentrated cattle feeding facilities is several tens of ppbv to hundreds of ppbv, (Rhoades et al., 2010) and several ppmv to tens of ppmv in poultry houses, (Wathes et al., 1997), this sensor
3
is sensitive enough for monitoring NH3 in air quality monitoring program related to animal feeding practices. The sensitivity of the CPR/PMMA dual coatings based fiber optic NH3 sensor is comparable with the BCPesolegel silica coating based fiber optic NH3 sensor (Huang and Tao, 2011). Two bent optical fiber sensor probes using polyaniline coating and Fe(III)-porphyrin-doped PMMA coating have also been tested for sensing NH3 in air. The sensitivity of these two sensor probes for monitoring NH3 in air is not as high as that of the pH indicator (BCP, CPR) based sensor probes. The detection limit of these two sensors estimated from our preliminary work is in ppmv range, which is comparable with the detection limit of reported fiber optic NH3 sensors using similar reagents (Yuan and El-Sherif, 2003; Jin et al., 2001; Morales-Bahnik et al., 1994). These sensors can also be used for air quality monitoring in some animal feeding facilities. 3.2. Cross-response to moisture change and methods of eliminating the interference Water vapor is a major component of air, and its concentration in outdoor air is continuously changing with time and location. Therefore, a sensor developed for air quality monitoring applications must be free of moisture cross response, or a method can be used to eliminate moisture cross response. Most reported fiber optic NH3 gas sensors, including the four sensors discussed in this paper, have cross response to moisture (Huang and Tao, 2011; Tao et al., 2006; Raimundo and Narayanaswamy, 2001). The mechanism of moisture cross response is different depending on sensing mechanism and transducer structure. For a fiber optic sensor using a sensing reagent immobilized polymer coating as a transducer, polymer coating’s microstructure and hydrophilic property, as well as the reactivity of sensing reagent with water vapor affect the sensor’s response to moisture change. For example, the BCPesolegel silica coating based sensor developed from our previous work has cross response to moisture. Fig. 3 shows calibration curves of the sensor for monitoring NH3 in air of different RH% (Tao et al., 2006). The slope of the calibration curves at different humidity level is almost the same. However, the interception of the calibration curves increases with the increase of RH% in air sample. We have previously described the mechanism of this moisture cross response (Tao et al., 2006; Xu et al., 2004). In brief, the moisture cross response of this sensor is caused by, 1) the porous solegel silica coating scatters evanescent wave (EW) of light guided inside the optical fiber; and 2) porous solegel silica coating absorbs water vapor
0.2
RH = 47%
Abs.= 0.0057x + 0.0327 (RH = 47%)
RH = 60%
0.18
RH = 70%
Abs. = 0.0058x + 0.0444 (RH = 60%) (R = 0.979) Abs = 0.0059x + 0.0637 (RH = 70%) (R = 0.9831) Abs= 0.0058x + 0.0772 (RH = 80% (R = 0.9844)
RH = 80%
(R = 0.9863)
0.16
Absorbance
0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0
5
10
15
20
NH3 concentration (ppmv) Fig. 2. Time response of a bent optical fiber NH3 sensor having a BCP-doped solegel silica coating exposed to air samples of different NH3 concentration.
Fig. 3. Calibration curves of a bent optical fiber NH3 sensor having a BCP-doped solegel silica coating for detecting NH3 in air of different humidity level.
