-
根据比尔-朗伯(Beer-Lambert)定律, 强度为I0(λ), 波长为λ的激光经过长为L的光程后, 在探测端得到的光强大小It(λ)可表示为:
$ {I_{\rm{t}}}\left( \lambda \right) = {I_{\rm{0}}}\left( \lambda \right)\left[ { - \alpha \left( \lambda \right)CL} \right] $
(1) 式中, C为目标气体所占总气体的体积分数; α(λ)为气体的吸收系数, 与激光的波长、被测气体的种类有关。对于低含量弱吸收, 满足α(λ)CL≪ 1, 则(1)式可化简为:
$ {I_{\rm{t}}}\left( \lambda \right) = {I_{\rm{0}}}\left( \lambda \right)\left[ {1 - \alpha \left( \lambda \right)CL} \right] $
(2) 根据谐波检测技术, 通过对激光器注入电流进行小信号调制, 激光器的输出波长和输出光强也受到调制, 则经过气体吸收后的光强为:
$ \begin{array}{*{20}{c}} {{I_{\rm{t}}}\left( \lambda \right) = \overline {{I_{\rm{0}}}\left( \lambda \right)} \left( {1 + A\cos \theta } \right)\left[ {1 - } \right.}\\ {\left. {\alpha \left( {{\lambda _0} + \alpha \cos \theta } \right)CL} \right]} \end{array} $
(3) 式中, λ0为激光中心波长, $ \overline{{{I}_{0}}\left( \lambda \right)}$为激光中心波长处对应的输出光强, A为光强调制深度, θ为调制角度, a为波长调制深度。对(3)式进行傅里叶级数展开, 可得到各次谐波成分, 其中各次谐波的信号幅度可表达为[14]:
$ {F_n}\left( \lambda \right) = \frac{1}{n}\int_{ - {\rm{ \mathsf{ π} }}}^{\rm{ \mathsf{ π} }} {{I_{\rm{t}}}\left( \lambda \right)\cos \left( {n\theta } \right){\rm{d}}\theta } $
(4) 式中, n为大于或者等于1的正整数。
理论上, 只要知道吸收系数的线型, 通过(3)式、(4)式就能算出谐波分量与气体体积分数的关系。
-
相比于甲烷等简单分子来说, 丙烷的分子结构相对复杂, 是典型的含有CH键的分子, 分子吸收线之间的间隔非常窄, 会相互交叉形成宽谱吸收[15], 其在近红外波段的吸收光谱来自PNNL数据库[16], 如图 1所示。可见丙烷在1686.00nm~1687.00nm约1nm宽的范围内有一个宽谱吸收峰, 为了对比在这个范围内空气中其它常见气体的吸收, 图中还列出了CO2, CH4, H2O这3种气体的吸收光谱, 除了CH4在1687.30nm附近有一个明显的窄带吸收尖峰外, 其它范围内这3种气体吸收都很弱, 因此, 可利用丙烷在该范围内的宽谱吸收峰作为探测的依据。
Figure 1. Absorption spectrum of propane, carbon dioxide, methane and water near 1686nm[16]
为了分析丙烷在1686.00nm~1687.00nm范围内的宽谱吸收线型, 采用宽谱光源发出连续光, 用光谱仪分别测量丙烷吸收前后的光强大小, 根据(1)式算出吸收系数α(λ), 可拟合得到一个随波长变化的洛伦兹线型, 如图 2所示。
$ \alpha \left( \lambda \right) = {\alpha _0} + \frac{{{k_0}{\gamma _{\rm{L}}}/{\rm{ \mathsf{ π} }}}}{{{{\left( {\lambda - {\lambda _0}} \right)}^2} + \gamma _{\rm{L}}^2}} $
(5) 式中, 拟合得到的γL为宽谱吸收峰的半峰半宽, 大小为0.21nm, 中心波长λ0=1686.42nm, 对应的吸收系数大小α(λ0)=0.117cm-1, 以及常数项α0=0.068cm-1, k0=0.032。
当激光器的输出波长对应着宽谱吸收峰中心波长λ0时, 通过(5)式拟合得到的吸收系数线型, 再代入到(3)式、(4)式中, 可推导得到一次谐波信号幅度F1及二次谐波信号幅度F2与气体体积分数C的关系:
$ {F_1} = {I_0}\left( \lambda \right)\left[ {\bar Q - \bar PC} \right] $
(6) $ {F_2} = {I_0}\left( \lambda \right)\left[ {\bar W - \bar UC} \right] $
(7) F1和F2与光功率I0(λ)大小有关, 且都与丙烷气体体积分数C成线性关系, 当系统的调制深度A和a一定时, 对于丙烷在1686nm的中心波长处, 有常数项$\overline{Q}=A, \overline{P}=\left( {{\alpha }_{0}}-k\frac{2\sqrt{{{m}^{2}}+1}-2}{{{m}^{2}}\sqrt{{{m}^{2}}+1}} \right)AL$, $\overline{U}=\left[ 2\left( 2+{{m}^{2}} \right)-2\times \sqrt{{{m}^{2}}+1}kl \right]/\left( {{m}^{2}}\sqrt{{{m}^{2}}+1} \right) $, W为与激光器非线性噪底相关的常数, 其中k=k0/(πγL), 调制系数m=a/γL。为消除光强I0(λ)随激光器本身的功率抖动或者气室灰尘等造成的不稳定因素, 把F2与F1作比, 可得到与光功率I0(λ)大小无关的式子:
$ \frac{{{F_2}}}{{{F_1}}} = \frac{{\bar W + \bar UC}}{{\bar Q - \bar PC}} = - \frac{{\bar U}}{{\bar P}} + \frac{{\bar T}}{{\bar Q - \bar PC}} $
(8) 式中, 常数$ \overline{T}=\frac{\bar{U}\bar{Q}+\bar{P}\bar{W}}{{\bar{P}}}$, 由此可见F2/F1与体积分数C成反比关系。一般, 如果分母项的常数P很小而分子项很大, 则当气体体积分数C很低且在一定范围内变化时, F2/F1随体积分数C的曲率变化也很缓慢, 在一定范围内可看成是线性变化。下面将通过数学推导作进一步解释。
对(8)式在某体积分数C0处进行泰勒级数展开:
