A chemometric approach to the calibration of LED-based optical gas

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Letter

A chemometric approach to the calibration of LEDbased optical gas sensors using HITRAN data Parvez Mahbub, John Leis, and Mirek Macka Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b01295 • Publication Date (Web): 30 Apr 2018 Downloaded from http://pubs.acs.org on May 1, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1 2 3

A chemometric approach to the calibration of LED-based optical gas sensors using HITRAN data Parvez Mahbub1,2,*, John Leis3, Mirek Macka1,4

4 5

1

6 7

2

8 9

3

10 11 12

4

13

*corresponding author: [email protected]

14

Australian Centre for Research on Separation Science (ACROSS) and School of Natural Sciences, University of Tasmania, Private Bag 75, Hobart 7001, Australia Institute of Sustainability and Innovation, Victoria University, Footscray Park Campus, Melbourne, Victoria 3011, Australia School of Mechanical and Electrical Engineering, University of Southern Queensland, Australia Department of Chemistry and Biochemistry, Mendel University in Brno, Zemedelska 1, CZ 613 00 Brno, Czech Republic

Abstract

15

Modelling the propagation of light from LED sources is problematic since the emission

16

covers a broad range of wavelengths and thus cannot be considered as monochromatic.

17

Furthermore, the lack of directivity of such sources is also problematic. Both attributes are

18

characteristic of LEDs. Here we propose a HITRAN (high-resolution transmission molecular

19

absorption database) based chemometric approach that incorporates not-perfect-

20

monochromaticity and spatial directivity of near-infrared (NIR) LED for absorbance

21

calculations in 1% to 6% methane (CH4) in air, considering CH4 as a model absorbing gas.

22

We employed the absorbance thus calculated using HITRAN to validate the experimentally

23

measured absorbance of CH4. The maximum error between the measured and calculated

24

absorbance values were within 1%. The approach can be generalised as a chemometric

25

calibration technique for measuring gases and gas mixtures that absorb emissions from

26

polychromatic or not-perfect-monochromatic sources, provided the gas concentration, optical

27

pathlength, as well as blank and attenuated emission spectra of the light source are

28

incorporated into proposed chemometric approach.

29

Keywords: HITRAN, chemometrics, calibration of gas sensor, not-perfect-

30

monochromaticity, spatial directivity, optical pathlength, emission spectra of LED 1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 13

31

Remote monitoring of hazardous gases using optical sensors based on the theories of

32

absorption spectroscopy has become a routine task in modern-day industrial environments

33

such as mining, landfills as well as in laboratories, office spaces, medical care facilities, and

34

livestock enclosures. The absorbance-based gas detection is mainly undertaken in the infrared

35

(IR) region due to the selective absorption of IR radiation by a wide range of gas molecules.

36

In this context, the absorption parameters of gas molecules (absorption cross-section and line

37

intensity) at specified pressures and temperatures are tabulated using spectroscopic databases

38

such as HITRAN* and employed for establishing HITRAN-based (theoretical) calculations of

39

absorbance of gas molecules at different concentrations and pathlengths using the

40

fundamental principle of absorption of light (known as the Bouguer-Beer-Lambert law). For

41

in-situ gas sensing by optical sensors, this is very straightforward for monochromatic light

42

sources such as lasers,2-3 particularly due to high degree of coherency, monochromaticity and

43

directivity of laser sources. Mizaikoff and co-workers employed quantum cascade lasers with

44

hollow waveguides for in-situ and laboratory measurement of different gases such as

45

benzene, toluene, xylene, methane, propane, butane, carbon dioxide, cyclopropane and

46

isobutylene.4-6 Although lasers are predominantly used for gas sensing in IR spectral region,

47

Light Emitting Diodes (LEDs), whose emission spectra is not strictly monochromatic, are

48

increasingly becoming a preferred choice as light sources, since LEDs facilitate portability,

49

low-cost and less energy consumption.7 The not-perfect-monochromaticity (polychromatic

50

emission spreading across a narrow wavelength range) and spatial directivity (the angle of

51

spread of radiation in 3D space) associated with the LEDs employed in optical sensors is

52

characterised by the bandwidth (full width at half of maximum emission intensity) and

