- Email: [email protected]
Lower detection limit enhancement for low concentration ammonia measurement
Accepted Manuscript Title: Lower Detection Limit Enhancement for Low Concentration Ammonia Measurement Author: H. Manap M.R. Mohamed M.S. Najib PII: DOI: Reference:
S0925-4005(16)32011-1 http://dx.doi.org/doi:10.1016/j.snb.2016.12.049 SNB 21423
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
8-1-2016 5-12-2016 9-12-2016
Please cite this article as: H.Manap, M.R.Mohamed, M.S.Najib, Lower Detection Limit Enhancement for Low Concentration Ammonia Measurement, Sensors and Actuators B: Chemical http://dx.doi.org/10.1016/j.snb.2016.12.049 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Lower Detection Limit Enhancement for Low Concentration Ammonia Measurement. H. Manap, M.R. Mohamed & M. S. Najib. Faculty of Engineering Technology, University of Malaysia, Pahang (UMP), Lebuhraya Tun Razak, 26300 Malaysia. Email: [email protected]
1
Highlights
We report the absorption cross section for ammonia gas in the UV region. We compare ammonia measured spectrum with theoretical data. We recommend the optimum integration time suitable for ammonia concentration measurement for smooth system execution time. We report a few techniques to enhance Lower Detection Limit We report the best response time for sensor system to execute.
Abstract: This paper describes an optical sensor system for quantifying ammonia at low concentration. An open path optical technique is used to measure ammonia concentration within the Ultraviolet region. Experimental results describing the operation of the sensor with wavelengths combination technique to enhance the Lower Detection Limit is presented. The results show the sensor is best measuring ammonia concentration at combination wavelengths (around 212 nm) with the Lower Detection Limit of 4.31 ppm and 1 s response time is achieved. Keywords: optical sensor; ammonia measurement; lower detection limit.
1. Introduction. Ammonia gas is toxic to both human and animal life alike and its maximum safe level is 25 ppm for long term exposure (8 hour) and 35 ppm for short term exposure (15 min) [1]. According to the European Environment Agency, (EEA) report Jan 2014, ammonia (NH3) emissions is primarily contributed by the agricultural sector. Only minimal amounts of ammonia emissions derived from other sectors such as industrial processes and road transport. There are many types of ammonia sensors which have their own advantages and disadvantages and have been discussed in details in previous report [2]. However not many sensor can detect very low concentration within a short duration which is less than 3 s. This is particularly true in sensors based on solid state devices such as semiconductors. In addition, an optical fibre based gas sensor can have many advantages in terms of low weight and small size [3], resistance to high temperature [3-4], no electromagnetic interference, and can have distributed measurement rather than a point sensor [5]. Another advantage of this ammonia sensor is it uses Ultraviolet (UV) as a light source. Barber et al [6] have mentioned that UV absorption is merely affected by water content, which makes this sensor in UV range plausible for operation within moisture-saturated samples. Also, in other report [7] it has been shown that the absorption spectrum for water in the UV range exists between 183 nm to 193 nm. Hence, there should be no cross sensitivity issues with water content since ammonia absorption lines in this work have occurred between 200 nm to 225 nm and this has been proven and reported previously [8]. On top of these advantages, an optimization of the optical sensor system is also needed to enhance the sensor performance. In this paper, we report a few methods to increase the ammonia optical sensor performance by achieving a better Lower Detection Limit (LDL) of 4.31 ppm. We also manage to reduce the response time of the sensor system from 2 s to 1 s. 2
2. Theory Different gas species absorb light at different characteristic wavelengths and for ammonia gas, it has its own specific gas spectrum. A comprehensive collection of absorption cross sections for gaseous molecules can be accessed from the MPI Mainz database including ammonia gas [9]. The data varies from source to source and they depend on temperature and wavelength range. Only one that suits with experimental condition parameter was selected and compared with the measured results. The data for NH3 from the MPI Mainz database was taken as the theoretical spectrum for NH3 gas and is shown in Figure 1 below. For NH3 absorption comparison with the theory, the Beer-Lambert Law has been utilised. The BeerLambert law described the linear relationship between absorbance and concentration of an absorbing species and its general form is shown in (1).
I e ( . N .l ) Io
(1)
Where I is the transmitted intensity, Io is the incident intensity, l(cm) is the distance that the light travels through the gas, σ (cm2/Molecule) is the absorption cross section and N (Molecules/cm3) is the gas concentration. The concentration of test gas, N in the Beer Lambert Law in equation (1) is given in unit Molecules/cm3, so we need to change the unit to ppm as normally the gas concentration is read in this unit. The ideal gas law (PV = nRT ) is also used in this unit conversion, where P = pressure (atm), V = volume (L), n = mass of substance (moles), R = ideal gas constant (0.082 atm L mol-1K-1), T = absolute temperature (ºC + 273) in degrees Kelvin. V RT n P
(0.082
22.4
atm L ) (273 K) mol K 1 atm
(2)
L mol
Standard temperature and pressure are defined as 1 atm and 0 ºC (273 K). Under these conditions, a mole of an ideal gas occupies a volume of 22.4 L. At times, gaseous concentrations are expressed using mixed units of mass per unit volume such as mg/m3. The relationship between ppm and mg/m3 depends on the density of the gas which depends on its pressure, temperature and molecular weight as shown in equation (3).
