4
21. M. W. Beckstead, R. D. Derr, and K. F. Price, Raket. Tekh. Kosmn., 8, No. 12 (1970). 22. B. V. Novozhilov, Nonsteady-State Combustion of Solid Rocket Fuels ~in Russian], Nauka, Moscow (1973). XeCI and KrF EXCIMER LASERS FOR DIAGNOSTICS OF FLAMES BY SPONTANEOUS RAMAN SCATTERING Ao N. Malov and S. Yu. Fedorov Flame diagnostic techniques based on Raman scattering of light are widely known. They include, among others, spontaneous Raman scattering spectroscopy and coherent antistokes Raman scattering (CARS) [1-3]. Raman scattering techniques are a unique instrument for measurements in flames, which are objects with a multispecies composition, a high level of density and temperature fluc- tuations, and intrinsic luminosity. Here full use can be made of the potential advantages of these methods: localization, noncontact, rapid response, the ability to measure the populations of vibrational and rotational levels of different molecules, and the ability to eliminate parasitic background from the measurement volume. These advantages are fully rea- lized in CARS systems, but these usually require complicated experimental procedures. Spon- taneous Raman scattering methods are more accessible, but the resulting signal is often too small for measuring the parameters of gaseous media with satisfactory accuracy. In recent years, however, the appearance of new monochromatic UV sources, i.e., excimer lasers, has extended the feasibility of simple experimental techniques. The reason for the high efficiency of excimer lasers in spontaneous Raman scattering systems compared to tra- ditional (argon, ruby, neodymium) laser sources is the strong wavelength dependence of the scattered signal ~%-4, as well as the greater sensitivity of photodetectors in the IFV range. Because of the increased useful signal it appears to be possible to measure molecular level populations at densities of i0 -17 cm -3 or higher during a single laser pulse with satisfactory accuracy, so that measurements can be made in nonstationary flows. Here we present spectra of spontaneous Raman scattering obtained under flame conditions using the most widely available XeCI and KrF excimer lasers. A simple pulsed spontaneous Raman scattering spectrometer for measuring the local densities of the prSncipal species in flames with high temporal resolution is described. Description of the Experimental Apparatus. An ultraviolet excimer oscillator-amplifier laser system developed at the Institute for Theoretical and Applied Mechanics, Siberian Branch, Academy of Sciences of the USSR was used [4]. The light from the laser 1 was focused by a lens L1 with a focal length of 112 mm (Fig. i)o An image of the focused beam was pro- jected onto the input slit of a DMR-4 double monochromator by a collecting system made up of lenses L2 and L3 (magnification "4• A diaphragm 5 mounted at the inlet slit restricted the length of the measurement volume to give dimensions of 1 • 0.5 • 0.5 mm. Light passing through the monochromator 8 struck the photocathode of an FEU-71 photomultiplier 7. The sig- nal from the photomultiplier was fed into a correlator 6 and on to an N-381 recorder 9 or a V7-16 general-purpose voltmeter i0. The correlator was synchronized with the laser pulse. The photomultiplier signal was integrated for I0 ~sec and the laser pulse duration was 30 nsec. In order to eliminate electrical noise from the laser, the scattered signal detection system was mounted in a shielded case. After the laser light passed through the measurement volume, it struck a photodiode 3 whose signal was recorded on a storage oscilloscope ii. This signal was used to normalize the laser energy. A burner with a nozzle diameter of 0.5 mm was mounted perpendicular to both the laser axis and the axis of the collecting system and could be moved in two directions with the aid of micrometer screws. Novosibirsk. Translated from Fizika Goreniya i Vzryva, Vol. 24~ No. 4, pp. August, 1988. Original article submitted July i, 1987. 54-58, July- 0010-5082/88/2404-0431 $12.50 1989 Plenum Publishing Corporation 431

XeCl and KrF excimer lasers for diagnostics of flames by spontaneous Raman scattering

Embed Size (px)

Citation preview

21. M. W. Beckstead, R. D. Derr, and K. F. Price, Raket. Tekh. Kosmn., 8, No. 12 (1970). 22. B. V. Novozhilov, Nonsteady-State Combustion of Solid Rocket Fuels ~in Russian], Nauka,

Moscow (1973).