53

divergent optical paths, respectively. Hence, absorbance calculations without incorporating

54

the not-perfect-monochromaticity and spatial directivity using spectral databases such as *

HITRAN is a molecular absorption database which compiles the spectroscopic parameters of different gases and gas mixtures and utilises computer codes to predict and simulate the transmission and emission of light in the atmosphere.1 2 ACS Paragon Plus Environment

Page 3 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


55

HITRAN would incorporate a systematic bias. Mayrwöger et al.8 reported HITRAN-based

56

absorbance calculation for gases where absorbance of CO2 was simulated using ray tracing

57

software (ZEMAX®) that required complex post-processing of ray tracing data. Such

58

simulations become complicated since the thermal IR emitters using heated filaments do not

59

feature a narrowly-formed radiation pattern in the study reported. In this context, LEDs offer

60

a better choice since the IR emission can be modulated in the kHz range, which substantially

61

reduces surface heating of the emitter. To date, there have been numerous illustrations of the

62

use of mid infrared (MIR) and near infrared (NIR) LEDs for sensing of various gases such as

63

CH4 and CO2.9-13 However, to the best of authors’ knowledge, no scientific investigations

64

involving LED-based optical gas sensors to date has demonstrated accurate calculation of

65

absorbance of gases from the fundamental principles of light absorption by gaseous

66

molecules, taking the not-perfect-monochromaticity and spatial directivity of LEDs into

67

consideration simultaneously. In most LED based gas sensing studies, a single-point

68

calibration and validation was reported,11-12 which did not provide any concrete evidence on

69

the accuracy of the sensor calibration over certain concentrations ranges. Therefore, while

70

LED emitters present the possibility of low-cost and rugged sources for IR gas detection, the

71

principles for accurate absorbance-based gas detection are not yet fully established. In this

72

short communication, we present as a proof of principle, a chemometric approach to

73

calculating of methane (CH4) absorbances using HITRAN data, taking not-perfect-

74

monochromatic emission and spatial directivity of the LED into consideration for the first

75

time, using a rapidly pulsed NIR LED-based CH4 sensor within a closed optical path.14 We

76

also demonstrate that our approach to calculation of the absorbance of gases for a wide

77

concentration range using the HITRAN database lays the foundation for simple and accurate

78

chemometric approach of calculating absorbance of gases or gas mixtures using LEDs.

3 ACS Paragon Plus Environment


79

Page 4 of 13

Not-perfect-monochromaticity of NIR LEDs

80

As LEDs’ emission spectra exhibit polychromaticity across a narrow wavelength range, the

81

emission maxima cannot be regarded as a single parameter for the purposes of calculation of

82

the absorbance of a species at different concentrations using the Bouguer-Beer-Lambert law.

83

In this study, we integrated the area under the attenuated emission spectra of the employed

84

LED (Lms16LED-R, Alfa Photonics, Latvia) in the range of 1% to 6% CH4 in air and

85

calculated the absorbance as the negative logarithm of the ratio of area under the attenuated

86

emission spectra compared to the blank emission spectra of the LED source. We derived the

87

LED attenuated emission spectra in 1% to 6% CH4 as a product of LED emission spectra in

88

blank and transmittance (i.e., 10

89

HITRAN generated absorption cross-section, gas concentration and optical pathlength as

90

illustrated in eqn.1. The absorbance values at a particular wavelength, for a specific HITRAN

91

absorption cross-section and gas concentration, were integrated as a function of optical

92

pathlength (which varies according to the directivity of the LED and discussed in the next

93

section). The total absorbance was then calculated according to eqn.2. The detailed procedure

94

of calculating the LED attenuated emission spectra in 1% to 6% CH4 is described in the

95

electronic supplementary information (ESI).

-absorbance

), where the absorbance was calculated from

=

× 10( ! "#$%&'()%* +&%$$ $,+()%* ×-"$ +%*+,*(&"()%* ×%'()+". '"(/.,*-(/) (1)

96 &," 7*8,& (/, 9:; "((,*7"(,8 ,

(2)

(A-F) in the ESI.