(
mg ppm 273 P(atm ) ) 3 L T ( K ) 1 atm m 22.4 mol
(3)
where ρ(mg/m3) is the density of the gas, ω is the molecular weight, P(atm) is the measured pressure and T(K) is the measured temperature of the test gas. Equation (3) is commonly used in many text books [10-13] and can be simplified as shown in (4) with assumption the test gas temperature is 298 K and at 1 atm pressure. 3
(
mg ppm ) 3 L m 24.4 mol
(4)
Before we can use (4) and read the concentration in ppm unit, we need to find the relation between the density of the gas, ρ(mg/m3) and the concentration of test gas, N(Molecules/cm3 ). Since N(Molecules/cm3) * (1000 mg/1g) * (1 cm3/10-6m3) * (1 mol/6.022x1023 Molecules) * (ω g/mol) = ρ(mg/m3) Hence the relation between N(Molecules/cm3) and ρ(mg/m3) is given by, N=
N A x 10-9
(5)
where N A is Avogadro’s constant. In order to use the gas concentration, N in parts per million, ppm, (4) and (5) were plugged in the Beer Lambert Law equation (1) above; Hence [ln ppm
I ][ 24.4] Io
N A l 10 9
(6)
Rearrange equation (6) and we get
I ][24.4] Io ppm N A l 10 9 [ln
(7)
Using (7), we can accurately calculate the absorption cross section,σ subject to gas concentration, ppm is known. This method of molecule gas absorption cross section calculation was similarly described in [14-16]. Once σ is determined at selected wavelength, then concentration of ammonia can be measured using equation (6).
3. Experimental Setup The experimental arrangement is shown in Figure 2. A Deuterium-Halogen lamp (DH-2000 from Ocean Optics) was used as a light source. The light was transmitted through a UVNS fibre (Ultra Violet Non Solarising) from CeramOptec Inc. with 644.4 nm core size. Two collimating lens were placed at both ends of the gas cell and were used to focus the incident and transmitted light. The transmitted light then travelled through another optical fibre at the other end of the test gas cell to the light detector. The light detector that was used in this experiment was an Ocean Optics HR2000 spectrometer. The spectrometer has a range from 200 to 650 nm and it provides resolution down to 0.65 nm (FWHM).
4
The spectrometer is interfaced with computer using SpectraSuite software. It is a specifically designed program provided by Ocean Optics in order to acquire the data from the spectrometer in real time. The transmitted intensity, I and the incident intensity, Io were recorded and equation (2) was used to get absorption cross section. These values were plotted against the wavelength and compared with the theory. Two mass flow controllers, (MFC) model F-201CV by Bronkhorst High-Tech were used to control the flowrate of the test gas. The MFCs were powered by 15 V power supply and were operated at the gas pressure above 1 bar. It can be controlled by software provided by Bronkhorst called FlowDDE V4.58. The maximum flow rate of each MFC is 1 l/min. Each of the gas has their own flow rate and the gas mixing calculations are based on the flow constant provided by Bronkhorst.
4. Results and Analysis. In our previuos report [17], it is proven that the absorption lines for ammonia gas measured are similar with theoretical result as shown in Figure 1. It is also reported [18] that this optical sensor system is capable of quantifying the ammonia gas concentration over a wide range of concentrations (0 – 50 ppm). However, the Lower Detection Limit for this sensor system, which was 9 ppm, is still unconvincing to detect ammonia in an open atmosphere such as at a farm where ammonia may be present at much lower concentrations. According to Webber et al [19], the highest ammonia concentration detected in the environmentally controlled chamber at a farm for a two-day period was 8 ppm. Additionally, the ammonia concentration detected on the first day by Webber et al was only at 3 ppm. 4.1 Optimisation of the Optical Sensor System According to Yamamoto et al [20], the ammonia concentration increases exponentially with increasing temperature, with higher concentrations during summer and lower concentrations during winter. Therefore an improvement towards a better Lower Detection Limit was required in order to deliver a sensor that is able to detect ammonia in various weather conditions throughout the year. However the Lower Detection Limit improvements must retain the current sensor performance such as good response time. The trade off between these two criteria (Lower Detection Limit and response time) is always present but is important in order to optimise the performance of this optical sensor system.The main factor that influenced the response time is the integration time which can be set by the user. In this experiment it is set to 2 s as it provides a decent execution time for the software to capture the data. 4.2 Wavelength Combination The initiative taken to reduce the value of the Lower Detection Limit is by combining the intensity measured at two adjacent wavelengths. The idea is to measure intensities at 212.5 nm and 212.7 nm concurrently instead of measuring intensities at a single wavelength, 212.7 nm. These measured intensities are averaged before they are inserted into concentration equation (6). By taking average, the noise and the standard deviation, σ are presumed to be lower which improves the Lower Detection Limit. Measuring 5
intensities at two adjacent wavelengths concurrently will not increase the response time of the optical sensor system. The result is shown in Figure 3. It shows that by averaging measured intensities at two adjacent wavelengths, the noise can be reduced. The peak to peak (P2P) value in Figure 3 (a) is 12.8 ppm which is 3.45 times higher than the P2P in Figure 3 (b) where intensities have been simultaneously measured at two wavelengths. Due to lower noise in Figure 3 (b), the standard deviation of the blank condition will also be reduced accordingly. Hence the Lower Detection Limit of the sensor is reduced to 2.25 ppm, which is 75 percent better than the previous Lower Detection Limit. The main reason for the Lower Detection Limit improvement was due to lower noise produced by the combined wavelengths. The lower noise was obtained because the peaks of the first wavelength, λ1 intensities were combined with the troughs of the second wavelength, λ 2 intensities that produced the averaged intensities which have lower noise. This is clearly shown in Figure 4 below where the averaged intensities for the combined wavelength demonstrate a lower noise. 4.3 Integration Time In another experiment,the integration time was set to 1 s to improve software execution time or response time. The intensities were measured at the same two wavelengths as discussed above. The result using the lower integration time produces a higher noise as shown in Figure 5. The P2P value was 6.7 ppm which is almost twice as high as the previous experiment where 2 s was set as the integration time. Therefore the Lower Detection Limit also increases from 2.25 ppm to 5.25 ppm which decreases the sensor performance. However, by reducing the integration time to 1 s, it has resulted in 50% decrease in software execution time thus making the sensor system more responsive.
