Self-calibration methodology by normalized intensity for wavelength
modulation spectroscopy measurement
Jie Shao
a,
n
, Jie Guo
a
, Liming Wang
a
, Chaofu Ying
a
, Zhen Zhou
b
a
Institute of Information Optics, Zhejiang Normal University, Jinhua 321004, PR China
b
Kunshan Hexin Mass Spectrometry Technology Co., Ltd., Kunshan 215311, PR China
article info
Article history:
Received 16 July 2014
Received in revised form
9 September 2014
Accepted 30 September 2014
Available online 7 October 2014
Keywords:
Absorption spectroscopy
Wavelength modulation spectroscopy
Diode laser
abstract
A methodology of self-calibration for concentration measurement based on wavelength modulation
absorption spectroscopy has been developed by normalized fixed intensity. Experimental results show
that the simple self-calibration method not only effectively improves the calibration accuracy, but also
greatly improves stabilization and reliability as compared with the popular method called self-calibration
by 1f normalized. In addition, the proposed system do not need any additional equipment when
comparing with the use of traditional wavelength modulation absorption spectrometer. The standard
deviation of a concentration measurement with different optical intensity aimed on the detector has
been found to be below 1.0%. The dependence of the concentration assessment on the laser intensity
fluctuation has also been investigated, which shows that the method of self-calibration could be applied
in the field, specially where the dust is easily splattered on the windows of the detector.
& 2014 Elsevier B.V. All rights reserved.
1. Introduction
The tunable diode laser absorption spectroscopy (TDLAS)
technique has become an important method of non-contact
gaseous measurement, which mainly includes direct absorption
spectroscopy (DAS) and wavelength modulation absorption spec-
troscopy (WMS) techniques. DAS has been used widely to measure
gas characteristics because of its simplicity and ability to absolute
measurements by integrating the spectral absorbance without
calibration [1–3]. However, due to it being easily affected by laser
intensity fluctuation and baseline fitting error that often hampers
the accuracy and limitation detected, the DAS technique cannot
meet the requirements in measuring concentrations of trace gas
[4,5]. The detected limitation and accuracy can be largely im-
proved using the WMS technique by modulating the wavelength
of a diode laser [6]. Therefore, when the TDLAS technique is
combined with wavelength modulation, it is more suitable for
in-situ measurement because of its compact, robust and high
sensitivity [7–9].
Traditional WMS [10,11] has a high sensitivity to gas concen-
tration measurements in practical environments, while it cannot
determine gas concentration directly from the harmonic signals
without calibration experiments. Therefore, in order to deduce the
absolute concentration, traditional WMS for concentration
measurements is the need to calibrate by a known mixture and
condition. However the technique is influenced by many factors,
including the intensity modulation (sometimes called residual
amplitude modulation (RAM)) which comes from the nonlinear
output of laser, the temperature and pressure of the targeted gas,
the parameters of setting for lock-in amplifier, the intensity of the
light aimed onto detector, the modulation amplitude, etc. [12–16].
Therefore, the WMS technique cannot directly deduce the magni-
tude of absorption which is important information for concentra-
tion measurement.
To recover the absolute concentration from the detected signals
of WMS, Hanson et al. [17,18] take advantage of first harmonic (1f)
signal amplitude to correct second harmonic (2f) signal. Such a
technique is called self-calibration by 1f normalized. And the
method is popular in many field applications for gas concentration
measurements. The detector signal is passed to several digital
lock-in amplifiers to isolate 1f and 2f signals. However, in experi-
ments using digital lock-in amplifiers, high-speed data card is
required to take. Two physical lock-in amplifiers are used, which
obviously increases the complexity of the device. Furthermore the
normalized 1f is the magnitude at the center of the spectrum,
artificial error is easily involved because we cannot exactly select
the position and the center wavelength will shift after long time
running. To overcome this issue we need use a solid etalon with a
known free spectral range (FSR) in experiment, the relative
wavelength can be calculated accurately in the experiment. When
the relative wavelength is known in practice, 1f
-normalization can
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/optcom
Optics Communications
http://dx.doi.org/10.1016/j.optcom.2014.09.073
0030-4018/& 2014 Elsevier B.V. All rights reserved.
n
Corresponding author.
E-mail address: shaojie@zjnu.cn (J. Shao).
Optics Communications 336 (2015) 67–72