Page 5 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


101

Insert Figure 1

102

In Fig. 1, we observe that the emission spectra of the NIR LED (Lms16LED-R, Alfa

103

Photonics, Latvia) is spread across the 1.4 to 1.8 µm range. In this wavelength range, we

104

acquired HITRAN generated absorptivity spectra of CH4 (known as absorption cross-section)

105

at 1 atm pressure and 25° C temperature. The calculated absorbance (which is a function of

106

absorption-cross section, gas concentration and optical pathlength) can vary significantly at a

107

particular wavelength, depending on the variation of the optical pathlength travelled by light

108

from emitter to detector. In the next section, we demonstrate the effects of spatial directivity

109

of the IR LED on the optical pathlength in an 80 cm long electropolished aluminium tube

110

used in an experimental CH4 sensor.14

111

Spatial Directivity of IR LEDs

112

The spatial directivity plot of LEDs is an established method developed by the Commission

113

Internationale de l’Éclairage (CIE) to indicate the angular spread of the emitted light.15 Figure

114

S2 in the ESI illustrates the spatial directivity of NIR LED employed in this study, where the

115

IR rays are emitted at angles from 0° to 70°.16 We calculated the optical pathlengths travelled

116

by the rays emitted at 1° intervals between 0° and 70° using trigonometric rules (for details,

117

see Fig. S3 and associated calculations in the ESI). For calculating optical pathlength, we

118

initially employed the viewing angle of the LED as a guide for considering each possible path

119

travelled by the rays emitted from the LED. The viewing angle is defined as the angle where

120

the measured light intensity is 50% of its maximum value.17 The 35° viewing angle of the

121

NIR LED in this study covers ~90% (according to Fig. S2 in ESI) of the emission. However,

122

considering the rays emitted at angles between 0° and 35° can only serve as a first estimate of

123

the HITRAN based absorbance calculation of CH4 as it resulted unsatisfactory agreement

124

between the experimentally measured and the calculated absorbance values (Fig. 2A). We

125

then included more light paths (~95%) into the HITRAN based absorbance calculations by



Page 6 of 13

126

considering rays emitted at angles between 0° and 60°. Nonetheless, the error in measured

127

absorbance was ±40% when compared to the calculated absorbances (Fig. 2B). Further, we

128

included more light paths (~99%) into the HITRAN based absorbance calculations by

129

considering rays emitted at angles between 0° and 70°, which reduced the error to ±1%

130

between the experimental and calculated absorbance values over the range of 1% to 6% CH4

131

as demonstrated in Fig.2C. Hence, a careful observation of the emitted rays in the spatial

132

directivity plot of LED is a crucial step in the chemometric approach using the HITRAN

133

database for accurate calculation of absorbance values.

134

Insert Figure 2

135

The simplified approach of including the maximum amount of light emitted from the LED

136

based on the spatial directivity plots as explained above yields an agreement with the

137

experimentally established calibration as shown in Fig. 2C. This approach is relatively

138

straightforward and does not require use of any specialised software (such as ZEMAX or

139

similar) as employed by Mayrwöger et al.8 for ray tracing.

140

Proof-of-Principle

141

In this study, the NIR LED emits according to vendor’s specification between 1400 nm to

142

1800 nm with maximum emission at 1660 nm (Fig. S1A in ESI). The bandwidth of this NIR

143

LED at half of its maximum emission is approximately 130 nm (from 1580 nm to 1710 nm)

144

which is significantly larger than other UV to near infrared LEDs used in other studies (e.g.,

145

see Table S1 in the supplementary information of Noori et al.20). We have used CH4 as a

146

model gas and in Fig.1, we observe that some considerable portion of the HITRAN generated

147

CH4 absorption cross-section peaks fell within the bandwidth of the LED employed. As the

148

HITRAN generated spectra of absorption cross-section change for different gases, the not-

149

perfect-monochromaticity of the LED plays a significant role in the absorption and hence,

150

needs to be considered in modelling approach for LED emission spectra calculation. In this 6 ACS Paragon Plus Environment

Page 7 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


151

study, we have demonstrated that the not-perfect-monochromaticity of LED sources can be

152

better modelled for the gas detection application by employing the chemometric approach of

153

calculating the emission spectra of the LED using eqn. 1. Additionally, the absorbance values

154

at a specific wavelength varied as a function of optical pathlength, and we have demonstrated

155

that the spatial directivity of IR LEDs can be incorporated via the proposed chemometric

156

approach.