$ \begin{array}{*{20}{c}} {\frac{{{F_2}}}{{{F_1}}} = \left( { - \frac{{\bar U}}{{\bar P}} + \frac{{\bar T}}{{\bar Q - \bar P{C_0}}}} \right) + \frac{{\bar T\bar P}}{{{{\left( {\bar Q - \bar P{C_0}} \right)}^2}}}\left( {C - {C_0}} \right) + }\\ {\frac{{\bar T{{\bar P}^2}}}{{{{\left( {\bar Q - \bar P{C_0}} \right)}^2}}}{{\left( {C - {C_0}} \right)}^2} + }\\ {\frac{{\bar T{{\bar P}^3}}}{{{{\left( {\bar Q - \bar P{C_0}} \right)}^4}}}{{\left( {C - {C_0}} \right)}^3} \cdots } \end{array} $
(9) 从级数的2次幂开始, 每一项的系数都比前一项多一个乘数因子:
$ \begin{array}{*{20}{c}} {X = \frac{{\bar P}}{{\bar Q - \bar P{C_0}}} = }\\ {{{\left\{ {{{\left[ {\left( {{\alpha _0} - k\frac{{2\sqrt {{m^2} + 1} - 2}}{{{m^2}\sqrt {{m^2} + 1} }}} \right)L} \right]}^{ - 1}} - {C_0}} \right\}}^{ - 1}}} \end{array} $
(10) 由于丙烷吸收很弱, α0和k都是小量, 调制系数m一般也很小。当丙烷在低体积分数情况下(C < 0.03), 泰勒展开处的C0也是一个小量, 因此可认为乘数因子X的分母项很大, 则X可认为是小量, 所以体积分数C在C0附近一定范围内, C-C0也是一个小量, 因此(9)式中2次幂以上都可作为小量而忽略不计, 则F2/F1与C在一定的体积分数范围内可看作是一种线性关系:
$ \begin{array}{*{20}{c}} {\frac{{{F_2}}}{{{F_1}}} = \left( { - \frac{{\bar U}}{{\bar P}} + \frac{{\bar T}}{{\bar Q - \bar P{C_0}}}} \right) + }\\ {\frac{{\bar T\bar P}}{{{{\left( {\bar Q - \bar P{C_0}} \right)}^2}}}\left( {C - {C_0}} \right)} \end{array} $
(11) 因此, 理论分析表明, 对于具有宽谱吸收峰的丙烷气体, 由于其拟合得到的洛伦兹线型和简单分子的相比多了常数项α0和k0, 其一次谐波和二次谐波信号大小都与丙烷气体体积分数有关。在实际系统探测中, 通过锁相放大或者窄带滤波技术, 获得经过气体吸收后的F2和F1信号, 在一定的体积分数范围内, 可通过两者间比值与体积分数的线性关系来进行计算。
基于近红外波段激光光谱吸收的丙烷探测研究
Research of propane detecting based on near-infrared laser spectral absorption
-
摘要: 为了对石油气挥发性有机化合物的主要成分进行实时监测,实现石油化工行业的安全生产,采用激光光谱分析技术、利用宽谱光源分析了丙烷在1686.00nm~1687.00nm波段的光谱吸收特征,获得了吸收系数随波长变化的洛伦兹线型,其半峰半宽为0.21nm。选择中心波长为1686.30nm的分布反馈式半导体激光器作为光源,在丙烷宽谱吸收峰范围内进行波长扫描,得到了一次谐波信号和二次谐波信号随丙烷体积分数的变化规律,并在丙烷的体积分数0.0050~0.0300范围内标定了二次谐波与一次谐波信号的比值与体积分数的线性关系。结果表明,实验系统有很好的稳定性与重复性,能够进行实时的丙烷在线检测。该研究为探测其它挥发性有机化合物气体提供了理论及实验参考。Abstract: In order to real-time detect the main component of volatile organic compounds of oil gas and realize production safety in petrochemical industry, laser spectrum analysis technique was applied. Firstly, the broadband optical source was utilized to analyze the spectral absorption features of propane between 1686.00nm and 1687.00nm. The relationship between absorption coefficient and wavelength was fit into Lorentz lineshape, with half width at half max of 0.21nm. A distributed feedback laser with center wavelength of 1686.30nm was chosen as the sigle mode light source, and laser wavelength scan was made within wide absorption spectrum of propane. The changing regularity of the first and second harmonic signal with volume fraction of propane was acquired. The linear relationship between the ratio of the second harmonic signal to the first harmonic signal and propane volume fraction was calibrated with propane volume fraction from 0.0050 to 0.0300. The results show that, the system has good stability and repeatability, and can be used to do propane online detection. The study can provide the reference of theory and experiment for detecting other volatile organic compound gases.