4.4 Modified Wavelength Combination Technique As mentioned earlier, the Lower Detection Limit deteriorated when the integration time was set to 1 s. The value was doubled from 2.25 ppm to 5.25 ppm. In this modified version, intensities were measured at four adjacent wavelengths which were 212.3, 212.5, 212.7, 212.9 nm and the integration time was retained at 1 s. The result shows that the noise is still significant as shown in Figure 6. The P2P value is 6.2 ppm and therefore there has been slight reduction from the previous experiment. The Lower Detection Limit was calculated to be at 4.31 ppm which corresponds to almost 18 percent improvement compared to 5.25 ppm in the previous experiment. 4.5 Concentration Measurement Once Detection Limit has been improved, a various concentrations of ammonia gas measurement took place. A few sets of ammonia concentrations (0, 10, 20, 30, 40 and 50 ppm) were achieved by releasing ammonia 6
and N2 at different flow rates which was actuated by controlling both MFCs simultaneously. The source of ammonia gas is from a type E tank (50 x 15 cm) with purity of 50 ppm NH3 in N2. These ammonia concentrations mesurement were compared with commercial ammonia sensor, Tetra and the result is shown in Figure 7. The commercial sensor used in this experiment was purchased from Crowcon Detection Instruments Ltd and the measurement principle is based on electrical properties changes after the sensor plate was exposed to ammonia. Based on the comparison graph, the optical system shows better accuracy compared to Tetra. For every concentrations set by the MFCs, the optical system shows good agreement with the target concentrations with an averaged error of 1.22 %. However, Tetra only shows similar concentration at 10 ppm and the reading differences become larger at higher concentrations as well as the reading errors.
5. Conclusions and Future Work. A novel optical measurement for NH3 gas at different integration time has been described and reported. In the first experiment where integration time was set to 2 s, the Lower Detection Limit was reduced to 2.25 ppm. This shows a significant improvement (75%) when wavelength combination technique is introduced. However when the integration time was lowered down to 1 s, only 18 percent Lower Detection Limit reduction was achieved even though the similar wavelength combination technique is used. The final Lower Detection Limit achieved for 1 s integration time is 4.31 ppm. Although the achieved value is not significant but the response time of this sensor system which is 1 s is improved by 100%. Thus, future work will focus on other methods in order to reduce P2P value and gain better Lower Detection Limit. Finally a full set of experimental tests along with in-situ experiments will be carried out in order to fully quantify the sensing system.
6. Acknowledgement. The authors would like to acknowledge the support of the University of Malaysia, Pahang (UMP) and the Ministry of Higher Education, Malaysia in providing fund for my research studies. Also the authors would like to thank the staff of Electrical Department, Faculty of Engineering Technology, University of Malaysia, Pahang, for their assistance and input during the research.