4
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
from air sample it exposed to, which enhances the EW-scattering and causes an interference signal. Two approaches have been investigated to eliminate this sensor’s moisture cross response problem. The first approach involves using a different sensing reagent and polymer combination. A coating on the surface of an optical fiber will scatter EW of light guided inside the optical fiber if it has multiple phases in microscopic scale. This can be a solid/gas phase structure, likes that in solegel silica coating. It can also be solid/solid multi-phases. Fig. 3 shows the example of moisture cross response of a sensor that has a solid/gas multi-phase coating. Fig. 4 shows the spectral response of a bent optical fiber NH3 sensor having a BCP immobilized PMMA coating as a transducer to air samples containing the same NH3 concentration, but different RH%. In addition to optical absorption spectrum caused by the reaction product of NH3 with BCP (peak absorption wavelength at around 600 nm), EWscattering signals in the broad wavelength range are also observed. This is an example of EW-scattering caused by a coating having micro solid (BCP)/solid (PMMA) heterogeneous structure. In this case, the sensing reagent is trapped as particles (solid phase) in the PMMA membrane (another solid phase). With the increase of relative humidity the solid reagent particles absorb more water, which causes an increase of particle size. As a result, EW-scattering signal increased. In order to make a fiber optic NH3 sensor which is insensitive to moisture change, the transducer coating has to be a single phase. In the case of using a polymer to immobilize a sensing reagent, this requires: 1) the polymer will form a continuous membrane coating; and 2) the sensing reagent has to be soluble in the polymer in solid phase. One such example is a BCP-doped polyvinyl alcohol (PVC) coating. PVC is a hydrophilic polymer due to the large number of eOH groups in the polymer chain. This polymer forms continuous membrane coating on fiber surface, which does not scatter light. BCP is hydrophilic, can be distributed as molecules in the PVC polymer to form a solid solution. Fig. 5 shows the response of a bent optical fiber NH3 sensor having a BCP-doped PVC coating to air samples of different RH% and its response to an air sample containing 1.0 ppmv NH3. This test result clearly demonstrates that the increase of NH3 concentration causes an absorption spectrum with peak absorption wavelength at 600 nm. The change of humidity from RH ¼ 34%e68% does not cause EW-scattering related baseline shift. Therefore, this sensor does not have cross response to moisture. The second approach for eliminating moisture caused EWscattering interference is monitoring sensing signals at two wavelengths. The signal at one wavelength is the sum of EW-scattering
Fig. 4. Spectral response of a bent optical fiber NH3 sensor having a BCP-doped PMMA coating exposed to NH3-containing air samples of different relative humidity. The spectral responses have two components: an EW-scattering signal in broad wavelength range and an absorption spectrum with peak absorption wavelength at 615 nm.
Fig. 5. Spectral responses of a bent optical fiber NH3 sensor having a BCP-doped PVA coating exposed to air samples of different humidity and an air sample containing 1.0 ppmv NH3.
signal and the absorbance of NH3/reagent reaction product. The sensing signal at a second wavelength is caused by EW-scattering only. Since the EW-scattering signal is independent of wavelength in broad wavelength range (Xu et al., 2004), a computer program can be developed to substrate the EW-scattering signal, and give out a sensing signal which is only related to trace NH3 in the air sample. Some sensing reagent immobilized in a polymer itself reacts with water vapor. The reaction product absorbs light and causes spectral response. It will be difficult to eliminate this kind of interference if the absorption wavelength of the waterereagent reaction product is close to that of the NH3ereagent reaction product. Fig. 6 shows such an example. Fe(III)-porphyrin reacts with both water and NH3 to form complexes, and the peak absorption wavelengths of the formed complexes are very close to each other. This phenomenon could happen to many fiber optic sensors using the complex formation of metal ions with NH3 as a sensing mechanism. It is inappropriate to use these sensors in air quality monitoring.
Fig. 6. Spectral response of a bent optical fiber NH3 sensor having an Fe(III)-porphyrin complex doped PMMA coating exposed to air samples of different humidity and NH3 concentration. These test results demonstrate the potential spectral interference of relative humidity change on the sensor for monitoring NH3 in air.