XeCI and KrF EXCIMER LASERS FOR DIAGNOSTICS OF FLAMES

BY SPONTANEOUS RAMAN SCATTERING

Ao N. Malov and S. Yu. Fedorov

Flame diagnostic techniques based on Raman scattering of light are widely known. They include, among others, spontaneous Raman scattering spectroscopy and coherent antistokes Raman scattering (CARS) [1-3].

Raman scattering techniques are a unique instrument for measurements in flames, which are objects with a multispecies composition, a high level of density and temperature fluc- tuations, and intrinsic luminosity. Here full use can be made of the potential advantages of these methods: localization, noncontact, rapid response, the ability to measure the populations of vibrational and rotational levels of different molecules, and the ability to eliminate parasitic background from the measurement volume. These advantages are fully rea- lized in CARS systems, but these usually require complicated experimental procedures. Spon- taneous Raman scattering methods are more accessible, but the resulting signal is often too small for measuring the parameters of gaseous media with satisfactory accuracy.

In recent years, however, the appearance of new monochromatic UV sources, i.e., excimer lasers, has extended the feasibility of simple experimental techniques. The reason for the high efficiency of excimer lasers in spontaneous Raman scattering systems compared to tra- ditional (argon, ruby, neodymium) laser sources is the strong wavelength dependence of the scattered signal ~%-4, as well as the greater sensitivity of photodetectors in the IFV range. Because of the increased useful signal it appears to be possible to measure molecular level populations at densities of i0 -17 cm -3 or higher during a single laser pulse with satisfactory accuracy, so that measurements can be made in nonstationary flows.

Here we present spectra of spontaneous Raman scattering obtained under flame conditions using the most widely available XeCI and KrF excimer lasers. A simple pulsed spontaneous Raman scattering spectrometer for measuring the local densities of the prSncipal species in flames with high temporal resolution is described.

Description of the Experimental Apparatus. An ultraviolet excimer oscillator-amplifier laser system developed at the Institute for Theoretical and Applied Mechanics, Siberian Branch, Academy of Sciences of the USSR was used [4]. The light from the laser 1 was focused by a lens L1 with a focal length of 112 mm (Fig. i)o An image of the focused beam was pro- jected onto the input slit of a DMR-4 double monochromator by a collecting system made up of lenses L2 and L3 (magnification "4• A diaphragm 5 mounted at the inlet slit restricted the length of the measurement volume to give dimensions of 1 • 0.5 • 0.5 mm. Light passing through the monochromator 8 struck the photocathode of an FEU-71 photomultiplier 7. The sig- nal from the photomultiplier was fed into a correlator 6 and on to an N-381 recorder 9 or a V7-16 general-purpose voltmeter i0. The correlator was synchronized with the laser pulse. The photomultiplier signal was integrated for I0 ~sec and the laser pulse duration was 30 nsec. In order to eliminate electrical noise from the laser, the scattered signal detection system was mounted in a shielded case. After the laser light passed through the measurement volume, it struck a photodiode 3 whose signal was recorded on a storage oscilloscope ii. This signal was used to normalize the laser energy. A burner with a nozzle diameter of 0.5 mm was mounted perpendicular to both the laser axis and the axis of the collecting system and could be moved in two directions with the aid of micrometer screws.

Novosibirsk. Translated from Fizika Goreniya i Vzryva, Vol. 24~ No. 4, pp. August, 1988. Original article submitted July i, 1987.

54-58, July-

0010-5082/88/2404-0431 $12.50 �9 1989 Plenum Publishing Corporation 431

i 4 1 4- [7 . , I . _ \ , , L " "

Fig. i. A diagram of the experimental apparatus: i) excimer laser, 2) view- ing dump, 3) photodiode, 4) power sup- ply, 5) diaphragm, 6) correlator cir- cuit, 7) photomultiplier, 8) double monochromator, 9) recorder, i0) gener- al-purpose voltmeter, Ii) storage os- cilloscope.

