1. Introduction

In the past, photoionization quantum yields and sorption coefficients in the range 10 to 100 nm ve been measured in windowless systems utilizing ferential pumping techniques with monochromators. is method suffers because the path lengths are not ell defined, which may explain why the agreement tween different investigators is not always satisctory [1]. Recently de Reilhac et al., [2] have proved the measurement of absorption coefficients the range 10 to 50 nm using absorption cells with iminum windows, which provide a well defined th length. However, so far not many measurements ve been made.

In the course of our study [3] of the vacuum ultralet photolysis of various organic compounds with re gas resonance lamps, it became necessary for to know very exactly the ionization quantum yields d the absorption coefficients of the different gases der investigation at the particular wavelengths of ht emitted by these lamps. Particularly for molecules hose absorption spectra contain structure narrow mpared with the slit function, the accuracy of the lues of absorption coefficients and ionization quanm yields measured with a monochromator will pend on the resolution of the monochromator. It is thus best for our purposes that we determine

Work supported in part by U.S. Atomic Energy Commission.
Figures in brackets indicate the literature references at the end of this paper.

the values of these constants with the same lamps that are used in the photolysis experiments. Although this method gives the ionization quantum yield and absorption coefficient only at selected wavelengths (the rare gas resonance lines), they are, for those using rare gas resonance lamps as photochemical light sources, the wavelengths of primary interest. These measurements also provide a rough check of the accuracy of the values reported for these constants at these energies as determined by other methods. In the case of the neon resonance lines it is more difficult to correlate our measurements with others, since the lamp emits two resonance lines. However, these lines are rather close together and the 73.6 nm line is about three times the intensity of the 74.4 nm line [4]. Especially for those substances in which there is no structure in the absorption curve, these measurements should be good.

Since the experiments reported in this paper were performed, a somewhat similar investigation has been reported by Bennett et. al., [5] on the absorption coefficients and ionization yields of a number of compounds at 58.4 nm. In general, there is good agreement between their results and ours.

2. Experimental Procedure

2.1. Helium and Neon Resonance Lamps

The details of constructing and filling the helium and neon resonance lamps have been given before [6].

429-270 O 71 - 6

The operational characteristics, such as the effect of the rare gas pressure and of the power of the microwave generator on the intensity of the lamp, have also been discussed. To summarize briefly, the lamps are enclosed glass constructions, filled with a low pressure of rare gas, fitted with thin aluminum windows (100200 nm thick), and operated with a microwave gener ator. The helium pressure was about 1.5 torr while the neon pressure was approximately 2 torr in the two respective lamps. The intensity of these lamps (~102 quantum/s) was very steady, with less than a five percent fluctuation over the time necessary for one set of measurements (1 to 2 h). Over a longer period of time there was a gradual decay in flux mainly due to a deterioration of the aluminum window. This decrease amounted to about 10 percent over a 24 h period.

2.2. Double Cell Arrangement

In order to obtain accurate values of ionization quantum yields and extinction coefficients, the double cell arrangement, shown in figure 1, was used. Each cylindrical chamber of the double cell is about 11 cm long and 5 cm in diameter and has a volume of approximately 200 cm3. The two compartments, both of which contain a set of parallel circular nickel electrodes, are separated by an aluminum window. The gas that is to be investigated is introduced into the chamber immediately adjacent to the lamp. The second chamber contains a gas that serves as a standard actinometer. In this way, the light transmitted through the material

in the first cell can be directly determined by a saturation current measurement in the second cell, or when the first cell is evacuated, the constancy of the intensity of the lamp can be checked.

In some cases, where the saturation ion current measurements show good plateaus, only a single cell is necessary. All of the measurements on the fluoroand chloromethanes were made in the single cell apparatus. Most of the other results were obtained in the double cell arrangement.

2.3. Actinometry

The standard actinometric gas can in principle be any gas for which the quantum yield of ionization at the wavelength of interest is known. In practice, experimental errors of measurement can be minimized if the actinometric gas gives a constant saturation current over a wide voltage range. This will be the case for gases which have a low absorption coefficient at the wavelength emitted by the lamp. When saturation ion currents are measured in a gas having a high absorption coefficient at the particular wavelength of light, the saturation current which is observed is usually obtained over a very short voltage range and. therefore, is not always well characterized. The plateau of the saturation current can be improved by adding to the absorbant gas some other gas which is transparent to the radiation being used. For example, neon is transparent to the helium resonance radiation and helium is transparent to neon resonance radiation, so each of these rare gases can be used as a diluent for strongly absorbing gases irradiated with the lamp | giving off the resonance lines of the other gas.

Rare gases in general are considered to have an ionization efficiency of unit when they absorb a photon of energy greater than their ionization energies [7] However those rare gases which are ionized at the helium and neon resonance lines also absorb this light very strongly. For instance the absorption coefficients for argon at the neon and helium lines are 900 and 975 atm1 cm respectively [8]. Thus the plateau of the saturation current measured in argon (fig. 2) is not well characterized. However, when helium is added to the cell with the argon, a short well defined plateau is obtained. The added helium acts as a moderator so that ion multiplication does not occur until higher applied voltages.

Hydrogen gas is used as a secondary standard because of its lower absorption coefficients (300 and 170 atm-1 cm at 74 and 58 nm respectively) [9]. A very well defined plateau is obtained for the saturation ion current measured in pure hydrogen (fig. 2) over a wide pressure range. The ionization quantum yield of hydrogen is 0.94 and 1.00 at the neon and helium resonance lines respectively.