7
7. References [1] Occupational exposure limits 2000 Guidance Note. EH40/2000, HSE Books, 2000. [2] H Manap, G Dooly, SO Keeffe, E Lewis, “Ammonia detection in the UV region using an optical fiber sensor” Sensors, IEEE, pp. 140-145, 2009. [3] Stewart, G., Jin W., Culshaw B, “Prospects for fibre optic evanescent field gas sensors using absorption in the near-infrared.” Sensors & Actuators B38: pp. 42-47, 1997. [4] O'Farrell, M., E. Lewis, C. Flanagan, W.B. Lyons, N. Jackman, "Design of a System that uses Optical Fibre Sensors and Neural Networks to Control a Large-Scale Industrial Oven by Monitoring the Food Quality Online" IEEE Sensors Journal, 5(6), pp. 1407- 1420, 2005. [5] Grattan, K. T. V. and T. Sun "Fiber optic sensor technology: an overview." Sensors and Actuators A: Physical Vol 82 Issues 1-3 pp. 40-61, 2000. [6] Barber, T. E., Fisher, W. G. and Watcher, E. A.: “On-line monitoring of aromatic hydrocarbons using a near-ultraviolet fiber-optic absorption sensor”, Environ. Sci. Technol. 29(6), pp. 1576–1580, 1995. [7] C.A. Cantrell, A. Zimmer, and G.S. Tyndall, "Absorption cross sections for water vapor from 183 to 193 nm," Geophys. Res. Lett. 24, pp. 2195-2198, 1997. [8] H. Manap, G. Dooly, S O'Keeffe, E. Lewis, "Cross-sensitivity evaluation for NH3 sensing using absorption spectroscopy in the UV region", Sensors & Actuators B: Chem 154 (2), pp. 226-231. 2011. [9] F.Z. Chen, D.L. Judge, C.Y.R. Wu, and J. Caldwell, "Low and room temperature Photoabsorption cross sections of NH3 in the UV region," Planet. Space Sci. 47, pp. 261-266, 1999. [9] E. Hawe, G. Dooly, C. Fitzpatrick, E. Lewis, Paul Chambers. “Measuring of exhaust gas emissions using absorption spectroscopy” Int. J. Intelligent Systems Tech. and Applications, Vol. 3, 2007. [10] Risk Assessment Principles for the Industrial Hygienist by M. A. Jayjock, J. Lynch, pp. 31, 2000. [11] Air monitoring for toxic exposures by Henry J. McDermott, Shirley A. Ness, pp. 157, 2004. [12] Industrial hygiene control of airborne chemical hazards by William Popendorf, pp. 26, 2006. [13] Occupational Hygiene Management Guide by Stanley E. Jones, H. Roland, Hosein, pp. 70, 1993. [14] Gerard Dooly, Elfed Lewis, Colin Fitzpatrick, and Paul Chambers, “Low Concentration Monitoring of Exhaust Gases Using a UV-Based Optical Sensor” IEEE Sensor Journals, Vol. 7, No. 5, 2007. [15] G. Dooly, E. Lewis and C. Fitzpatrick, “On-board monitoring of vehicle exhaust emissions using an ultraviolet optical fibre based sensor” Journal of Optics ,A Pure Application, Opt. 9, S24–S31, 2007. [16] E. Hawe, G. Dooly, C. Fitzpatrick, E. Lewis, P. Chambers,” Measuring of exhaust gas emissions using absorption spectroscopy” Int. J. Intelligent Systems Tech. & Applications, Vol. 3, 2007. [17] Manap H., Dooly, G., O'Keeffe, S., Lewis, E., "Ammonia detection in the UV region using an optical fiber sensor," Sensors, IEEE, pp.140-145, Oct. 2009.
8
[18] O'Keeffe, S.; Manap, H.; Dooly, G.; Lewis, E."Real-time monitoring of agricultural ammonia emissions based on optical fibre sensing technology" ,Sensors, IEEE, pp. 1140-1143, 2010. [19] Michael E Webber, Tyson MacDonald, Michael B Pushkarsky, C Kumar N Patel, Yongjing Zhao, Nichole Marcillac and Frank Mitloehner, “Agricultural ammonia sensor using diode lasers and photoacoustic spectroscopy”, Meas. Sci. Technol. 16, pp. 1547–1553, 2005. [20] Noriko Yamamoto, Hideaki Nishiura, Takahisa Honjo, Yoshiki Ishikawa and Koji Suzuki, “A longterm study of atmospheric ammonia and particulate ammonium concentrations in Yokohama, Japan”, Journal of Atmospheric Environment, Vol. 29, Issue 1, pp. 97-103, 1995.
9
Biography Hadi Manap was born in Kuantan , Malaysia in 1973. He received his first degree in Mathematics and Physics from University of Malaya in 1998. Then he got his Masters degree in Process Plant Management from University of Technology, Malaysia in 2004 and appointed as a teaching staff at Faculty of Electrical & Electronic Engineering, University Malaysia Pahang (UMP). He was sponsored by UMP for PhD studies in the University of Limerick under the supervision of Prof Elfed Lewis and he was awarded his PhD in 2011. Currently he is a senior lecturer in the Faculty of Engineering Technolgy, UMP.
M. S. Najib received his first degree in Electrical and Electronic Engineering from International Islamic University, Malaysia (IIUM) in 2003. Then he further his Master degree in Automation and Control in Newcastle University, UK and graduated in 2005. In 2010, he continue his PhD studies at University of Technology MARA (UiTM) in the field of instrumentation and graduated in 2014. Currently he is a senior lecturer in University of Malaysia, Pahang (UMP) and his research is about an e-nose.
M.R Mohamed is a senior lecturer at the Universiti Malaysia Pahang (UMP) and an active member of the Sustainable Energy & Power Electronics Research (SuPER) Cluster, which he leads one of the sub-cluster i.e. The Sustainable Energy Group (SEG). His research interests include energy storage, electric vehicles, power electronics and drives systems, renewable energy, control systems and engineering education.
10
Fig. 1. Ammonia spectrum, Chen et al, Planet Space Sci. 47 (1999) 261.
Fig. 2: Laboratory experimental setup.
11
Fig. 3: Peak to peak (a) before and (b) after wavelength combination.
Fig. 4 : Peaks and Troughs combination.
12
Fig. 5: Data for 1 s integration time.
