Home Search Collections Journals About Contact us My IOPscience

Development of a Lidar System for Measuring Methane Using a Gas Correlation Method

This content has been downloaded from IOPscience. Please scroll down to see the full text.

1999 Jpn. J. Appl. Phys. 38 6130

(http://iopscience.iop.org/1347-4065/38/10R/6130)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 134.117.10.200
This content was downloaded on 26/06/2014 at 12:58

Please note that terms and conditions apply.

Jpn. J. Appl. Phys. Vol. 38 (1999) pp. 6130–6132
Part 1, No. 10, October 1999
°1999
c Publication Board, Japanese Journal of Applied Physics

Development of a Lidar System for Measuring Methane Using a Gas Correlation Method
Atsushi M INATO, MD. Mahbubul Alam J OARDER, Satoru O ZAWA, Minoru K ADOYA1 and Nobuo S UGIMOTO2
Faculty of Engineering, Ibaraki University, 4-12-1 Nakanarusaswa, Hitachi 316-8511, Japan
1 NEC Guidance and Electro-Optics Division, 1-10 Nissincho, Fuchu 183-8501, Japan
2 National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba 305-0053, Japan

(Received April 27, 1999; accepted for publication July 1, 1999)

We developed a laser long-path absorption lidar system using a gas correlation method for measuring atmospheric methane.
This technique uses a broad-band laser and a gas correlation cell. We developed a potassium titanyl arsenate (KTA) optical
parametric oscillator at 3.416 µm for the laser source. The optical round-trip path length is 20 m. The experiment was carried
out using this system. It was estimated that the error in the density of methane is 4.4 ppm for measurement within 1 s.
KEYWORDS: long-path absorption measurement, infrared laser, optical parametric oscillator, atmospheric methane

laser and generates a signal wave of 1.5 µm and idler wave

1. Introduction of 3.4 µm. The laser power of the idler wave is 10 µJ. The
The background level density of atmospheric methane is repetition rate is 1 kHz. The pulse width is 11 ns. The spec-
approximately 1.6 ppm. The precise measurement of this gas tral bandwidth of the laser is 50 GHz, which is wider than the
is very important to understand global warming. On the other spectral bandwidth of the absorption lines of methane.
hand, methane is also an explosive gas. It explodes when the We carried out temperature, vibration and shock tests. We
concentration exceeds 5% (50000 ppm). Therefore, monitor- placed the laser head in a thermostatic bath. For temperatures
ing of methane is necessary for the safety of environments from 0◦ C to 50◦ C, we measured the pulse peak power and
containing this gas. pulse width. The change in central wavelength in this temper-
Many measurement techniques have been developed to ature range was 0.6 nm.
date.1–8) Most of these techniques require a single-mode laser Laser characteristics observed before and after following
or a high-accuracy scanning of the laser wavelength. How- vibration and shock tests.
ever, in general, controlling these lasers is difficult for field (A) Vibration test
measurements, where vibrations and changes in temperature 5–14 Hz: Amplitude 2.5 mm
are inevitable. 14–200 Hz: Maximum acceleration 2 G
Previously we reported a measurement technique for atmo- (B) Shock test
spheric trace species using a gas correlation method.9–11) In Pulse waveform: Sinusoidal half-wave
this technique, a multimode laser, having a broad-band spec- Acceleration: 10 G
trum, and a reference cell containing the target gas are used. Duration: 11 ms
Neither single-mode lasing nor scanning of the laser wave- The results indicate that this laser is reliable for field mea-
length are required. This means that we can construct a mea- surements.
surement system which is good for field measurements where To develop an accurate measurement system for a short op-
vibrations and changes in temperature are inevitable. tical path length, parameters such as the laser wavelength, the
Here, we report on the development of a laser source and length of the cell and the pressure of methane in the cell, are
a measurement system for measuring atmospheric methane, selected.11) The round-trip optical path length is 20 m. The
and evaluate the error in density of methane. length of the cell is 1 cm. One hundred percent methane at
1 atm is contained in the cell.
2. Configuration of the Developed System Figure 2 shows the atmospheric absorption spectrum
Figure 1 shows a schematic of our system. Laser pulses around the laser wavelength. In this wavelength region, there
are transmitted in air, scattered by the wall of a building, and are some absorption lines of methane but the effect of absorp-
collected using a 15-cm telescope. The diverging laser beam tion of other molecules such as H2 O is negligible.
is collimated and transmitted. The diameter of the laser beam The laser wavelength was roughly tuned by monitoring the
after the collimation lens is approximately 3 cm. The col- signal wave using a spectrum analyzer. Then, it was finely
lected laser pulses are divided by a beam splitter and received tuned to the absorption lines around 3416 nm by monitoring
by two InSb detectors (EG&G J10D). A gas correlation cell, the absorption of the idler wave with the methane cell. The
which contains 100% methane, is placed before detector B. absorption of the methane cell is 68%.
The gas correlation cell works as an optical filter which cuts The received laser powers by detector A and B, Pa and Pb ,
the spectral component absorbed by methane. are expressed by
In our previous paper, we proposed the monitoring of four Z
signals to reduce the effects of atmospheric turbulence and Pa (N ) = Aa T0 (t) P0 (ν) exp −(N Lσ (ν)) dν, (1)
fluctuation of laser power.11) Because the laser power is very Z
low, we opted to monitor only two signals, as shown in Fig. 1, Pb (N ) = Ab T0 (t) P0 (ν) exp
to receive the laser power effectively.
We developed a multimode potassium titanyl arsenate − [(Ncell L cell + N L)σ (ν)] dν, (2)
(KTA) optical parametric oscillator (OPO) as the laser
source.12) The OPO is pumped using a LD-pumped YAG where Aa and Ab are the optical efficiencies of the measure-
6130
 Jpn. J. Appl. Phys. Vol. 38 (1999) Pt. 1, No. 10 A. M INATO et al. 6131