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
3.3. Potential interference of CO2 to fiber optic NH3 sensors In ambient air, CO2 concentration is usually hundreds to thousands of times higher than that of NH3. In addition, CO2 concentration in cattle feeding facilities also continuously changes with time and location. Therefore, in order to develop a fiber optic sensor for monitoring NH3 in air quality program it is necessary to investigate potential interference from CO2 in air. Many fiber optic NH3 sensors use NH3’s basic property to generate a sensing signal. For example, pH indicators have been used in several reported fiber optic sensors as sensing reagents (Huang and Tao, 2011; Tao et al., 2006; McMurray and Albadran, 1999). The reported fiber optic sensors using polyaniline are also based on NH3’s basic property (Yuan and El-Sherif, 2003; Jin et al., 2001). Because CO2 is an acidic oxide it will react with pH indicators and polyaniline in the existence of water. Therefore, it is possible that the change of CO2 concentration in air causes a change of sensing signal and interferes such fiber optic sensor’s function of detecting NH3 in air. On the other hand, CO2 is a very weak coordinating ligand. Optical fiber NH3 sensors based on the formation of coordinating compounds of NH3 with metal compounds are usually free of CO2 interference. Fig. 7 is the spectral response of the BCP-doped solegel silica coating based fiber optic NH3 sensor exposed to NH3-containing air samples of different CO2 concentration. In this case, the change of CO2 concentration from 350 ppm to 1350 ppm does not cause significant interference to the sensor’s function of sensing NH3 in air. Although this sensor uses a pH indicator as a sensing reagent, the pH indicator is distributed on the surface of solegel silica particles surface. Silica particle surface is covered with SieOH, which is a weak acid. The pH indicator molecules on solegel silica particle surface exist as acidic form. The change of trace CO2 concentration in air samples does not change the form of pH indicator molecule on solegel silica particle surface, and thus, does not interfere the sensor’s function of monitoring NH3 in air. The response of PCR/PMMA dual layers coated bent optical fiber sensor probe to CO2 in air was also investigated. It was found that the change of CO2 concentration in air will cause interference of a freshly-made PCR/PMMA probe, but does not cause interference once the probe was exposed to air over-night.It is possible that during the long-time air exposure the PCR in the coating reacted with CO2 (350 ppm) in air, and changed to acidic form. Therefore,
further change of CO2 concentration in air does not cause interference to the sensor’s function of sensing NH3 in air. It was reported that the emission rate of CO2 in cattle feeding facilities is about 200 times that of NH3 (Leytem et al., 2011). Carbon dioxide concentration in cattle feeding facilities changes in the range from 400 ppmv to 700 ppmv. In considering these facts and the test results of this work, it can be conclude that the change of CO2 concentration in AFOs will not interfere the fiber optic sensors for monitoring NH3.
3.4. Effect of temperature on fiber optic NH3 sensor’s response Temperature is another factor that needs to be taken into consideration when designing a fiber optic NH3 sensor for applications in air quality monitoring because: 1) air temperature is continuously changing with time and location; and 2) temperature affects chemical equilibrium constant. Almost all the reported fiber optic NH3 sensors are based on NH3 reaction with a sensing reagent inside a coating on fiber surface. Such a reaction can be described by following equation:
NH3 þ SR5SRNH3
(1)
Where, SR is sensing reagent (metal compound or pH indicating compound) in the coating and SRNH3 is reaction product formed inside the coating. The sensing signal is proportional to the concentration of reaction product, SRNH3. The reaction is reversible and is in equilibrium with gas phase NH3 concentration. The equilibrium constant of this reaction can be expressed as:
Kc ¼
½SRNH3 ½NH3 ½SR
(2)
Because the sensing reagent is usually in large excess compared to trace NH3 in gas phase, the concentration (or quantity) of reaction product, SRNH3, and thus the sensing signal in absorbance, can be expressed as:
Abs: ¼ K½SRNH3 ¼ Kc0 ½NH3
(3)
Where Kc0 is a combined constant including reaction equilibrium constant, absorption coefficient, interaction path-length of EW with the reaction product on fiber’s surface and sensing reagent concentration. Equation (3) indicates that a linear calibration curve (absorbance against NH3 concentration) should be obtained for such a fiber optic NH3 sensor at a decided temperature (constant Kc0 value), and the sensing signal is proportional to NH3 concentration in gas phase. From thermochemistry, the equilibrium constant of a chemical reaction is related to temperature as described by following equation: o 1 DHrxn
lnKc ¼
Fig. 7. Spectral response of a bent optical fiber NH3 sensor having a BCP-doped solegel silica coating as a transducer exposed to air samples of different CO2 concentration. The change of CO2 concentration in air does not cause interference to this sensor’s function of monitoring NH3 in air.
5
R
T
þ
DSorxn R
(4)
o is the standard enthalpy change of the reaction, and Where, DHrxn DSorxn is the standard entropy change of the reaction. Therefore, if a fiber optic NH3 sensor is calibrated at different temperatures, the slope of the calibration curves, which is proportional to Kc, should have a relationship with temperature similar to that described by Equation (4). The effect of temperature on fiber optic NH3 sensor was investigated using a bent optical fiber sensor probe having a BCP-doped solegel silica coating as an example. This sensor has been calibrated for sensing NH3 in air at different temperatures. Fig. 8 shows the obtained calibration curves. A plot of ln Kc0 against reciprocal of
6
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7
function of monitoring trace NH3 in air. This interference can be eliminated through calibrating the sensor at different temperatures to find out the relationship of slope of calibration curves with temperature, and use a temperature sensor together with the fiber optic NH3 sensor in air quality monitoring. From this work it can be concluded that among the four tested fiber optic NH3 sensors developed from our works, the fiber optic NH3 sensor using a bent optical fiber probe coated with BCP-doped solegel silica is most appropriate for monitoring NH3 in animal feeding facilities. In further work we will build prototype sensors and test the longterm stability of the sensors in a real cattle feedlot. Fig. 8. Calibration curves of a bent optical fiber NH3 sensor for detecting NH3 in air of different temperature. Temperature change affects the slope of the calibration curves.