Fig. 6: Concentration measurement using four wavelengths.
13
Fig 7 : Concentration measurement & comparison with Tetra sensor.
14
S0925-4005(16)32011-1 http://dx.doi.org/doi:10.1016/j.snb.2016.12.049 SNB 21423
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
8-1-2016 5-12-2016 9-12-2016
Please cite this article as: H.Manap, M.R.Mohamed, M.S.Najib, Lower Detection Limit Enhancement for Low Concentration Ammonia Measurement, Sensors and Actuators B: Chemical http://dx.doi.org/10.1016/j.snb.2016.12.049 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Lower Detection Limit Enhancement for Low Concentration Ammonia Measurement. H. Manap, M.R. Mohamed & M. S. Najib. Faculty of Engineering Technology, University of Malaysia, Pahang (UMP), Lebuhraya Tun Razak, 26300 Malaysia. Email: [email protected]
1
Highlights
We report the absorption cross section for ammonia gas in the UV region. We compare ammonia measured spectrum with theoretical data. We recommend the optimum integration time suitable for ammonia concentration measurement for smooth system execution time. We report a few techniques to enhance Lower Detection Limit We report the best response time for sensor system to execute.
Abstract: This paper describes an optical sensor system for quantifying ammonia at low concentration. An open path optical technique is used to measure ammonia concentration within the Ultraviolet region. Experimental results describing the operation of the sensor with wavelengths combination technique to enhance the Lower Detection Limit is presented. The results show the sensor is best measuring ammonia concentration at combination wavelengths (around 212 nm) with the Lower Detection Limit of 4.31 ppm and 1 s response time is achieved. Keywords: optical sensor; ammonia measurement; lower detection limit.
1. Introduction. Ammonia gas is toxic to both human and animal life alike and its maximum safe level is 25 ppm for long term exposure (8 hour) and 35 ppm for short term exposure (15 min) [1]. According to the European Environment Agency, (EEA) report Jan 2014, ammonia (NH3) emissions is primarily contributed by the agricultural sector. Only minimal amounts of ammonia emissions derived from other sectors such as industrial processes and road transport. There are many types of ammonia sensors which have their own advantages and disadvantages and have been discussed in details in previous report [2]. However not many sensor can detect very low concentration within a short duration which is less than 3 s. This is particularly true in sensors based on solid state devices such as semiconductors. In addition, an optical fibre based gas sensor can have many advantages in terms of low weight and small size [3], resistance to high temperature [3-4], no electromagnetic interference, and can have distributed measurement rather than a point sensor [5]. Another advantage of this ammonia sensor is it uses Ultraviolet (UV) as a light source. Barber et al [6] have mentioned that UV absorption is merely affected by water content, which makes this sensor in UV range plausible for operation within moisture-saturated samples. Also, in other report [7] it has been shown that the absorption spectrum for water in the UV range exists between 183 nm to 193 nm. Hence, there should be no cross sensitivity issues with water content since ammonia absorption lines in this work have occurred between 200 nm to 225 nm and this has been proven and reported previously [8]. On top of these advantages, an optimization of the optical sensor system is also needed to enhance the sensor performance. In this paper, we report a few methods to increase the ammonia optical sensor performance by achieving a better Lower Detection Limit (LDL) of 4.31 ppm. We also manage to reduce the response time of the sensor system from 2 s to 1 s. 2
2. Theory Different gas species absorb light at different characteristic wavelengths and for ammonia gas, it has its own specific gas spectrum. A comprehensive collection of absorption cross sections for gaseous molecules can be accessed from the MPI Mainz database including ammonia gas [9]. The data varies from source to source and they depend on temperature and wavelength range. Only one that suits with experimental condition parameter was selected and compared with the measured results. The data for NH3 from the MPI Mainz database was taken as the theoretical spectrum for NH3 gas and is shown in Figure 1 below. For NH3 absorption comparison with the theory, the Beer-Lambert Law has been utilised. The BeerLambert law described the linear relationship between absorbance and concentration of an absorbing species and its general form is shown in (1).
I e ( . N .l ) Io
(1)
Where I is the transmitted intensity, Io is the incident intensity, l(cm) is the distance that the light travels through the gas, σ (cm2/Molecule) is the absorption cross section and N (Molecules/cm3) is the gas concentration. The concentration of test gas, N in the Beer Lambert Law in equation (1) is given in unit Molecules/cm3, so we need to change the unit to ppm as normally the gas concentration is read in this unit. The ideal gas law (PV = nRT ) is also used in this unit conversion, where P = pressure (atm), V = volume (L), n = mass of substance (moles), R = ideal gas constant (0.082 atm L mol-1K-1), T = absolute temperature (ºC + 273) in degrees Kelvin. V RT n P
(0.082
22.4
atm L ) (273 K) mol K 1 atm
(2)
L mol
Standard temperature and pressure are defined as 1 atm and 0 ºC (273 K). Under these conditions, a mole of an ideal gas occupies a volume of 22.4 L. At times, gaseous concentrations are expressed using mixed units of mass per unit volume such as mg/m3. The relationship between ppm and mg/m3 depends on the density of the gas which depends on its pressure, temperature and molecular weight as shown in equation (3).