This spectrometer could, therefore, operate in two regimes: (i) taking a spectrum during pulsed periodic operation of the laser and scanning the spectrum with the monochroma- tor while recording the signal on a recorder, and (2) recording the amplitude of the signal and normalizing to the laser energy when the laser is operated in a single pulse regime, tuning the monochromator to the wavelength range of interest and recording the readout with a voltmeter and recording oscilloscope. The laser pulse energy was -I00 mJ for the XeCI laser (308 nm) and ~50 mJ for the KrF laser (249 nm). The collecting system had an effi- ciency of ~0.02.

Results of the Measurements and Discussion. Raman scattering spectra were taken in air under standard conditions, in a jet of hydrogen, and in a hydrogen-air diffusion flame.

Figure 2a shows the characteristic Raman scattering spectrum of air. The Q-branches of the vibrational bands of 0= and N 2 were recorded here. The spectrometer resolution was ~3 nm, and was determined by the slit widths and dispersion of the monochromator and by the width of the laser line. At this resolution the recorded lines correspond to the appa- ratus function of the spectrometer. Their intensities are proportional to the populations of the vibrational levels of the molecules. Figure 2b shows a fragment of the spectrum from a jet of hydrogen in air taken 5 mm from the nozzle cutoff and roughly on the axis of the jet. Here and in the following the excess pressure in front of the nozzle is 0.2 atm.

In Fig. 2c the Raman scattering spectrum was taken at the same point in a hydrogen-air diffusion flame. The spread in the magnitude of the signal at the lines is explained by fluctuations in the instantaneous densities of the molecules in the flow and, to a lesser extent, by the spread in the laser pulse energy (• Nevertheless, from the spectra it is possible to evaluate qualitatively the changes in the densities of the components. For example, there is a clear reduction in the densities of H= and N= and an increase in the density of H20 in the flame compared to the cold jet.

It was found that the XeCI laser radiation excites fluorescence in the OH molecules present in the flame. This fluorescence is a result of the coincidence of the XeCI laser line (308 nm) and the wavelength of the (0, 0) band of the 2Z-2n transition of the OH mole- cule [5]. In the spectrum from the flame (Fig. 2c) there is a signal produced by laser in- duced fluorescence (LIF) of OH with a long wavelength edge around 1800 cm -I from the Ray- leigh line. The luminescence spectrum of OH is shown in Fig. 2c for comaprison. Compared to that spectrum, the LIF spectrum is more intense by roughly i0 ~ times and its center is shifted into the Stokes region.

Using the LIF of OH from an XeCI laser to obtain information on the states of OH might be a subject for a separate study, but in the present work this makes it difficult to detect Raman scattering spectra. Its presence hinders the detection of spontaneous Raman scatter- ing spectra of molecules with vibrational energies below 1800 cm -I, including 02 and CO 2. Molecules whose vibrational energy exceeds 1800 cm -I (N 2, H20, H 2, etc.) are recorded in the usual way. This shortcoming of the spectrometer can be avoided by using a KrF laser. It has roughly the same energy per pulse and is soemwhat more efficient in a spontaneous Raman scattering system because of its shorter wavelength. The materials for the optical elements in the system, however, must be chosen more carefully in order to reduce their fluorescence. Figure 2d shows a spectrum taken with the KrF laser in a hydrogen flame at the same distance from the nozzle cutoff. Each of these spectra was taken in ~5 min. The time to record the signal at each point of the spectrum was determined by the duration of

432

a I b H 2

~p ~ 4156 H

N '5$1

o: N: 1556 2551

~ ~ ,Y652

o 1ooo 2000 ?ooo 4ooo H~

_ , ~ Jl Ill 14~ie 253; OH 0(308 nm ). 1556 l ,

0 I [ I l I

0 2000 4000 0 1000 2000 ,3"000 P~ cm- z

Fig. 2. Raman scattering spectra of air (a), a jet of hy- drogen in air (b), and a hydrogen-air diffusion flame (c, d): (!) luminescence (without attenuation); (2) fluores- cence (attenuation by "i0~); i = 308 nm (a-c), 249 (d).

the laser pulse (30 nsec). The spectrometer was operated in this regime to search for lines and to tune to the required lineo

Further measurements were taken in the single pulse regime. Information was obtained on the intensity of vibrational-rotational bands of certain molecules in a single pulse and this was used with a calibration to compute the density of the molecules at the given point in the flame.