-
Key words:
- spectroscopy /
- propane detection /
- tunable diode laser /
- near infrared /
- volatile organic compound
-
Figure 1. Absorption spectrum of propane, carbon dioxide, methane and water near 1686nm[16]
-
[1] DU Z H, ZHAI Y Q, LI J Y, et al. Techniques of on-line monitoring volatile organic compounds in ambient air with optical spectroscopy[J]. Spectroscopy and Spectral Analysis, 2009, 29(12):3219-3203(in Chinese). [2] LIN W H, GAO Z H, YANG Y, et al. NO2 detection based on laser spectrum differential method[J]. Laser Technology, 2014, 38(6):835-838(in Chinese). [3] HE Y, ZHANG Y J, KAN R F, et al. Open-path online monitoring of ambient atmospheric CO2 based on laser absorption spectrum[J]. Spectroscopy and Spectral Analysis, 2009, 29(1):10-13(in Chinese). [4] SUN K, XIE L, JU Y, et al. Compact fiber-optic diode-laser sensor system for wide-dynamic-range relative humidity measurement[J]. Chinese Science Bulletin, 2011, 56(32):3486-3492. doi: 10.1007/s11434-011-4710-x [5] YANG B, HE G Q, LIU P J, et al. Research progress of tunable diode laser absorption spectroscopy for combustion diagnostics[J]. Laser Technology, 2011, 35(4):503-510(in Chinese). [6] QU J M, FENG X Q, LIU Y C, et al. Source characteristics of VOCs emissions from vehicular exhaust in the pearl river delta region[J]. Acta Scientiae Circumstantiae, 2014, 34(4):826-834(in Chinese). [7] WANG Y T, LIU J, ZHANG J C, et al. A methane gas sensor with optic fiber based on frequency harmonic detection technique[J]. Measurement and Control Technology, 2003, 22(11):19-21(in Chinese). [8] SKELDONA K D, McMILLANB L C, WYSE C A, et al. Application of laser spectroscopy for measurement of exhaled ethane in patients with lung cancer[J]. Respiratory Medicine, 2006, 100(2):300-306. doi: 10.1016/j.rmed.2005.05.006 [9] LI M X, LIU J G, KAN R F, et al. Design of real-time measurement of atmospheric CO and CH4 based on tunable diode laser spectroscopy system[J]. Acta Optica Sinica, 2015, 35(4):0430001(in Chinese). doi: 10.3788/AOS [10] HARRISON J J, BERNATH P F. Infrared absorption cross sections for propane(C3H8) in the 3μm region[J]. Journal of Quantitative Spectroscopy & Radiative Transfer, 2010, 111(9):1282-1288. [11] SONG S F, XING W R, LIU M. Theory and research advancement of quantum cascade lasers[J]. Laser & Infrared, 2013, 43(9):972-976(in Chinese). [12] LIU J Q, ZHAI S Q, KONG N, et al. Quantum cascade infrared photodetectors[J]. Inflared and Laser Engineering, 2011, 40(8):1397-1402(in Chinese). [13] CHAN K, ITO H, INABA H. All-optical remote monitoring of propane gas using a 5-km-long, low-loss optical fiber link and an InGaP light-emitting diode in the 1.68μm region[J]. Applied Physics Letters, 1984, 45(3):220-222. doi: 10.1063/1.95189 [14] XIN Z. Diode-laser absorption sensors for combustion control[D]. Stanford, USA: Stanford University, 2005: 5-22. [15] GENG H, ZHANG Y J, LIU W Q, et al. Acquisition method of high resolution spectra of ethanol vapor in near-IR range[J]. Journal of Atmospheric and Environment Optics, 2012, 7(1):57-62(in Chinese). [16] SHARPE S W, JOHNSON T J, SAMS R L, et al. Gas-phase databases for quantitative infrared spectroscopy[J]. Applied Spectroscopy, 2004, 58(12):1452-1461. doi: 10.1366/0003702042641281