(
mg ppm 273 P(atm ) ) 3 L T ( K ) 1 atm m 22.4 mol
(3)
where ρ(mg/m3) is the density of the gas, ω is the molecular weight, P(atm) is the measured pressure and T(K) is the measured temperature of the test gas. Equation (3) is commonly used in many text books [10-13] and can be simplified as shown in (4) with assumption the test gas temperature is 298 K and at 1 atm pressure. 3
(
mg ppm ) 3 L m 24.4 mol
(4)
Before we can use (4) and read the concentration in ppm unit, we need to find the relation between the density of the gas, ρ(mg/m3) and the concentration of test gas, N(Molecules/cm3 ). Since N(Molecules/cm3) * (1000 mg/1g) * (1 cm3/10-6m3) * (1 mol/6.022x1023 Molecules) * (ω g/mol) = ρ(mg/m3) Hence the relation between N(Molecules/cm3) and ρ(mg/m3) is given by, N=
N A x 10-9
(5)
where N A is Avogadro’s constant. In order to use the gas concentration, N in parts per million, ppm, (4) and (5) were plugged in the Beer Lambert Law equation (1) above; Hence [ln ppm
I ][ 24.4] Io
N A l 10 9
(6)
Rearrange equation (6) and we get
I ][24.4] Io ppm N A l 10 9 [ln
(7)
Using (7), we can accurately calculate the absorption cross section,σ subject to gas concentration, ppm is known. This method of molecule gas absorption cross section calculation was similarly described in [14-16]. Once σ is determined at selected wavelength, then concentration of ammonia can be measured using equation (6).
3. Experimental Setup The experimental arrangement is shown in Figure 2. A Deuterium-Halogen lamp (DH-2000 from Ocean Optics) was used as a light source. The light was transmitted through a UVNS fibre (Ultra Violet Non Solarising) from CeramOptec Inc. with 644.4 nm core size. Two collimating lens were placed at both ends of the gas cell and were used to focus the incident and transmitted light. The transmitted light then travelled through another optical fibre at the other end of the test gas cell to the light detector. The light detector that was used in this experiment was an Ocean Optics HR2000 spectrometer. The spectrometer has a range from 200 to 650 nm and it provides resolution down to 0.65 nm (FWHM).
4
The spectrometer is interfaced with computer using SpectraSuite software. It is a specifically designed program provided by Ocean Optics in order to acquire the data from the spectrometer in real time. The transmitted intensity, I and the incident intensity, Io were recorded and equation (2) was used to get absorption cross section. These values were plotted against the wavelength and compared with the theory. Two mass flow controllers, (MFC) model F-201CV by Bronkhorst High-Tech were used to control the flowrate of the test gas. The MFCs were powered by 15 V power supply and were operated at the gas pressure above 1 bar. It can be controlled by software provided by Bronkhorst called FlowDDE V4.58. The maximum flow rate of each MFC is 1 l/min. Each of the gas has their own flow rate and the gas mixing calculations are based on the flow constant provided by Bronkhorst.
4. Results and Analysis. In our previuos report [17], it is proven that the absorption lines for ammonia gas measured are similar with theoretical result as shown in Figure 1. It is also reported [18] that this optical sensor system is capable of quantifying the ammonia gas concentration over a wide range of concentrations (0 – 50 ppm). However, the Lower Detection Limit for this sensor system, which was 9 ppm, is still unconvincing to detect ammonia in an open atmosphere such as at a farm where ammonia may be present at much lower concentrations. According to Webber et al [19], the highest ammonia concentration detected in the environmentally controlled chamber at a farm for a two-day period was 8 ppm. Additionally, the ammonia concentration detected on the first day by Webber et al was only at 3 ppm. 4.1 Optimisation of the Optical Sensor System According to Yamamoto et al [20], the ammonia concentration increases exponentially with increasing temperature, with higher concentrations during summer and lower concentrations during winter. Therefore an improvement towards a better Lower Detection Limit was required in order to deliver a sensor that is able to detect ammonia in various weather conditions throughout the year. However the Lower Detection Limit improvements must retain the current sensor performance such as good response time. The trade off between these two criteria (Lower Detection Limit and response time) is always present but is important in order to optimise the performance of this optical sensor system.The main factor that influenced the response time is the integration time which can be set by the user. In this experiment it is set to 2 s as it provides a decent execution time for the software to capture the data. 4.2 Wavelength Combination The initiative taken to reduce the value of the Lower Detection Limit is by combining the intensity measured at two adjacent wavelengths. The idea is to measure intensities at 212.5 nm and 212.7 nm concurrently instead of measuring intensities at a single wavelength, 212.7 nm. These measured intensities are averaged before they are inserted into concentration equation (6). By taking average, the noise and the standard deviation, σ are presumed to be lower which improves the Lower Detection Limit. Measuring 5
intensities at two adjacent wavelengths concurrently will not increase the response time of the optical sensor system. The result is shown in Figure 3. It shows that by averaging measured intensities at two adjacent wavelengths, the noise can be reduced. The peak to peak (P2P) value in Figure 3 (a) is 12.8 ppm which is 3.45 times higher than the P2P in Figure 3 (b) where intensities have been simultaneously measured at two wavelengths. Due to lower noise in Figure 3 (b), the standard deviation of the blank condition will also be reduced accordingly. Hence the Lower Detection Limit of the sensor is reduced to 2.25 ppm, which is 75 percent better than the previous Lower Detection Limit. The main reason for the Lower Detection Limit improvement was due to lower noise produced by the combined wavelengths. The lower noise was obtained because the peaks of the first wavelength, λ1 intensities were combined with the troughs of the second wavelength, λ 2 intensities that produced the averaged intensities which have lower noise. This is clearly shown in Figure 4 below where the averaged intensities for the combined wavelength demonstrate a lower noise. 4.3 Integration Time In another experiment,the integration time was set to 1 s to improve software execution time or response time. The intensities were measured at the same two wavelengths as discussed above. The result using the lower integration time produces a higher noise as shown in Figure 5. The P2P value was 6.7 ppm which is almost twice as high as the previous experiment where 2 s was set as the integration time. Therefore the Lower Detection Limit also increases from 2.25 ppm to 5.25 ppm which decreases the sensor performance. However, by reducing the integration time to 1 s, it has resulted in 50% decrease in software execution time thus making the sensor system more responsive.