Fig. 1. Gas correlation laser long-path absorption system for measurement

of methane.

Fig. 3. Received signals by detector A and detector B as a function of the

concentration of methane in the calibration cell. The density of methane
in the cell is transferred to the optical thickness.

The signals of the two detectors are recorded using a digital

oscilloscope (HP54616B) and transferred to a Windows com-
puter via a GPIB interface. The outputs of the photovoltaic
InSb detectors are directly connected to the 1 MÄ input of the
oscilloscope. The sampling rate of oscilloscope is 100 MHz.
To collect data efficiently, eight shots of signals are averaged
in the digital oscilloscope. It takes about 1 s to record the av-
eraged data.
3. Evaluation of the System
The signal-to-noise ratio of the developed system was eval-
Fig. 2. Synthesized absorption spectrum of the atmosphere around uated first. The standard deviation of the received laser power,
3416 nm. The path length is 100 m. 1Pa , was measured as
1Pa /hPa i = 0.035, (5)
ment system, including the reflectance of a hard target. T0 (t) where the standard deviation is normalized by the average of
represents the atmospheric turbulence effect. T0 (t) is a func- the received laser power, hPa i. The fluctuation of this value
tion of time. P0 (ν) represents the spectrum of the trans- depends on the fluctuation of laser power, the atmospheric
mitted laser. N and Ncell are the density of methane in air turbulence and the electric noise. Next, the standard deviation
and in the cell, respectively. σ (ν) is the absorption cross of f (N ), which is the ratio of Pa and Pb , was measured.
section of methane. L is the round-trip optical path length.
1 f (N )/h f (N )i = 0.025. (6)
L cell is the length of the methane cell.The received signal de-
creases as the optical thickness of methane increases. The Here, the effect of atmospheric turbulence is cancelled out.
laser spectrum has a component which is not absorbed by Next, the methane cell was removed and the standard devia-
methane. Consequently, as the optical thickness of methane tion of f (N ) was measured.
increases, the gradient of the received signal decreases. When
1 f NoCell (N )/h f NoCell (N )i = 0.018 (7)
the Ncell L cell is much larger than N L, the following relation
is established. Because the spectrum of each pulse changes, fluctuation of
dPa /dN > dPb /dN (3) f (N ) increases when the methane cell is placed in front of
detector B.
Based on the difference between dPa /dN and dPb /dN , we By using a calibration cell, the error of the derived density
can derive the density of methane. The ratio of Pa to Pb is of methane was evaluated. The calibration cell, which con-
expressed as a function of the density of methane. tains 100% methane, was placed in the outgoing path. The
Z length of this cell is 1 cm. For various pressures of methane
Aa P0 (ν) exp −[σ (ν)(Ncell L cell + N L)]dν
in this cell, the received signals are measured. Figure 3 shows
f (N ) = Z . (4) the results. In the range of density of methane shown in
Ab
P0 (ν) exp −(σ (ν)N L)dν Fig. 3, Pa and Pb are linear with respect to optical thickness
of methane. The gradient of Pa is much larger than that of Pb .
In eq. (4), the effect of atmospheric turbulence is canceled The error of the derived density of methane, 1N , is ex-
out. pressed as
6132 Jpn. J. Appl. Phys. Vol. 38 (1999) Pt. 1, No. 10 A. M INATO et al.