In order to obtain good saturation current plateaus. as well as meaningful absolute values for the saturation ion currents, the investigator must also be aware of the effects on the measured currents of such parameters as lamp flux and pressure of the absorbant gas. These effects have been discussed in detail previously

10], and thus will only be summarized here. If the umber of quanta emitted per unit area of the window s too high, no saturation current can be obtained; electron multiplication occurs before the plateau of the saturation ion current is reached. It has been shown hat if the pressure of the sample gas is too high, all the ions are not collected at the electrodes and the measured current is too low. Thus the investigator should take care to find the optimum conditions of flux and pressure for each individual gas, in order to obtain meaningful saturation ion current measure

ments.

*Standard.

n.d.- not determined.

(References are for literature values that are given in parentheses.)
"Ausloos, P., and Lias, S. C., Rad. Res. Rev. 1, 75 (1968).

Watanabe, K., Matsunaga, F. M. Sakai, H., Appl. Opt. 6, 391 (1967).
Samson, J. A. R., J. Opt. Soc. Am. 54,6 (1964).

Metzger, P. H., and Cook, G. R., J. Chem. Phys. 41,642 (1964).

Bennett, S. W.. Tellinghuisen, J. B., and Phillips, L.. F.. J. Phys. Chem. 75, 719 (1971). Schoen, R. L., J. Chem. Phys. 37, 2.032 (1962).

* Cook, G. R., and Metzger, P. H., J. Opt. Soc. Am. 54,968 (1964).

Huffman, R. E., Can. J. Chem. 47, 1823 (1969).

Cairns, R. B., and Samson, J. A. R., J. Geophys. Rev. 70, 99 (1965).

values reported by Bennett et al. [5], at the helium resonance line are also included in this table, as well as some earlier results obtained with windowless monochromators. In general, for the determinations of ionization quantum yields at the neon resonance lines, excess helium was added, while at the helium line, excess neon was added; as mentioned above, the plateau of the saturation ion current can be improved by adding a nonabsorbing diluent to the gas of interest. For the ionization quantum yield measurements at the argon resonance lines, the standard of comparison was the saturation ion current generated in nitric oxide. The ionization quantum yield of nitric oxide is known over this wavelength region [11].

state and that one can estimate an effective rotational temperature from the data. The rotational temperatures T are found to be 3570 K for the 2-1 band and 3371 K for the 1-0 band (excluding the initial part of the curve). The vibrational temperature Tr for R23 is found to be 3275 K. This is in agreement with earlier work which indicates that lower temperatures are usually observed for the vibrational mode than for the rotational mode in such flames [21] near the reaction

zone.

Since there is some atmospheric CO absorption for wavenumbers greater than 2300 cm-1, the apparent intensity data for some of the 1-0 lines in this region will be less than their true intensity. In addition, there is evidence of self-absorption in the strongest 1-0 lines as shown in figure 3. Accordingly, data were heavily weighted in favor of the 2-1 band in choosing the value of 3500 K for TR; no correction was needed for selfabsorption.

The calculated equilibrium temperature for C2H2-O2 flame in which the fuel-oxygen ratio is twice that of a stoichiometric mixture is 3400 K [22]. The equilib rium constant K for the reaction:

as given by Gaydon [23], is 0.34 at 3000 K and 1.58 at 3500 K. Therefore, one would expect little undissociated CO2 at 3500 K. Indeed very little CO2 emission was observed from the flame, which supports the estimate of 3500 K.

P'X and a may be determined for a particular emission line by taking the quotient of its integrated intensity for two path lengths. If a mirror is placed behind the emitting volume doubling the path for the second measurement, then the ratio of integrated intensities should be

This criterion depends upon the fact that consistent measurements could be made for recorder deflections > 1 percent of full scale.

The measured values of RL/R12 are shown in column 4 of table 3. Corresponding values of P'X are shown for a = 1.0 and 1.4. Column 7 of table 3 lists values of P' calculated from eq (3).

14

The integrated emissivity for individual lines was determined from absorption measurements, as previously described. Figure 7-A shows the recorder tracing for such measurements on P20 1-0, P14 2-1 and Ps 3-2. In B and C of figure 7 are traces of R1 and R2 respectively for each of these lines. A total of four runs was used to determine the ratio R1/R2 for each line. This ratio was used to correct the experimental values of the integrated emissivity for each line measured with the absorber in the path, since these measurements were made with the mirror behind the flame. Table 3 contains the experimentally determined values of R= Se(v)dv and R/R12 for each of the lines investigated.

In addition to the measurement of R* for isolated 1-0 emission lines, studies were also made of two overlapped pairs of 1-0 and 2-1 lines, namely, P18(1-0) -P12 (2-1) and R28 (1-0)-R39 (2-1). R* was measured for these near coincident pairs in the normal manner. The R* computed to represent the data then took the form

A

1 — exp (− P' · 2X · V (a, v) dv

P'X can be determined for a chosen value of a by comparing the two path measurements with the values calculated using the above expression. Penner [1] has calculated the above ratio for a Voigt line shape V(v, a) and for values of P'X and a which we can expect to encounter for CO at 3500 K.

Trial calculations of expression (18) using the previ ously described program demonstrated that the calcu lated ratio depends upon the interval Av over which the integration is carried out since small but finite contributions to (18) occur at large distances from the line center for a > 1.

In order to correspond to our measured values, a family of curves of R1/R2 (P'X) was computed for various values of a. The integrations were carried out for the range of Av for which the emissivity € >0.01.

TRANSMISSION

EMISSION