157

Although we achieved excellent agreement between the calculated absorbance and the

158

experimentally measured absorbance values for 1% to 6% CH4 by considering rays emitted

159

between 0° and 70° that captured ~99% of the emission from LED, still there could be some

160

insignificant amount of light from outside the limits of this angular spread of emission which

161

may not be captured or detected. In this regard, we note the work of Moreno and co-

162

workers,18, 19 who demonstrated three basic electromagnetic radiation zones: near-field, mid-

163

field and far-field regarding the modelling of propagation of light from LEDs. Please check

164

ESI for detailed discussions on the effects of such extraneous electromagnetic regions on the

165

propagation of light from the LED.

166

Conclusions

167

Although LEDs are relatively low-cost, low-power light sources and facilitate portability of

168

optical gas sensors, the not-perfect-monochromaticity and spatial directivity of LED emitters

169

are yet to be considered simultaneously in absorption-based gas measurements. Through

170

incorporating the attenuation of the LED emission spectra across a narrow absorption band

171

and the directivity of the LED into HITRAN based absorbance calculations, this study

172

establishes a framework for experimental calibration curves used for IR LED-based optical

173

sensors. The significance of this approach is that it is applicable to a wide range of gas

174

sensors employing polychromatic or not-perfect-monochromatic LEDs as light sources and to

175

gases or gas mixtures that absorb in UV, visible or infrared wavelengths. This study has



Page 8 of 13

176

established that the considerations of not-perfect-monochromaticity and spatial directivity of

177

LEDs are desirable to develop chemometric approach to enhancing accuracy in such sensors.

178

Acknowledgements

179

This research did not receive any specific grant from funding agencies in the public,

180

commercial, or not-for-profit sectors. MM acknowledges his ARC Future Fellowship

181

(FT120100559) to support his contributions to this study.

182

Figure Captions

183 184 185 186

Figure 1 NIR LED attenuated emission spectra (1400-1800 nm) in 6% 12CH4 (natural abundance 98.82%) calculated from the HITRAN generated absorption cross sections of CH4 within the 1400-1800 nm wavelength range, gas concentrations and optical pathlength (see ESI for HITRAN simulation details).

187 188 189 190 191 192 193

Figure 2 Absorbance values of pure CH4 experimentally measured by rapid pulsing of NIR LED compared to the HITRAN calculated absorbance values of different CH4 concentrations (% CH4 converted to Molecule.cm-3) considering rays emitted between (A) 0° to 35°, (B) 0° to 60° and (C) 0° to 70° from the NIR LED. The HITRAN data show unsatisfactory agreement (A), within ±40% of the experimentally measured CH4 absorbance data (B) and within ±1% of the experimentally measured CH4 absorbance data (C). The error bars represent general experimental error of ±1%.


Page 9 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


194

Figures

195 196

Figure 1



197

199

200

201

0.012

A

0.012

Experimental

0.010

HITRAN

0.008

Linear (Experimental)

0.006

Linear (HITRAN)

y = 0.157x - 0.0003 R² = 0.9983

0.004 y = 0.0108x + 5E-07 R² = 1

0.002 0.000

202

B

Experimental

Absorbance, a.u.

198

Absorbance, a.u.

0.010

y = 0.157x - 0.0003 R² = 0.9983

HITRAN

0.008


0.006

Linear (HITRAN)

0.004 y = 0.0656x + 2E-05 R² = 1

0.002 0.000

0%

1%

2%

3%

4%

5%

6%

0%

1%

2%

Methane, %

3%

4%

5%

6%

Methane, %

203

0.012

205

206

207

C

Experimental

204

Absorbance, a.u.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Page 10 of 13

0.010

HITRAN

0.008


y = 0.157x - 0.0003 R² = 0.9983

Linear (HITRAN)

0.006

y = 0.1484x + 9E-05 R² = 0.9999

0.004 0.002

208

0.000 209

Figure 2

0%

1%

2%

3%

4%

5%

6%

Methane, % 10

ACS Paragon Plus Environment

Page 11 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257


References (1) (2) (3)