4.4 Modified Wavelength Combination Technique As mentioned earlier, the Lower Detection Limit deteriorated when the integration time was set to 1 s. The value was doubled from 2.25 ppm to 5.25 ppm. In this modified version, intensities were measured at four adjacent wavelengths which were 212.3, 212.5, 212.7, 212.9 nm and the integration time was retained at 1 s. The result shows that the noise is still significant as shown in Figure 6. The P2P value is 6.2 ppm and therefore there has been slight reduction from the previous experiment. The Lower Detection Limit was calculated to be at 4.31 ppm which corresponds to almost 18 percent improvement compared to 5.25 ppm in the previous experiment. 4.5 Concentration Measurement Once Detection Limit has been improved, a various concentrations of ammonia gas measurement took place. A few sets of ammonia concentrations (0, 10, 20, 30, 40 and 50 ppm) were achieved by releasing ammonia 6
and N2 at different flow rates which was actuated by controlling both MFCs simultaneously. The source of ammonia gas is from a type E tank (50 x 15 cm) with purity of 50 ppm NH3 in N2. These ammonia concentrations mesurement were compared with commercial ammonia sensor, Tetra and the result is shown in Figure 7. The commercial sensor used in this experiment was purchased from Crowcon Detection Instruments Ltd and the measurement principle is based on electrical properties changes after the sensor plate was exposed to ammonia. Based on the comparison graph, the optical system shows better accuracy compared to Tetra. For every concentrations set by the MFCs, the optical system shows good agreement with the target concentrations with an averaged error of 1.22 %. However, Tetra only shows similar concentration at 10 ppm and the reading differences become larger at higher concentrations as well as the reading errors.
5. Conclusions and Future Work. A novel optical measurement for NH3 gas at different integration time has been described and reported. In the first experiment where integration time was set to 2 s, the Lower Detection Limit was reduced to 2.25 ppm. This shows a significant improvement (75%) when wavelength combination technique is introduced. However when the integration time was lowered down to 1 s, only 18 percent Lower Detection Limit reduction was achieved even though the similar wavelength combination technique is used. The final Lower Detection Limit achieved for 1 s integration time is 4.31 ppm. Although the achieved value is not significant but the response time of this sensor system which is 1 s is improved by 100%. Thus, future work will focus on other methods in order to reduce P2P value and gain better Lower Detection Limit. Finally a full set of experimental tests along with in-situ experiments will be carried out in order to fully quantify the sensing system.
6. Acknowledgement. The authors would like to acknowledge the support of the University of Malaysia, Pahang (UMP) and the Ministry of Higher Education, Malaysia in providing fund for my research studies. Also the authors would like to thank the staff of Electrical Department, Faculty of Engineering Technology, University of Malaysia, Pahang, for their assistance and input during the research.