long-path absorption measurement system for atmospheric

1N = 1/(d f (N )/dN )1 f (N ), (8)
methane. By employing a gas correlation method, we suc-
where d f (N )/dN is the gradient of f (N ) and 1 f (N ) is the cessfully constructed a compact and stable system. We de-
standard deviation of the ratio of Pa to Pb . termined that a value for the error of methane of 4.4 ppm is
In Fig. 3, d f (N )/dN is approximately 2.84 × 10−4 achieved in 1 s measurement.
(1/ppm·m). Because 1 f (N ) was 0.025, the error of the mea- There exist numerous absorption lines of other gases, such
sured density of methane is estimated to be 88 ppm·m from as H2 O, CO2 and N2 O, in the infrared range which is covered
eq. (5). For a path length of 20 m, an accuracy of 4.4 ppm is by OPO.14) The laser long-path absorption technique using
expected. This value is much lower than the level of explo- the gas correlation method is easily applicable for measure-
sion. ments of other gases.
To measure the background level of methane, it is neces-
sary to achieve a higher level of accuracy. To improve the
accuracy of measurement, it is necessary to extend the path 1) H. Inaba, T. Kobayashi, M. Hirama and M. Hamza: Electron. Lett. 15
(1979) 749.
length, and to improve the signal-to-noise ratio. 2) H. Sano, R. Koga, M. Kosaka and K. Shinohara: Jpn. J. Appl. Phys. 20
In a previous paper, we reported another laser long-path (1981) 2145.
absorption experiment for CO2 .13) In this case, the round-trip 3) R. T. Menzies, C. R. Webster and E. D. Hinkley: Appl. Opt. 22 (1983)
path length was 1 km. The laser power was 100 µJ and the 2655.
4) K. Uehara and H. Tai: Appl. Opt. 31 (1992) 809.
laser wavelength was approximately 2 µm. The signal-to- 5) R. A. Baumgartner and R. L. Byer: Opt. Lett. 2 (1978) 163.
noise ratio was 33 in measurement within 1 s. In the experi- 6) K. Uehara: Appl. Phys. B 38 (1985) 37.
ment, we used a retroreflector as an optical target to improve 7) E. R. Murray, J. E. Van der Laan and J. G. Hawley: Appl. Opt. 15 (1976)
the signal-to-noise ratio. 3140.
8) N. Menyuk and D. K. Killinger: Appl. Opt. 26 (1987) 3061.
For methane at 3.4 µm, if we can achieve the same signal- 9) N. Sugimoto, A. Minato and T. Kobayashi: CLEO Pacific Rim (1996)
to-noise ratio in a 1 km path, the estimated error of density p. 35.
of methane would be 0.1 ppm. This value is adequate for the 10) A. Minato, Y. Tsutsumi, T. Kobayashi and N. Sugimoto: CLEO Pacific
measurement of background methane. Rim (1997) p. 208.
11) A. Minato, T. Kobayashi and N. Sugimoto: Jpn. J. Appl. Phys. 37 (1998)
To achieve a high signal-to-noise ratio, it is necessary to 3610.
monitor all four signals proposed in our previous paper to re- 12) M. Kadoya and K. Asaba: NEC Research anf Development 37 (1996)
duce the effects of fluctuation of the laser spectrum.11) 361.
13) N. Sugimoto and A. Minato: Appl. Opt. 32 (1993) 6827.
14) N. Sugimoto, A. Minato, K. Ozawa, Y. Saito and A. Nomura: Jpn. J.
4. Conclusion
Appl. Phys. 34 (1995) 2329.
In this paper, we reported the development of a new laser