(4) (5) (6) (7) (8) (9) (10) (11) (12) (13)

(14)

(15) (16)

(17)

(18) (19)

Rothman, L. S.; Gordon, I. E.; Babikov, Y. et al., J. Quant. Spectros. Radiat. Transfer. 2013, 130, 4-50. Zeninari, V.; Parvitte, B.; Joly, L. et al., Appl. Phys. B: Lasers Opt. 2006, 85, 265–272. Fleisher, A. J.; Hodges, J. T. Advances in Spectroscopic Methods for Gas Sensing, In proceedings CCQM-GAWG Workshop 2016 Standards and Measurements for Clean Air, Instituto Português da Qualidade (IPQ) Caparica, Portugal, 2016, www.bipm.org/ws/CCQM/CLEANAIR/Allowed/October.../GAWG161014_07.p df, accessed 1 July 2017. Wilk, A.; Carter, J. C.; Chrisp, M. et al. Anal. Chem. 2013, 85, 11205−11210. Ribessi, R. L.; Neves, T. A.; Rohwedder, J. J. R. et al. Analyst, 2016, 141, 5298– 5303. Young, C. R.; Menegazzo, N.; Riley, A. E. et al. Anal. Chem. 2011, 83, 61416147. Macka, M.; Piasecki, T.; Dasgupta, P. K. Annu. Rev. Anal. Chem. 2014, 7, 183207. Mayrwöger, J.; Hauer, P.; Reichl, W.; Schwödiauer, R.; Krutzler, C.; Jakoby, B. IEEE Sens. J., 2010, 10, 1691-1698. Johnston, S. F. Meas. Sci. Technol. 1992, 3, 191. Fanchenko, S.; Baranov, A.; Savkin, A.; Sleptsov, V. IOP Conf. Ser. Mater. Sci. Eng. 2016, 108, 012036. Massie, C.; Stewart, G.; McGregor, G.; Gilchrist, J. R. Sensor. Actuat. B-Chem. 2006, 113, 830–836. Leis, J.; Buttsworth, D.; Snook, C.; Holmes, G. IEEE Trans. Instrum. Meas., 2014, 63, 3088-3095. Okajima, H.; Kakuma, S.; Uchida, K.; Wakimoto, Y.; Noda, K. Measurement of methane gas concentration using an infrared LED, In proceedings 2006 SICEICASE International Joint Conference, Busan, South Korea, 2006, pp 1652-1655. Noori, A. (2018). Radiometric analysis of LEDs and the use of rapidly pulsed infra-red LEDs for portable sensing of gases, PhD Thesis, Hobart, University of Tasmania. CIE (2007). Measurement of LEDs, Technical Report Commission Internationale de l’Éclairage CIE 127: 2007, 2nd edition. ISBN 978 3 901 906 58 9 LED Microsensor (2017). Beam divergence data of NIR LED in TO-18 package with cap, LED Microsensor NT, Latvia, http://lmsnt.com/leds1600-5000/beamdivergence/, accessed 5 July 2017. Van. Heek, B. (2014) White Paper of Viewing Angle Measurement, Barco N.V. & invidis consulting GMBH 81671 Munich, Germany, 2014 pp.8, https://www.barco.com/~/media/downloads/white%20papers/2016/whitepaper%2 0viewing%20angle_v2%20docx.pdf, accessed 1 Feb 2018. Sun, C.-C., Chien, W.-T., Moreno, I., Hsieh, C.-C., Lo, Y.-C. Opt. Express, 2009, 17, 13918-13927. Moreno, I., Sun, C.-C. (2008). LED array: where does far-field begin? In Proc. SPIE 7058, Eighth International Conference on Solid State Lighting, 70580R, Ferguson, I. T., Taguchi, T., Ashdown, I, E., Park, S.-J. (Editors), 2 September 2008, doi: 10.1117/12.795944 11 ACS Paragon Plus Environment


258 259

Page 12 of 13

(20) Noori, A., Mahbub, P., Dvořák, M., Lucieer, A., Macka, M. Sensor. Actuat. BChem., 2018, 262, 171-179.


Page 13 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

260


For TOC Only

261 262


A chemometric approach to the calibration of LED-based optical gas

Recommend Documents