7
7. References [1] Occupational exposure limits 2000 Guidance Note. EH40/2000, HSE Books, 2000. [2] H Manap, G Dooly, SO Keeffe, E Lewis, “Ammonia detection in the UV region using an optical fiber sensor” Sensors, IEEE, pp. 140-145, 2009. [3] Stewart, G., Jin W., Culshaw B, “Prospects for fibre optic evanescent field gas sensors using absorption in the near-infrared.” Sensors & Actuators B38: pp. 42-47, 1997. [4] O'Farrell, M., E. Lewis, C. Flanagan, W.B. Lyons, N. Jackman, "Design of a System that uses Optical Fibre Sensors and Neural Networks to Control a Large-Scale Industrial Oven by Monitoring the Food Quality Online" IEEE Sensors Journal, 5(6), pp. 1407- 1420, 2005. [5] Grattan, K. T. V. and T. Sun "Fiber optic sensor technology: an overview." Sensors and Actuators A: Physical Vol 82 Issues 1-3 pp. 40-61, 2000. [6] Barber, T. E., Fisher, W. G. and Watcher, E. A.: “On-line monitoring of aromatic hydrocarbons using a near-ultraviolet fiber-optic absorption sensor”, Environ. Sci. Technol. 29(6), pp. 1576–1580, 1995. [7] C.A. Cantrell, A. Zimmer, and G.S. Tyndall, "Absorption cross sections for water vapor from 183 to 193 nm," Geophys. Res. Lett. 24, pp. 2195-2198, 1997. [8] H. Manap, G. Dooly, S O'Keeffe, E. Lewis, "Cross-sensitivity evaluation for NH3 sensing using absorption spectroscopy in the UV region", Sensors & Actuators B: Chem 154 (2), pp. 226-231. 2011. [9] F.Z. Chen, D.L. Judge, C.Y.R. Wu, and J. Caldwell, "Low and room temperature Photoabsorption cross sections of NH3 in the UV region," Planet. Space Sci. 47, pp. 261-266, 1999. [9] E. Hawe, G. Dooly, C. Fitzpatrick, E. Lewis, Paul Chambers. “Measuring of exhaust gas emissions using absorption spectroscopy” Int. J. Intelligent Systems Tech. and Applications, Vol. 3, 2007. [10] Risk Assessment Principles for the Industrial Hygienist by M. A. Jayjock, J. Lynch, pp. 31, 2000. [11] Air monitoring for toxic exposures by Henry J. McDermott, Shirley A. Ness, pp. 157, 2004. [12] Industrial hygiene control of airborne chemical hazards by William Popendorf, pp. 26, 2006. [13] Occupational Hygiene Management Guide by Stanley E. Jones, H. Roland, Hosein, pp. 70, 1993. [14] Gerard Dooly, Elfed Lewis, Colin Fitzpatrick, and Paul Chambers, “Low Concentration Monitoring of Exhaust Gases Using a UV-Based Optical Sensor” IEEE Sensor Journals, Vol. 7, No. 5, 2007. [15] G. Dooly, E. Lewis and C. Fitzpatrick, “On-board monitoring of vehicle exhaust emissions using an ultraviolet optical fibre based sensor” Journal of Optics ,A Pure Application, Opt. 9, S24–S31, 2007. [16] E. Hawe, G. Dooly, C. Fitzpatrick, E. Lewis, P. Chambers,” Measuring of exhaust gas emissions using absorption spectroscopy” Int. J. Intelligent Systems Tech. & Applications, Vol. 3, 2007. [17] Manap H., Dooly, G., O'Keeffe, S., Lewis, E., "Ammonia detection in the UV region using an optical fiber sensor," Sensors, IEEE, pp.140-145, Oct. 2009.
8
[18] O'Keeffe, S.; Manap, H.; Dooly, G.; Lewis, E."Real-time monitoring of agricultural ammonia emissions based on optical fibre sensing technology" ,Sensors, IEEE, pp. 1140-1143, 2010. [19] Michael E Webber, Tyson MacDonald, Michael B Pushkarsky, C Kumar N Patel, Yongjing Zhao, Nichole Marcillac and Frank Mitloehner, “Agricultural ammonia sensor using diode lasers and photoacoustic spectroscopy”, Meas. Sci. Technol. 16, pp. 1547–1553, 2005. [20] Noriko Yamamoto, Hideaki Nishiura, Takahisa Honjo, Yoshiki Ishikawa and Koji Suzuki, “A longterm study of atmospheric ammonia and particulate ammonium concentrations in Yokohama, Japan”, Journal of Atmospheric Environment, Vol. 29, Issue 1, pp. 97-103, 1995.
9
Biography Hadi Manap was born in Kuantan , Malaysia in 1973. He received his first degree in Mathematics and Physics from University of Malaya in 1998. Then he got his Masters degree in Process Plant Management from University of Technology, Malaysia in 2004 and appointed as a teaching staff at Faculty of Electrical & Electronic Engineering, University Malaysia Pahang (UMP). He was sponsored by UMP for PhD studies in the University of Limerick under the supervision of Prof Elfed Lewis and he was awarded his PhD in 2011. Currently he is a senior lecturer in the Faculty of Engineering Technolgy, UMP.
M. S. Najib received his first degree in Electrical and Electronic Engineering from International Islamic University, Malaysia (IIUM) in 2003. Then he further his Master degree in Automation and Control in Newcastle University, UK and graduated in 2005. In 2010, he continue his PhD studies at University of Technology MARA (UiTM) in the field of instrumentation and graduated in 2014. Currently he is a senior lecturer in University of Malaysia, Pahang (UMP) and his research is about an e-nose.
M.R Mohamed is a senior lecturer at the Universiti Malaysia Pahang (UMP) and an active member of the Sustainable Energy & Power Electronics Research (SuPER) Cluster, which he leads one of the sub-cluster i.e. The Sustainable Energy Group (SEG). His research interests include energy storage, electric vehicles, power electronics and drives systems, renewable energy, control systems and engineering education.
10
Fig. 1. Ammonia spectrum, Chen et al, Planet Space Sci. 47 (1999) 261.
Fig. 2: Laboratory experimental setup.
11
Fig. 3: Peak to peak (a) before and (b) after wavelength combination.
Fig. 4 : Peaks and Troughs combination.
12
Fig. 5: Data for 1 s integration time.
Fig. 6: Concentration measurement using four wavelengths.
13
Fig 7 : Concentration measurement & comparison with Tetra sensor.
14