This work is supported by USDA National Institute of Food and Agriculture through award# 2010-65112-20521.
1/T (K -1) 0 0.00326 -1
0.00328
0.0033
0.00332
0.00334
0.00336
0.00338
References
ln Kc'
-2 -3 -4 ln Kc' = 29027 (1/T) - 99.711
2
(R = 0.8955)
-5 -6 Fig. 9. A plot of ln Kc0 (Kc0 is the slope of calibration curves in Fig. 8) against the reciprocal of temperature (1/T). This linear relationship agrees with Equations (3) and (4), which provides a method for calibrating the sensor for detecting NH3 in air of changing temperature.
temperature (1/T) is showing in Fig. 9. A linear relationship of ln Kc0 with (1/T) can be established as:
lnKc0 ¼ 29027
Acknowledgment
1 99:71 R2 ¼ 0:8955 T
(5)
This relationship agrees with Equations (3) and (4). This gives a method to correct temperature effect on fiber optic NH3 sensors. A fiber optic NH3 sensor can first be calibrated in the laboratory at different temperatures to find out the relationship of calibration curve’s slope with temperature before deploying the sensor to for field applications. A temperature sensor can be deployed together with the fiber optic NH3 sensor in the field. The NH3 concentration in air can be calculated from the sensor’s absorbance signal and temperature using Equations (5) and (3) with a computer program.
4. Conclusions From the results of this work, it can be concluded that: 1) Fiber optic NH3 sensors using bent optical fiber probes having pH indicators as sensing reagents can achieve low-ppbv level detection limit, which is sensitive enough for monitoring trace NH3 in air of animal feeding facilities. 2) The change of air CO2 concentration does not affect pH-indicator based fiber optic sensor’s function for monitoring NH3. Carbon dioxide in air also does not affect the function of metal-complex formation based fiber optic NH3 sensors. 3) Humidity level change in air affects some fiber optic sensor’s function for monitoring NH3 in air. This interference can be eliminated by using appropriate sensing reagent/polymer combinations to form solid solution coatings on optical fiber surface, or by measuring optical sensing signals at two wavelengths and using a computer program to obtain a sensing signal caused by NH3 only. 4) Air temperature change affects all fiber optic NH3 sensor’s
Cole, N.A., Clark, R.N., Todd, R.W., Richardson, C.R., Gueye, A., Greene, L.W., McBride, K., 2005. Influence of dietary crude protein concentration and source on potential ammonia emissions from beef cattle manure. Journal of Animal Science 83, 722e731. Gates, R.S., Casey, K.D., Wheeler, E.F., Xin, H., Pescatore, A.J., 2008. U.S. broiler housing ammonia emissions inventory. Atmospheric Environment 42, 3342e3350. Guo, H., Tao, S., 2007. “Silver nanoparticles doped silicate nanocomposites coated on an optical fiber for ammonia sensing”. Sensors & Actuators 123, 578e582. Huang, Y., Tao, S., 2011. An optical fiber sensor probe using a PMMA/CPR coated bent optical fiber as a transducer for monitoring trace ammonia. Journal of Sensor Technogies 1, 29e35. Jin, Z., Su, Y., Duan, Y., 2001. Development of a polyaniline-based optical ammonia sensor. Sensors & Actuators B72, 75e79. Leytem, A.B., Dungan, R.S., Bjorneberg, D.L., Koehn, A.C., 2011. Emissions of ammonia, methane, carbon dioxide, and nitrous oxide from dairy cattle housing and manure management systems. Journal of Environmental Quality 40, 1383e1394. Maeda, Y., Takenaka, N., 1993. Chemiluminescence determination of trace amounts of ammonia and halogen species in the environment. Proceedings of SPIE e The International Society for Optical Engineering 1715 (Optical Methods in Atmospheric Chemistry), pp. 185e193. Malyshev, V.V., Pislyakov, A.V., 2003. Dynamic