Figure 3a shows the distributions of the densiteis of H 2 and N 2 in the perpendicular cross section of the flame normalized to the laser pulse energy. The calibration was based on the signals from known densities (atmospheric density for N 2 and 100% for H2). Figure 3b shows the distribution of H20 at the same cross section. This profile was not calibrated. The ordinate is the signal amplitude in photoelectrons. The time resolution of the density measurement is 30 nsec. At those points where the confidence interval is not plotted, it is less than the size of the plot point~ The actual scatter is considerably greater than the instrumental accuracy of the measurements. (Several measurements were made at each point in the flame~ This scatter is explained by the fluctuations in the density of the constituents in the flame characteristic of a turbulent flow. It illustrates the possibil- ity of using this system for measuring the instantaneous densities of molecules and studying the turbulence parameters. Similar density profiles of N2, H2, and 0 2 at the same cross section of the flame, but time averaged using a cw argon laser, have been given elsewhere [6]. Our experiments reveal the same flame structure, but provide additional information on the local density fluctuations.

These measurements can also be used to estimate the limiting sensitivity that can be attained with this system. The signal obtained from the atmospheric density of N2, for ex- ample, was ~500 photoelectrons. However, no use was made of the following possibilities: (i) reverse mirrors were not used for the laser and scattered radiation, although they might increase the useful signal by a factor of 4 and (2) the maximum pulse energy was not ob- tained from the laser [4]. (Instead of 500 mJ, only i00 mJ was obtained from the XeCI la- ser.) This would yield a fivefold increase in the signal~ A further slight increase in the

433

II~ Cm "3

2.10 ~ _

I.i0 ~g

0 -5

0 I~ pe 600-

r

r 400-

r ~ 2oo

o r o oi -5 ~o~ o 2

o 0 o

_R ~ ~ o * j

~8 ~o

-J - 1 0 1 3 ' ~x, mn

1

-I

§

b

# +

§ ++ + § .§ ~ |. �9

I I I I T

-I 0 I J" ~x, nnn

Fig. 3. Density distributions of (a) N 2 (i) and H2 (2), and (b) H20 molecules (3) in the perpendicular cross section of the flame at a distance of 5 mm from the nozzle cutoff. The • interval owing to the statistical character of the optical signal is shown.

signal could be achieved by more careful matching of the thickness of the laser beam waist to the width of the entrance slit of the monochromator. By making use of these additional possibilities, therefore, the signal from atmospheric density N= molecules could reach at least 500 x 4 • 5 = 104 photoelectrons per pulse. The measurement error at this signal level is 1% and a density of ~2.1017 N 2 molecules/cm ~ could be measured with an accuracy of 10%.

The authors thank Yu. I. Krasnikov for help in constructing the correlator circuit.

2~

3o

4. 5. 6~

LITERATURE CITED

M. Lapp and C. M. Penney (eds.), Laser Raman Gas Diagnostics, Plenum, New York (1974). S. A. Akhmanov and N. I. Koroteev, Nonlinear Optical Methods for Scattered Light Spec- troscopy [in Russian], Nauka, Moscow (1981). R. Goulard (ed.), Combustion Measurements, Academic Press, Washington (1976). A. N. Malov and A. M. Razhev, Zh. Tekho Fiz., 55, 664 (1985). A. Gaydon, The Spectroscopy of Flames [Russian translation], IL, Moscow (1959). A. L. Rudnitskii, S. Yu. Fedorov, and Yu. A. Yakobi, in: Optical Methods for Studying Gaseous Flows and Plasmas [in Russian], ITMO, Minsk (1982).

434