properties and sensitivity of semiconductor metal-oxide thick-film sensors to various gases in air gaseous medium. Sensors & Actuators B96, 413e434. Marquis, B.T., Vetelino, J.F., 2001. A semiconducting metal oxide sensor array for the detection of NOx and NH3. Sensors & Actuators B77, 100e110. McGinn, S.M., Janzen, H.H., Coates, T., 2003. Atmospheric pollutants and trace gases: atmospheric ammonia, volatile fatty acids, and other odorants near beef feedlots. Journal of Environmental Quality 32, 1173e1182. McMurray, H.N., Albadran, J., 1999. Colorimetric and fluorimetric polymer membrane gas-sensing materials. MRS Bulletin 24, 55e59. Morales-Bahnik, A., Czolk, R., Ache, H.J., 1994. An optochemical ammonia sensor based in immobilized metalloporphyrins. Sensors & Actuators B18-19, 493e496. Mukhtar, S., Mutlu, A., Lacey, E.E., Parnell, C.B., 2009. Seasonal ammonia emissions from a free-stall dairy in central Texas. Journal of the Air & Waste Management Association 59, 613e618. Phillips, V.R., Lee, D.S., Scholtens, R., Garland, J.A., Sneath, R.W., 2001. A review of methods for measuring emission rates of ammonia from livestock buildings and slurry or manure stores, Part 2: monitoring flux rates, concentrations and air flow rates. Journal of Agricultural Engineering Research 78, 1e14. Raimundo, I.M., Narayanaswamy, R., 2001. Simultaneous determination of relative humidity and ammonia in air employing an optical fiber sensor and artificial neural network. Sensors & Actuators B74, 60e68. Renard, J.J., Calidonna, S.E., Henley, M.V., 2004. Fate of ammonia in the atmosphere e a review for applicability to hazardous releases. Journal of Hazardous Materials 108, 29e60. Rhoades, M.B., Parker, D.B., Cole, N.A., Todd, R.W., Caraway, E.A., Auvermann, B.W., Topliff, D.R., Schuster, G.L., 2010. Continuous ammonia emission measurements from a commercial beef feedyard in Texas. Transactions of the ASABE 53, 1823e1831. Roberts, J.M., Langford, A.O., Goldan, P.D., Fehsenfeld, F.C., 1988. Ammonia measurements at Niwot Ridge, Colorado and Point Arena, California using the tungsten oxide denuder tube technique. Journal of Atmospheric Chemistry 7, 137e152. Tao, S., Fanguy, J.C., Xu, L., 2006. Optical fiber ammonia sensing probes using reagent immobilized porous silica coating as transducers. Sensors & Actuators B115, 158e163.
Y. Huang et al. / Atmospheric Environment 66 (2013) 1e7 Tao, S., Gong, S., Fanguy, J.C., Hu, X., 2007. The application of a light guiding flexible tubular waveguide in evanescent wave absorption optical sensing. Sensors & Actuators B120, 724e731. Tao, S., Fanguy, J.C., Sarma, T.V.S., 2008. A fiber optic sensor for monitoring trace ammonia in high temperature gas samples with a CuCl2-doped porous silica optical fiber as a transducer. IEEE Sensors Journal 8, 2000e2007. Varghese, O.K., Gong, D., Dreschel, W.R., Ong, K.G., Keat, G., Grimes, C.A., 2003. Ammonia detection using nanoporous alumina resistive and surface acoustic wave sensors. Sensors & Actuators B94, 27e35.
7
Wathes, C.M., Holden, M.R., Sneath, R.W., White, R.P., Phillips, V.R., 1997. Concentrations and emission rates of aerial ammonia, nitrous oxide, methane, carbon dioxide, dust and endotoxin in UK broiler and layer houses. British Poultry Science 38, 14e28. Xu, L., Fanguy, J.C., Soni, K., Tao, S., 2004. Optical fiber humidity sensor based on evanescent wave scattering. Optics Letters 29, 1191e1193. Yuan, J., El-Sherif, M.A., 2003. Fiber-optic chemical sensor using polyaniline as modified cladding material. IEEE Sensors Journal 3, 5e12. Yuan, J., Feng, B., Lu, J., Liu, H., Liu, D., 1994. Effects of NH3 on N2O formation and destruction in fluidized bed coal combustion. Journal of Thermal Science 3, 278e282.